[time-nuts] Best Rubidium Frequency Standard

[Mark Sims]

In many ways the 5065A is the probably the most repairable of all the units 
(closely followed by the FRK family and the M100).   They all use parts that 
are mostly still available and the circuitry is accessible.  You can assume 
that the lamp (and maybe some of the microwave parts) in any Rb oscillator is 
unobtainium.  
Most failures seen in these systems tend to be in the power related parts 
(electrolytic caps and heater / lamp driver transistors).  Total rubidium lamp 
failures are rare.  Lamps with degraded output due to rubidium depositing on 
the lamp walls can usually be refurbished with a heat gun.
If you don't need the ultimate in performance the telecom Rb's are a very good 
value.   They can be had for prices where repair is not an issue... just 
replace them.  At one time LPROs could be had for less than $50 and they were 
usually in good operating condition.   The bulk packs of the TEMIC units on 
Ebay approach that level today.  As time has passed the availability has gone 
down,  prices have gone up,  and the quality what is available had gone down.   
  
___
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] Best Rubidium Frequency Standard

[Mark Spencer]

I'd be curious in knowing what the correct settings to discipline a prs 10 are, 
even pointers and hints would be  very welcome.  I was never very happy with 
the performance of mine in that configuration vs simply letting it free run and 
periodically re syncing it (via the one pps input.)



All the best
Mark Spencer


Sent from my iPhone

> On Mar 12, 2016, at 3:11 PM, Charles Steinmetz  wrote:
> 
> jerry wrote:
> 
>> Is there any model-suffix or other identifier that would inform whether a 
>> used PRS-10 has the PPS sync input feature??
> 
> Not that I am aware of.  Not even the "customer part number" is a reliable 
> guide -- I have had several examples each of various CPNs, and some did have 
> the sync input feature and some did not.  (I've also seen examples of other 
> CPNs, *none* of which had the feature.  My sample hasn't been large enough to 
> conclude that none of them does, however.)
> 
> Re: how useful the feature is, if it is present: the PRS10 PLL is extremely 
> adjustable, and there are LOTS of combinations of settings that won't work 
> worth a damn with any particular external input.  So, finding a good 
> combination can be difficult, and finding the best combination can be elusive 
> at best to maddening (or worse).  Read the manual carefully, and pay 
> attention both to the default settings and to the hints that SRS provides.  
> With the right settings, a PRS10 *does* work extremely well with the PPS 
> input from a GPS.  They do generally take several days or more to lock, 
> because of the long time constants involved.
> 
> 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] Best Rubidium Frequency Standard

[Charles Steinmetz]

I wrote:

With the right settings, a PRS10 *does* work extremely well with the 
PPS input from a GPS.  They do generally take several days or more 
to lock, because of the long time constants involved.


Bob replied:

I would call having to wait a few days for it to lock a bit of a 
disadvantage. Even more so for those with an antenna challenged 
environment that gives them dropouts every few hours.


Well, one is certainly free to use shorter time constants to achieve 
lock faster.  The very long TCs simply allow one to exploit the 
exemplary stability of the PRS10 for performance much better than 
what an OCXO-based GPSDO can deliver.  If you are content with the 
stability of, say, a TBolt, you can replicate that performance by 
adjusting the PRS10 loop to match TBolt dynamics -- in which case it 
will lock in a comparable time.


WRT holdover, the long TC and the inherent stability of the PRS10 
mean that it will stay very close to the GPS PPS even over long 
holdover periods, so re-acquiring lock does not take nearly as long 
as acquiring it initially.  For the same reason, a PRS10 set up for 
maximum stability can acquire lock even if there are holdover periods 
during the acquisition process (in both cases, assuming that the GPS 
does not output "bad" PPS pulses when it is not locked to GPS).


In principle, one might be able to begin the process by setting the 
PRS10 loop "tighter," then changing the loop constants in one or more 
steps after it achieves initial lock.  I have not tried this, and do 
not know if changing the loop programming on the fly upsets the PRS10 
phase.  If not, it should work (and one could even program a BBB, 
'uino, or other small processor to do it automatically).


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.

Re: [time-nuts] quartz thermometers

[jimlux]

On 3/12/16 7:35 AM, Bob Camp wrote:

Hi

The conventional MCXO uses an AT cut crystal rather than an SC. It
runs on the fundamental and the third overtone. You can build one
with just about any normal fundamental crystal. There are also people
doing OCXO’s sort of the same way. Back when the MCXO came out, it
was about the only way to get a TCXO that had < 1x10^-8 sort of
stability. Now you can get that with a cheap (as in not $5,000 each)
commercial IC based TCXO.



And heck, if you can tolerate the 1.5W warm up power, you can get ppb
sorts of accuracy in a OCXO at 250mW. (Vectron EX421)  They're a few 
hundred dollars.


The Qtech MCXO is about 100 mW, and somewhat bigger than the EX421 (not
much bigger)

The vectron MX-503 is in the 50 ppb range at 40 mW (I don't know how it 
does the internal compensation.. maybe it uses some other kind of 
temperature sensor)


___
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] Best Rubidium Frequency Standard

[timenut]

Bert, Bob, Charles, Corby, Eric, Luciano, Mark, Rob, . . .

Thanks for the excellent advice. It has been very helpful. Let me see if I can
summarize what different people are saying.

   1. The absolute best rubidium standard is the HP 5065A. It has excellent
  short term stability. The only disadvantages are - a) the price, B) it
  is harder to use for disciplining via GPS, c) it is sufficiently old
  that life span could be an issue and d) has poor temperature control and
  suffers from pressure sensitivity.

   2. The second best is the PRS-10. Provided it is working. Its phase noise
  is generally very good, and above 100Hz may even be better than the HP
  5065A but is poor under 1Hz. It is difficult to repair, but that problem
  appears to be shared by everything except the HP 5065A which is still
  through-hole. Other than its performance, it has some advantages - a) it
  is still in production and so can be repaired by SRS, B) it has a
  built-in 1 pps locking circuit that can be used to lock to GPS and c)
  given (a) and the 22 year design, life span issues are not a significant
  problem. It also has several disadvantages - a) that same locking
  circuit may not always be present and B) that locking circuit does not
  perform as well as one would wish for GPS 1 pps signals (although
  pre-applying the sawtooth correction apparently helps a lot as does
  pre-filtering the GPS signal for an effectively larger time constant).

   3. The Efratom M-100 and FRM rubidium standards are also highly
  recommended. Their main disadvantages are a) the same repair issues that
  most of these units have, B) the potential life span limitations and c)
  they are smaller units and not as good as either the PRS-10 or HP 5065A.
  They do have the service manual available. I don't know about
  schematics.

   4. The LPRO-101 is far from the worst of the lot. As far as repair, there
  is a repair guide that people have put together but no official service
  manual with schematics.

   5. The FE-5650A / FE-5680A may also be acceptable, with the same
  disadvantages of life span and difficultly repairing but have very
  high noise levels.

So, as I interpret the above, if I can score a working HP 5065A for an
affordable cost then it would be the way to go. While possible, that is not
likely. Otherwise, a working PRS-10 with a functioning 1 pps input is very
good, again assuming affordable cost which is much more likely.

The ability to personally repair either may be limited. For the HP 5065A there
are no parts available - especially lamps - and the only way to fix one is to
use a parts unit. That gets very expensive and could take many months finding
an appropriate parts unit. For the PRS-10, firmware issues and the heavy use
of SMD would make repair more difficult. But, for it, at least factory repair
is available for a reasonable cost.

It may be possible to get some of the other rubidiums somewhat cheaper, but
all have their disadvantages. Most share the repair issue and life span issues
of the HP 5065A and the SMD construction of the PRS-10. Most of them would
also need to have their output cleaned up when used in an application
requiring low phase noise. Buying some of them in lots may enable repair via
parts units for a reasonable cost. That might be thing to do just to have some
as comparisons.

I plan on more research, especially with respect to the PRS-10, but with
respect to what is currently known, have I missed anything? Any additional
feedback? Other units that are worthy of consideration (and actually
available, of course)?

I have asked SRS about some of the PRS-10 issues. I will everybody know
what they say (assuming that they respond).


Mike


___
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] Best Rubidium Frequency Standard

[Charles Steinmetz]

jerry wrote:

Is there any model-suffix or other identifier that would inform 
whether a used PRS-10 has the PPS sync input feature??


Not that I am aware of.  Not even the "customer part number" is a 
reliable guide -- I have had several examples each of various CPNs, 
and some did have the sync input feature and some did not.  (I've 
also seen examples of other CPNs, *none* of which had the 
feature.  My sample hasn't been large enough to conclude that none of 
them does, however.)


Re: how useful the feature is, if it is present: the PRS10 PLL is 
extremely adjustable, and there are LOTS of combinations of settings 
that won't work worth a damn with any particular external input.  So, 
finding a good combination can be difficult, and finding the best 
combination can be elusive at best to maddening (or worse).  Read the 
manual carefully, and pay attention both to the default settings and 
to the hints that SRS provides.  With the right settings, a PRS10 
*does* work extremely well with the PPS input from a GPS.  They do 
generally take several days or more to lock, because of the long time 
constants involved.


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.

