Re: [time-nuts] Rb Oscillator - rather fundamental question

[Richard (Rick) Karlquist]

David C. Partridge wrote:

Cough - the rubidium clock or oscillator does have an intrinsic frequency,
which is the rubidium hyperfine transition of 6 834 682 610.904 324 Hz, it's
just that the frequency generated by the transition in question isn't used
to DEFINE the second, so by definition, it must be secondary.  Only a
Caesium clock is a primary standard, as the second is DEFINED to be the time
taken for 9,192,631,770 cycles of the radiation corresponding to the
transition between the two hyperfine levels of the ground state of the
caesium 133 atom.[1].

Unless of course they changed the rules recently ...

[1] 

Dave


Well, what you said is true as far as it goes, but not the whole story.
The fact that a clock is based on cesium does not necessarily mean it
is a primary standard.  For example the "chip scale atomic clock" uses
cesium and is a secondary standard.  OTOH, certain experimental clocks
based on atoms such as rubidium, mercury, etc could be considered
primary standards in spite of the definition of the second.

It's not the type of atom, but the type of clock that is crucial.
"Cesium" usually refers to an atomic beam clock and "Rubidium" usually
refer to a gas cell device.  In an atomic beam, the atoms are, on the
average, unperturbed, and will transition at exactly the 9192... 
frequency in the definition of the second.  Except that they are offset

from this frequency by a known amount due to the C-field.  In a gas
cell device, the atoms are perturbed by the buffer gas which results
in a unknown frequency shift from the 6834... frequency.  You have
to remove this offset by comparing to a primary standard.

We used to say that in theory you could build a cesium beam standard
from a kit of parts on a desert island having no other clocks, and when 
you turned it on, it would be on the correct frequency (within a

tolerance) guaranteed by design/physics.  There is no way you
could do this with a rubidium or cesium gas cell standard
to any kind of accuracy associated with atomic clocks (it would only be
in the general neighborhood of 6834...)

That is the difference between primary and secondary standards.
Another difference is that secondary standard have "aging" and
primary standards don't.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5071A question

[Richard (Rick) Karlquist]

When the 5071A product line was sold to Symmetricom ~4 years ago,
the production manager and his team of 15 moved with the product.
The manager left Symmetricom a few years ago, and recently most
of the rest of the team left Symmetricom.  The 5071A will now
be made on the east coast at the facility that made Symmetricom's
other cesium products.  A few ex-HP'ers may assist in some capacity.
These guys are top notch and there is at least some hope that
they won't lose the recipe.  Everyone is watching to see what
happens.  (Note:  I am not involved with Symmetricom).

Beyond any problems having to do with Symmetricom, the design is
getting very long in the tooth in terms of components.  The boards
I designed have, for example, 1206 passives and discrete transistors.
They were, at least, surface mount, which was still a novelty in
1990.  Eventually, something will have to be done along the
lines of redesign.  I hate to think what would happen in that case.

Rick Karlquist N6RK


Christopher Hoover wrote:
We time nuts, and time labs worldwide too, recognize the 5071A as an 
amazingly accurate and precise instrument as well as an extraordinary 
industrial product with these same qualities repeated many times over 
and over.


Most of us know the product was transferred from HP/Agilent to Symnetricom.

My question is as follows.  Is the 5071A (or at least its tube?) still 
in production at Symnetricom to HP/Agilent specifications and quality?


If not, we've really lost something.

-ch


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to 
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

and follow the instructions there.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 74AC gates phase noise

[Richard (Rick) Karlquist]

Bruce Griffiths wrote:


Only if the noise figure of the following amplifier is 4dB or so.
With no extra amplification is used one only needs a signal level of 
+1dBm to achieve a phase noise floor of -178dBc/Hz if the output is 
extracted through the crystal in such a way that the thermal noise of 
the load dominates.


Read US Patent 4283691, which explains how the 10811 works.  The
situation is far more complicated than the simple analysis above.
If you play your cards right, you can get much better phase noise
than what you have indicated.  The thermal noise of the load does
not enter into it.  Unfortunately, the very low noise first stage
in the 10811 is degraded by the emitter follower after it.  As I
previously stated, you can bypass these additional stages if you
want a lower phase noise floor.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] nubie querie

[Richard (Rick) Karlquist]

Don Collie jnr wrote:
I`m not sure that questions like these is welcome on this list, but here goes anyway : 


1/ What are the the 10 sources of the most constant [invariant] frequencies 
known to man, in order of decreacing constancy? Four immediately come to mind.


I vaguely remember reading that pulsars have some fantastic
stability like 1E-20.  I don't remember how they established
this.



2/ What is the mechanism in an oscillator, that is responsible for phase noise? 
[In only one sentence please]


Thermal (Johnson) noise in the resonator and thermal and shot noise in 
the active device.



If this is known, then it becomes easier to design low phase noise oscillators.


Perhaps, but just knowing the mechanism doesn't tell you that the
phase noise is inversely proportional to loaded Q, which you would
also need to know to optimize phase noise.




I hope you enjoy these questions,.Don.
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] nubie querie

[Richard (Rick) Karlquist]

For a simple crystal oscillator the two word answer might be "Leeson's model".  That of course is a cop out since it clearly defines multiple things that contribute to phase noise. 


Bob


And Leeson's model is basically the same as Edson's model, circa 1950
(Edson:  Vacuum Tube Oscillators) with 1/f noise added.  However, 
crystals have "excess noise" above and beyond simple thermal noise. 
There is no model, you just have to characterize the crystal.


Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP10811 taxonomy?

[Richard (Rick) Karlquist]

It is entirely possible that a 10544 could have excellent aging and beat
a 10811.  The SC cut doesn't improve aging.  The other "disadvantages"
of the 10544 in terms of electronics also don't affect aging.  The
main advantage of the 10811 is that it is much better from a cold start
in an instrument.  Not something that affects time-nuts users.

Rick Karlquist N6RK

Robert Atkinson wrote:

Hi Charles,
See http://www.leapsecond.com/museum/10811a/90027-1.pdf for details of the 
10811D/E versions and options. It's been appended to this copy of the A/B 
manual http://www.hparchive.com/Manuals/HP-10811AB-Manual.pdf
 
Hope this helps.
 
Robert G8RPI.


--- On Thu, 11/3/10, Charles P. Steinmetz  wrote:


From: Charles P. Steinmetz 
Subject: [time-nuts] HP10811 taxonomy?
To: "Discussion of precise time and frequency measurement" 
Date: Thursday, 11 March, 2010, 4:07


Has anyone compiled a taxonomy of the apparently numerous variants of 10811 
OXOs?  (Distinguishing features, rated performance, used in what instruments, 
etc.)

Also, I found Mark's observation that 10544s may drift less than 10811s interesting.  I 
own several 10544As and have experience with perhaps several dozen others, and think very 
highly of them, but haven't seen enough 10811s for a meaningful comparison 
("meaningful" being a relative term, when the sample sizes are as low as 
dozens).  What have others observed?

Best regards,

Charles



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



  
___

time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5370B OCXO

[Richard (Rick) Karlquist]

Bob Camp wrote:

Hi

Tough to believe that HP worried a lot about SKU inflation back when they
did the 5370 :)

I'm assuming that the 5370 was a Santa Clara design. That would put the
counter designers down the hall from the oscillator factory. Unlikely that
there was a communications gap about what could or could not be done.

You may well be correct though. Setting up and managing another part is the
most likely reason why not to add a couple more tests.

Bob


I was at Santa Clara Division in the 1980's.  They worried about both
the paper work for extra part numbers and the inventory problem. 
Especially  the latter.  There was always some manager whose stock

options depended on keeping inventory down (this was during the
Japanese "just in time" fad).

Also, if you wanted a part number, you had to obtain one from the 
official keeper of numbers, and she would usually give you a lecture

about the perils of "wasting numbers".

The counter designers were in a different bldg than the oscillator 
factory, which were at opposite ends of the 55 acre site.   More

importantly, they reported to different managers.  So
you shouldn't assume that communication was great.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low cost 6+ GHz Prescaler

[Richard (Rick) Karlquist]

Samuel DEMEULEMEESTER wrote:

Hi Rick,

My goal is not to design a killer prescaler useable for all applications,
but just to get the same specs than the original HP/Agilent prescaler board,
and why not a bit better. The original HP board is a very basic design :
4-stage of MMIC and a good-old MB510 prescaler. Nothing more. With moderns
parts, I was able to get higher performances for a cheaper (10x cheaper)
price. My design is already working as expected but I want to add some
features for ppl who what to use the board as a "generic" prescaler. In any
case, that's still a prescaler board designed for a counter.


The point is that I was the designer of the prescaler board with the 
MB510, and it isn't a very good prescaler.  The specs don't reveal the

problem with noisy sources.  The MB510 was something like $2, so I am
not sure what you are referring to in terms of 10x cheaper.  It is
not clear that "modern" parts are any better.  At the time of the 5334B,
well over 20 years ago, 10 GHz prescalers were already available.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP 11729C versus 11848A

[Richard (Rick) Karlquist]

What the 10811 production line did was to compare two 10811's
to each other by driving a high level mixer.  Anzac AM-123
amplifiers were used to increase the output level of the 10811's.
You can homebrew the AM-123 if you read the patent and can
get a 2N5109/2N5943 type of transistor.  Amplify the mixer
output with an LT1028/LT1128 type of low voltage noise op amp.

The ll729 is a microwave downconverter that is basically irrelevant
to measuring 10811's.  It didn't need to be very good because
an 8662 was used, which meant that the phase detector was
always working with noisy inputs (compared to the 10811).

Rick Karlquist
N6RK

Adrian wrote:
I tried to measure phase noise of a 10811A, but found out that the 
specified PN is below the noise floor of my 11729C.


Can anyone tell why the (phase detector method) PN noise floor is so 
much different between the two units?


11729C at 100 Hz -126 sBc/Hz (-133 dBc/Hz typ.)
11729C at 1 kHz -135 dBc/Hz (-140 dBc/Hz typ.)

11848A at 100 Hz -150 dBc/Hz (-160 dBc/Hz typ.)
11848A at 1 kHz  -160 dBc/Hz (-170 dBc/Hz typ.)

Some 25 dB is quite a difference, isn't it?

Basically, both units apper to be not that much different, except that 
the 11729C has an IF amp and power splitter between the input and the PD 
L port, while on the 11848A the L input is fed directly into the mixer. 
There are some differences in the LNA circuits, but that shouldn't be 
responsible for the huge noise floor difference.


11929C requires 0 dBm (-5...+10 dBm) 'L' (MW Test Signal) input level, 
that is amplified by the IF amp to >+10 dBm at the mixer input. Btw. the 
IF amp saturates at input levels grater than -50 dBm. For the 'R' input 
(5-1280 MHz), the manual specifies -1...+1 dBm.


For the 11848A, the L input is +15...+23 dBm, and 0...+23 dBm at the R 
input. Below +15 dBm L and R, the system degrades considerably. Reducing 
'L' to +7 dBm adds 10 dB to the noise floor. Reducing 'R' below +15 dBm 
adds directly to the noise floor. So, reducing it to 0 dBm would add 15 
dB to the noise floor.


So, it looks like the 11729C phase detector is more like a +10 dBm 
mixer, while the 11848A has a +17...+23 dBm mixer.


Replacing the 11729C PD with a ultra high level mixer should get the 
noise floor close to 11848A specs. It would just require to feed L and R 
directly into the mixer rather than using the instrument inputs.


Any thoughts / experiences referring to this?

Adrian

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to 
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

and follow the instructions there.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Flood of low end priced VNAs on FleeBay

[Richard (Rick) Karlquist]

We have a bunch of sweepers at work, and many of the them have
died and can't be fixed.  The only way they can be repaired is
to cannibalize one to fix another, assuming they don't have
the same bad module.  We have given away an 8510 to a school
and have others gathering dust.

Rick

jimlux wrote:

Steve Rooke wrote:

Just noticed there is a flood of lower end priced (<$5k & <<$5k) VNAs
on FleeBay just in case anyone is looking for one. No need to thank
me, just buy me one as well :)

Steve



Interesting... be aware that a lot of those 8510s aren't the whole 
analyzer, just the display or RF section, without the source (sweeper). 
ANd not just any sweeper will do.. you need one that the 8510 can 
control over GPIB and that has the right sync in and out.


A full 8510 has 3 or 4 boxes (depending on whether it has the 
s-parameter test set).  And, they're getting mighty long in tooth and 
tricky to fix.  I've got a cranky one at work that locks up 
occasionally, and we've been trying to figure out why.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to 
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts

and follow the instructions there.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] RF Prescaler for 53131A/53132A/53181A counters - update !

[Richard (Rick) Karlquist]

Samuel DEMEULEMEESTER wrote:


Right now, the performance is really good up to 3.5 GHz :

50 MHz : -7 dB
100 MHZ : -15 dB
250 MHz : -26 dB
500 MHz : -30 dB
1 GHZ : -32 dB
2 GHz : -32 dB
3 GHz : -30 dB
 


The real deal on the performance of prescalers is the ability to count
noisy sources.  If you did these tests with a pristine source,
they don't mean much unless all you ever want to measure are
pristine sources.  You also didn't mention how many significant
digits of accuracy you verified.  A marginal prescaler might
look good for the first half dozen or so digits but make errors beyond
that.  I like to see at least 10 digits that are not only stable
but correct; they can be stable but incorrect.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] A different timenuts interest

[Richard (Rick) Karlquist]

Hal Murray wrote:

Several years ago, I found a web site for a commercial place that made them 
for museums.  (I forget why I was looking for that sort of stuff.)  You might 
find interesting stuff/ideas via google but I didn't find a similar site with 
a bit of searching.




The Museum of Science and Industry in Chicago had one when I lived
there in the 1960's.

Rick N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Homebrew Rubidium oscillator, jitter and other tales :-)

[Richard (Rick) Karlquist]

ulm...@vaxman.de wrote:


blinked. This problem was eventually solved by driving the LED with a discrete
transistor instead of a free 74AC14 gate and decoupling this driver with an
RC-combination.


CMOS logic gates have a totem pole output that is famous for "overlap" 
where both transistors on briefly turned on at the same time, resulting 
in large current spike from the power supply.  It would be interesting
to see if this was the problem, or whether it was the "antenna" 
connected to the output.


Rick N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 2.5V reference IC in HP E1938 oscillator.

