Magnus and Azelio,
(About Pendulum or Fluke or Philips PM6681 Counter or equivalent CNT-81)
Here's a link to the thread where Magnus shared info in 2015.
https://time-nuts.febo.narkive.com/6WTFfsyN/pm6680-or-53131a-for-timepod
Your post is about half way down -- 2015-11-18 22:18:04 UTC
Yesterday I dug out my GPIB-capable PC and sent a couple commands to my
Fluke PM6681.
First I tried a basic one:
*IDN?
and got
PHILIPS, PM6681, 0, MAIN V1.09 26 JAN 2001 / GPIB V1.13 26 JAN 2001
So connection is good. Interestingly the *IDN command description says
the PM6681 will return its SN but the SN field here is 0. Oh well, not
important.
Then I sent *PUD?
and got
#261 CALIBRATED: 2006-11-07, CALPLS: 4.25 ns, TMP: +22 °C[LF]
where the [LF] at the end is not literal, it represents the line feed
char 0x0a.
So in addition to the CALPLS value, it looks like they also save the TMP
in centigrade when the test was run. I wonder if the counter uses that?
So I hope if my counter ever lost this cal value, I could send it this
command:
:SYST:UNPR; *PUD #261 CALIBRATED: 2006-11-07, CALPLS: 4.25 ns, TMP: +22 °C
I don't plan to try that now. If it ain't broke don't fix it.
There is one odd thing I see though. The last two of #261 is supposed to
say the string length is 61. But it isn't. I count it as 53 chars. I
don't know if this matters but the counter gave that number to me. In
the Programming Manual description page for *PUD, it gives a couple
examples and the #2nn values shown do have lengths that match their
string lengths.
If it is useful to anyone, I made a version of just the *PUD command
description from the Programming Manual and put it here:
www.xertech.net/pm6681/PUD_cmd.pdf
I also made a version of the Interpolater calibration process page from
the Service Manual. It can't really be used since it is obsolete and the
old DOS program seems unobtanium. It may give a few hints what they were
up to.
I put it here:
www.xertech.net/pm6681/interpolate_cal.pdf
So thanks for pointing out that the *PUD command saved string is what
you lose if the memory backup battery dies. Reading and saving the value
is what I hoped for and now I've done it.
If anyone has a PM6681 counter or equivalent with the "Cal.Lost"
message, sending my string above might be good enough to get it working,
though maybe not optimum.
On 11/9/2020 1:58 AM, Magnus Danielson wrote:
Thanks for the memory refresh.
You can read the string using PUD?
Do that and keep the result. PUD and PUD? is the magic in the counter,
the rest is software and hardware outside of the counter for calibration.
In order to write, you need to move the calibration jumper inside.
Cheers,
Magnus
On 2020-11-08 23:01, Azelio Boriani wrote:
Old story about the PM6681 (18 Nov 2015, thread: "PM6681 and Timelab")
where a sort of calibration procedure is described: the PUD command is
NOT a calibration command.
PM6681 programming manual, page 9-127: PUD Protected User Data...This
is a data area where the user may write ANY data up to 64
characters...
If the user can write any data, how can it be a calibration command or
calibration data area?
Better watch out those 3V coin cells, we will never get the real
calibration commands/procedure. I have tried with the disassembled
firmware, no way. The visible strings of GPIB commands are all
described in the programming manual, so nothing useful.
On Sat, Nov 7, 2020 at 3:31 AM Magnus Danielson <mag...@rubidium.se> wrote:
Hi Rex,
I need to dig in the archive to refresh my memory. I don't recall
precisely, but I think I recalled that the manual indirectly describes
the calibration data string.
I have learned a few things from Pendulum, but I did not have the right
tools at hand to set things up.
There was a more recent setup that could use more modern generators, but
the trick was still the same. You lock the generator and counter to the
same frequency, then you set the generator to a small offset frequency
from 10 MHz, which is 9.999 MHz as I recall it. This slowly sweeps
through all the phase-relationships between the reference oscillator and
the counter input, thus sweeps the interpolator phase. It then chooses
the calibration constant giving the lowest RMS error, as this is the
best compensation for the hardware min-point. All this is free from
memory. Then that value with calibration date is written into memory. If
I recall correctly 2.21 ns is a typical value.
I have PM6681 in need of calibration, and as I recall it I was able to
program it enough for the calibration error warning did not show up.
This not to say it was actually calibrated.
At some point I will return to that project. The generator I used did
not support that offset frequency, but I have others that do. Also, my
crapiola GPIB programming needs attention. My intention is to share the
fruits of this project when it comes to that. The lab has been in
shambles for too long, but shaping up slowly and nicely.
Cheers,
Magnus
On 2020-11-05 09:30, Rex wrote:
Hi Magnus,
Just catching up on list messages and saw this one from you.
I have a Fluke PM 6881 counter. I don't think I've ever seen a
description of a method for reading/restoring these battery backed up
calibration constants. I looked for a way, as losing them is something
I've worried about. Not that it has happened and I did replace the
battery once.