Re: [time-nuts] HP 5065A flaws and improvements

[Poul-Henning Kamp]

In message 

Re: [time-nuts] HP 5065A flaws and improvements

[Bob Camp]

Hi

> On Mar 12, 2016, at 4:56 PM, Stewart Cobb  wrote:
> 
> Bob Camp wrote:
> 
>> The 5065 has great ADEV numbers.
>> In “as delivered” condition it has horrid
>> TC and pressure sensitivity.
> 
> That's the first I've heard of any drawbacks to the 5065.  Can you give
> more details on this?

The manual talks about a 5x10^-11 temperature stability over 0-50C. Various 
members have reported data that suggests a stability about 10X worse than this. 
The manual spec is worse than what you see on most telecom Rb’s over that 
narrow a range. The actual performance compared to a telecom Rb in the same 
environment appears to be worse. 

Pressure sensitivity is not listed. All Rb’s are sensitive to pressure to some 
degree. Lab data from various experiments often shows up with a note “pressure 
change” to explain a bump in the data. 

Bob


> 
> The "as delivered" implies that there may be improvements available? Any
> details on improvements?
> 
> I'm aware of Corby's optical filter trick, which improves SNR and hence
> ADEV, but that would not affect sensitivity to temperature or pressure.
> 
> Cheers!
> --Stu
> ___
> 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] Best Rubidium Frequency Standard

[Bob Camp]

Hi


> On Mar 12, 2016, at 6:11 PM, Charles Steinmetz  wrote:
> 
> jerry wrote:
> 
>> Is there any model-suffix or other identifier that would inform whether a 
>> used PRS-10 has the PPS sync input feature??
> 
> Not that I am aware of.  Not even the "customer part number" is a reliable 
> guide -- I have had several examples each of various CPNs, and some did have 
> the sync input feature and some did not.  (I've also seen examples of other 
> CPNs, *none* of which had the feature.  My sample hasn't been large enough to 
> conclude that none of them does, however.)
> 
> Re: how useful the feature is, if it is present: the PRS10 PLL is extremely 
> adjustable, and there are LOTS of combinations of settings that won't work 
> worth a damn with any particular external input.  So, finding a good 
> combination can be difficult, and finding the best combination can be elusive 
> at best to maddening (or worse).  Read the manual carefully, and pay 
> attention both to the default settings and to the hints that SRS provides.  
> With the right settings, a PRS10 *does* work extremely well with the PPS 
> input from a GPS.  They do generally take several days or more to lock, 
> because of the long time constants involved.

I would call having to wait a few days for it to lock a bit of a disadvantage. 
Even more so for those with an antenna challenged environment that gives them 
dropouts every few hours.

Bob

> 
> 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] HP 5065A flaws and improvements

[Stewart Cobb]

Bob Camp wrote:

> The 5065 has great ADEV numbers.
> In “as delivered” condition it has horrid
> TC and pressure sensitivity.

That's the first I've heard of any drawbacks to the 5065.  Can you give
more details on this?

The "as delivered" implies that there may be improvements available? Any
details on improvements?

I'm aware of Corby's optical filter trick, which improves SNR and hence
ADEV, but that would not affect sensitivity to temperature or pressure.

Cheers!
--Stu
___
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] Best Rubidium Frequency Standard

[Bob Camp]

Hi

If you dig into the modern GPS Rb’s they put the 5065 to shame. They are also 
don’t seem to show up on eBay ….

Bob

> On Mar 12, 2016, at 4:50 PM, Bruce Griffiths  
> wrote:
> 
> The ultimate solution to this conundrum is to roll ones own Rubidium standard 
> to achieve performance beyond the HP5065A e.g.:
> https://doc.rero.ch/record/32317/files/2318.pdf
> 
> Bruce
> 
> On Saturday, March 12, 2016 01:46:29 PM Bob Camp wrote:
>> Hi
>> 
>> Another example of the “stored in magic memory” issues on some modern Rb’s:
>> 
>> To improve the yield of Rb’s cells, a DDS can be used to tune a bit one way
>> or the other off of the normal resonance frequency. The reasons why they
>> are off are a bit involved, the fact is they do get manufactured that way.
>> To line things up with these cells, some sort of coarse tune word is dumped
>> into the flash / eprom / whatever. That DDS now knows where to find the
>> resonance of the cell and it has. This all works fine until you swap a cell
>> between boards. The new cell may (or may not) line up with the old cell.
>> Things may (or may not) lock up properly.
>> 
>> Once you get past the “blown capacitor”, “no output on the regulator” or
>> “dead power transistor” sort of fixes …. the newer Rb’s are not
>> particularly easy to care for and feed. Indeed, the Temex units I mentioned
>> earlier probably all could be fixed with detailed work and access to the
>> code. I stop when the simple “replace these 4 caps” stuff does not bring
>> them to life. As long as that gets more than 60% running, I’m happy.
>> 
>> Bob
>> 
>>> On Mar 12, 2016, at 12:14 PM, cdel...@juno.com wrote:
>>> 
>>> Mike,
>>> 
>>> If you are planning to buy "suspect units" with the intent to repair I
>>> would steer clear of the PRS10.
>>> 
>>> It is true you can pay for a factory repair but having the schematics and
>>> theory of operation only helps for particular repairs.
>>> 
>>> This is because a lot of the alignment parameters are stored in memory
>>> and there are no instructions on these alignments and how to store them.
>>> 
>>> An example: I had a unit come in for repair. It had no output. The
>>> customer, a professor, said a student applied a large DC to the output.
>>> The output is direct from an RF transformer. It burnt open the winding.
>>> 
>>> I thought, how hard can it be?
>>> 
>>> I popped off a good output board from a junker and installed it. I now
>>> had an output but the unit would not operate properly. Part of the output
>>> go to other circuitry that would need to be aligned to match. No way to
>>> do that or store the alignment! Bummer.
>>> 
>>> Luckily the transformer was a metal can type. I opened it up and rewound
>>> the tiny wire output coil and that restored the unit to operation.
>>> 
>>> After that I sold my pile of defunct PRS10 units and don't plan to
>>> purchase any more.
>>> 
>>> For ease of repair the HP 5065A is first, then the FRKL and H, these have
>>> bog standard thru hole circuitry and the manuals are excellent. Another
>>> of the worst is the M100 as its cards are conformal coated and the lamp
>>> oven assy. is potted! As you mentioned most of the later telecomm units
>>> are surface mount and most have some sort of microprocessor involved.
>>> 
>>> Cheers,
>>> 
>>> Corby
>>> 
>>> ___
>>> 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] Best Rubidium Frequency Standard

[Bruce Griffiths]

The ultimate solution to this conundrum is to roll ones own Rubidium standard 
to achieve performance beyond the HP5065A e.g.:
https://doc.rero.ch/record/32317/files/2318.pdf

Bruce

On Saturday, March 12, 2016 01:46:29 PM Bob Camp wrote:
> Hi
> 
> Another example of the “stored in magic memory” issues on some modern Rb’s:
> 
> To improve the yield of Rb’s cells, a DDS can be used to tune a bit one way
> or the other off of the normal resonance frequency. The reasons why they
> are off are a bit involved, the fact is they do get manufactured that way.
> To line things up with these cells, some sort of coarse tune word is dumped
> into the flash / eprom / whatever. That DDS now knows where to find the
> resonance of the cell and it has. This all works fine until you swap a cell
> between boards. The new cell may (or may not) line up with the old cell.
> Things may (or may not) lock up properly.
> 
> Once you get past the “blown capacitor”, “no output on the regulator” or
> “dead power transistor” sort of fixes …. the newer Rb’s are not
> particularly easy to care for and feed. Indeed, the Temex units I mentioned
> earlier probably all could be fixed with detailed work and access to the
> code. I stop when the simple “replace these 4 caps” stuff does not bring
> them to life. As long as that gets more than 60% running, I’m happy.
> 
> Bob
> 
> > On Mar 12, 2016, at 12:14 PM, cdel...@juno.com wrote:
> > 
> > Mike,
> > 
> > If you are planning to buy "suspect units" with the intent to repair I
> > would steer clear of the PRS10.
> > 
> > It is true you can pay for a factory repair but having the schematics and
> > theory of operation only helps for particular repairs.
> > 
> > This is because a lot of the alignment parameters are stored in memory
> > and there are no instructions on these alignments and how to store them.
> > 
> > An example: I had a unit come in for repair. It had no output. The
> > customer, a professor, said a student applied a large DC to the output.
> > The output is direct from an RF transformer. It burnt open the winding.
> > 
> > I thought, how hard can it be?
> > 
> > I popped off a good output board from a junker and installed it. I now
> > had an output but the unit would not operate properly. Part of the output
> > go to other circuitry that would need to be aligned to match. No way to
> > do that or store the alignment! Bummer.
> > 
> > Luckily the transformer was a metal can type. I opened it up and rewound
> > the tiny wire output coil and that restored the unit to operation.
> > 
> > After that I sold my pile of defunct PRS10 units and don't plan to
> > purchase any more.
> > 
> > For ease of repair the HP 5065A is first, then the FRKL and H, these have
> > bog standard thru hole circuitry and the manuals are excellent. Another
> > of the worst is the M100 as its cards are conformal coated and the lamp
> > oven assy. is potted! As you mentioned most of the later telecomm units
> > are surface mount and most have some sort of microprocessor involved.
> > 
> > Cheers,
> > 
> > Corby
> > 
> > ___
> > 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] quartz thermometers

[Magnus Danielson]

Hi,

I think you meant to say that the B-mode is anywhere between +6 dB or -6 
dB from the C-mode.