[Richard (Rick) Karlquist]

Hi, I was the designer of the board, but I don't remember the
part number of the reference.  I will try to consult my paper
schematic when I get a chance, if no one else can help you.
I do remember that I originally used some convenient reference
which seemed OK from the data sheet, but turned out to be
too noisy.  I changed it to a different one.

When asking questions such as this, it is helpful to know why
you need to know the part number.

Rick Karlquist N6RK

On 11/9/2014 12:54 PM, Rob040 . wrote:

Dear time-nuts,
Can anyone enlighten me by telling which IC has been used for creating the 
+2.5V reference voltage on the oscillator board of the E1938?It's the only IC 
in the diagram where the type number is not mentioned. In a post of years back 
I found LT1121, but that one is a voltage regulator and is used on the control 
board.
Cheers! 
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 10811

[Richard (Rick) Karlquist]

On 11/17/2014 5:54 PM, Bert Kehren via time-nuts wrote:

I ground one side of the tuning diode and use the 2 to 12 V as the external
  OCXO for my FRK's along with increasing the time constant. I have not
verified it but I think removing the zener Voltage should also improve  ADEV.
Bert Kehren



The choice for the Zener diode came from my old boss at HP,
who was very knowledgeable about using discrete zener diodes
as low noise references.  According to him, this particular
part number has very respectable noise.  This is just something
you have to know experientially, there is no theory of zener
noise AFAIK.  You might try measuring the noise of the 6.2V
reference voltage directly at baseband, and then multiplying
by the 1 Hz/volt sensitivity.  Let us know your results.

Rick Karlquist N6RK
(Now retired from HP/Agilent/Keysight)


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 10811

[Richard (Rick) Karlquist]

Yes, 5.1V Zeners (or something like that) have a "magic" zero tempco if
you put a conventional diode in series with it.  I used to know
stuff like that during the Jurassic period.

However, the diode in the 10811 is ovenized, so that is not so 
important. 6.2V was chosen to get +/- 5V tuning range, which was 
probably a spec inherited from the 10544.  Everything was done for a 
reason relevant to HP, which may or may not be a reason relevant to you.


The frequency of the 10811 is more sensitive to the temperature of
the oscillator transistor (a selected 2N5179) than the diode.  This
inspired the E1938A.

Rick

On 11/18/2014 2:31 PM, Brooke Clarke wrote:

Hi Rick:

When working on Tunnel Diode amplifiers we used (AFAICR) 5.1 V Zener
diodes to stabilize the lower voltage that drive the diode.
5.1V was supposed to have excellent temperature characteristics in terms
of repeatability (don't remember if low noise was part of the selection
criteria).
http://www.prc68.com/I/Aertech.shtml#TDA
The boards with the terminals have the Zener and a custom compensation
network using both Veco (spelling?) (-TC) and Balco (+TC) and fixed
resistors so that the gain stays constant over mil temperature ranges.

Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Richard (Rick) Karlquist wrote:



On 11/17/2014 5:54 PM, Bert Kehren via time-nuts wrote:

I ground one side of the tuning diode and use the 2 to 12 V as the
external
  OCXO for my FRK's along with increasing the time constant. I have not
verified it but I think removing the zener Voltage should also
improve  ADEV.
Bert Kehren



The choice for the Zener diode came from my old boss at HP,
who was very knowledgeable about using discrete zener diodes
as low noise references.  According to him, this particular
part number has very respectable noise.  This is just something
you have to know experientially, there is no theory of zener
noise AFAIK.  You might try measuring the noise of the 6.2V
reference voltage directly at baseband, and then multiplying
by the 1 Hz/volt sensitivity.  Let us know your results.

Rick Karlquist N6RK
(Now retired from HP/Agilent/Keysight)


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] MIT 2 inch cesium fountain, optically pumped

[Richard (Rick) Karlquist]

See:

http://newsoffice.mit.edu/2014/portable-atomic-clocks-1112

Any comments?

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low Additive Phase Noise 10 MHz Amps

[Richard (Rick) Karlquist]

I did some checking around for low noise buffer amps earlier
this year.  They needed to have 200 MHz bandwidth, so this
isn't directly applicable to 10 MHz.  I also needed isolation.
About the only information in print is from the usual suspects
at NIST.  They wrote a series of papers taking a fairly classic
discrete design and refining it.  Check FCS proceedings.  My
idea was to take ideas from 10 MHz and extend them to 200 MHz.
I didn't see any really profound ideas in the NIST papers.
There is a reproducibility problem because the original discrete
devices may not be available, or NIST might have used special
hand picked devices.

BTW, I cringe when I see the term "additive phase noise".
Phase noise, as all time nuts know, is NOT ADDITIVE NOISE,
as in AWGN.  It is multiplicative.  The correct term, IMHO,
is "residual phase noise".  What additive noise refers to
is the classic noise figure type noise involving small signals.
Again, as all time nuts know, low NF is necessary but not
sufficient for low phase noise.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits 10% discount in December

[Richard (Rick) Karlquist]

On 11/27/2014 7:07 AM, Tim Shoppa wrote:

For a hobbyist doing things a few at a time, what advantage is there to
buying RF transformers made by Mini-circuits etc., vs winding them using
commonly available ferrite cores/binocular cores?

If I needed to do a production run of 1000+ boards with tiny SMT
transformers, sure, no problem buying them from mini-circuits or a
distributor etc. But for hobbyist stuff seems far more flexible to wind
them onesy-twosy using not so tiny cores and windings selected for the
particular application.

Tim N3QE


You need the tiny cores to get the performance of the MiniCircuits
transformers.  You just can't get the same bandwidth using macro sized
"binocular" cores.  Now, if you don't need a lot of bandwidth, then
what you are saying could make sense.  Another issue is stray 
capacitance.  Considerably lower with a tiny core.


I have spent many hours characterizing MiniCircuits transformers
beyond the data sheet specs, and dissecting them to learn how they
do it.  They really do have a lot of rocket science in them.  In
terms of the engineering I am buying (especially in a one-off 
application) they are ridiculously cheap.  And I say that as a fairly 
knowledgeable transformer designer in my own right.


I do keep binocular cores around for higher power transformers, and
for "emergencies" when I need a transformer "yesterday".

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits 10% discount in December

[Richard (Rick) Karlquist]

On 11/27/2014 9:09 AM, Tim Shoppa wrote:

On Thu, Nov 27, 2014 at 10:28 AM, Bob Camp  wrote:


The main complaint is the difficulty of getting the correct cores. I seem
to have a few dozen bags of cores.



The mainline distributors (Allied, Newark, Mouser, etc.)  have excellent
selection of Fair-Rite and other cores. Admittedly to a neophyte the
equivalence of Fair-Rite or Laird part numbers to an Amidon-style number
may not be evident.


I recently needed some binocular cores for a transformer for a client.
After checking all the distributors, I had to buy 500 of them.  Minimum
quantity.  So much for one-off hobbyist projects.  The reason why I
did not use a MiniCircuits transformer for this client is that the
impedance was much less than 50 ohms.  This is one area that
MiniCircuits really does not address.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits 10% discount in December

[Richard (Rick) Karlquist]

On 11/27/2014 11:03 AM, Didier Juges wrote:

Another reason is reproducibility. If you or someone else wants


 to reproduce your design, using a well defined and available

 commercial part makes it much easier to achieve the same

 performance, particularly for RF components.


Didier KO4BB


Exactly right.

I wrote an article on receiving loops and showed a design
with a 50:5 (turns ratio) transformer wound on a toroid.
Again, this is not available from MiniCircuits.  I have
wasted time dealing with numerous dumb questions about
"can I use XXX core that I have laying around the lab?"
or can I use a different gauge wire to wind it, etc.
So many people complained about the shipping cost to
buy one core that I stocked the cores and included them
with PC boards I was already selling to reduce the shipping
cost to "zero".

BTW, 73 material would NOT work in this application.
I was asked about that multiple times even though the
article specifically said it would not work and explained
why.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits 10% discount in December

[Richard (Rick) Karlquist]

On 11/28/2014 10:08 AM, Dave M wrote:

Rick,
Thanks for the brief review of MiniCircuits stuff (I'm not connected
with them in any way except as a customer).
Since you've characterized some of their parts, perhaps you could help
answer a question that someone else posted, and one that I would like to
have answered as well.
Have you measured the effects of DC current in the windings of an RF
transformer, such as is seen if the transformer is in the collector
circuit of an amplifier?  If so, could you provide a generalization of
the effects, such as changes in frequency response, losses, etc.?

Many Thanks!,
Dave M



The very tiny cores on MiniCircuits transformers will start to saturate
at hundreds of mA.  The effect is that the magnetizing inductance drops,
which matters more at low frequencies than high frequencies.  I try
to avoid feeding DC to an amplifier through a transformer winding.
Instead I use a separate RF choke for that.  However, it would probably
work OK for, say, up to 25 mADC for a small signal transistor, but
why take a chance.

If you are using a DC feed through a transformer winding, be careful
not to accidently short circuit it, causing the full available current
from the power supply to flow through the transformer.  This can
actually magnetize the core and permanently damage it.  Saturation
via DC is much more deleterious than saturation via AC.

It is easy to calculate the flux density using Ampere's law, which
is one of the four Maxwell's equations.  H = I/(2piR).  Since R
(radius) is in the denominator, cores saturate from the inside
first before the whole core is saturated.  Multiply H by mu,
(as any time nut knows) to get B.  If R is 1 mm, and I is 628 mA,
then H = 10 ampere turns per meter.  If mu-relative is 1000, then
B = 4piX10^-7 X 1000 X 10 = 125 mT.  That is a hefty 1250 Gauss.
Some materials may be affected at 1/10 this flux density.

Now a days, a lot of RF is differential, in which case you are
free to feed DC through the output transformer without worrying
about this issue.

I worked for several companies where those 6 hole cylindrical chokes
were ubiquitous.  I specifically characterized those and established
a maximum current rating of only 100 mA.  Of course, many production
designs exceeded this limit and "worked" anyway.  I actually observed
someone try to put 20A through one of these.  The tantalum capacitors
on the "cold" side of the bead actually exploded due to RF current.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits 10% discount in December

[Richard (Rick) Karlquist]

On 11/28/2014 1:04 PM, Bob Camp wrote:


If you do need to run substantial current through a choke core, the larger 
binocular cores with a half turn through them are a better choice.

Still useless for 20A  (or even 2A)  though …

Bob



The binocular cores come in several hole sizes.
All other things being equal, current handling
capacity is directly proportional to hole size.

One thing to watch out for with putting DC thru
binocular cores happens in push pull RF power amplifiers.
The output transformer is usually a binocular
core on steroids, or its equivalent constructed
with beads or sleeves, etc, threaded over a single
"turn" made from brass tubes connected together
at the end away from the transistors.

In cheap (illegal) CB amplifiers, you will frequently see
+13.6 VDC connected to the junction of the brass
tubes, as if it were a center tap.  It actually isn't
a center tap in terms of core saturation, and the DC
currents to the transistors are unmitigated in terms of
magnetizing the core.  Although the cores are larger,
so are the currents, and these amplifiers just live with
the degradation including the magnetization.  This
occurs because each core sees only a half-turn.  If
you replace the tubes with a 2 turn wire primary, then
the problem goes away, but of course then the amplifier
would never work as high as 27 MHz, which is does normally
only by resonating stray PC board trace inductance with
peaking capacitors on the transformer.  This forms a
two stage step up structure.  If you "improve" the layout
to get rid of the trace inductance, the amplifier no longer
works!  See Motorola AN-762.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] How can one measure ADEV of a good oscillator?

[Richard (Rick) Karlquist]

On 11/30/2014 11:09 PM, Dr. David Kirkby (Kirkby Microwave Ltd) wrote:

I think I have a flaw in my understanding of this.

How can something like an SR620 measure the ADEV of an oscillator,  if the
oscillator is of a similar or better than the reference fed into the SR620?


What HP did with the 10811 was to make a few special crystals that
were 500 Hz off frequency and build them into oscillators.  These
oscillators were mixed with the DUT and the 500 Hz beat note was then
squared up and its ADEV measured with a frequency counter.  After
measuring a bunch of production line oscillators, they could establish
a minimum ADEV that would be attributed to the offset oscillator.  If
this level of performance wasn't good enough, other offset crystals
could be tried until a "golden" crystal was found.



> I was thinking it might be possible if one has 3 oscillators and 3 time
interval counters to perhaps solve 3 simultaneous equations. I can't prove
that, but it seems intuitively correct.


In theory this makes sense, however, it would require a high offset 
crystal and a low offset crystal to do a 3 way round robin.  There

wasn't enough need to go to the trouble of having 2 crystal designs.

There is an NBS paper written maybe 40 years ago explaining the magic
of the beat note method.

Rick Karlquist N6RK






I must be missing something!

Also I have seen graphs of both Allan variance and Allen deviation.  Both
are typically 10^-12 for a decent oscillator, but given the variance and
standard deviation are related by a square root, they can't both be around
10^-12.  I would expect to see values of 10^-6 or 10^-24, but I don't see
such dramatic differences from 10^-12.

If I see numbers around 10^-12 on an OCXO,  is that the Allen variance or
Allen Deviation?

Dave
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] How can one measure ADEV of a good oscillator?

[Richard (Rick) Karlquist]

On 12/1/2014 4:08 PM, Bob Camp wrote:



Others did a similar thing by simply taking production OCXO’s to the limit of 
their EFC range. That let you do a coarse sort to find some number of “likely” 
units. Next step was to pop  a few of them open and short this or that out to 
get a reasonable beat note. Numbers in the 10 to 20 Hz range were pretty 
common. After that a single mixer setup followed by a simple limiter got things 
good enough to screen the production lots.

Bob



I am fairly sure they would have done that if it was workable,
to avoid special crystals.  One rather huge problem with a
10 Hz beat note is that you are never going to measure ADEV
at short (<1 second) sample times that way.

Rick N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Minicircuits specs

[Richard (Rick) Karlquist]

Please Gerhard, more details on your choke
(medium size red Amidon core & two 220 uH Siemens chokes).
Maybe I can use it for 160 meter antennas.

Your T1-1 measurements make sense according to
my experience with these things.  The "-6"
series (T1-6, etc) has larger cores and should
withstand more DC.  (I like to take apart
MCL stuff to see what is inside; very enlightening).
I always use the -6 series for 160 meter work.

