HP3458 - OHMF and DELAY vs APER and NPLC
Hi Attila,
I don't know if you saw the last couple messages I sent a while back, or
if they didn't show up on list. Here's one again, just in case. The
other will follow if I can find it - it especially addresses some of the
resistor issues.
Yes, those are very nice cables - I looked them up. But, they are still
banana plugs, relying on spring contact. From the description, I'd guess
there are two contacts in series - one from the plug core to the banana
or basket spring, and the other from there to the inner surface of the
jack. Another type is machined directly from springy material, with
cross cuts, and profiled to compress on insertion - these have only one
contact involved, so can be the best of the banana plug class. The fancy
materials and plating help greatly with contact issues, but may still
not be enough. I didn't see any specs on the datasheet other than "low
EMF," so I wonder if maybe they offer related low EMF jack/binding posts
too. Maybe the combination is specified with numbers. Also, what does HP
say in the 3458A manual about interconnect and fixturing?
Regarding dielectric absorption, what is the reference node? If you
connect the wire pairs at the resistor, there is very low resistance and
voltage drop between them at the "open" (3458A) end, so I would think
the insulation between them shouldn't matter too much. The other part of
dielectric absorption would be from each wire pair to the other, and to
everything else beyond. This should be dominated mostly by very small
(due to distance of separation) air capacitance, in series with the the
C from the conductors to the insulation/air boundary. I suppose that you
may be at the resolution level where the effects are noticeable, but my
gut feel is that the wire insulation characteristics are still minute in
scale. The main thing is to maximize the distance between the wire
pairs, and to everything else.
It appears the measuring process includes AC, with the various auto-cals
and zeroing going on, so the resistor and wiring would be jumping to
different voltages in the process. Is it possible to just set up for
pure DC, with constant current in the resistor? This may help to
separate the capacitive from DC effects. Speaking of timing, as you
noted, the few seconds of effects are about right for dielectric
absorption, but also a good profile for thermal or contact effects too.
Especially if the test current to the resistor is being turned on and
off during the process, everything may respond over some time span in
seconds.
One last thing occurred to me. DMMs typically have a switch or post to
connect the low input (guard) to earth ground. Have you tried different
settings for this? I would think grounded should be best, but maybe not.
Maybe you have some subtle ground loops in the system, that could add DC
offsets or AC from the activity. If the grounding choice makes a
difference, check the CMRR specs. That invokes one more last, last thing
- the DUT itself. I don't know what it looks like, but I'd guess it's in
a metal can. If so, there's of course parasitics to it from the
resistor. Is the can grounded, guarded, or isolated? Could there be
significant leakage current into the can? Could the can be filled with
tar or other potting material? this could give you yet another
dielectric absorption issue to worry about.
Ed
On 1/30/2022 4:59 AM, Attila Kinali wrote:
Hoi Ed,
On Fri, 28 Jan 2022 15:02:46 -0800
ed breya eb@telight.com wrote:
I'm not familiar with the 3458A, so don't know how the connections are
made up front. Presuming it's working right, I would suspect the
interconnect setup. The thermal voltages and contact noise may be the
main problem, especially using any sort of banana plugs.
These are not "any sort" of banana plugs. These are low-EMF from Pomona.
I doubt that wires would perform significantly better. But it's worth
a shot.
Put the
source current and high sense on one piece, and the low source and sense
on the other. In operation, the Kelvin connections tie each twin lead
pair together at the resistor, so even crappy insulation won't matter.
This I doubt. The effect I am seeing looks awfully like what I would
expect from a settling effect that takes a few seconds. And dielectric
absorption is one of those that are exactly in that time-range.
Before changing anything around, a little experiment may shed some
light. Instead of doing resistance measurements, set it to measure the
voltage on the resistor, leaving the whole setup as it was. In the most
sensitive ranges, with all those digits, I'd think you'll see the
effects of the input bias current, the thermal voltages, and some
resistor noise. The thermals should be easy to spot - just puff some air
at the connections, or wave your hand around.
That's a good idea. I'll set it up next.
Attila Kinali
volt-nuts mailing list -- volt-nuts@lists.febo.com -- To unsubscribe
send an email to volt-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
Hi Attila,
Here's the other one.
I got to thinking some more about the DUT, and what's inside, and the
parasitics. I found ref [2] that you showed in the OP, about the HP
resistance standards. It's very informative, covering a lot of the
construction details. One thing that struck me is the discussion of
transient performance beginning on page 12, noting that chemical
contamination in older (non-HP) types is known to cause an
electro-chemical cell effect that can last for many seconds. I think you
said your resistor is an old type, similar to the HP ones. I wonder if
maybe this is the cause of trouble.
To eliminate all possible DUT problems, I'd suggest setting it up with
just a regular, medium precision 10 K resistor, like a T9 with wire
leads. For quick, temporary use, you can scrape off the lead plating
down to bare copper, then tightly twist the lead and copper wires to the
3458A, then crimp a splice lug on to crush them together for a gas-tight
seal. Even a (steel-insert) wire nut will do to hold it together for a
while, as long as there's good Cu-Cu contact between all the parts. Then
you can suspend the little DUT up away from everything, and mostly
eliminate the parasitics, packaging, and other issues. If it works with
this, but not the original DUT, you'll know what the deal is.