Anyway, adjusting the 8,8 MHz and 10,7 MHz chains should help. Maybe put 
a 47 nF in replacement of C5 of 100 nF? Possibly lower the "high" (which 
will become low after the above mod) mode-suppressor somewhat.


Cheers,
Magnus

On 03/12/2016 07:27 PM, Bob Camp wrote:

Hi

The “suppression” of the B mode is variable. Depending on the exact rotation 
and contour and plating on the blank, it can be anywhere from 6 db above the 
desired B mode to 6 db below. Early crystals ran in the “above” range. As the 
world got better at working out the variables, crystals generally run in the 
“below” range today.

Early on, putting an oscillator on the C mode was indeed as simple as dropping 
it into a standard AT cut third overtone circuit (you still need the 
fundamental trap). Up came the unit on the “thermometer” B mode. It was simple 
enough and (at the time) obvious enough .. no need to comment on how to do it.

With a modern crystal, randomly chosen from a lot, getting onto the B mode is 
fairly tough. You need to reverse the selectivity of the trap. The trap (or 
something like it) is still in there to keep the few percent of crystals that 
need it happy. The issue being that when something “goes off” a whole batch 
need the trap. That’s a big hole in the production schedule.

Bob


On Mar 12, 2016, at 11:24 AM, Richard (Rick) Karlquist  
wrote:

Does anyone have a documented way to do this?

I asked the original designers of the 10811
(Burgoon and Wilson) about this a few years
ago and they couldn't remember what the circuit
mods were.  I got hold of their lab notebooks
and it wasn't in there either.  What I remember
them telling me 35 years ago was that you
remove the complicated mode suppressor between
the base and the emitter and replace it with
the equivalent capacitor.  I tried this, and
guess what:  it still oscillated in mode C!
All these years no one knew the mode suppressor
wasn't necessary.  I tried to redesign the
mode suppressor to suppress mode C.  There
wasn't any good way to make this work.  I
finally had to finagle the tuned circuit
in the collector to force it to mode B.

My E1938A circuit does not use a mode suppressor
as such.  The natural selectivity of the
tank circuit is sufficient.

Can anyone add anything?

Rick

On 3/12/2016 5:33 AM, Magnus Danielson wrote:

Hi,

For the 10811 you can modify it to change mode and then use that mode to
measure and trim the temperature oven.

There exists crystal oscillators where the 10 MHz is a traditional
SC-cut mode and then a 30 MHz mode is exercised which measures the
crystal temperature. In the Microprocessor Controlled Crystal Oscillator
(MCXO) one then measure the difference in frequency and uses this to
re-synthesize a correction on the 10 MHz. The benefit is that it is the
temperature of the oscillating crystal that is being measured.
Naturally, it could be used for oven control and/or EFC control too.

The MCXOs exists in manufacturing, but whenever you ask about them they
just wonder what military project are you working on.

I'd love to experiment with this form of temperature sensing one day,
when I have time... if that ever happens...

Cheers,
Magnus

On 03/12/2016 10:21 AM, ken hartman wrote:

Interestingly, the use of AC-cut crystals (high linear tempco of
frequency)
is found in the development of OCXOs. Using a reference AC-cut
resonator -
in place of the final AT/SC resonator - one can learn much about the
thermal  characteristics of the oven loop performance. While not a
precise
temp sensor, it is a high sensitivity  indicator of  temperature
variations
of the resonator.

On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins 
wrote:


It may be that the need for that kind of resolution died out.

The next step up from quartz thermometry is resistance thermometry.
The linearization equation for platinum has enough terms to make it
uncertain around .01 C.
Temperature calibration baths usually use platinum resistance sensors.

It may be that the triple point of water does not have the certainty to
reach '0.0001C'

Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
as an employee.

Bill Hawkins


-Original Message-
From: Alan Ambrose
Sent: Friday, March 11, 2016 11:42 AM

Hi,

I hope this is still relevant and not too off-topic...but since it
involves crystals and tempco...

Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
and '0.0001C resolution' seem to have been a thing from the mid-60's to
the mid-80's:

http://www.hparchive.com/Journals/HPJ-1965-03.pdf

There still appear to be some manufacturers making the crystals:

http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf

Anyone know why they died out? Did a better technology replace them?

TIA, Alan

___
time-nuts mailing list -- 

Re: [time-nuts] Best Rubidium Frequency Standard

[Jerry]

Charles,

Is there any model-suffix or other identifier that would inform whether a used 
PRS-10 has the PPS sync input feature??

jerry

-Original Message-
From: time-nuts [mailto:time-nuts-boun...@febo.com] On Behalf Of Charles 
Steinmetz
Sent: Saturday, March 12, 2016 1:23 AM
To: Discussion of precise time and frequency measurement
Subject: Re: [time-nuts] Best Rubidium Frequency Standard

Mike wrote:

>4. The PRS-10 can handle the GPS synchronization simply by feeding the
>   GPS 1 pps signal to it. That eliminates a lot of secondary 
> effort needed to
>   get the disciplining up and running. That would minimize both the cost
>   and effort. Control of both the PRS-10 and an LEA U-Blox can 
> be done via
>   an USB to RS-232 adapter (or two). This feature is something that none
>   of the alternative rubidiums appear to have.

NOTE if you are buying a used PRS10:  Like many oscillator products, 
the PRS10 has been sold in numerous "special" versions (with specific 
features as ordered by the OEM customers).  Many of these special 
versions do NOT have the PPS sync input feature.

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] Best Rubidium Frequency Standard

[Bob Camp]

Hi

Another example of the “stored in magic memory” issues on some modern Rb’s:

To improve the yield of Rb’s cells, a DDS can be used to tune a bit one way or 
the other off of the normal resonance frequency. The reasons why they are off 
are a bit involved, the fact is they do get manufactured that way. To line 
things up with these cells, some sort of coarse tune word is dumped into the 
flash / eprom / whatever. That DDS now knows where to find the resonance of the 
cell and it has. This all works fine until you swap a cell between boards. The 
new cell may (or may not) line up with the old cell. Things may (or may not) 
lock up properly. 

Once you get past the “blown capacitor”, “no output on the regulator” or “dead 
power transistor” sort of fixes …. the newer Rb’s are not particularly easy to 
care for and feed. Indeed, the Temex units I mentioned earlier probably all 
could be fixed with detailed work and access to the code. I stop when the 
simple “replace these 4 caps” stuff does not bring them to life. As long as 
that gets more than 60% running, I’m happy.

Bob

> On Mar 12, 2016, at 12:14 PM, cdel...@juno.com wrote:
> 
> Mike,
> 
> If you are planning to buy "suspect units" with the intent to repair I
> would steer clear of the PRS10.
> 
> It is true you can pay for a factory repair but having the schematics and
> theory of operation only helps for particular repairs.
> 
> This is because a lot of the alignment parameters are stored in memory
> and there are no instructions on these alignments and how to store them.
> 
> An example: I had a unit come in for repair. It had no output. The
> customer, a professor, said a student applied a large DC to the output.
> The output is direct from an RF transformer. It burnt open the winding.
> 
> I thought, how hard can it be?
> 
> I popped off a good output board from a junker and installed it. I now
> had an output but the unit would not operate properly. Part of the output
> go to other circuitry that would need to be aligned to match. No way to
> do that or store the alignment! Bummer.
> 
> Luckily the transformer was a metal can type. I opened it up and rewound
> the tiny wire output coil and that restored the unit to operation.
> 
> After that I sold my pile of defunct PRS10 units and don't plan to
> purchase any more.
> 
> For ease of repair the HP 5065A is first, then the FRKL and H, these have
> bog standard thru hole circuitry and the manuals are excellent. Another
> of the worst is the M100 as its cards are conformal coated and the lamp
> oven assy. is potted! As you mentioned most of the later telecomm units
> are surface mount and most have some sort of microprocessor involved.
> 
> Cheers,
> 
> Corby
> 
> ___
> 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] Best Rubidium Frequency Standard

[Bert Kehren via time-nuts]

Corby
the M100 have conformal coatings but non of the M100's that I have tested  
had potted Lamp modules.
Conformal coating is a pain but so far all we have worked on is the coil  
section and scraper and brass brush did do the job.
 will send you off list pictures of a bad disassembled lamp  module.
Bert Kehren
 
 
In a message dated 3/12/2016 1:00:53 P.M. Eastern Standard Time,  
cdel...@juno.com writes:

Mike,

If  you are planning to buy "suspect units" with the intent to repair I
would  steer clear of the PRS10.

It is true you can pay for a factory repair  but having the schematics and
theory of operation only helps for particular  repairs.

This is because a lot of the alignment parameters are stored  in memory
and there are no instructions on these alignments and how to  store them.

An example: I had a unit come in for repair. It had no  output. The
customer, a professor, said a student applied a large DC to the  output.
The output is direct from an RF transformer. It burnt open the  winding.

I thought, how hard can it be?

I popped off a good  output board from a junker and installed it. I now
had an output but the  unit would not operate properly. Part of the output
go to other circuitry  that would need to be aligned to match. No way to
do that or store the  alignment! Bummer.