Thanks, 73

Rick Karlquist N6RK

On 12/7/2014 10:14 PM, Gerhard Hoffmann wrote:

Am 29.11.2014 um 20:01 schrieb Gerhard Hoffmann:



DG8SAQ VNWA, my best DC block from PSPL because it also has the
lowest lower corner. It took some experimenting to find a choke that
would not modify S21 and carry the current. Finally: a medium size
red Amidon core & two 220uH Siemens chokes with a looong air gap.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Linear voltage regulator hints...

[Richard (Rick) Karlquist]

On 12/8/2014 4:53 PM, Bert Kehren via time-nuts wrote:

We solved that problem by attaching the regulators to the FRK back plate
which is with fan control is kept within 0.01C. We did not do it for that
purpose but found that we needed some more heat in order to keep the fan in an
  optimum fan speed. Rb, OCXO and Fan power transistor are on the back
plate.  Works for us.
Bert Kehren



When I designed the E1938A, I chose a reasonable looking voltage
reference IC and put it inside the oven.  I figured that was
the end of it as far as tempco was concerned.  It was, but it
turned out that the *noise* of the reference was unacceptable
and of course the oven does nothing to fix that.  I had to
switch to a lower noise reference.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Linear voltage regulator hints...

[Richard (Rick) Karlquist]

Hittite introduced a "low noise" reference a
few years ago, but it was only low noise when
filtered with a big cap.  IOW, the cap did
all the heavy lifting and the IC was nothing
special.  Good marketing, bad engineering.

Rick

On 12/9/2014 4:30 AM, Bob Camp wrote:

Hi


On Dec 8, 2014, at 10:35 PM, Brooke Clarke  wrote:

Hi Rich:

Did you use the 723 or . . . .?
As far as I know the 723 is supposed to have low noise.


There are more modern parts with lower close in noise. Linear Technology has a 
number of them. Your filter caps take out the broadband stuff, close in is all 
that really matters.

Bob



Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Richard (Rick) Karlquist wrote:



On 12/8/2014 4:53 PM, Bert Kehren via time-nuts wrote:

We solved that problem by attaching the regulators to the FRK back plate
which is with fan control is kept within 0.01C. We did not do it for that
purpose but found that we needed some more heat in order to keep the fan in an
  optimum fan speed. Rb, OCXO and Fan power transistor are on the back
plate.  Works for us.
Bert Kehren



When I designed the E1938A, I chose a reasonable looking voltage
reference IC and put it inside the oven.  I figured that was
the end of it as far as tempco was concerned.  It was, but it
turned out that the *noise* of the reference was unacceptable
and of course the oven does nothing to fix that.  I had to
switch to a lower noise reference.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Linear voltage regulator hints...

[Richard (Rick) Karlquist]

I don't remember, but it wasn't a 723.

Rick

On 12/8/2014 7:35 PM, Brooke Clarke wrote:

Hi Rich:

Did you use the 723 or . . . .?
As far as I know the 723 is supposed to have low noise.

Mail_Attachment --
Have Fun,

Brooke Clarke
http://www.PRC68.com
http://www.end2partygovernment.com/2012Issues.html
http://www.prc68.com/I/DietNutrition.html
Richard (Rick) Karlquist wrote:



On 12/8/2014 4:53 PM, Bert Kehren via time-nuts wrote:

We solved that problem by attaching the regulators to the FRK back plate
which is with fan control is kept within 0.01C. We did not do it for
that
purpose but found that we needed some more heat in order to keep the
fan in an
  optimum fan speed. Rb, OCXO and Fan power transistor are on the back
plate.  Works for us.
Bert Kehren



When I designed the E1938A, I chose a reasonable looking voltage
reference IC and put it inside the oven.  I figured that was
the end of it as far as tempco was concerned.  It was, but it
turned out that the *noise* of the reference was unacceptable
and of course the oven does nothing to fix that.  I had to
switch to a lower noise reference.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Linear voltage regulator hints...

[Richard (Rick) Karlquist]

On 12/9/2014 1:30 PM, ed breya wrote:


buried zener for lowest noise - this eliminates all the low voltage
references and three-terminal etc regulators that use band-gap
references. The down side is that the good kind of reference ICs will
need a higher (like 10V and up) operating voltage than may be available,
so that complicates it.


Great post, Ed.  I might add that my understanding of band gap
regulators is that they rely on amplifying a small DC difference
in voltage between two transistors.  This also amplifies the SUM
of the noise of the respective inputs, which jacks up the noise to
much more than a good zener.  Because of physics, no band gap
reference will ever be low noise.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Linear voltage regulator hints...

[Richard (Rick) Karlquist]

On 12/9/2014 3:02 PM, Bob Camp wrote:
 Here's an overnight ADEV plot against the new Cs of where we are in 
the project.  Red is ADEV.  Green is the TIC.  Blue is the output of the 
GPSDO to Channel A and the Cs to Channel B of my 5335A measuring TI, 
using the 1PPS from my GPSDO to trigger the external gate.  I see that 
the Cs phase has drifted down slightly vs my GPSDOe.  I wonder if this 
is an indication of a calibration problem with the Cs or some subtle 
issue on my board?


Looks perfectly normal to me. Your Cs is sensitive to magnetic field. Have you 
“zeroed” it out? No of course not, nobody does. (almost nobody …).

Bob


This reminds me of the great CBT demagnetizer debate at HP.
Even Len Cutler didn't think this was necessary, at least
in the 5071, and possibly the 5061.  When Len says something
is overkill, you can be sure of it.  Anyway, we still had
to support a CBT demagnetizer for customers who wanted it.
The customer (with money to spend) is always right.  The
5071 measures the C-field with a Zeeman line measurement
and adjusts it if necessary.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 100 MHz VCXOs

[Richard (Rick) Karlquist]

On 12/11/2014 4:20 PM, Bob Camp wrote:

Hi

Those OCXO’s were made to the spec’s of an OEM customer. The spec’s are owned 
by that customer and can not be released without authorization from them. 
Anybody who wants to stay in the business would be a bit crazy to release 
somebody else’s intelectual property to the public. That’s not an uncommon 
issue with surplus OCXO”s. Rouglly 99.9% of them are built to customer spec.

Bob



The reminds me of when I worked for HP.
We had a division that made diodes, which
all had a part number of the form 5082-.
There was a special series of part numbers
of the form 5082-5XXX that never appeared
in the catalog.  I remember I was working
on a step recovery diode multiplier and was
trying to reverse engineer various multipliers
that were designed into commercial equipment.
I got the HP part number of the diode
which was of the form 5082-5XXX from the
manual and found out that there was no data
available on these part numbers in the HP
internal system.  I called the Microwave
Semiconductor Division and got the story
Bob gave about OEM spec's.  I said, sure I
understand that, but I work for HP.  I was
told that you have to work for MSD, not just
HP to get access to this information, and
in any event, they would NOT sell us these
diodes, even as an internal transfer.

Rick Karlquist N6RK
HP 1979-1999 (then Agilent)
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] New Years Eve TV countdown

[Richard (Rick) Karlquist]

We have DirecTV with some receivers standard
definition and others High Definition.  The
delay is considerably greater on the HD version.
Even OTA HD is delayed considerably, as noted
if you try to listen to a football game on the
radio while watching.  Sometimes you hear "touchdown"
before the ball is even snapped on TV.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 3048A Expert Needed

[Richard (Rick) Karlquist]

On 1/20/2015 7:59 AM, Martyn Smith wrote:

Hello,

I have both the HP3048A and Timepod Phase Noise test Sets.

I have two low noise 100 MHz PLL's.  The 100 MHz output is locked to a
10 MHz reference input.




However, the 3048A won't achieve lock.  The 100 MHz PLL's loop bandwidth
(about 0.2 Hz) is too slow for the 3048A to achieve lock.

Does anyone have any ideas?




Martyn


I believe this is not specifically related to anything going
on in the 3048, but rather that you can't lock together the
two 100 MHz oscillators you have in general.  Whenever I have
run into that, the fix is to simply divide down both of the
100 MHz outputs by the same number N.  Now feed those into an
external phase detector and make a phase locked loop independent
of the 3048.  Then measure the noise using the 3048 in residual
mode.  The purpose of dividing down is to reduce the phase
fluctuations to a level that is within the linear range of the
phase detector.

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5>10 doubler

[Richard (Rick) Karlquist]

On 1/28/2015 11:28 AM, Charles Steinmetz wrote:

Gerhard wrote:


It is a different game when you want to notch away sub/harmonics.


One problem with using crystals as traps (notch filters) is that the
series resistance of a crystal is several orders of magnitude higher
than that of a good series-resonant LC -- generally in the 50-100 ohm
range.  So, although the notch is very narrow, it will not be very deep
unless it is in a high-impedance circuit.  For example, in a 50 ohm


It is very straightforward to use LC networks to transform the
impedance of the crystal to a much lower value and get around this
problem.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5>10 doubler

[Richard (Rick) Karlquist]

On 1/29/2015 5:41 PM, Alexander Pummer wrote:

And the narrow notch for the harmonic is not required anyway, since the
fundamental is fare enough, therefore a high Q  LC trap will work
better, also with the setting of the biasing af the active devices the

Alex KJ6UHN


When I designed the 5071A RF chain, I used five cascaded frequency
doublers to go from 10 MHz to 320 MHz.  I definitely used traps
to reduce the 10, 30, and 40 MHz spurs (using 10->20 MHz as an
example).  It was no easy thing because I could only use coils of 
moderate Q (less than 50) and I needed at least 80 dB suppression.  You 
might wonder why I needed to reduce 40 MHz spurs in the 20 MHz

output.  It turns out (little known fact) that the if I drove
the 20->40 MHz doubler with 20 MHz contaminated with 40 MHz
harmonics, it would degrade the spectral purity of the 40 MHz
output.  Strange but true.

The 5071 filters are basically cascaded notch filters, as
opposed to band pass filters.  Doing this allowed me to
have zero adjustments.  Previous atomic clocks used narrow
high-Q filters that had to be tuned up, and were temperature
sensitive.  The production engineers had to constantly stay
on top of these filters because they were so temperamental.
OTOH, the 5071 filters just work.  There was never even a
production change to them AFAIK.  The key to getting the notch
filters to work was to use 2% components, and use two coils
and or two capacitors together to get around the fact that
the standard values are quantized to 10%.  Additionally, I
measured each tank circuit in situ on the PC board and tweaked
it to take into account the de facto board parasitics.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Usefulness of high end counters for ADEV plots of oscillators

[Richard (Rick) Karlquist]

First of all, the oven oscillator option of the 53230
is no where near as stable in ADEV as a 10811 for example.
The counter itself is 1 or 2 orders of magnitude better
than the built in timebase.  So don't waste your money
on the OCXO option when you, as a time nut, undoubtably
already own something like a 10811.

When I was still with Agilent, I made innumerable measurements
of ADEV with a 53230 down into the low parts in 10^11, which
was the DUT ADEV, not limited by the counter.  I vaguely
remember measuring a 10811 as a sanity check and using
the internal OCXO (not knowing any better).  After wasting
a lot of time, I eventually measured one 10811 against
another and discovered that the ADEV floor was down to
1E-12 at least, and that the internal OCXO was junk.

About the only good thing about the 53230 is that it
is a self contained box that makes ADEV measurements and
displays them in real time without requiring an external
computer with software.  In 1974, HP made a computing
counter (5360 I think) that did this.  Customers loved this
box, but the HP engineers hated the box.  Therefore, no
HP/Agilent counter ever did ADEV again, until the product
line was offshored to engineers who didn't know any better
and put the feature back in.

For serious ADEV measurements, you want two DUT's offset by
a few 100 Hz and mixed, as originated in the HP 5490 system.

Rick Karlquist N6RK

On 1/30/2015 8:38 AM, James via time-nuts wrote:


  Though I'm new to the list I've lurked for quite a while


 and from reading various posts I am in a slightly confused

 state as to whether buying an expensive counter (eg Keysight 53230A

 or a Tek fca 3100) will be useful as a measurement tool for

 developing a GPSDO.



Given a one shot measurement resolution of 50 psecs (on the Tek which is a 
pendulum CNT91) means that the uncertainty is around 50E-12 at 1 sec or 5 x 
10^-11 or 10^-10 in round numbers for a ADEV at 1 sec? For this noise floor to 
get well below 10^-11 (the sort of ADEV of an OCXO) requires the interval to be 
increased to nearer to 100 seconds?

So does this mean that an expensive counter allows useful ADEV plots from 100 
seconds on but not the lower time frames? (By useful I mean able to measure 
down to and below 10^-11 not down to 10^-14!)

The extra cost of the 53230A over the 3100 gets down to 20 psecs so possibly 
reduces the period to a bit less than 100 seconds but still above 10 seconds 
probably?

Does paying extra for an OCXO gain significantly on this basis?

Have I got the basic numbers right, and if one of my main aims is to have a 
good instrument for playing with GPSDO development will investing in such an 
expensive (for an individual hobbyist) instrument buy me a useful measurement 
capability or would it just be good for measuring long term frequency and say 
1pps jitter from the GPS?

Sorry for the long post,

James







___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 10544A vs. 10811

[Richard (Rick) Karlquist]

The other big difference is that the 10811 uses
an SC cut crystal instead of an AT cut crystal.
From a cold start, the SC achieves a given stability
much faster than an AT cut.  If you are just going
to run the oven continuously (likely mode for time
nuts), this isn't any big deal to you.

The reason why the phase noise is better is not
so much due to the SC cut crystal, but rather to
the grounded base output buffer in the 10811.

A 10811 is guaranteed to work in any 10544 socket
in HP equipment.  They had to do this so that they
could stop making 10544's as soon as the 10811 was
introduced.

Rick Karlquist N6RK

On 2/1/2015 12:54 AM, Bruce Griffiths wrote:

Since the 10544A uses a PWM oven controller there are significant oven PWM 
frequency related sidebands.The PN noise floor of the 10811A (-160dBc/Hz) is 
significantly lower than that of the 10544A (-145dBc/Hz).
Bruce


  On Sunday, 1 February 2015 2:57 PM, Bill  wrote:


  I have a choice. Can I assume the 10811 is the better OCXO for phase
noise and ADEV compared with the 10544A?

Thanks and regards...Bill
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 10544A vs. 10811

[Richard (Rick) Karlquist]

On 2/1/2015 2:31 PM, Magnus Danielson wrote:

As I recall it, the 10544 is a BT-cut and not AT-cut.
We discussed this a few years back, and even checked the cold
temperature before heating up, and it matched BT-cut and not AT-cut.

Anyway, go with the 10811.