Ed
volt-nuts mailing list -- volt-nuts@lists.febo.com -- To unsubscribe
send an email to volt-nuts-leave@lists.febo.com
To unsubscribe, go to and follow the instructions there.
Hoi Ed,
On Sat, 12 Feb 2022 14:07:28 -0800
ed breya eb@telight.com wrote:
I don't know if you saw the last couple messages I sent a while back, or
if they didn't show up on list. Here's one again, just in case. The
other will follow if I can find it - it especially addresses some of the
resistor issues.
Yes, I have seen them, thank you.
I am still in the process of confirming or ruling out the points
you have mentioned.
As I wrote before, changing the cables for some PTFE insulated ones
did not change anything. Grounding I have not yet tried, but I doubt
it will do anything. The settling of the resistor is indeed something
that is going on, but I am not quite sure what the process there ought
to be or how long its time constant is. I have also indication that
different DELAY settings with APER 1 might not lead to that different
results, if I measure continuously (instead of doing just once every 3h
as I have been doing). But I need more data.
I also have some indication that ACAL is responsible for quite a bit
of variations between measurements. The cause and extend of which
I cannot currently say, due to lack of data.
Attila Kinali
--
The driving force behind research is the question: "Why?"
There are things we don't understand and things we always
wonder about. And that's why we do research.
-- Kobayashi Makoto
I have measured TCR of 4040B unit I have here.
It's not great, came out about +8.15ppm/K alpha, -0.57ppm/K^2 beta.
Published my results and data here: https://xdevs.com/review/ln4040b/
Included photo of wiring connection for reference.
On 2/12/2022 2:15 PM, Attila Kinali wrote:
Moin,
On Wed, 9 Feb 2022 00:50:58 -0500
Illya Tsemenko illya@xdevs.com wrote:
Any improvement on your 3458A results?
A bit. I now know what isn't the cause: The cable and EMF.
I found some silver plated, PTFE insulated AWG 30 wire (AlphaWire 2841/1)
and substituted it for the Pomona leads (Yes, not ideal, but that's the
only PTFE insulated wire I could find).
I did the same run of different NPLC, APER and DELAY settings as before
and the numbers are virtually the same. Additionally I also did a run with
APER 1 and larger DELAY numbers to see how that changes things (plot attached)
On some more long term measurement, the behaviour looks virtually the same.
There is no change that wouldn't be within the noise of the measurement.
I also covered everything in cardboard to shield it from all air currents.
While this inproved my temperature readings quite a bit (noise went down by
a factor of 5-7), it did not change the ohm reading noise at all (again,
all within noise/uncertainty).
But things are slow as the whole thing needs hours to stabilize. So each
time I change anything in the setup or the way how I measure, I let it
run for a couple of days to see what it did.
I've got L&N 4040-10k resistor and logged it for a bit last night.
This standard was just random ebay special, nothing fancy.
Coincidentally, that's the same type of resistor I am measuring.
To mitigate tempco standard was placed in peltier active air bath,
programmed at +23.00°C.
I whish I could do that as well, but my current budget does not allow
to buy a commercial air/oil bath and fire regulations prevent me from
using a home-brewn one.
Overall noise of readout is within ±0.35ppm peak to peak.
That's about 1/3 of the readout noise I have for the same settings.
My current working hypothesis is, that the resistor has some significant
temperature gradient/change based behaviour, which would explain some of
the large jumps I see, which seem to be always around when temperature was
changing quickly (as in faster than 0.01-0.02°C/h).
I will further investigate and report back once I have a better understanding
of what is going on.
Attila Kinali
Good afternoon!
On Wed, 9 Feb 2022 00:50:58 -0500
Illya Tsemenko illya@xdevs.com wrote:
Any improvement on your 3458A results? I've got L&N 4040-10k resistor
and logged it for a bit last night. This standard was just random ebay
special, nothing fancy.
After a few months of measurment I can report the following things:
My HP 1kΩ and the L&N 100Ω resistor both behave like they should
in terms of temperature coefficient (within the bounds of my
measurement accuracy) and in terms of DELAY, APER and NPLC.
The L&N 10kΩ resistor, while its tempco seems to be about right,
has a long settling time. Even with a 10s DELAY, and a 1s APER
measurement, the measured resistance still varies by about 0.3ppm
when changing DELAY +/-1s. While the HP 1kΩ and L&N 100Ω do not
show any such behaviour.
Does that mean that my L&N 10kΩ is defect? Or is this a general
problem with that type of resistance standards?
Or in other words: Would it be worth it to get another L&N 10kΩ
off ebay or would I face the same problems with that one as well?
Other 10kΩ resistance standards seem to be quite rare or so
expensive that I'd rather buy a new one from Fluke or IET.
Another interesting tidbit is, that the first resistance measurement
after an ACAL (a complete one) can be off by as much as 10ppm in
either direction. All subsequent ones are within the expected
noise band. At least with APER, haven't done enough measurements
with NPLC to say anything about that.