Luckily the transformer was a metal can type. I  opened it up and rewound
the tiny wire output coil and that restored the  unit to operation.

After that I sold my pile of defunct PRS10 units and  don't plan to
purchase any more.

For ease of repair the HP 5065A is  first, then the FRKL and H, these have
bog standard thru hole circuitry and  the manuals are excellent. Another
of the worst is the M100 as its cards  are conformal coated and the lamp
oven assy. is potted! As you mentioned  most of the later telecomm units
are surface mount and most have some sort  of microprocessor  involved.

Cheers,

Corby

___
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] Best Rubidium Frequency Standard

[James Robbins]

Any one have any experience with the Symmetricom Militarized 8130A Rubidium?  
An update to a M100?  Thanks.

Jim Robbins
N1JR
___
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] quartz thermometers

[ken hartman]

I believe - if I understand correctly - that Magnus refers to realizations
based on what has been referred to as "the Schodowski patent"

Dual mode quartz thermometric sensing device

SS Schodowski
US Patent 4,872,765

as referenced here:

Resonator self-temperature-sensing using a dual-harmonic-mode crystal
oscillator

SS Schodowski
Frequency Control, 1989., Proceedings of the 43rd Annual Symposium on, 2-7

Not easy to do, but truly a measure of resonator temperature variations.





On Sat, Mar 12, 2016 at 7:33 AM, Magnus Danielson <
mag...@rubidium.dyndns.org> wrote:

> Hi,
>
> For the 10811 you can modify it to change mode and then use that mode to
> measure and trim the temperature oven.
>
> There exists crystal oscillators where the 10 MHz is a traditional SC-cut
> mode and then a 30 MHz mode is exercised which measures the crystal
> temperature. In the Microprocessor Controlled Crystal Oscillator (MCXO) one
> then measure the difference in frequency and uses this to re-synthesize a
> correction on the 10 MHz. The benefit is that it is the temperature of the
> oscillating crystal that is being measured.
> Naturally, it could be used for oven control and/or EFC control too.
>
> The MCXOs exists in manufacturing, but whenever you ask about them they
> just wonder what military project are you working on.
>
> I'd love to experiment with this form of temperature sensing one day, when
> I have time... if that ever happens...
>
> Cheers,
> Magnus
>
>
> On 03/12/2016 10:21 AM, ken hartman wrote:
>
>> Interestingly, the use of AC-cut crystals (high linear tempco of
>> frequency)
>> is found in the development of OCXOs. Using a reference AC-cut resonator -
>> in place of the final AT/SC resonator - one can learn much about the
>> thermal  characteristics of the oven loop performance. While not a precise
>> temp sensor, it is a high sensitivity  indicator of  temperature
>> variations
>> of the resonator.
>>
>> On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins 
>> wrote:
>>
>> It may be that the need for that kind of resolution died out.
>>>
>>> The next step up from quartz thermometry is resistance thermometry.
>>> The linearization equation for platinum has enough terms to make it
>>> uncertain around .01 C.
>>> Temperature calibration baths usually use platinum resistance sensors.
>>>
>>> It may be that the triple point of water does not have the certainty to
>>> reach '0.0001C'
>>>
>>> Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
>>> as an employee.
>>>
>>> Bill Hawkins
>>>
>>>
>>> -Original Message-
>>> From: Alan Ambrose
>>> Sent: Friday, March 11, 2016 11:42 AM
>>>
>>> Hi,
>>>
>>> I hope this is still relevant and not too off-topic...but since it
>>> involves crystals and tempco...
>>>
>>> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
>>> and '0.0001C resolution' seem to have been a thing from the mid-60's to
>>> the mid-80's:
>>>
>>> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>>>
>>> There still appear to be some manufacturers making the crystals:
>>>
>>> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>>>
>>> Anyone know why they died out? Did a better technology replace them?
>>>
>>> TIA, Alan
>>>
>>> ___
>>> 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] quartz thermometers

[Bob Camp]

Hi

The “suppression” of the B mode is variable. Depending on the exact rotation 
and contour and plating on the blank, it can be anywhere from 6 db above the 
desired B mode to 6 db below. Early crystals ran in the “above” range. As the 
world got better at working out the variables, crystals generally run in the 
“below” range today. 

Early on, putting an oscillator on the C mode was indeed as simple as dropping 
it into a standard AT cut third overtone circuit (you still need the 
fundamental trap). Up came the unit on the “thermometer” B mode. It was simple 
enough and (at the time) obvious enough .. no need to comment on how to do it.

With a modern crystal, randomly chosen from a lot, getting onto the B mode is 
fairly tough. You need to reverse the selectivity of the trap. The trap (or 
something like it) is still in there to keep the few percent of crystals that 
need it happy. The issue being that when something “goes off” a whole batch 
need the trap. That’s a big hole in the production schedule. 

Bob

> On Mar 12, 2016, at 11:24 AM, Richard (Rick) Karlquist 
>  wrote:
> 
> Does anyone have a documented way to do this?
> 
> I asked the original designers of the 10811
> (Burgoon and Wilson) about this a few years
> ago and they couldn't remember what the circuit
> mods were.  I got hold of their lab notebooks
> and it wasn't in there either.  What I remember
> them telling me 35 years ago was that you
> remove the complicated mode suppressor between
> the base and the emitter and replace it with
> the equivalent capacitor.  I tried this, and
> guess what:  it still oscillated in mode C!
> All these years no one knew the mode suppressor
> wasn't necessary.  I tried to redesign the
> mode suppressor to suppress mode C.  There
> wasn't any good way to make this work.  I
> finally had to finagle the tuned circuit
> in the collector to force it to mode B.
> 
> My E1938A circuit does not use a mode suppressor
> as such.  The natural selectivity of the
> tank circuit is sufficient.
> 
> Can anyone add anything?
> 
> Rick
> 
> On 3/12/2016 5:33 AM, Magnus Danielson wrote:
>> Hi,
>> 
>> For the 10811 you can modify it to change mode and then use that mode to
>> measure and trim the temperature oven.
>> 
>> There exists crystal oscillators where the 10 MHz is a traditional
>> SC-cut mode and then a 30 MHz mode is exercised which measures the
>> crystal temperature. In the Microprocessor Controlled Crystal Oscillator
>> (MCXO) one then measure the difference in frequency and uses this to
>> re-synthesize a correction on the 10 MHz. The benefit is that it is the
>> temperature of the oscillating crystal that is being measured.
>> Naturally, it could be used for oven control and/or EFC control too.
>> 
>> The MCXOs exists in manufacturing, but whenever you ask about them they
>> just wonder what military project are you working on.
>> 
>> I'd love to experiment with this form of temperature sensing one day,
>> when I have time... if that ever happens...
>> 
>> Cheers,
>> Magnus
>> 
>> On 03/12/2016 10:21 AM, ken hartman wrote:
>>> Interestingly, the use of AC-cut crystals (high linear tempco of
>>> frequency)
>>> is found in the development of OCXOs. Using a reference AC-cut
>>> resonator -
>>> in place of the final AT/SC resonator - one can learn much about the
>>> thermal  characteristics of the oven loop performance. While not a
>>> precise
>>> temp sensor, it is a high sensitivity  indicator of  temperature
>>> variations
>>> of the resonator.
>>> 
>>> On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins 
>>> wrote:
>>> 
 It may be that the need for that kind of resolution died out.
 
 The next step up from quartz thermometry is resistance thermometry.
 The linearization equation for platinum has enough terms to make it
 uncertain around .01 C.
 Temperature calibration baths usually use platinum resistance sensors.
 
 It may be that the triple point of water does not have the certainty to
 reach '0.0001C'
 
 Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
 as an employee.
 
 Bill Hawkins
 
 
 -Original Message-
 From: Alan Ambrose
 Sent: Friday, March 11, 2016 11:42 AM
 
 Hi,
 
 I hope this is still relevant and not too off-topic...but since it
 involves crystals and tempco...
 
 Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
 and '0.0001C resolution' seem to have been a thing from the mid-60's to
 the mid-80's:
 
 http://www.hparchive.com/Journals/HPJ-1965-03.pdf
 
 There still appear to be some manufacturers making the crystals:
 
 http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
 
 Anyone know why they died out? Did a better technology replace them?
 