Cheers,
Magnus

On 02/01/2015 07:51 PM, Richard (Rick) Karlquist wrote:

The other big difference is that the 10811 uses
an SC cut crystal instead of an AT cut crystal.


I stand corrected.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] OXCO insulation

[Richard (Rick) Karlquist]

On 2/25/2015 4:32 AM, Attila Kinali wrote:

On Wed, 25 Feb 2015 10:16:58 +0100
Attila Kinali  wrote:


Actually, you should put the temp sensor close to the heater, not the crystal.
The delay between the actuator (heater) and the feedback (temperature sensor)
defines the dead time. The presence of a crystal somewhere in the system is
of no importance to the oven controller, it's just additional thermal weight.


Bob Camp just reminded me of Rick Karlquist's papers on this topic:

"A low Profile High-Performance Crystal Oscillator for Timekeeping
Applications", by Karlquist, Cutler, Ingman, Johnson, Parisek, 1997
http://www.karlquist.com/osc.pdf

"The Theory of Zero-Gradient Crystal Ovens",
by Karlquist, Cutler, Ingman, Johnson, Parisek, 1997
http://www.karlquist.com/oven.pdf

More can be found on his homepage: http://www.karlquist.com/



A quick summary of the insulation issue in OCXO's:

Virtually any kind of foam, or just still air has similar
insulating properties in an OCXO.  Exotic kinds of insulation
like aerogel, or vacuum, are impractical because you have to
get rid of the heat generated by the electronics.  The E1938A
oscillator described in the above references achieved a thermal
gain in excess of 1,000,000 (in a single stage oven) in prototype 
testing, and routinely achieved many 100's or thousands on a production 
basis and had only 1/4 inch of foam that was nothing special thermally.

We also were able to get high thermal gain with a still air
experiment.

At HP, the main issues with foam were outgassing and mechanical
fatigue, like a mattress getting flat and the foam had to withstand
the oven temperature continuously.  Another issue was what
happens if the oven runs away and the foam burns and produces
toxic gases.  Generally, we are talking manufacturability issues.
The one thing the choice of foam had little to do with was performance.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Obscure HP T/F instruments in ebay.fr

[Richard (Rick) Karlquist]

Before the Keysight split, there was an Agilent
museum at HQ in Santa Clara.  It was packed full
of interesting old HP stuff and even had a part
time archivist.  I'm now retired and don't know
what became of this museum in the split.
I feel I got out while the getting was good.

Rick Karlquist N6RK
HP/Agilent 1979-2014

On 3/19/2015 9:01 PM, Tom Van Baak wrote:

If that is the case, then this stuff belongs to a museum and not on ebay. IMHO.


Hi Attila ,

I completely understand how you feel, but this happens all the time with niche 
collections. You just can't find a brick and mortar museum interested in taking 
all that inventory. How many people would travel to city X in country Y to see 
a collection of electronics made by company Z? So these collections tend to 
last only as long as the original pioneer behind them is active. Once they are 
gone, there's a good chance that it all ends up on eBay, scattered around the 
globe. At least it doesn't end up in recycling or the trash.

Checking current vs. completed auctions for that seller, you'll note that a 
large number of the good or exotic items have already been sold. I noted that 
high value items like hp rubidium and cesium standards apparently never made it 
to eBay, suggesting some cherry picking occurred before the collection went out 
for bid.

I once thought "HP should have their own museum". But then they split into Agilent, 
then Symmetricom bought out their T&F line, then they became Keysight, then Symmetricom 
became Microsemi. With these companies, there isn't strong technical, moral, or business 
justification to allocate office space and resources to host dusty museums that might only 
attract tens or hundreds of people a year. They are rightly focused on current and future 
products, leaving us bottom feeders and nostalgic historians to collect and display the old 
stuff in our own homes, or on the web.

For me the greatest museum loss occurred when "The Time Museum" in Rockford, IL 
closed in 1999. This was the best collection of clocks in the world, 1500 pieces from an 
ancient Egyptian water clock to a vintage hydrogen maser and everything in between. But 
the heirs of the founder were not into Time or into Museums. So it went to a massive 
international auction (Sotheby's) and was scattered for all of time.

/tvb

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/11/2015 5:01 AM, J. L. Trantham wrote:
Unless the design has been changed, the 10811 option
for the 53132 has poor short term stability and
degrades the performance of the 10811 by something
like an order of magnitude.  I complained about
this when the counter first came out 25 years
ago but no one would listen.  At the time I had
recently transferred out of counter R&D to work
on the 5071A.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/12/2015 1:03 AM, Adrian Godwin wrote:

What caused that degradation ? I'm interested in dos/don'ts for best use of
a 10811.



I don't remember the details much after 25 years, but
basically they have a distribution amplifier that
allows for internal or external 10 MHz and what I
remember is that I looked at the schematic and
concluded that no one with a background in precision
time and frequency would design it that way.  And
it turned out that the person who designed it did
not have any such background.  I vaguely require
some measurement that had disappointing results that
caused me to want to look at the schematic of it
in the first place.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/12/2015 2:22 PM, Magnus Danielson wrote:

Hi,

The buffer transistors has not AC-bypass of the emitter resistance, so
that the DC current becomes large and thus contributes flicker noise.

The comparator at the bottom isn't doing a beutifull work of squaring
things up without contributing noise, considering the sine output of the
10811.

Was that it, Rick?

Cheers,
Magnus



The resolution of page 13 is poor, and it seems to be a bitmap instead
of a vector file.  The fuzzy thing in the lower right corner looks
like it might be a comparator.  I think this was the smoking gun.

There was a saying by H.L. Menken to the effect that for every
complex problem, there is a simple, obvious, invalid solution.

Squaring up a 10811 with a comparator is a perfect example of this
principle.  Non-time-nuts always seem to gravitate to this design.

Of course you're right, any comparator will add jitter to a 10811.
The faster they are, the more jitter they add.

I noticed that the standard 10 MHz oscillator is built with
an ECL line receiver.  Another example of Menken's saying.
This is a TERRIBLE oscillator design, but one that would appeal
to the non-initiated.  I built one of these oscillators in 1976
at the suggestion of my boss.  After seeing how bad it was, I
quietly designed it out and never used it again.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Distribution Amps

[Richard (Rick) Karlquist]

On 4/12/2015 3:04 PM, Perry Sandeen via time-nuts wrote:

List,



For me it was simpler to buy asurplus HP 5087A for best offer which turned out 
to be $300 delivered.



The 5087 series is ancient technology that has mediocre performance.
I remember looking at the circuit designs in the series in case
I wanted to leverage them for the 5071A.  There was nothing
that would be of interest for the 5071A.  If the 5087 meets
your needs, then fine.  But many time nuts need something
better.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 5 to 10 MHz doublers

[Richard (Rick) Karlquist]

The 5071A doublers I designed use MCL ASK-1 mixers.
The LO and RF ports are connected in series.  This
arrangement is self limiting.  So you drive them
fairly hard and the output is level.  The IF port needs
to see a DC short circuit of course.  This was
essential in the 5071A since there were 5 doublers
in cascade.   The input impedance is about 30 ohms
and you need a fair amount of drive, like 30 mA peak
to peak. I tested and rejected other approaches like
putting the ports in parallel or driving them with
a quadrature hybrid.  The series connection just
worked better. It's going on 25 years, but this is what I remember.

The filtering is another non-trivial story.  Buying
DBM's doesn't make that job any easier.

Rick Karlquist N6RK

On 4/12/2015 3:10 PM, Perry Sandeen via time-nuts wrote:

List,



There has been a lot ofinformation on the list for 5 MHz frequency doublers.



As it is easier for me to buythan make, would using something like a 
Mini-Circuits surplus DBM’s be anacceptable way to go?



Regards,



Perrier

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/12/2015 6:30 PM, Attila Kinali wrote:


It might be that I'm already too sleepy, but I don't see why
a faster comparator would add more jitter. Actually, my intuition
(which is clearly wrong) would say the contrary. So, which effect
does increase the jitter with comparator speed?


The faster the comparator, the greater its analog bandwidth.
Thus there is more total noise to cause jitter.  The DC to
daylight comparator is the opposite of the John Dick (JPL)
paper on zero crossing detectors in PTTI around 1990.  John
teaches that you use the MINIMUM bandwidth amplifier to
square up a sine wave.


BTW: If anyone here has any good text to read on oscillator design,
please let me know. I'm collecting those :-)

Attila Kinali


Start by reading everything by:

Marv Frerking
Mike Driscoll
John Vig

Oh, and I wrote a few papers on oscillators myself :-)

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/13/2015 3:11 AM, John Miles wrote:


A comparator with less open-loop gain was what they needed.  Somebody at HP 
really liked ECL line receivers, though.  Those were very noisy at HF, but this 
had little or nothing to do with their bandwidth (see my other post.)

To square up a 10 MHz signal from an OCXO it's hard to beat a simple diff amp 
with a pair of bipolars, a la Wenzel.

-- john, KE5FX
Miles Design LLC



The HP counters all used ECL line receivers for the A and B
channel to convert the input frequency signal to be counted
into a digital square wave.  Naive engineers then ape'd this
for use on the timebase clock.

Engineers at HP who actually knew what they were doing,
such as Tom Falkner, did use differential pairs.  However,
HP being a huge company, the word did not necessarily get
disseminated to all the other HP engineers.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Ultra High Stability Time Base Options for 53132A

[Richard (Rick) Karlquist]

On 4/13/2015 12:14 PM, Magnus Danielson wrote:


Oh yes. Some people say that you should not overcomplex things. My
experience is that oversimplifying them can cause a long stretch of
complex problems and complex workarounds making the total solution more
expensive in development, customer relations and more complex than
starting with a more advanced solution, that actually attempts to
address the design issues. Ah well.



This is extremely good advice.  The ultimate example of the
oversimplified design is the Muntz TV set, where few parts are
used, but they all interact with each other in mysterious
ways that depend on unknown unspecified parameters.  The ultimate 
example of the overcomplicated design is the Japanese VCR, circa 1980.

Schematic looks like it was designed by committee.  The parts
count has become bloated to the point of redundancy.  Neither
is desirable.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] New +/- 1 sec in 100 days mech clock

[Richard (Rick) Karlquist]

Perhaps it is not a good analogy, but I think of
the cesium beam tube in the 5071A.  The plans
alone are very non-trivial.  Then there are
a bunch of proprietary machining details that
I can't disclose, that are way beyond the
merely having access to a CNC tool.  The
systematic error due to the CBT is below
something like 1 part in 10^14, which is
1 second in 3 million years.  Perhaps that
is in some sense equivalent to Harrison's
1 second in 0.3 years.  7 orders of magnitude
difference.  As many time-nuts are probably
aware of, most if not all cesium clocks that are better
than the 5071A have reversible beams to
cancel out CBT assymmetry.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP5328A & HP5328B option 040

[Richard (Rick) Karlquist]

On 4/23/2015 7:17 AM, Alberto di Bene wrote:

On 4/23/2015 9:20 AM, VK2DAP wrote:


/1) Generally speaking, would it be correct to say that when a product
model number changes from A to B,//
//that represents an improvement or major update to a product?/


I remember having read (don't recall where...) that the HP5328B was a
cost-reduction step wrt the A model...


Oh, yes, now I remember.  The 5328B, 5334B, and one other counter
(maybe the 5335) were part of a cost reduction program that was
internally called the "killer B's".  I was the project manager
for the 5334B (see my previous posting).

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP5328A & HP5328B option 040

[Richard (Rick) Karlquist]

On 4/23/2015 12:20 AM, VK2DAP wrote:

Dear time-nuts,

I have a question about the HP5328A and HP5328B universal counters.

1) Generally speaking, would it be correct to say that when a product model 
number changes from A to B, that represents an improvement or major update to a 
product?

2) I am interested in the "delay" option that is mentioned in the user manual 
for the HP5328A (option 040). My question is simple. Why does this option not feature in 
the HP5328B, but only as option 040 on the HP5328A?




Now you have asked a very interesting question.  Since you
are new to time-nuts, you probably don't know I worked
for HP/Agilent/Keysight for 35 years.  It would be
a gross oversimplification to assume that an A/B change
is an improvement, although in some cases that may
be true.  Often it has more to do with certain parts
becoming unavailable.  You should also know that there
is typically a 5 year support life after the product
goes out of production.  It is very common that they
will increment the suffix to get the 5 year clock
running so they no longer have to support very old
instruments.  This was certainly the case with the
5061B cesium standard.  The nixie displays were
unobtainium and we couldn't support the 5061A because
of this.  I don't know specifically about the 5328A
vs B.  However, I was the project manager for the
5334B counter.  The way that came about was that I
just happened to notice that there were various design
aspects of the 5334A that wasted a lot of money.  I
didn't work in the counters section at the time, but
nevertheless I annoyed the R&D manager by pointing
out these money leaks.  I guess he got tired of hearing
me complain and one day he offered my the job of project
manager on a 5334B model.  We needed to reduce cost
because we were losing military contracts to Racal-Dana.
I changed certain design details in the B model where
I could save money.  The idea was to simply keep the
performance the same and not add features.  There were
many things I inherited from the A model that I left
alone if I couldn't reduce the cost.  We were also on
a very tight schedule.  This prevented me from replacing
the 4 separate microprocessors in the 34A with a single
one.  (Very long story as to why this was)

Now to get to your question about why a feature in the
A version would not carry through to the B version.
The 5334A had an option of a digital voltmeter, which was
put in essentially "because we could", but then justified
after the fact by claiming that customers wanted it
because some of them ordered the option.  I never thought
this made sense and used my authority as 5334B project manager
to get rid of it in the 5334B.  Since HP also sold
voltmeters, they could always buy a voltmeter from
our voltmeter division.  There are a bunch of reasons
why the 5328B would lose a feature that the 5328A had,
but the point is that this doesn't break any "rule".

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

[time-nuts] OT: PDF editing (was: Re: Ultra High Stability Time Base Options for 53132A)

[Richard (Rick) Karlquist]

On 5/4/2015 5:25 PM, J. L. Trantham wrote:

Charles,

Thanks for the help.  I need to learn how to add text to .PDF documents.



Go to:

www.tracker-software.com

and download (for free):

PDF-Xchange viewer.



Rick Karlquist N6RK


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

On 4/26/2015 3:51 AM, Bryan _ wrote:

All:

P
 I was looking at the project from David partridges web site 
http://www.perdrix.co.uk/FrequencyDivider/index.html



-=Bryan=-   
___



This is a comparator based circuit.  This will give
you worse performance than just about anything else,
but it may be good enough anyway.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

On 5/6/2015 3:24 PM, Bob Camp wrote:

Hi

A standard input on a frequency counter is not a very demanding thing in the 
hierarchy of
TimeNut signals. You can drive any of them with some pretty simple logic gate 
based
circuits. No need to spend a lot of money.