Is doing this described in one of the manuals? Sounds like it is
through GPIB? I'd greatly appreciate any pointers to info or other
details you might provide.
I did see, in the service manual, a short description of a method for
Interpolator calibration that seems to be for making these
calibrations. Seems if the saved cal values get lost, the counter will
display "CaL.LOSt". The cal procedure is driven by an old DOS program
(that I've never found) and requires a: PM5768 Pulse gen, PM5193 LF
Sig Gen, good 10 MHz, all GPIB controlled from the program. Never saw
more details but sounds messy. If there is description of GPIB
commands for reading/setting cal values, I missed them.
thanks for mentioning this and anything more you can provide
-Rex
On 10/29/2020 5:37 PM, Magnus Danielson wrote:
Hi,
I second this. You can read the calibration data out of the counter and
save. I've done some experiments with that, but nothing conclusive, but
I blame my lack of patience and not a proper setup.
Do replace the battery, it is cheap and relatively easy to do.
Would you loose this calibration, through a little GPIB commands one can
write a fake value in. This will however not produce the best resu. The
calibration routine actually runs an off beat frequency and then test
different values, and look for least RMS value, because it is the
calibration point. I've not had time to replicate all that, but I did
manage to write the fake value in and at one time get rid of the CAL
LOST warning.
Cheers,
Magnus
On 2020-10-29 14:17, Azelio Boriani wrote:
For those who have the PM6681 (aka CNT81): check the 3V memory backup
cell and replace it before the dreaded calibration lost (CAL LOST)
will appear on the LCD. Replace the coin cell with great care (with
the counter powered up), see the service manual for the procedure.
<https://archive.org/details/FLUKE_PM6681_Service_Manual>
On Fri, Oct 23, 2020 at 6:17 PM Magnus Danielson
<mag...@rubidium.se> wrote:
Hi,
On 2020-10-22 19:13, Attila Kinali wrote:
On Thu, 22 Oct 2020 11:50:08 +0000
Giorgio Barinetti <gior...@barinetti.it> wrote:
Choices are many, but I'll try to avoid the "older" machines lile
5370 or 5335. The 531xx series seems nice ( money apart )
But again : which one between the 3 ? 53131, 53132 or 53181 ?
Maybe try to get hold of one of the Philips (later licensed to
Fluke) PM6680 or PM6681? These are more common in Europe than
in the US, so the big US dominated websites/forums/.. don't
mention them that often. Solid devices that can be had as low
as 300€ if you are willing to wait, 500-800€ is the usual going
price. The SR620 is the workhorse that drives a lot of the
time and frequency metrology worldwide and can be had new and
used (new on http://thinksrs.com goes for 800-2000€ used).
If you go for a new one, I would consider looking at the
Pendulum CNT-90 and CNT-91. (Pendulum is the company that
took over Philips frequency counter business and the CNT-90
is the continuation of the PM668x line, also sold as PM6690
by Fluke)
Let me correct on the history and geniology there.
Philips had a instrument making side called Philips Industrier
Järfälla
that did a range of measurement instruments. Later they joined forces
with Fluke. Later Philips felt that the business unit was a bad fit to
stay in Philips, so they sold it off to become a separate company
which
became Pendelum. Pendelum was really the business unit with people etc
through that process, and the Fluke relation and rebranding continued.
Naturally Pendelum moved out of the Philips Industrier Järfälla office
over to Bälstabro (both locations in north of Stockholm) as it was
sold
off. Pendelum also managed to rebrand their counters to Tektronix,
which
mainly consisted of cosmetic changes to get the look and feel.
Pendelum
was operated for many years like this, some of their production in
Pajala, where as other where done in Bälstabro. Later they reshaped
the
production so that it moved to Poland where it remains. Pendulum was
sold to Spectracom and was operated as a subsidary for a while, until
they shut operation down.
The CNT-80/81 (PM 6680 and PM6681) production went on as long as they
had the timing ASIC. The CNT-90 (100 ps) was developed to the
CNT-91 (50
ps), where the later replaced the CNT-81 (50 ps). They aimed to do the
CNT-92, but could not at that time do it with the same technical
setup.
They also had the Wander Meter WM-10 which aided in testing telecom
sync. After some testing, I suggested they would broaden the
product to
handle more signals and that is when they mostly firmware upgraded
it to
the WM-11, and first time I tested the WM-11 the second '1' was cut
out
from another '10' print and put over the '0'. Also, it was still under
development. Several years later they where still going back to my
list
of feedback. They later did a revamped this into a new instrument with
even improved capabilities such as interfaces boards. This was later
sold off to Calnex which sells it as Sentinel, which is a great
product
for telecom operators.
I still have contact with some of the Philips/Fluke/Pendelum staff and
owner.
Cheers,
Magnus
_______________________________________________
time-nuts mailing list -- time-nuts@lists.febo.com
To unsubscribe, go to
http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
and follow the instructions there.