 TIA, Alan
 
 ___
 time-nuts 

Re: [time-nuts] Best Rubidium Frequency Standard

[Bert Kehren via time-nuts]

Having used FRK for over 35 years I am partial to FRK/M100. More  later.
We have had some bad experience with PRS-10 failing lamp module and others  
have shared the same observation. Clear sign of oxidation and we know the 
unit  has never been exposed to water!
Back to FRK. Previously mentioned bigger is better also heatsink on all  
Rb's are a must Fan control is better. The biggest drawback of FRK and M100 is 
 availability of lamps. At one time you could buy them for $ 300. I have 
spares  for our team. If you buy one make sure lamp voltage is at least 8 
Volt. Many of  the M 100's have not seen continuous duty.
The question is what do you want. We are frequency nuts and use FRK and  
M100 with active Fan control and keep the backplate to 0.01 C. Typical aging 
is  below 5 E-14 per day over a 600 day period. An other plus is to the best 
of my  knowledge M 100 have the only measured pressure sensitivity test 
data. Data  would also apply to FRK's.
Quite a few time-nuts have controlled these Rb's with a Shera contoller I  
included. The weakness of Shera is the AD1861 audio DAC. At one time we did 
a  detailed analysis of available DAC's, went as far as designing and 
building  a test board capable of auto testing multiple DAC's using the LTC2400 
and  decided for Rb's the best  DAC is the 16 bit LTC1655.
If not M100/FRK with what is out there I would go with Skip's RDR   Fe5650. 
(no commercial interest) We have a couple of them. From what I can tell  
there is a high degree of similarity with the 5680A different packaging. There 
 is a lot of data available, there was at one time a FE5680A frenzy after I 
did  my first post based on some recommendation from Germany where they 
enjoyed  popularity at least a year before us. What happened to our German time 
nuts?  What comes now out of China I would not touch, my second order was 
so bad I had  to return it. I doubt that Skip's units ever saw China, he may 
want to comment  on it. As I mentioned in my first post and confirmed 
subsequently there are  frequency spikes on the output and a cleanup is 
mandatory, 
I have a plot it  was done with a HP5065A and time pod, maybe  byTom? 
Cooling is low cost  with a Laptop heatpipe/fan. Frequency control can be done 
with RS232 or a GPSDO.  Cash Olsen is ready to ship first set of units for $65 
plus shipping.
Bert Kehren
 
 
In a message dated 3/12/2016 11:02:12 A.M. Eastern Standard Time,  
kb...@n1k.org writes:

Hi

The  PRS-10 generally shows up in versions that do not have 
the external pps  input working. Even on the ones that *do* have 
a pps, it is not one that  works well with the pps from a GPS. Simply
put, that option is not worth  spending money on. 

Manually adjusting an Rb against a GPS is a bit  time consuming, but 
fairly easy to do.You need observations spread over  days and you need
to keep track of what you have done. Once adjusted it  should run quite 
well for a long time. Weekly / monthly / quarterly checks  should be all it 
needs.

All modern Rb’s are based on firmware.  different versions have different 
code in them. Swapping those files  around is *not* often successful. Repair
can easily involve parts that no  longer are made. A lot of PRS-10’s simply 
get  scrapped…



For $129 you can get a batch of 5 Temex Rb’s.  They probably will need two
dollars worth of caps replaced. You likely will  find that one or two of 
them work
“as is”. 

Bob


> On  Mar 11, 2016, at 8:50 PM, time...@metachaos.net wrote:
> 
>  Bob,
> 
> After all of the excellent feedback here, my current  thinking is that I 
can
> get a parts / repair PRS-10 for somewhere near  $150. There are several 
reasons
> that this might be the best  option...
> 
>   1. It appears to be about the best that  I can afford from what I have 
seen
>  of the testing  that others have done. The HP 4065A is completely 
out of
> the question. The only one I see available is $1340, plus shipping.  A
>  brand new PRS-10 is $1495. So the choice there  would be a very old 
unit
>  that might work for a few  months or even a few years, but probably 
not
>  very  long - with no parts available without using a parts unit. Or, 
a
>   unit nearly as good that has an expected life span of 20+ years  and
>  which is still in production and would be  under warranty for the 
first
>  year or so.
>  
>   2. There are complete schematics available for the PRS-10  which gives 
me a
>  leg up on attempting a repair.  There are also detailed circuit 
operational
>   descriptions in the service manual. That helps too.
>  
>   3. Worst case, SRS has a repair service available at, at  least in the 
past,
>  a reasonable cost (reportedly a  fixed price of $200). Now you need 
an
>  RMA quote,  but the price probably hasn't changed that much. If the
> repair is something that I can't handle - e.g. the rubidium tube is  
bad,
>  then it can still be repaired for a total  cost of around $350 to 
$400.
> 

[time-nuts] Best Rubidium Frequency Standard

[cdelect]

Mike,

If you are planning to buy "suspect units" with the intent to repair I
would steer clear of the PRS10.

It is true you can pay for a factory repair but having the schematics and
theory of operation only helps for particular repairs.

This is because a lot of the alignment parameters are stored in memory
and there are no instructions on these alignments and how to store them.

An example: I had a unit come in for repair. It had no output. The
customer, a professor, said a student applied a large DC to the output.
The output is direct from an RF transformer. It burnt open the winding.

I thought, how hard can it be?

I popped off a good output board from a junker and installed it. I now
had an output but the unit would not operate properly. Part of the output
go to other circuitry that would need to be aligned to match. No way to
do that or store the alignment! Bummer.

Luckily the transformer was a metal can type. I opened it up and rewound
the tiny wire output coil and that restored the unit to operation.

After that I sold my pile of defunct PRS10 units and don't plan to
purchase any more.

For ease of repair the HP 5065A is first, then the FRKL and H, these have
bog standard thru hole circuitry and the manuals are excellent. Another
of the worst is the M100 as its cards are conformal coated and the lamp
oven assy. is potted! As you mentioned most of the later telecomm units
are surface mount and most have some sort of microprocessor involved.

Cheers,

Corby

___
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] quartz thermometers

[Richard (Rick) Karlquist]

Does anyone have a documented way to do this?

I asked the original designers of the 10811
(Burgoon and Wilson) about this a few years
ago and they couldn't remember what the circuit
mods were.  I got hold of their lab notebooks
and it wasn't in there either.  What I remember
them telling me 35 years ago was that you
remove the complicated mode suppressor between
the base and the emitter and replace it with
the equivalent capacitor.  I tried this, and
guess what:  it still oscillated in mode C!
All these years no one knew the mode suppressor
wasn't necessary.  I tried to redesign the
mode suppressor to suppress mode C.  There
wasn't any good way to make this work.  I
finally had to finagle the tuned circuit
in the collector to force it to mode B.

My E1938A circuit does not use a mode suppressor
as such.  The natural selectivity of the
tank circuit is sufficient.

Can anyone add anything?

Rick

On 3/12/2016 5:33 AM, Magnus Danielson wrote:

Hi,

For the 10811 you can modify it to change mode and then use that mode to
measure and trim the temperature oven.

There exists crystal oscillators where the 10 MHz is a traditional
SC-cut mode and then a 30 MHz mode is exercised which measures the
crystal temperature. In the Microprocessor Controlled Crystal Oscillator
(MCXO) one then measure the difference in frequency and uses this to
re-synthesize a correction on the 10 MHz. The benefit is that it is the
temperature of the oscillating crystal that is being measured.
Naturally, it could be used for oven control and/or EFC control too.

The MCXOs exists in manufacturing, but whenever you ask about them they
just wonder what military project are you working on.

I'd love to experiment with this form of temperature sensing one day,
when I have time... if that ever happens...

Cheers,
Magnus

On 03/12/2016 10:21 AM, ken hartman wrote:

Interestingly, the use of AC-cut crystals (high linear tempco of
frequency)
is found in the development of OCXOs. Using a reference AC-cut
resonator -
in place of the final AT/SC resonator - one can learn much about the
thermal  characteristics of the oven loop performance. While not a
precise
temp sensor, it is a high sensitivity  indicator of  temperature
variations
of the resonator.

On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins 
wrote:


It may be that the need for that kind of resolution died out.

The next step up from quartz thermometry is resistance thermometry.
The linearization equation for platinum has enough terms to make it
uncertain around .01 C.
Temperature calibration baths usually use platinum resistance sensors.

It may be that the triple point of water does not have the certainty to
reach '0.0001C'

Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
as an employee.

Bill Hawkins


-Original Message-
From: Alan Ambrose
Sent: Friday, March 11, 2016 11:42 AM

Hi,

I hope this is still relevant and not too off-topic...but since it
involves crystals and tempco...

Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
and '0.0001C resolution' seem to have been a thing from the mid-60's to
the mid-80's:

http://www.hparchive.com/Journals/HPJ-1965-03.pdf

There still appear to be some manufacturers making the crystals:

http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf

Anyone know why they died out? Did a better technology replace them?

TIA, Alan

___
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] quartz thermometers

[Larry McDavid]

In case you missed the significance of Tom's clarification below, note 
that the thermal conductivity of helium is 6 times that of air and the 
heat capacity of helium is 5 times that of air. These are big differences!


We once built a GCMS for NASA; two were landed on Mars. Modules of the 
GCMS were heated to 225° C in operation by various foil and wire wound 
heaters. Tested in one atmosphere of air or in 25 millibar CO2 (Mars 
atmosphere), these heaters worked well. At one point during organic 
cleaning, I happened to run the heaters in helium at slightly lower than 
atmospheric pressure; the heaters could not be heated to 225° C because 
of the characteristics of helium.


Helium has unusual properties!

Larry


On 3/12/2016 3:03 AM, Tom Van Baak wrote:

Neville,

Apparently it's not vacuum mounted, but helium filled. Lots of good info here:

http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1965-03.pdf

/tvb

- Original Message -
From: "Neville Michie" 
To: "Discussion of precise time and frequency measurement" 
Sent: Saturday, March 12, 2016 2:19 AM
Subject: Re: [time-nuts] quartz thermometers



It always puzzled me that quartz crystals would be considered prime temperature 
sensors.
I can see that an instrument could be built that reliably showed many decimal 
places of reading,
but I could never accept that a vacuum mounted quartz crystal would be closely 
enough
thermally coupled to whatever was having its temperature measured.

...