Bob




Logic gate, yes.  Comparator, no.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

On 5/8/2015 2:19 PM, Bob Camp wrote:


On May 7, 2015, at 11:09 AM, Attila Kinali  wrote:

On Wed, 06 May 2015 18:09:03 -0700
"Richard (Rick) Karlquist"  wrote:


A standard input on a frequency counter is not a very demanding thing in the 
hierarchy of
TimeNut signals. You can drive any of them with some pretty simple logic gate 
based
circuits. No need to spend a lot of money.


Logic gate, yes.  Comparator, no.


This reminds me a lot of a similar discussion a couple of weeks ago.
(where the issue boiled down to noise bandwidth)

What is the problem with a comparator vs a logic gate?
What makes the logic gate supperior?

Attila Kinali



The comparator as a squarer circuit is folklore that unsophisticated
users want to believe in, because it is seemingly the easiest way
to get the job done.  Wouldn't it be wonderful to be able to put in
any signal from -30 dBm to +15 dBm and get a perfect square wave
out with no effort?  Unfortunately, what a comparator looks like
is a very high gain differential amplifier that is slew rate limited.
The threshold voltage input must be extremely low noise or it
will introduce jitter.  Even if the input pin is clean, there is
internal noise.  Driving it will a low level signal will produce
a jittery output for obvious reasons.  The trouble is that if you
drive it with a high level signal, the jitter doesn't go away because
the input stage is already in saturation.  Also, the effective noise
figure of the comparator is usually high.  Making the comparator
faster exacerbates the problem.  Read papers on "zero crossing 
detectors" such as John Dick's 1990 paper in PTTI and you will

see that a comparator is the exact opposite architecture from
the optimum one.  I hope that clears up the question.

Regarding logic gates:  it is not so much that there is something
magic about gates; actually ECL gates are lousy.  It is just that
comparators are so bad that almost anything else is better.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

The only gates that seem to do very well are high speed (as in 74AC or faster)
silicon CMOS. You need to run them with a fairly clean supply and feed them
with a p-p input that matches the supply voltage. Other than that, not a lot
of magic. Are they ideal - surely not. Will they hit 2x10^-13 ADEV at 1 second 
and
drop from there as tau increases - yes they will.

Bob



Interesting.  Back in the stone age in 1990, 74AC was state of the art.
As soon as it came out, I used it exclusively for everything that didn't
require ECL, for which I used ECLinPS exclusively.

In the 5071A, we used one to square up 80 MHz for the DDS board.  You
are exactly right:  put in a huge sinewave obtained by good old
fashioned analog anplifiers and let the 74AC do its thing.  80 MHz
is pretty much flat out for a 74AC series. The 80 MHz came from a
10 to 80 MHz multiplier running from the 10811.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

On 5/20/2015 11:22 AM, Magnus Danielson wrote:



The older HP counter manuals explained it very nicely too, as they
illustrated the slew-rate & amplitude noise to time-noise conversion.

What do amazes me is the fact that I've yet to see a counter input
channel which takes care to square up the signal properly, they rather
provide the comparator after the obvious damping and AC-blocking
conditioning. I can't even recall that there where much such shaping as
a side-product.


The counter front ends seem to be modeled after scope front ends
and scope triggering circuits, where you can adjust the triggering
level.  Any jitter in the triggering would normally only affect
the interpolator.  The interpolators in general were no great shakes,
so the triggering wasn't the limiting factor.



Now, remind me why ECL is lousy, I can't recall there being very high
gain in them, but fairly high bandwidth and they stay in the linear
operation region.




Magnus
___


ECL is bad because the voltage swing is low; because as you say,
a lot of the circuitry is in the active region all the time, and
because the current source in the emitters generates a lot of
noise.

In the early 1990's, I thought I had proved that the high ECL
noise was mostly common mode and that you could reduce it
20 dB by using a transformer to couple the output.  Alternately,
a good differential amplifier with high CMRR would do the trick.
I had actual measurements to back up this theory.

Subsequently, other people tried to reproduce this and could not.
By that time, I had moved on and didn't have the bandwidth to
continue to own the problem.

It would make a nice project for some time-nut to prove or disprove
my hypothesis regarding ECL.

ECL line receivers as squarers are not as bad as comparators, but
are much noisier than 74AC.

Rick Karlquist N6RK

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Using CPLD/FPGA or similar for frequency divider

[Richard (Rick) Karlquist]

I used a CPLD in a 900 GHz (that's right 900 GHz) optical
sampling scope timebase.  It was great because you just
write a 17 bit counter in VHDL and there it is.  You
don't have to know anything about building digital
hardware any more (40 years of experience wasted).
Nobody cares about look ahead carry, etc.
I cleaned up the timing with conventional logic, so
I don't know what the jitter of the CPLD was.
We needed jitter in the low fs, so I am sure
the CPLD was not OK without cleaning up, but
then that was a lunatic fringe project.

Rick Karlquist N6RK

On 6/2/2015 6:13 AM, David C. Partridge wrote:

Is this a sensible thing to consider doing?  Or would I be better sticking to 
AC/HC/AHC/LVC logic?

Regards,
David Partridge

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Nature: Hyper-precise atomic clocks face off to redefine time

[Richard (Rick) Karlquist]

Can someone explain to me how this is going to work in
light of the fact that each clock is in a different
gravitational field?  Or is accuracy not the measurement,
but rather stability?  No, that can't be because any
lab that wants to measure stability merely needs to build
two or three copies of their favorite clock and insure
against synchronization.  They in principle shouldn't
need to compare against a dissimilar type of clock.
Therefore, we are back to the gravity issue.

When we worked on the 5071A, we barely had enough sensitivity
to notice a few parts in 10^13 between Santa Clara and
Boulder (~5000 feet).

Rick Karlquist N6RK

On 6/3/2015 12:18 AM, Hal Murray wrote:

Nice picture: A strontium-ion optical clock housed at the National Physical
Laboratory in Teddington, UK.

Over the past decade, various laboratories have created prototype optical
atomic clocks, which use different elements such as strontium and ytterbium
that emit and absorb higher-frequency photons in the visible spectrum. This
finer slicing of time should, in principle, make them more accurate: it is
claimed that the best of these clocks gain or lose no more than one second
every 15 billion years (1E18 seconds) -- longer than the current age of the
Universe -- making them 100 times more precise than their caesium
counterparts. Optical clocks are claimed to be the best timekeepers in
existence, but the only way to verify this in practice is to compare
different models against each other and see whether they agree.

Starting on 4 June, four European laboratories will kick off this testing
process -- the National Physical Laboratory (NPL) in Teddington, UK; the
department of Time-Space Reference Systems at the Paris Observatory; the
German National Metrology Institute (PTB) in Braunschweig, Germany; and
Italy's National Institute of Metrology Research in Turin. Between them, the
labs host a variety of optical clocks that harness different elements in
different experimental set-ups.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Using CPLD/FPGA or similar for frequency divider

[Richard (Rick) Karlquist]

The counter only had to run at ~50 MHz, on account of our
mode locked laser ran at that frequency.  I don't remember
what the CPLD was rated at.

Rick

On 6/5/2015 8:19 PM, Hal Murray wrote:


rich...@karlquist.com said:

I used a CPLD in a 900 GHz (that's right 900 GHz) optical sampling scope
timebase.  It was great because you just write a 17 bit counter in VHDL and
there it is.  You don't have to know anything about building digital
hardware any more (40 years of experience wasted). Nobody cares about look
ahead carry, etc.


Is that really true?  Or perhaps, what fraction of the digital design space
does it apply to?

How fast was your counter running?  How fast would it run?  Was it a simple
counter or was there enable/up/down/load type gating involved?

What would you have done if you needed to run a bit faster?  Could you buy a
faster chip?  How much more could you get with tricky logic?

I agree that modern tools and parts have allowed a lot more people to build
digital circuits.




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Modulation Domain Analysis

[Richard (Rick) Karlquist]

That's interesting.  I worked for the HP Santa Clara Division
from 1979 until just before it was closed in 1998.  I
forget who "invented" MDA at SCD, but it was hyped like
it was some new concept and I never heard anything about
the HP9540.

Many times someone would come to me and ask me about
some "new" bright idea they had, and I would tell them
"Yes, I can confirm that your idea is excellent, because
I read the original paper on it that was published in
19XX."  It is interesting that people would often get
mad at me, as if it is my fault they reinvented the wheel.

If only I known about your HP Journal article, I could
have throw it up to the "innovators" at SCD.

Before I worked for HP, an HP Journal article came out
about fractional-N synthesizers, and everyone at Zeta
Labs was anxious to use the technology in the Zeta
Labs designs.  Except one guy, who pointed out that
he had invented frac-N 11 years previously, and he
called it "digiphase."  I've never heard anyone at
HP ever acknowledge that guy.

Rick (now retired from HP/Agilent/Keysight)

On 6/16/2015 12:54 PM, Robert Gilchrist Huenemann wrote:

I stumbled onto the time nuts list from a posting on modulation domain analysis 
a couple of weeks ago. I am enjoying the discussion.

I want to comment on modulation domain analysis, or phase digitizing. This is a 
technique that uses a period mode frequency counter, or two such counters back 
to back, to recover the modulation history of a frequency modulated waveform.

This technique was first used in the HP9540 automated transceiver test system. 
This system was described in the August 1973 HP Journal. The HP9540 used a 
single HP5326 period mode counter with a 10 MHz clock. At that time, no counter 
was available with a higher clock frequency.

A breadboard system was assembled as part of the HP9540 development effort 
which used two HP5326 counters back to back. To insure that alternate periods 
were measured, the second HP5326 ran off the gate output of the first. However, 
it was realized that the characteristics of the HP9540 and its specific 
application were such that two counters were not required. Please refer to my 
HP Journal article for details.

The HP9540 was developed at HP's Automatic Measurement Division. This division 
was disbanded in 1974.

Modulation Domain Analysis and Phase Digitizing were terms that came into use 
with the later development of specialized stand alone instruments that combined 
computational capability, back to back period mode counters, higher clock 
frequencies, interpolation and algorithms for various measurements. All of 
these were worthwhile improvements on the basic technique first used in the 
HP9540.

I would be happy to answer questions. Thank you for allowing me to post this 
information.


Robert Gilchrist Huenemann, M.S.E.E.
120 Harbern Way
Hollister, CA 95023-9708
831-635-0786
bo...@razzolink.com
https://sites.google.com/site/bobhuenemann/
Extra Class Amateur Radio License W6RFW
IEEE Life Member 01189471

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/17/2015 8:22 AM, Jim Lux wrote:

I'm looking for some representative data for inexpensive microwave VCOs
(in the 2.5-6 GHz range, in general).  Not in a locked loop situation,


If the phase noise data you have goes to a low enough frequency to
get below the 1/f corner (which is the case for the example you cited)
then it is a very safe bet that the noise will go up by 30 dB/decade
below that.

Having said that, if an ordinary engineer had asked me this question,
I would think that he needed some coaching on how to clean up the
VCO with a synthesizer of sufficiently wide loop bandwidth.  However,
you are very knowledgeable, so I will assume you are going to do
that and just want to predict the phase noise after clean up.  The
trick (as most time nuts know) is to use a small enough capacitor
in the loop filter so that you get clean up at a 40 dB/decade rate
so you can actually make some headway against the 30 dB/decade
1/f slope.

I have been through this exercise innumerable times and also taught
it to many others, and it seems to be very predictable.

In the unlikely event you use the VCO open loop, you'll have lots
of problems with microphonics, power supply noise, and even magnetic
fields from power transformers, as well as load pulling and thermal
drift.  Making microwave oscillators that can be used open loop 
(especially inexpensive ones) is definitely a lost art.  It died with 
the HP8640 sig gen.


Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/17/2015 8:22 AM, Jim Lux wrote:

I'm looking for some representative data for inexpensive microwave VCOs
(in the 2.5-6 GHz range, in general).  Not in a locked loop situation,



If you are working up to 2.5 GHz, you can get a low power
chip for $2 from Analog Devices that has a VCO and synthesizer.
For about $15, you can get a 4.4 GHz chip from ADI.  National
Semi and Hittite now part of ADI) also make these sorts of things.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/18/2015 1:20 PM, Magnus Danielson wrote:


The trick is to convert the 2nd degree loop to a 3rd degree loop, which
then allows for a 12 dB/oct slope, to counteract the 9 dB/oct slope.


No this is not correct.  A very conventional Type 2 loop, where the
loop filter consists of an integrator with series RC for the feedback,
and a resistor for the input, will suppress noise at 12 dB/octave with
the corner at w=1/RC.  A 3rd order loop is rarely helpful because
the noise is typically limited by phase detector noise floor.
Of all the many hundreds of PLL's I have designed, I don't remember
ever using 3rd order.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/17/2015 11:36 PM, Attila Kinali wrote:


Do you have any recomendation, where an ordinary engineer could
read up on this topic?

Attila Kinali



There is always Floyd Gardners, Phase Lock Techniques.

However, a better tutorial would be the one written by
HP's Dieter Scherer which was published in Microwaves
& RF Magazine (or possibly Microwave Journal).  I believe
the same content was available from HP as an Ap Note or
something.

Can anyone give Attila pointer to this document?
I had a paper copy for many years but I am pretty
sure it was a victim of a major cleanup after I
retired.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/20/2015 6:25 AM, Jim Lux wrote:



This makes a good case for the "30dB/decade very close in"



Somewhat had asked about "how" close in the 30 dB/decade is
good for.  There is a reference about this issue.  The
book Edson: Vacuum Tube Oscillators has what I believe is
the first published calculation of oscillator phase noise.
Way before Kurakowa and Leeson.  Edson has a formula that
gives the 3 dB bandwidth of the oscillator signal.  It's
not really a bright line.  This bandwidth is typically
a small fraction of a Hz for the oscillators Edson is
talking about.  However, for an unsophisticated 10 GHz
oscillator, it might amount to something that matters.

I think Edson also had a paper in Proc. IRE around 1950
that might have discussed this.

An interesting anecdote about Edson's book.  When I got
out of college I worked on RF oscillator design for
Boeing Electronic Products and learned about what was then
a very obscure topic called phase noise.  I was no expert,
but could at least say the right buzz words.  At the time,
the only book about synthesizers was Phaselock Techniques,
which I studied extensively.  It didn't have much on
phase noise.

In 1975, I interviewed for a job at another company
designing marine radios.  They wanted to replace crystals
with a synthesizer.  The hiring manager I interviewed with
was just about at the end of his rope because most
applicants couldn't even spell phase noise.  At some point,
I realized I had a lock on the job.  My new boss taught me all about
phase noise using Edson as a text book.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Close in phase noise of microwave VCOs

[Richard (Rick) Karlquist]

On 6/20/2015 1:16 PM, Alex Pummer wrote:

Actually a YIG, even "standalone" has very good phase noise performance,
as long as the tuning current is quite, once upon the time HP made some
cheaper version of the 856x-es spectrum analyzers [ perhaps that was the
95xx ] they had the first LO just "standing alone" no PLL,  and drifted
away, but the phase-noise performance was not so bad. These YIGs are to
find on e-bay sometimes,
73
KJ6UHN
Alex


According to R&D Engineers who worked in the division that
made spectrum analyzers, the division manager issued an order
that he didn't want any new products to cannibalize the sales
of the flagship 8566.  The 8590 series especially was dumbed down
and crippled, such as you say, with a drifty unlocked LO.

This sort of nonsense actually happened quite frequently, such that
when someone wanted to get funding to develop a new product,
they had to have a story as to why it wouldn't cannibalize
other products.  Many meritorious ideas were killed due to
this issue.

Of course, as we know, if you don't cannibalize your own products,
your competitors will.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] how to find low noise transistors

[Richard (Rick) Karlquist]

On 7/18/2015 2:16 AM, Attila Kinali wrote:


I always wonder how you figure out whether a transistor is low noise
or not. What part of the datasheet hints at which transistors have low
noise and which have not? Even if it's just try and measure, how
do you find good candidates to measure?

Attila Kinali



For a BJT operating above the 1/f noise corner, and at non-microwave
frequencies, the noise properties depend only on RF current gain and
base spreading resistance.  See "Low noise electronic design" by
Motchenbacher and Fitchen.  RF (not DC) current gain can be measured the
usual ways, but base spreading resistance has to be inferred from
noise figure measurements made with low source resistance.  The
RF current gain is the real fundamental noise property of the device
that you cannot change.  Fortunately, it can be determined from the
data sheet, if not directly, then by calculating it from DC current
gain and F-sub-t, based on the operating frequency.  The low frequency
current noise (above the 1/f corner) is simply equal to the shot
noise of the DC base current.  The low frequency voltage noise is
the sum of the Johnson noise that a resistor would have if its
value were the sum of the base spreading resistance and half of
r-sub-e.  Where r-sub-e is the "emitter resistance",IE the effective on 
resistance of the transistor.  Base spreading resistance can be

overcome by using a sufficiently high source impedance and/or
paralleling devices (if you can tolerate the additional capacitance).

At frequencies such as 100 kHz and 10 MHz, it is very easy to get
a noise figure well below 1 dB with a BJT, so it should be no great
problem to find a suitable device.

Even lower noise figures are available with JFET's, which have
noise current equal to the shot noise of gate current, which is
specified.  The resulting noise current is negligible for most
devices.  This leaves the noise voltage, which is just the
Johnson noise of a resistor equal to the channel resistance.
By scaling to larger devices and/or paralleling devices, this
can be reduced to arbitrarily low values.  The limiting factor
is the substantial capacitance of JFET's.  This limits them
to about 1 to 10 MHz, before high beta BJT's dominate.  I have
observed noise figure of below 0.2 dB in JFET's at 2 MHz.

Below 50 to 100 MHz, MOSFET's and ePHEMT's have excessive 1/f
noise and are a non starter.  Above the 1/f corner, it is easy
to get noise figures of a few tenths of a dB with ePHEMT's.

All of this discussion doesn't address 1/f noise, which could
be an issue in oscillators and low phase noise amplifiers.
For that purpose, you are back to characterizing devices yourself.
Putting negative feedback around the transistor can alleviate
this by reducing upconversion of noise.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] how to find low noise transistors

[Richard (Rick) Karlquist]

On 7/20/2015 8:12 AM, Charles Steinmetz wrote:

Rick wrote:


Base spreading resistance can be overcome
by using a sufficiently high source impedance


This sounds like the all-too-common noise figure fallacy (increasing
input impedance to get a lower NF).  All this does is raise the source
impedance's contribution to the total noise -- it doesn't reduce the
amplifier input noise voltage, and it increases the amplifier noise due
to input noise current.  The result is more output noise, not less, and
a reduced signal to noise ratio, despite the improved NF.

Even if you use a transformer to raise the source impedance, the signal
voltage increases by the turns ratio while the impedance increases by
the square of the turns ratio (thus, the output noise due to the
amplifier input noise current flowing through the source impedance also
increases as the square of the turns ratio).

Best regards,

Charles



When you raise the source impedance, you also have to reduce the
collector current.  Your analysis didn't take that into account.

Refer to page 83 of the first edition of "Low Noise Electronic Design".
Equation e. states that optimum noise figure is a function of the ratio 
between base spreading resistance and (beta)(r-sub-e).  If base 
spreading resistance is high, you make r-sub-e high by reducing 
collector current.  Equation f. states that doing that will increase

optimum source resistance.

Rick Karlquist N6RK


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] how to find low noise transistors

[Richard (Rick) Karlquist]

OTOH, the other cure for high base spreading resistance is
to simply parallel multiple devices.  This avoids the bad
side effects you mention.  The other key noise parameter
in a BJT is RF current gain, and this cannot be "cured"
by any circuit design tricks.

Rick Karlquist N6RK

On 7/23/2015 8:29 PM, Charles Steinmetz wrote:

Rick wrote:


optimum noise figure is a function of the ratio between base spreading
resistance and (beta)(r-sub-e).  If base spreading resistance is high,
you make r-sub-e high by reducing collector current.


I replied:


reducing transistor current to raise the noise resistance causes
undesirable collateral effects (including reduced bandwidth, which
increases phase noise due to baseband noise modulation of transistor
capacitances and generally increases nonlinearity).


I should also have mentioned:

Reducing transistor current also frequently reduces beta (sometimes by a
large factor, depending on the transistor's beta vs. current curve and
where you are on it).  This directly affects (beta)(r-sub-e) and,
therefore, directly reduces the noise figure.  I've pasted in the beta
vs. collector current graphs for the ubiquitous 2N3904 and 2N4401 to
illustrate this.  Some transistors are better than these over a useful
range of collector currents, others are much worse.  The beta of PNPs,
which are generally quieter than NPNs, also generally falls off faster
with reduced collector current.

Note that these are static (DC) curves, which are good approximations
for the 1/f region.  The curves in the white noise region, even at
relatively low frequencies like 1 kHz, generally fall off faster than
this as current is reduced, so the effect of reduced beta on in-band
noise figure is greater.

Best regards,

Charles




___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] how to find low noise transistors

[Richard (Rick) Karlquist]

On 7/24/2015 11:58 AM, Brooke Clarke wrote:

Hi Charles:

Does hFE (DC) have much relevance to this?  Would hfe (AC) be the
important one?


Only insofar as DC current gain is an upper bound on AC current gain.
If your operating frequency is less than f-sub-t divided by beta,
then DC current gain = AC current gain.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP Panel colors

[Richard (Rick) Karlquist]

There used to be a color at HP called "mint gray" FWIW.

Rick

On 8/13/2015 1:58 AM, tim...@timeok.it wrote:


I am interested in knowing the exact name or number  of the colors used by HP 
for the HP5065A front panels.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] 10811 unsoldered fuse

[Richard (Rick) Karlquist]

The thermal fuse plugs into pin sockets.  It cannot be
soldered for the obvious reason that the solder
would melt it...at 109 degrees as it is marked.
My suggestion would be to jumper it out of the circuit.

Rick Karlquist N6RK

On 10/19/2015 12:31 PM, Dimitri.p wrote:

How common is it to find undetected missing solder on 10811 parts after
all these years?

Dimitri


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

The oscillator transistor and buffer amplifier are basically
the same as the HP 10811, except for the absence of mode
suppressors.  The difference here is that the oscillator
self limits in the oscillator transistor, whereas the 10811
has ALC.  The discontinuous operation of the transistor,
as explained by Driscoll some 45 years ago, is undesirable
because it increases the load resistance the crystal sees.
The 2 transistor "Driscoll oscillator" fixes this problem
by using an additional stage that limits instead of the
oscillator transistor.  This has been widely used for
decades.  It is interesting to note that the 10811 ALC
works by varying the DC bias current in the oscillator
transistor.  This is in contrast to the elaborate DC
bias current stabilization here.

I have demonstrated that the close in phase noise in
the 10811 is entirely due to the flicker noise of the
crystal.  The only place where the 10811 circuit comes
into play is beyond 1 kHz from the carrier, where the
Burgoon patent circuit (which apparently has prior art
from Ulrich Rhode) reduces the phase noise floor.  I
have built two different oscillator circuits for 10811
crystals and have measured the flicker noise as being
the same as the intrinsic noise of the crystal.

Thus, obsessing over noise in oscillators circuits may
be overkill, unless you are planning to use a much
better crystal (BVA, etc).  OTOH, it might be advantageous
to improve the reverse isolation by adding additional
grounded base buffer stages.  There are various NBS/NIST
papers where several grounded base stages are cascaded.
I did this in the HP 10816 rubidium standard.

It is good to see time-nuts learning about oscillator
circuit by building them.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

On 10/26/2015 9:15 PM, Bruce Griffiths wrote:

The 10811A ocxo uses an oscillator of this type albeit with a lower crystal 
current, an overtone crystal. However the output stages spoil the PN 
floor..Cascaded transformer coupled CB stages are somewhat quieter.
Bruce



That's right. Burgoon (10811 designer) told me he had to meet other
requirements besides noise floor.  He had a special
one-off version of the 10811 without these compromises
that he built to provide a reference source to use
for phase noise measurements.  In the 10816 rubidium,
I used 3 common base transistors in cascade as
the output buffer and got similar results to
his special version.





Driscoll developed various high frequency crystal oscillators employing MMICs 
RF splitters together with a crystal, various matching circuits and a diode 
limiter.



Yes, he did, but long before that work he championed
his 2 transistor circuit that was extensively copied
by many other designers.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

On 10/27/2015 10:11 PM, Bruce Griffiths wrote:


The answer to this conundrum is surely that the equation for PN doesn't apply 
directly in this case
for offset frequencies outside the crystal bandwidth.
The Crystal actually bandpass filters the signal and PN noise generated by 
oscillator.
For offset frequencies outside the crystal bandwidth the oscillator generated 
PN is greatly attenuated
so that the noise of the buffer amplifier chain (CB stage plus output 
amplifiers) dominates.
In calculating the noise floor of the buffer amplifier chain the fact that the 
crystal has
a high impedance at these frequencies should be taken into account.

Bruce



Bruce has it exactly right.  At offset frequencies beyond 1 kHz,
the source impedance for the grounded base is very high due to
the crystal impedance being very high.  As Burgoon explains,
this condition suppresses base recombination noise, and the
only noise mechanism that is significant is the collector shot
noise.  (To minimize shot noise, don't run more DC collector
current than necessary).

I read Ulrich Rohde's 1977 article showing this circuit,
before I started working at HP in 1979.  When I got to HP,
and learned about the 10811, I pointed this article out to Burgoon.
It turns out he independently reinvented the circuit,
but he was apparently the first person to realize the noise implications
of the circuit.

This buffer circuit, extended to multiple grounded base stages
in cascade for additional reverse isolation, makes so much sense that
every oscillator where phase noise floor or reverse isolation
is important should be using it, IMHO.  (Burgoon's patent expired long
ago).

A brief comment about the collector-base capacitor in the 10811 Colpitts
circuit:  this has the usual Colpitts function at 10 MHz, but it
also prevents the 2N5179 from oscillating at 1 GHz.  It must be
installed very close to the transistor or else the 2N5179 will be
unstable.  I discovered this when I copied the schematic, but
not the layout, of the 10811 for use in the 10816 rubidium.

The 2N5179 in the 10811 is selected for minimum beta and Ft at
20 mA, which is the start up condition due to the ALC being
at full gain.  It has a special HP part number, so you wouldn't
know this just looking at the parts list.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

On 10/28/2015 10:38 AM, Hal Murray wrote:


rich...@karlquist.com said:

The 2N5179 in the 10811 is selected for minimum beta and Ft at 20 mA, which
is the start up condition due to the ALC being at full gain.  It has a
special HP part number, so you wouldn't know this just looking at the parts
list.


How much of a difference does that selection step make?

I'd expect the parts within a batch to be very similar, more so for mature
parts.




All I can tell you is that Burgoon found a non-zero number of
2N5179's that wouldn't start. Knowing the way things were done,
he probably got a response from the vendor to the effect that
it was simply an unspecified parameter and they only guarantee
JEDEC specs and the transistor(s) he found were not a fluke.
HP greatly discouraged the batch qualification paradigm, although
the did resort to it when justified.  It was not justified for
the 10811.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

Do you have a specific URL for "hacking oscillators"?  I can't
find it on Rubiola's web site.

Rick

On 10/28/2015 1:32 PM, Gerhard Hoffmann wrote:

Am 28.10.2015 um 19:22 schrieb KA2WEU--- via time-nuts:

This oscillator seems to have been more a frequency standard then a noise
standard. Today's 10 MHz oscillators are different/better, such a
crystal is
no  longer available/made.

Yes. Rubiola gives it the credit of being able to be mass-produced, and
it _was_
one successful product. There is a section in "hacking oscillators" on it;
my copy of the book is 200 miles away right now.

regards,

Gerhard, DK4XP


(see www.rubiola.org)
___

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Low noise quartz crystal oscillator by Bruce Griffiths

[Richard (Rick) Karlquist]

Driscoll wrote a lot about oscillators over the years.
I couldn't find anything specific to discontinuous operation.
Do you have a titel of a paper related to this?


What Driscoll was talking about was self limiting in a
transistor.  That is discontinuous operation, although
Driscoll doesn't call it that.  His earliest papers on
this circuit go back to around 1972, and are in either
UFFC proceedings and/or FCS.  Many later papers cite
these.




The 2 transistor "Driscoll oscillator" fixes this problem
by using an additional stage that limits instead of the
oscillator transistor.