--
Best wishes,

Larry McDavid W6FUB
Anaheim, California  (SE of Los Angeles, near Disneyland)
___
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] quartz thermometers

[Bob Camp]

Hi

The conventional MCXO uses an AT cut crystal rather than an SC. It runs on the 
fundamental and the third overtone. You can build one with just about any 
normal fundamental crystal. There are also people doing OCXO’s sort of the same 
way. Back when the MCXO came out, it was about the only way to get a TCXO that 
had < 1x10^-8 sort of stability. Now you can get that with a cheap (as in not 
$5,000 each) commercial IC based TCXO. 

The weak point in the MCXO is the “dual frequency” oscillator. If one part in 
one arm ages, that “pulls” the frequency difference. Any change in the 
frequency difference has the potential of messing up the compensation. The 
whole process of making sure that the deltas are minimized is a major focus in 
their design and manufacture. 

Bob

> On Mar 12, 2016, at 8:33 AM, Magnus Danielson  
> wrote:
> 
> Hi,
> 
> For the 10811 you can modify it to change mode and then use that mode to 
> measure and trim the temperature oven.
> 
> There exists crystal oscillators where the 10 MHz is a traditional SC-cut 
> mode and then a 30 MHz mode is exercised which measures the crystal 
> temperature. In the Microprocessor Controlled Crystal Oscillator (MCXO) one 
> then measure the difference in frequency and uses this to re-synthesize a 
> correction on the 10 MHz. The benefit is that it is the temperature of the 
> oscillating crystal that is being measured.
> Naturally, it could be used for oven control and/or EFC control too.
> 
> The MCXOs exists in manufacturing, but whenever you ask about them they just 
> wonder what military project are you working on.
> 
> I'd love to experiment with this form of temperature sensing one day, when I 
> have time... if that ever happens...
> 
> Cheers,
> Magnus
> 
> On 03/12/2016 10:21 AM, ken hartman wrote:
>> Interestingly, the use of AC-cut crystals (high linear tempco of frequency)
>> is found in the development of OCXOs. Using a reference AC-cut resonator -
>> in place of the final AT/SC resonator - one can learn much about the
>> thermal  characteristics of the oven loop performance. While not a precise
>> temp sensor, it is a high sensitivity  indicator of  temperature variations
>> of the resonator.
>> 
>> On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins  wrote:
>> 
>>> It may be that the need for that kind of resolution died out.
>>> 
>>> The next step up from quartz thermometry is resistance thermometry.
>>> The linearization equation for platinum has enough terms to make it
>>> uncertain around .01 C.
>>> Temperature calibration baths usually use platinum resistance sensors.
>>> 
>>> It may be that the triple point of water does not have the certainty to
>>> reach '0.0001C'
>>> 
>>> Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
>>> as an employee.
>>> 
>>> Bill Hawkins
>>> 
>>> 
>>> -Original Message-
>>> From: Alan Ambrose
>>> Sent: Friday, March 11, 2016 11:42 AM
>>> 
>>> Hi,
>>> 
>>> I hope this is still relevant and not too off-topic...but since it
>>> involves crystals and tempco...
>>> 
>>> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
>>> and '0.0001C resolution' seem to have been a thing from the mid-60's to
>>> the mid-80's:
>>> 
>>> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>>> 
>>> There still appear to be some manufacturers making the crystals:
>>> 
>>> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>>> 
>>> Anyone know why they died out? Did a better technology replace them?
>>> 
>>> TIA, Alan
>>> 
>>> ___
>>> 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] Best Rubidium Frequency Standard

[Bob Camp]

Hi

The PRS-10 generally shows up in versions that do not have 
the external pps input working. Even on the ones that *do* have 
a pps, it is not one that works well with the pps from a GPS. Simply
put, that option is not worth spending money on. 

Manually adjusting an Rb against a GPS is a bit time consuming, but 
fairly easy to do.You need observations spread over days and you need
to keep track of what you have done. Once adjusted it should run quite 
well for a long time. Weekly / monthly / quarterly checks should be all it 
needs.

All modern Rb’s are based on firmware. different versions have different 
code in them. Swapping those files around is *not* often successful. Repair
can easily involve parts that no longer are made. A lot of PRS-10’s simply 
get scrapped…



For $129 you can get a batch of 5 Temex Rb’s. They probably will need two
dollars worth of caps replaced. You likely will find that one or two of them 
work
“as is”. 

Bob


> On Mar 11, 2016, at 8:50 PM, time...@metachaos.net wrote:
> 
> Bob,
> 
> After all of the excellent feedback here, my current thinking is that I can
> get a parts / repair PRS-10 for somewhere near $150. There are several reasons
> that this might be the best option...
> 
>   1. It appears to be about the best that I can afford from what I have seen
>  of the testing that others have done. The HP 4065A is completely out of
>  the question. The only one I see available is $1340, plus shipping. A
>  brand new PRS-10 is $1495. So the choice there would be a very old unit
>  that might work for a few months or even a few years, but probably not
>  very long - with no parts available without using a parts unit. Or, a
>  unit nearly as good that has an expected life span of 20+ years and
>  which is still in production and would be under warranty for the first
>  year or so.
> 
>   2. There are complete schematics available for the PRS-10 which gives me a
>  leg up on attempting a repair. There are also detailed circuit 
> operational
>  descriptions in the service manual. That helps too.
> 
>   3. Worst case, SRS has a repair service available at, at least in the past,
>  a reasonable cost (reportedly a fixed price of $200). Now you need an
>  RMA quote, but the price probably hasn't changed that much. If the
>  repair is something that I can't handle - e.g. the rubidium tube is bad,
>  then it can still be repaired for a total cost of around $350 to $400.
> 
>  In that case the total cost might be a bit higher than a working unit.
>  Or, perhaps not. The two working TSD12s that I see on eBay (apparently
>  the same as the PRS-10 except the lock flag is set to 3 instead of 1 and
>  which can be reset) are $285 and $512. And I don't see any working
>  PRS-10s at all. If this option is necessary then the cost would be
>  split, making it easier to swallow. And, SRS calibrates the unit when
>  they fix it at no extra cost (as long as it is requested before 
> shipping).
> 
>   4. The PRS-10 can handle the GPS synchronization simply by feeding the GPS
>  1 pps signal to it. That eliminates a lot of secondary effort needed to
>  get the disciplining up and running. That would minimize both the cost
>  and effort. Control of both the PRS-10 and an LEA U-Blox can be done via
>  an USB to RS-232 adapter (or two). This feature is something that none
>  of the alternative rubidiums appear to have.
> 
>   5. The PRS-10 has free windows software to assist in programming and
>  controlling it.
> 
> The main negative is that I have never repaired an SMD product before. I don't
> know if my skills are up to it. However, I can practice on junk before I
> actually attempt to make a repair. And that is certainly a skill that will be
> needed in the future!
> 
> 
> Mike
> 
>> HI
> 
>> None of this is a simple slam dunk.
> 
>> The 5065 has great ADEV numbers. In “as delivered” condition It has horrid
>> TC and pressure sensitivity. It also is > 10X more expensive than a lot of 
>> the other devices.
> 
>> The units that *can* be disciplined are rarely set up to do so properly off
>> of a GPS source. If you want to run one as a 24/7 GPSDO, that may be a 
>> significant issue.
> 
>> All of the “modular” Rb’s require heat sinking. For a reasonable lifespan,
>> either a fairly healthy heatsink or  something smaller plus a fan is called
>> for. The impact on lifespan can easily be 10:1. 
> 
>> If your target is something like a microwave radio, many Rb’s are
>> “challenged” in terms of phase noise and/or spurs. Some sort of cleanup will 
>> be needed for almost all of them.
> 
>> The salvage process that some parts have been through is a bit brutal.
>> That’s both good news and bad news. It lets you buy a “kit of 5” Rb’s to
>> repair yourself for very little money. With luck you will get two or three
>> working and still not have spent $200 on the batch (delivered). On the other

Re: [time-nuts] MTI 260-0624-D OCXO

[Bob Camp]

Hi

Judging from your data, the pot takes the oven from at or above the inflection 
temperature of the crystal down to a point well below the crystal’s lower turn 
temperature. That’s about all it needs to do. 