Is the Burgoon patent you are refering to US4283691
"Crystal oscillator having low noise signal extraction circuit" ?


Yes.

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Time syncing WiFi routers using FM radio

[Richard (Rick) Karlquist]

On 11/12/2015 12:04 AM, Hal Murray wrote:


I think it was HP that measured the signal in the Silicon Valley area.  NBS
published and distributed the offset.

Does anybody remember that booklet?  Did I get the story reasonably accurate?



When I was hired by HP in 1979, my new boss (who was the head of the
Precision Frequency Sources R&D section) gave me a tour of the
Santa Clara Division site, including the "Metrology Room" which
was a glassed in area strategically located at the intersection
of the main east-west and north-south corridors in the complex.
Right next to the windows was a rack with several HP 5061A cesium
standards and various other non-descript rack chassis.  There
was also a small TV set that was rack mounted.  This was part
of the system to compare the cesium standards to the TV signal.
My boss described this as "the most accurate clock on the west
coast".  Passersby were invited to set their watch according to
the front panel display on the 5061A.  There was also a 5087A
distribution amplifier that supplied 10 MHz to my R&D section as
well as the various production lines.  This was referred to as
the "house standard".  My memory is a little hazy about the
transition from the TV signal to Loran C, which was fairly new
at that time.  Later, they built a large time display for the
lobby that was synchronized to the 5061A in the metrology room.

Interestingly, the Santa Clara site came with built in wall
clocks that were supposed to be centrally synchronized, as was
common in the 1960's.  These were not connected in any way
to the "most accurate clock on the west coast" and were frequently
off by several minutes and were not even consistent among
themselves.  So the joke was that we had both the most accurate clock 
and the least accurate clock.


Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] modern electronics education/jobs (was:

[Richard (Rick) Karlquist]

On 11/12/2015 1:01 PM, William Schrempp wrote:


has failed. I hear old machinists complaining about new machinists who can't
drill a hole if the drill-press isn't computer-controlled. And in my work,
nurse education, I see students who can't be bothered to learn how to take a
manual blood-pressure, because a machine can now do it (sort of). Much to
ponder here. . . .



Bill Schrempp



This reminds me of a summer job I had as a lab assistant between my
freshman and sophomore years at college.  There were a couple of
journeyman machinists with Bridgeport mills.  They didn't let me
use them, but they did patiently teach me how to use the drill
press, taps, hacksaw, etc to make simple parts that didn't require
their skills.  They told me that, in Germany, a kid training to be
a machinist would start out by being given a file, a pair of calipers,
and a rough block of metal.  His task was to make a perfect cube with 
sides of exactly 1 cm by 1 cm.  Only after mastering that, would
he be allowed to move onto more advanced equipment.  Fortunately, the 
machinists just told me this story to scare me, but they didn't make me 
file a perfect cube.  They did tell me I needed to learn to drill holes

with 0.005 inch accuracy using a machinist's scale and a center
punch to lay them out.

Rick
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Phase microstepper designs?

[Richard (Rick) Karlquist]

On 12/9/2015 2:26 AM, Attila Kinali wrote:

Moin,

I just tried to figure out how phase microsteppers are usually build,
but, beside the time-nuts discussion from 10 years ago and US patents
US4358741 and US4417352 my search turned out empty. I am pretty sure
that I used the wrong search terms and there should be lots of documentation
out there. Can someone give me a hint what to look for?

Attila Kinali


My extensive experience with digging up old articles (before the 
internet, the library was like my second office) is that when

the trail goes cold on some topic for no apparent
reason, it turns out that there is a (non-apparent) good reason
why no one is writing about the topic any more.  I usually
got to the bottom of this by asking experts in the industry.

My general impression about phase microsteppers based on when I
was involved in the time industry is that they didn't have a
very good reputation and were an inferior solution that was
used as an expediency because it it could be "bolted on"
to an existing system that wasn't originally designed to be
tuned.  I think the concensus was that direct digital synthesis
is the better way to go, assuming you can design it into the
architecture in the first place.  DDS is a pretty well developed
technology at this time that is supported by good OTS chips,
so it is not clear what the value proposition would be for
phase microsteppers.

When we were designing the 5071A 25 years ago, it was clear
that the method of tuning by varying the C field was not
acceptable for a state of the art standard.  There are
multiple reasons for this decision.   At the time,
there were a few DDS chips, but we had to roll our own DDS
due to the stringent and specialized requirements.  It took
a whole PC board for this circuit at the time.  The possibility
of using a phase microstepper was never even considered AFAIK.

Are you aware of some redeeming value that phase microsteppers
have that would make us want to investigate them?   Am I
being too harsh on them?  Maybe you can champion them and they
will make a come back :-)

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] GPSDO and oscillator steering - EFC vs DDS schemes?

[Richard (Rick) Karlquist]

It's been 20 years since I presented that paper in San Francisco
at FCS and I had just about forgotten about it.  It is
flattering to realize that people are still reading it now.

It might be useful for the discussion here if I explained why I
wrote the paper.  We had recently completed the 5071A, which,
unlike previous cesium standards, did not use the C field for fine
tuning.  This strategy turned out to be a real winner, provided
that we had a solution to offsetting the frequency.  The
5071A architecture provides something of a free lunch in this
respect because the synthesizer doesn't have to use the
10811 as a 10 MHz reference and then generate "almost" 10
MHz.  That is problematic for reasons I will explain below
that are probably familiar to many time nuts.  In the 5071A,
we multiply the 10811 to 9280 and then offset it about 87
MHz using a phase locked crystal oscillator that is driven
by the DDS.  This avoids multiplying up the DDS and the
XO cleans up DDs spurs, except very close in ones.  By
using a custom DDS, we could play our cards right to avoid
close in spurs.  The mix from 9280 to 9192.63277 is done
using a sideband, so the "mixer" is "free".  The CBT is
a "free" filter to remove the "carrier" and opposite sideband.
If you want more details, you can read my 1992 paper about
it.

None of this gives us a leg up on a bolt-on circuit that
makes almost 10 MHz from exactly 10 MHz.  That is the worst
thing you can try to do with a DDS in terms of spurs.  The
often seen paradigm of cleaning up a DDS with a phase locked
crystal oscillator is limited because (at least in 1995) DDS
chip sets use RF DACs limited to 10 or 12 bits.  Even now,
RF DAC's are limited to an advertised 16 bits, but still only 10
or 12 effective bits.  So if you want really low spurs, you
can't get them with a DDS.  The XO clean up doesn't work for
close in spurs, inside the PLL bandwidth.  And if and stuck
with an almost but not quite unity synthesis ratio, you are
painted into a corner.

I had always been fascinated with the 5100 direct synthesizer.
In theory, this is extensible to arbitrarily fine steps and
arbitrarily low spurs.  An interesting feature of the 5100
was that the finest resolution "decade" was actually a
continuously tunable variable frequency oscillator.  This
was feasible because the oscillator was divided by my a
huge divisor and could provide a useful output with "infinite"
resolution.  What I did was de-construct the 5100 architecture
and reduce it to a skeleton architecture that took advantage
of the near unity ratio to simplify the block diagram.  The
continuously tunable oscillator was replaced by a DDS.  It
took advantage of the availability of cheap 10.7 MHz filters.
Even if I could have chosen any center frequency I wanted,
these turned out to be about right, assuming you are working
at 10 MHz.  If you wanted to make a 5 MHz version, or a 100 MHz
version, you would have to get custom made filters.

About that time, HP started getting into the "smart clock"
business.  Although I was able to add varactor tuning to the
E1938A oscillator without degrading its stability, it
created a new problem of how to generate the DC tuning voltage.
To help with this, I put a 2.5V voltage reference inside
the E1938A.  The initial reference I chose turned out to be
noisy and I had to replace it with a lower noise chip.
Both the tuning voltage and the reference were brought
out of the E1938A "hockey puck" with dedicated return
lines, to avoid ground loops.

This still left the problem of coming up with a very stable
and accurate DC DAC.  IMHO, the smart clock designers never
really solved this problem.  They just got by with a marginal
system.  They didn't have the space to build an offset
synthesizer like the one in my 1995 paper.

A few years ago, Agilent introduced an arbitrary waveform
synthesizer using a very advanced RF DAC code-named "Griffin"
that was also proposed to be used in various other instruments.
This chip could be used to make an interesting offset
synthesizer, but the cost and complexity would make the
architecture in my 1995 paper look really attractive.
The Griffin can generate any frequency, but if you only
need a very narrow band of frequencies, all that capability
goes to waste.

Rick Karlquist N6RK
(now retired from Agilent/Keysight)


On 12/8/2015 8:32 AM, Attila Kinali wrote:

Moin,

I've been digging through some stuff and stumbled (again) over Rick's
paper on high resolution, low noise DDS generation[1] and got confused.
The scheme is very simple and looks like to be quite easy and reliably
to implement. If I understood it correctly, the critical points are the
DDS, its sideband generation and the LO/RF feedthrough in the mixers.
Nothing that is not known and nothing that is too difficult to handle
(the 10.7MHz filter get rid of most of the feedthrough already and
there has been a lot written on how to design DDS for specific applications).

What puzzled me is, why this has not been used

Re: [time-nuts] Phase noise from Allan Deviation ?

[Richard (Rick) Karlquist]

On 12/13/2015 7:15 PM, Tom McDermott wrote:


It brings up a question:  Is it possible to estimate the phase noise of that
internal crystal from the ADEV measurements?  There are a bunch of
papers that go the other way:  from Phase Noise to Adev.   Searching
brings up only one paper that goes from ADEV to Phase Noise but it's text
does not seem to be readily available.  It apparently models the oscillator
as a couple of well known error models.

-- Tom, N5EG]


In general, you cannot determine phase noise from ADEV,
even though you can determine ADEV from phase noise.
This is just a mathematical reality.

Mike Fischer (of HP) presented papers at the 1977 and
1978 that show conversions between PN and ADEV for
individual noise processes, where each process has a
specific slope of amplitude vs frequency.  The only
time you can go from ADEV to PN is if you can isolate
a process.

In the specific case of crystal oscillators, in general,
they follow a flicker noise of frequency process model
close to the carrier (within 100 Hz).  You can often
assume that ADEV is dominated by this process and therefore
translate it to an equivalent PN.  The way to tell if
ADEV is dominated by this process is that it will be
independent of tau, for tau of 0.1 sec or less.

On the IEEE UFFC site, there is a tutorial on crystal oscillator
design by Mike Driscoll (you don't have to be a member to
access it).  I believe this covers this topic.  You go
from ADEV to "frequency noise" and then to phase noise

You can "practice" this calculation on any crystal oscillator
that has published ADEV and phase noise.   It is of course
extremely easy to screw it up :-)   What I have found
is that most crystal oscillators seem to obey the flicker
model.

I have been able to measure flicker noise on crystals that
were not installed in an oscillator, and then install them
in an oscillator and the ADEV turned out to be what
was predictable from the phase noise.  It really works!

Rick Karlquist N6RK

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] DMTD - analog multiplier vs. diode mixer ?

[Richard (Rick) Karlquist]

On 1/5/2016 12:07 PM, Bruce Griffiths wrote:

The noise of such Gilbert cell based analog multipliers far exceeds that of the 
traditional mixer.
Bruce


Read Gilbert's paper or Gray and Meyers analog IC textbook and
you will see that the whole theory of operation of these
depends on keeping the signal levels in them very small,
especially if linearity (actually translinearity) is
important.  They always have current sources in the
emitters that contribute a lot of noise.  So you have
small signals and large noise.  The IC's that are
designed to be DC coupled have even more sources of
extra noise.

IMHO, they only make sense in low performance applications
where the lack of transformers is important or in DC
coupled applications.  The only time I have used an
analog multiplier IC was in Costas loop to demodulate
QPSK from weather satellite.  It needed to be DC coupled.

Rick Karlquist N6RK





 On Wednesday, 6 January 2016 9:01 AM, Poul-Henning Kamp 
 wrote:


  My little HP5065 project is continually running into the jitter of
my HP5370B counter which is annoying me, so I'm looking int DMTD.

Everybody seems to be using traditional diode-mixers for DMTD,
and to be honest I fail to see the attraction.

Why wouldn't a analog multiplier like AD835 be better idea ?

What am I overlooking ?


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] DMTD - analog multiplier vs. diode mixer ?

[Richard (Rick) Karlquist]

On 1/7/2016 4:35 AM, Attila Kinali wrote:


How about using the Gilbert Cell as "digital" mixer,
ie driving the currents hard from one branch to the other
and replacing the current sources by resistors?

How much would that improve the noise? Would it be still much
worse than the diode mixer?

Attila Kinali


You can drive the Gilbert cell as hard as you want, but
the active region is only about 100 mv so the extra
drive voltage doesn't help.  It is the same as if you
drove it with a 100 mV square wave.  Somewhat better
than a sine wave, but not a game changer.

You can of course try to replace the emitter current source
with a resistor, which works to the extent that you can
afford to throw away voltage across the resistor, but
you will never get a very high impedance.  No OTS IC's
are designed this way.  What would be better would be
to use an inductor.  A true noiseless current source.
Again no OTS IC's are designed this way.

You would have to homebrew the whole mixer from discretes.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] DMTD - analog multiplier vs. diode mixer ?

[Richard (Rick) Karlquist]

On 1/7/2016 3:11 PM, Bob Camp wrote:

Hi

If your intention is to run a mixer with saturated inputs …. just run
an X-OR gate. It will handle the high level signals much better than
an over-driven analog part.



Bob



If you look at the schematic of an XOR gate IC and compare it
to the schematic of, for example, an MC1496 mixer, you will
see a lot of similarity.  If the gate is of the ECL type,
it will have the addition of emitter followers, but that
it a minor detail of implementation.  I'm not sure there
is a huge difference.  ECL is a great logic family in
general (self-confessed ECL-phile here :-) but it is
probably the worst for phase noise, compared to the
saturating logic types.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

On 1/9/2016 12:44 PM, Attila Kinali wrote:


The purpose of the input circuit is to convert the RF input signal
into a low-jitter square wave that can drive the PIC clock input.
The circuit is closely based on the one published by Wenzel at
http://www.wenzel.com/documents/waveform.html, with modifications
suggested by Bruce Griffiths and Ulrich Bangert. The revised circuit
works with inputs as low as -20dBm.


This circuit is very similar to one that was championed by Tom
Faulker of HP/Agilent at the now closed Spokane site.  Tom
measured the circuit at about -171 dBc/Hz.  He was very knowledgeable
about this topic, so we can believe the number.