Bob

> On Mar 11, 2016, at 11:19 PM, time...@metachaos.net wrote:
> 
>> Hi
> 
>> Anything electrical (duty cycle/ harmonics / output level / etc) 
>> is designed in on a modern part done in high volume. 
> 
>> The only candidates for a pot on top of the oven are:
> 
>> Inner oven temperature
>> Outer oven temperature
>> EFC
> 
>> You have already ruled out EFC (it would be a coarse set so
>> visible on a normal counter). 
> 
>> That leaves inner oven and outer oven. 
> 
>> Why inner oven - to adjust for the data from your temperature run. The parts
>> are at best coarse set based on crystal temperature. 
> 
>> Why outer oven - same as above or to set for inner / outer separation
>> and clearance after the thermal stabilization in the sealed can. 
> 
>> If there is no apparent change in frequency (and from the location of
>> the pot) that suggests the outer oven. Regardless of which oven is being 
>> set, the
>> adjustment is likely a “move it 1.25 turns" sort of thing.  You then 
>> re-run the temperature test to see if it is in spec yet. 
> 
>> Bob
> 
> 
> Bob,
> 
> In a separate communication, Angus has stated...
> 
>> On the two that I tested, it appeared to adjust the temperature of
>> both of the ovens.
> 
>> One was a 5 MHz 260-0545-B. The outer oven was at 78.2 degC and at
>> least around that point the temp changed at about 1 degC per turn. The
>> EFC adjustment on this one was not quite enough to bring the frequency
>> to exactly 5 MHz, but the oven temp did look to be close to correct.
> 
>> The other was a 16.384 MHz 260-0546-G which was unused old stock, but
>> had overall poor stability. It turned out that the oven temp on it was
>> 1-3/8 turns from the turnover point, but when or how that happened I
>> don't know.
> 
>> It would be interesting to hear from somebody who knows how this type
>> of thing is actually adjusted in production.
> 
> I finally got my 260 (sort of) repaired and have done some measurements.
> 
> Generally, I can confirm what Angus was saying, to the limits of my ability to
> measure. One of the things that I can measure is the source voltage. I have a
> hacked together 12v supply. When hooked up to the 260, it initially reads just
> over 9v. As it warms up, it eventually tops out just under 10v. There appears
> to be roughly four different points, around 9.1v, 9.35v, 9.8v and 10v. The 260
> starts at around 9.1v, quickly moves up to 9.35v and after a while to 9.8v. It
> will then sometimes move briefly up to 10v and back down to 9.8v.
> 
> I interpret this as different heating stages - probably both ovens on, one
> oven on, some sort of maintenance heating point and no heaters on.
> 
> What I find as I adjust the pot (after letting the 260 warm up) is that
> turning the pot counter clockwise the supply voltage will go up to 10v and
> turning it clockwise it will drop to 9.35v or 9.1v. So, from that it would
> appear that you are right in that the temperature set points are being
> adjusted by the pot. Whether only one is adjusted or both are changed as Angus
> says isn't clear.
> 
> I also noticed that the 260 consistently has an adjustment range of +/-2e-7,
> regardless of where the frequency is set, with 0v applied to the adjustment
> pin giving a frequency 4Hz higher than when the 5.65v reference voltage is
> applied to the adjustment pin. It will adjust further if the reference pin is
> hooked up to the supply voltage.
> 
> So, I decided to get as much information as I could. I turned the pot
> clockwise until it "ticked" once each turn. That dropped the temperature to
> the minimum. I then turned it counter clockwise 1/2 turn at a time and
> recorded each temperature / frequency reading. I did this until it "ticked"
> once each turn. That took nearly three days. To obtain the temperature, I
> used Kapton tape to tape a temperature sensor for my multimeter to the casing
> of the outer oven. This is not necessarily the hottest spot, nor does it
> actually measure the temperature in either the inner or outer oven. Further,
> the multimeter's precision is only around 1 degree C. The accuracy is unknown,
> but it should be repeatable and roughly linear. I used my (uncalibrated)
> 2465BCT to measure the frequency. This is limited to 7 digits without a
> calibrated frequency standard and is probably slightly off. Still, that gave
> me a good bit of data. Here is the raw data.
> 
>   605.01
>   615.02
>   615.03
>   625.03
>   625.04
>   635.05
>   635.06
>   645.06
>   655.07
>   655.08
>   665.08
>   665.09
>   675.09
>   675.10
>   685.10
>   695.10
>   705.11
>   715.11
>   725.12
>   735.12
>   74

Re: [time-nuts] Best Rubidium Frequency Standard

[timenut]

Joseph,

There are a few data points that lead to this conclusion...

   1. Look at the images on eBay for various PRS-10 and TSD-12 auctions. You
  will see that they all have the same customer part number 143-44101-xx
  where xx is 04, 08 or 10. Further, you will find a PRS-10 and a TSD-12
  with the exact same customer part number of 143-44101-08. This indicates
  that the TSD-12 is a "drop-in" replacement for the PRS-10 for that
  customer.

   2. In the timenut's archive, there is a user who has / had both. He
  reported that the TSD-12 self-reports as a PRS12, and additionally
  have the same firmware version number.

   3. The PRS-10 and TSD-12 are physically and electrically identical based on
  the various eBay pictures. There are two variations that I see. The
  first is that on apparently newer devices the back side has four
  mounting holes not present on the older devices. The second is that on
  some devices the Cannon connector is on the bottom rather that on the
  back end. All of the TSD versions (-11 and -12) appear to have the
  mounting holes. Like the PRS-10, some have the Cannon connector on the
  back end and some have it on the side.

   4. The same user reported that all of the PRS-10 devices he examined had
  the default locking mode of 1 and the TSD-12 devices had a default
  locking mode of 3. However, the LM command allows the locking mode to be
  set, and there is another command that will make that change to the
  power up state.

   5. Since the package is physically identical, the firmware is physically
  identical, if you look at the service manual, there are not a lot of
  things that could be changed or omitted. Almost everything is in
  software, including the 1 pps locking (which hijacks the trim pot to do
  its dirty work).

   6. The only option listed on the SRS site for new devices is an option that
  has a reduced aging rate (at an extra $495). There could be some other
  minor options like that, but effectively it appears as if the TSD-12 (or
  -11) is interchangable with the PRS-10 if the LM command is used to set
  the default locking mode.


Mike


>>   Or, perhaps not. The two working TSD12s that I see on eBay (apparently
>>   the same as the PRS-10 except the lock flag is set to 3 instead of 1 
>> and
>>   which can be reset)

> Would you elaborate on this? I Googled for TSD12 and except for the
> auction listings, found nothing.

> Joe Gray
> WW5JG



-- 
Best regards,
 Timenutmailto:time...@metachaos.net

___
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] quartz thermometers

[Bob Camp]

Hi

They were not cheap beasts to keep running. The cal process was an “HP only” 
thing and when you looked at what came back each year, any data you took over 
the last few months needed to be re-evaluated. After a while, people started 
poke into this and that. They would do a wonderful job sitting in a constant 
temperature bath. Like any crystal, they aged and this created some level of 
drift (well beyond  0.0001C per year). The more troublesome problem was 
hysteresis. If you cycled the probe between temperatures, it might take a day 
to get back “close” to the original temperature (but still not quite to 
0.0001C). The final gotcha was mechanical shock bumps in normal use. Apparently 
this also created some offsets as well.

All of that added up to a frequency vs temperature curve that “moved around” at 
a level well past 0.0001C. Each time you sent it back, the coefficients got 
re-modeled. Use if for a while and the numbers don’t match the probe so much 
anymore. Just as with a SPRT, a triple point cell really needed to be part of 
the setup. For whatever reason, HP backed out of the business rather than get 
more complex. Fluke took a different approach (they bought Hart)  and stayed in 
the business. There are certainly a number of other outfits making fancy 
temperature gear.  

Bob


> On Mar 11, 2016, at 12:41 PM, Alan Ambrose  wrote:
> 
> Hi,
> 
> I hope this is still relevant and not too off-topic...but since it involves 
> crystals and tempco...
> 
> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals and 
> '0.0001C resolution' seem to have been a thing from the mid-60's to the 
> mid-80's:
> 
> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
> 
> There still appear to be some manufacturers making the crystals:
> 
> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
> 
> Anyone know why they died out? Did a better technology replace them?
> 
> TIA, Alan
> ___
> 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] quartz thermometers

[Magnus Danielson]

Hi,

For the 10811 you can modify it to change mode and then use that mode to 
measure and trim the temperature oven.


There exists crystal oscillators where the 10 MHz is a traditional 
SC-cut mode and then a 30 MHz mode is exercised which measures the 
crystal temperature. In the Microprocessor Controlled Crystal Oscillator 
(MCXO) one then measure the difference in frequency and uses this to 
re-synthesize a correction on the 10 MHz. The benefit is that it is the 
temperature of the oscillating crystal that is being measured.

Naturally, it could be used for oven control and/or EFC control too.

The MCXOs exists in manufacturing, but whenever you ask about them they 
just wonder what military project are you working on.


I'd love to experiment with this form of temperature sensing one day, 
when I have time... if that ever happens...


Cheers,
Magnus

On 03/12/2016 10:21 AM, ken hartman wrote:

Interestingly, the use of AC-cut crystals (high linear tempco of frequency)
is found in the development of OCXOs. Using a reference AC-cut resonator -
in place of the final AT/SC resonator - one can learn much about the
thermal  characteristics of the oven loop performance. While not a precise
temp sensor, it is a high sensitivity  indicator of  temperature variations
of the resonator.

On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins  wrote:


It may be that the need for that kind of resolution died out.

The next step up from quartz thermometry is resistance thermometry.
The linearization equation for platinum has enough terms to make it
uncertain around .01 C.
Temperature calibration baths usually use platinum resistance sensors.

It may be that the triple point of water does not have the certainty to
reach '0.0001C'

Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
as an employee.

Bill Hawkins


-Original Message-
From: Alan Ambrose
Sent: Friday, March 11, 2016 11:42 AM

Hi,

I hope this is still relevant and not too off-topic...but since it
involves crystals and tempco...

Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
and '0.0001C resolution' seem to have been a thing from the mid-60's to
the mid-80's:

http://www.hparchive.com/Journals/HPJ-1965-03.pdf

There still appear to be some manufacturers making the crystals:

http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf

Anyone know why they died out? Did a better technology replace them?