This is good, because the cited Wenzel document would give me
no confidence whatsoever if that was all I had to go on.  It
reads like it was written by some marketing guy (as opposed to
R&D) who knows just enough to be dangerous.  Other than the
circuit in question, the rest of the document is full of
unreliable information.  Such as how line receivers make great
sine wave to square wave converters.  They are terribly noisy.
IE, the document contains a kernel of truth.  It also has no
quantitative information about the circuit in question or
any other ones discussed.  It's disappointing to see this
published by an otherwise excellent outfit like Wenzel.

The modifications make sense IMHO.  I suspect that the 2N3906's
are good for two reasons:  the low f-t reduces noise bandwidth
and the high current gain reduces noise current.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Divider circuit for Rubidium Standard

[Richard (Rick) Karlquist]

No, it was just word of mouth within the company.
Somewhere I have a piece of notebook paper
on which Tom drew the circuit.  We did have
internal forums where papers where presented,
but this was never even published internally.
As with all forums, a lot of stuff happens
outside the official forums in the hallways,
or over lunch, etc.  Ya gotta network.
There was probably a great sigh of relief when
I retired, and wouldn't be annoying coworkers
with endless questions any more.

Tom's contribution wasn't the circuit per se.
I used this same basic circuit at Zeta Labs
40 years ago.  I ever remember the PNP transistors
we used: the Fairchild 2N5771.  Those were the days...
Tom's contribution was pointing out the fact that it was
better than the other circuits out there,
and the fact that he gave a number.  It has
been a never ending battle to disabuse low
information designers of line receivers,
comparators, and other low performance solutions.

If you want to do some reading, I vaguely remember
that NIST published some papers on sine to square
at UFFC or PTTI.  Start with Fred Walls for your
search.  Also remember that this circuit, as good
as it is, is a poor mans substitute for the multistage
zero crossing detectors discussed on this forum many
times.

Rick

On 1/10/2016 2:24 AM, Attila Kinali wrote:

Hi Rick,

On Sat, 9 Jan 2016 14:45:43 -0800
"Richard (Rick) Karlquist"  wrote:


This circuit is very similar to one that was championed by Tom
Faulker of HP/Agilent at the now closed Spokane site.  Tom
measured the circuit at about -171 dBc/Hz.  He was very knowledgeable
about this topic, so we can believe the number.


Is this documented anywhere publicly? I would be very interested
to read this.

Attila Kinali


___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] DMTD - analog multiplier vs. diode mixer ?

[Richard (Rick) Karlquist]

Phase frequency detectors (starting with the legendary MC4044)
being made out of flip flops, had metastability and/or race
conditions.  Motorola showed a block diagram made of gates,
as if it were combinatorial logic, but because of the feedback,
it is actually a state machine, as described in the MC4044
data sheet.  It had a dead zone around zero phase that came
to light when Fairchild introduced the competing 11C44 PFD
using Eric Breeze's patent to fix the dead zone.  The
11C44 data sheet showed their dead zone, vs Brand M.
Even that improved chip still had a "funny" zone, it just
never went to zero gain.

Fast forward to today, we are now seeing PFD's made
with samplers.  They too have a bunch of issues with
phase noise floors.  None of them come close to a mixer.

In the 5071A, I used a mixer as a phase detector that
had some flip flops only used for acquisition, so they
were non players in terms of phase noise.  I still think
I would do that even if I had to do over 25 years later.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] New Member + Basic Questions

[Richard (Rick) Karlquist]

On 1/14/2016 12:35 PM, Nathan Johnson wrote:

What does the group think of the HP 8660? Just scored a broken one too cheap to
pass up. I know it's not gonna be the last signal generator I buy, but for under
$100 shipped it should be an interesting project.
Nathan KK4REY



When I worked at HP, I had the chance to discuss this product
with various engineers who were involved in its development.
It frankly wasn't one of the better products in the line.
It has a high broadband noise floor.  The follow on
product, the 8662, is MUCH better.  No comparison.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] low noise multiplication to 100 MHz

[Richard (Rick) Karlquist]

It is interesting that the HP8662A multiplies 10 MHz to 640 MHz,
in steps of 2X.  But there is a crystal filter at 80 MHz to
clean up the wideband noise of the 10811.  In the 11729, they
filter the 640 MHz from the 8662 with a SAW filter, again to
eliminate multiplied up wideband noise.  It's going to be
tough to replace a 100 MHz OCXO by multiplying.  You have
the additional problem that 100 MHz is not a power of 2
multiplication.  I think you're stuck waiting for your
OCXO's.

Rick Karlquist N6RK



On 1/21/2016 6:43 AM, jimlux wrote:

My tiny 100 MHz low noise OCXOs are unexpectedly delayed at the mfr, and
I'm looking at alternative schemes.
One is to get 10 or 20 MHz OCXOs (typically in stock) and multiply them
up. I've got the Wenzel ap notes on 2diode and using HCMOS (and I've
used the packaged Wenzel multipliers), and I think I have some spare
board real estate on another board.

The 2diode multiplier describes using 1n5711 or 1n914, but I was
wondering if anyone has run this sort of multiplier up to 100 MHz?

What sort of symmetry does the resulting waveform have (yeah, it's
basically a filtered sinewave, because you're picking a harmonic, but
I've been surprised before)?




I'm driving an FPGA and a couple of ADCs.  The ADCs have differential
input that is 10kohms with 9pF in parallel offset from ground in the
usual way (we're using a transformer and appropriate bias resistors).
Not a 50 ohm load, in any case.  And it wants a clock that is high for
about 47.5% to 52.5% in one mode and much wider (30%-70% in another).. I
need to check.

The FPGA is less critical noise-wise, and has a AD8138 buffer in any
case, which can fix a variety of evils.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] what is acceptable harmonic content & level for a 10Mhz standard?

[Richard (Rick) Karlquist]

On 1/26/2016 11:52 AM, walter shawlee 2 wrote:

I have been working on a compact portable 10Mhz bench standard
using both an FE FE5680A Rb oscillator and an Oscilloquartz ovenized


It is important for a 10 MHz source to launch a pure sine wave
and also to have an accurate 50 ohm impedance at 10 MHz and its
harmonics.  This is because a sine wave is the only waveform
having the property that it is immune from distortion due to
reflections from poorly matched loads.  IE, sine wave with
multipath resolves again to a sine wave.  It is important to
have a broadband 50 ohm output impedance to absorb reflections
from loads, especially load-generated harmonics.  You didn't
ask, but you also want a lot of reverse isolation to avoid
pulling the OCXO.  All of these were addressed in the design
of the output section of the 5071A cesium standard.  It was
really quite non-trivial.  Specifically, the harmonics are
down over 80 dB.

Rick Karlquist N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP5370B & HP5345B Front-End IC Redesign Effort

[Richard (Rick) Karlquist]

On 1/22/2016 2:14 PM, Mathew Breton wrote:

I was gifted an HP 5370B with the usual problem: front-end problems, probably 
due to overstress. It is currently up and running again with a set of 5345A 
series A3/A4 boards as I wasn't able to get a cheap pair of 5088-706x hybrid 
ICs.
This sounds like a common problem. As a result, I'm designing an open-source 
drop-in (hopefully) replacement. My hat is off to the original IC designer, as 
it is not a trivial effort due to the wide input signal common-mode range, and 
very tight trigger timing requirements. Other items (like the E-ECL) output) 
are also adding a bit of extra effort.
I'm hoping that someone(s) might be interested in working with me on it. I 
would like to have my assumptions and math checked before I start the detailed 
design phase, and perhaps contribute some better ideas.
In addition, it would be really helpful if someone could run a few rise-time dispersion 
tests on an instrument with a working "B"-series A3/A4 PCB set (my unit 
obviously doesn't qualify).
Regards,
Mat Breton  


I would like to mention that the "father" of the 5370,
David Chu, retired from Agilent a few years ago.  He
might be receptive to giving you some advice about
your project.  He is still very sharp technically.
If there is sufficient interest,  I might be able
to arrange for an introduction.  It would be helpful
if we had a show of hands on time nuts as to how many
people on time nuts are interested in this board.
David was one of the best engineers in the history
of the Santa Clara division and the fact that the
5370 lives on is a testimony to how far ahead of
its time the design was, some 40 years ago.

Rick Karlquist N6RK
HP Santa Clara Division 1979-1998
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP Equipment Running Hot as Heck...

[Richard (Rick) Karlquist]

Vintage HP equipment often had HP made power transformers,
which ran all the time whether the equipment was on or off.
The core loss while idling could be fairly high.  The
core was usually just below saturation, so that if the
power switch was on 100V, it would really get hot.

There was some logic analyzer model that HP engineers
referred to as the "logic furnace" because it ran
so hot.

Rick Karlquist N6RK

On 1/27/2016 4:22 PM, Burt I. Weiner wrote:

Don,

Something to be careful of... Be sure the mains voltage switch is set
properly. I had a HP-3336A that ran hot like that. Like you, I put
bigger and bigger heat sinks on the regulator, but all that seemed to do
was make the bigger heat sinks as "hot as heck" also. What I discovered
was that the mains voltage selector was set to 100 volts. I reset it to
120, (or was it 117) and it ran much cooler.

Burt, K6OQK


From: "Don Latham" 

Someone has already probably said, watch out for switching regulators.
BTW,
almost all the Hp instruments I have from the 80's era run hot as heck.
I have
put on fans and piggybacked more fins (and more fins, and more fins...).  A
Military version of the 5328A counter I have has what sounds like a leaf
blower in it, with a proportional controller added.
If you do some work with switchers, I'm sure the list would be very
interested!
Don

Burt I. Weiner Associates
Broadcast Technical Services
Glendale, California  U.S.A.
b...@att.net
www.biwa.cc
K6OQK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to
https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.



___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP 10811

[Richard (Rick) Karlquist]

On 2/11/2016 2:56 PM, ws at Yahoo via time-nuts wrote:

Joe

The inner oven voltage needs to be stable! To better than 0.1V.
Unlike the single oven unit, the inner oven on the dual oven unit runs fine
at 15 Volts. It draws a couple hundred ma after warm-up.



It's been a long time, but IIRC, it was common knowledge
within HP that the standard 10811 oven will run on 15 Volts,
but it just takes longer to warm up.  The most extreme
example was the 5334A, which had such a bogus power supply
design that the voltage sagged to +12V during warmup.
Now the oven won't work correctly on only 12V, but it
will work to the extent that it warms itself up.  Once
the current cuts back, the voltage goes back up to over
15V and it works fine.  When I started the 5334B project,
the power supply was the first thing to get redesigned.

Rick N6RK
___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] Calibration procedures - what is normal?

[Richard (Rick) Karlquist]

On 2/12/2016 12:14 PM, Joseph Gray wrote:

I sent my HP 3457A in for cal. I should be getting it back next week.
I won't mention where I sent it, but it wasn't Keysight (I don't like
that name). I recently changed the SRAM battery and purposely did not


I left Agilent just before the split, but I don't know
anyone there who liked that name.  Or the logo :-)

You need to understand that the mission of service depots at
Agilent was mainly to be profit centers.  They were not there
to make customers happy in order to enhance Agilent sales.
At best, they needed to do warranty repairs to support equipment
under warranty, but even the whole warranty thing was a way
to make extra money, not to sell instruments in the first
place.  What you describe is perfectly consistent with my
experience using them as an internal customer.  BTW, they
charge internal customers the same high prices they charge
for external customers.  There is an attitude that it is not
worth making reliable products because they can make so
much money fixing them.  Consider yourself lucky you got
the extra effort.  They remind me of car dealer service
departments, in terms of the business model.  Specifically,
the "tune up" racket or the XXX,000 mile "service" racket.

Rick
N6RK



Not knowing what is normally done in the cal lab, I assumed that the
entire procedure as listed in the service manual would be done. It
seems that I was wrong.

In the end, the lab decided not to charge me for the extra time
involved. I thanked them for that.

My question is, do any cal labs (including Keysight) normally perform
the zero and full scale procedures as listed in the service manual?

Joe Gray
W5JG

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP Reliability

[Richard (Rick) Karlquist]

A few clarifications:

Before 1999, HP had a Medical Division that made
equipment you saw in hospitals and a Scientific
Instrument Division that made chemical analysis
equipment used in medical laboratories (and also
other laboratories).  IIRC, both began as
acquisitions.  The Agilent spin off in 1999
started a trend of additional divestiture.  The
Medical Division was sold to Philips, so former
HP hospital monitors were rebranded Philips.
After the Keysight spinoff, the remaining Agilent
company has two main parts:  chemical analysis,
descended from the old Scientific Instrument
Division, and life sciences, which grow organically
with the invention of gene arrays

What is now Keysight underwent massive reductions
in force after the dot com bust.  Around the
same time they started off shoring the Rohnert
Park manufacturing complex to Malaysia.  The
combination of the two eliminated something like
80% of the jobs in Sonoma county.  Some sites
like Liberty Lake (Spokane) closed completely.

This was when we started to see the quality
plummet.  They lost the recipe when they off
shored.  Instruments would arrive DOA, or would
fail after a few months.  Some had annoying
problems that would come and go, and they couldn't
seem to fix them.  When Windows XP expired, there
was a big crisis to upgrade to Windows 7 that
Malaysia fumbled the ball on.  (Many years ago,
the powers that be decided to use Windows internally
in place of Unix).

I wish Keysight well, but at the time I left in 2014,
and from what I've heard since, it didn't look encouraging.

Rick

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

Re: [time-nuts] HP Reliability

[Richard (Rick) Karlquist]

On 2/14/2016 11:20 AM, William H. Fite wrote:

They don't wonder; they know very well. But they're stuck. Consider
oscilloscopes. Why pay for a Keysight or Tectronix or LeCroy or, God
forbid, a Rohde & Schwarz when, for the vast majority of applications, a
Rigol will give you everything you need at 1/N the cost?



When I worked for HP/Agilent, there were countless examples of
low cost instrument prototypes developed at the central research
lab (HP Labs, and successors) that were killed by the manufacturing
divisions because they would "cannibalize sales" of the incumbent
product line.  The other excuse given was that it would "divert
resources" from the incumbent product line.  Or that the sales
force wouldn't sell it because there wasn't enough commission in it.
And we couldn't sell it direct because that would ruffle too many
feathers in the field.  I helped develop a product for which we had 
large volume purchase orders from several customers and they still 
killed it.


Rick

___
time-nuts mailing list -- time-nuts@febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.