TIA, Alan

___
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] quartz thermometers

[Tim Shoppa]

I swear I started a similar thread here on this mailing list, many years
ago, but google is not helping me find it. Maybe it was well over a decade
ago and on a Usenet group?

>From that thread, I remember learning that retrace/hysteresis quickly
becomes the limiting factor for quartz thermometers. e.g.
http://www.ieee-uffc.org/frequency-control/learning/Hysteresis_final.html

http://www.ieee-uffc.org/frequency-control/learning/hysteresis/fig_3.gif

Tim N3QE

On Fri, Mar 11, 2016 at 12:41 PM, Alan Ambrose 
wrote:

> Hi,
>
> I hope this is still relevant and not too off-topic...but since it
> involves crystals and tempco...
>
> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
> and '0.0001C resolution' seem to have been a thing from the mid-60's to the
> mid-80's:
>
> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>
> There still appear to be some manufacturers making the crystals:
>
> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>
> Anyone know why they died out? Did a better technology replace them?
>
> TIA, Alan
> ___
> 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] quartz thermometers

[Poul-Henning Kamp]

In message , Neville Michie wri
tes:

>but I could never accept that a vacuum mounted quartz crystal would
>be closely enough thermally coupled to whatever was having its
>temperature measured.

That is precisely why HP could make the three and four digit claim
for their quartz thermometers:  Compared to resitive temperature
sensors, practically no energy is dissipated in/by the crystal.

-- 
Poul-Henning Kamp   | UNIX since Zilog Zeus 3.20
p...@freebsd.org | TCP/IP since RFC 956
FreeBSD committer   | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
___
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] quartz thermometers

[Tom Van Baak]

Neville,

Apparently it's not vacuum mounted, but helium filled. Lots of good info here:

http://www.hpl.hp.com/hpjournal/pdfs/IssuePDFs/1965-03.pdf

/tvb

- Original Message - 
From: "Neville Michie" 
To: "Discussion of precise time and frequency measurement" 
Sent: Saturday, March 12, 2016 2:19 AM
Subject: Re: [time-nuts] quartz thermometers


> It always puzzled me that quartz crystals would be considered prime 
> temperature sensors.
> I can see that an instrument could be built that reliably showed many decimal 
> places of reading,
> but I could never accept that a vacuum mounted quartz crystal would be 
> closely enough 
> thermally coupled to whatever was having its temperature measured.
> Sensors like thermistors, thermocouples and platinum resistors can be made of 
> the right shape and size to 
> thermally couple to solids and liquids and so can make successful measurement 
> systems.
> Many important measurements rely on having no disturbance of the physical 
> system by the 
> use of the thermometer.
> 
> cheers,
> Neville Michie
> 
> 
>> On 12 Mar 2016, at 8:21 pm, ken hartman  wrote:
>> 
>> Interestingly, the use of AC-cut crystals (high linear tempco of frequency)
>> is found in the development of OCXOs. Using a reference AC-cut resonator -
>> in place of the final AT/SC resonator - one can learn much about the
>> thermal  characteristics of the oven loop performance. While not a precise
>> temp sensor, it is a high sensitivity  indicator of  temperature variations
>> of the resonator.
>> 
>> On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins  wrote:
>> 
>>> It may be that the need for that kind of resolution died out.
>>> 
>>> The next step up from quartz thermometry is resistance thermometry.
>>> The linearization equation for platinum has enough terms to make it
>>> uncertain around .01 C.
>>> Temperature calibration baths usually use platinum resistance sensors.
>>> 
>>> It may be that the triple point of water does not have the certainty to
>>> reach '0.0001C'
>>> 
>>> Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
>>> as an employee.
>>> 
>>> Bill Hawkins
>>> 
>>> 
>>> -Original Message-
>>> From: Alan Ambrose
>>> Sent: Friday, March 11, 2016 11:42 AM
>>> 
>>> Hi,
>>> 
>>> I hope this is still relevant and not too off-topic...but since it
>>> involves crystals and tempco...
>>> 
>>> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
>>> and '0.0001C resolution' seem to have been a thing from the mid-60's to
>>> the mid-80's:
>>> 
>>> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>>> 
>>> There still appear to be some manufacturers making the crystals:
>>> 
>>> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>>> 
>>> Anyone know why they died out? Did a better technology replace them?
>>> 
>>> TIA, Alan
>>> 
>>> ___
>>> 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] quartz thermometers

[Neville Michie]

It always puzzled me that quartz crystals would be considered prime temperature 
sensors.
I can see that an instrument could be built that reliably showed many decimal 
places of reading,
but I could never accept that a vacuum mounted quartz crystal would be closely 
enough 
thermally coupled to whatever was having its temperature measured.
Sensors like thermistors, thermocouples and platinum resistors can be made of 
the right shape and size to 
thermally couple to solids and liquids and so can make successful measurement 
systems.
Many important measurements rely on having no disturbance of the physical 
system by the 
use of the thermometer.

cheers,
Neville Michie


> On 12 Mar 2016, at 8:21 pm, ken hartman  wrote:
> 
> Interestingly, the use of AC-cut crystals (high linear tempco of frequency)
> is found in the development of OCXOs. Using a reference AC-cut resonator -
> in place of the final AT/SC resonator - one can learn much about the
> thermal  characteristics of the oven loop performance. While not a precise
> temp sensor, it is a high sensitivity  indicator of  temperature variations
> of the resonator.
> 
> On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins  wrote:
> 
>> It may be that the need for that kind of resolution died out.
>> 
>> The next step up from quartz thermometry is resistance thermometry.
>> The linearization equation for platinum has enough terms to make it
>> uncertain around .01 C.
>> Temperature calibration baths usually use platinum resistance sensors.
>> 
>> It may be that the triple point of water does not have the certainty to
>> reach '0.0001C'
>> 
>> Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
>> as an employee.
>> 
>> Bill Hawkins
>> 
>> 
>> -Original Message-
>> From: Alan Ambrose
>> Sent: Friday, March 11, 2016 11:42 AM
>> 
>> Hi,
>> 
>> I hope this is still relevant and not too off-topic...but since it
>> involves crystals and tempco...
>> 
>> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
>> and '0.0001C resolution' seem to have been a thing from the mid-60's to
>> the mid-80's:
>> 
>> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>> 
>> There still appear to be some manufacturers making the crystals:
>> 
>> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>> 
>> Anyone know why they died out? Did a better technology replace them?
>> 
>> TIA, Alan
>> 
>> ___
>> 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] quartz thermometers

[ken hartman]

Interestingly, the use of AC-cut crystals (high linear tempco of frequency)
is found in the development of OCXOs. Using a reference AC-cut resonator -
in place of the final AT/SC resonator - one can learn much about the
thermal  characteristics of the oven loop performance. While not a precise
temp sensor, it is a high sensitivity  indicator of  temperature variations
of the resonator.

On Fri, Mar 11, 2016 at 11:44 PM, Bill Hawkins  wrote:

> It may be that the need for that kind of resolution died out.
>
> The next step up from quartz thermometry is resistance thermometry.
> The linearization equation for platinum has enough terms to make it
> uncertain around .01 C.
> Temperature calibration baths usually use platinum resistance sensors.
>
> It may be that the triple point of water does not have the certainty to
> reach '0.0001C'
>
> Disclaimer: I only worked with industrial sensors from Rosemount, Inc.
> as an employee.
>
> Bill Hawkins
>
>
> -Original Message-
> From: Alan Ambrose
> Sent: Friday, March 11, 2016 11:42 AM
>
> Hi,
>
> I hope this is still relevant and not too off-topic...but since it
> involves crystals and tempco...
>
> Quartz thermometers (e.g. the HP 2804A) with their 'linear cut' crystals
> and '0.0001C resolution' seem to have been a thing from the mid-60's to
> the mid-80's:
>
> http://www.hparchive.com/Journals/HPJ-1965-03.pdf
>
> There still appear to be some manufacturers making the crystals:
>
> http://www.statek.com/products/pdf/Temp%20Sensor%2010162%20Rev%20B.pdf
>
> Anyone know why they died out? Did a better technology replace them?
>
> TIA, Alan
>
> ___
> 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] Best Rubidium Frequency Standard

[Charles Steinmetz]

Mike wrote:


   4. The PRS-10 can handle the GPS synchronization simply by feeding the
  GPS 1 pps signal to it. That eliminates a lot of secondary 
effort needed to

  get the disciplining up and running. That would minimize both the cost
  and effort. Control of both the PRS-10 and an LEA U-Blox can 
be done via

  an USB to RS-232 adapter (or two). This feature is something that none
  of the alternative rubidiums appear to have.


NOTE if you are buying a used PRS10:  Like many oscillator products, 
the PRS10 has been sold in numerous "special" versions (with specific 
features as ordered by the OEM customers).  Many of these special 
versions do NOT have the PPS sync input feature.


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.

Re: [time-nuts] quartz thermometers

[Poul-Henning Kamp]

In message , Alan Ambrose writes:

>Anyone know why they died out? Did a better technology replace them?

Worse but cheaper technology replaced them.

-- 
Poul-Henning Kamp   | UNIX since Zilog Zeus 3.20
p...@freebsd.org | TCP/IP since RFC 956
FreeBSD committer   | BSD since 4.3-tahoe
Never attribute to malice what can adequately be explained by incompetence.
___
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.