Attila Kinali
--
The driving force behind research is the question: "Why?"
There are things we don't understand and things we always
wonder about. And that's why we do research.
-- Kobayashi Makoto
Moin, Moin Attila,
first off, as you obviously descend from Northern Germany, and are currently located @ MPI Saarbrücken, I propose that you afterwards send me your 10kOhm standard for characterization, in my cool lab near Frankfurt. Out there it was up to 38°C, downstairs, only stable 22.4°C.
I think I’m able to determine its resistance @ 22°C with ppm uncertainty, and 0.2ppm StD @ 16 samples of APER 1.
1:1 transfer uncertainty is about 0.1 .. 0.2ppm. FLUKE has validated these characteristics in their 3458AHFL datasheet.
I’m performing such stable and precise 4W, OCOMP measurements regularly with my 3458A, on 6 precision 10KOhm resistors, as well on a 1kOhm.
I have read the whole thread.. sorry that I respond very late as I was locked out due to a wrong address.
There are several pitfalls in your setup and with the basic principles of Offset Compensation methods in general for 10k and 100k Ohm range resistance measurements.
There are a few guys who already have explained some of these effects on the eevblog forum, I cite two threads here:
https://www.eevblog.com/forum/metrology/hp3458a-ocomp-delay-related-problems/msg1465111/#msg1465111
Here are several topics I’d like to explain briefly.
-
emf voltages and cables
The Offset Compensation method (of 3458A or other long range DMMs) is intended to remove all emfs.
Therefore, it’s not necessary at all to use low emf cables, jacks, aor resistors. Manganine has higher emfs compared to others, so no problem.
The discussion here is misleading concerning the root cause of your noise and instability problems -
PVC cables have too low an isolation resistance. For 10kOhm, these introduce errors on the order of ppms when you drill plus and minus, but 10ppms when measuring 100kOhm.
-
The 3458A, as well as other DMMs are quite susceptible to external noise in your lab.
Therefore, you MUST use drilled 4W cables, which implies using PTFE.
These MUST have a shield around, which is connected to GUARD of the DMM (not grounded), and best be connected to the case of your standard resistor. -
You have to delete all noise sources in your lab, i.e. all switch mode power supplies in lamps, PSUs, chargers, etc.
Also get rid of any wireless emitters.
I assume, that is the main source of your instabilities -
You need to do your measurements with the 3458A at a very constant room temperature. My basement lab is constant to +/- 0.2°C over a week or so. Therefore I can perform DCV measurements stable to 0.1ppm over 24h before I have to redo an ACAL. I did not yet try such Ohm stability measurements, but this mode is definitely much less stable.
-
The OCOMP switches the measuring current on and off, as explained in those two threads. This leads to polarization currents, or dielectric absorption, every time when when the current is switched. That leads to a lowering of the measurement by at least 1ppm for 10k and 5ppm for 100k, depending on the cables / isolation used. Therefore, a delay of 1 sec for 10k and 5sec for 100k is necessary. This has to be tested. Increase the delay from zero, until the measured value does not increase any more.
1k and lower, as well as 1M and higher should be made with small delay of 30msec, as no such effect arises. These are universal parameters, also valid for all other DMMs with a similar OCOMP function.
The optimal setting of the 3458A can be found in the links. It’s basically APER 1, DELAY 1 in your case.
usage of NPLC 50, instead is counter productive, as this only increases the measurement time unnecessarily. -
use statistics, limited to 16 samples for one data point, see links.
The mean value will meet the transfer uncertainty of 0.1ppm and stability ~ Standard Deviation of 0.2ppm.
If you use much more samples, medium term instabilities kick in, which have nothing to do with your test problem.
These are temperature and timely drifts of your DMM and of your DUT. Think of the Allan Deviation statistics, what it is about. -
Characterize your 3458A for Time and T.C. stability of its resistance mode.
The 40k resistor of my 3458A has < 0.3ppm/K T.C., and is stable to a few ppm over several years. I already have a VHP101 inside, maybe your old instrument is not as stable.
Anyhow, the 3458A is no reference standard, therefore, you should make long term stability measurements in relation to another external reference (Volt, Ohm)Frank
Moin Frank!
On Thu, 21 Jul 2022 18:58:08 +0000
frank.stellmach@freenet.de wrote:
first off, as you obviously descend from Northern Germany,
Actually.. I'm not :-P
and are currently located @ MPI Saarbrücken, I propose that you
afterwards send me your 10kOhm standard for characterization, in
my cool lab near Frankfurt. Out there it was up to 38°C,
downstairs, only stable 22.4°C.
Thanks a lot for the offer. I might take you up on this.
But I don't think it's a temperature issue. The 10kΩ resistor
takes ages to settle. I've been using APER 1, DELAY 1,
AZERO ON, OCOMP ON and varied DELAY from 1 to 10. Between DELAY 1
and DELAY 10 is almost 10ppm of difference. To get below 0.1ppm
I would need a DELAY value in the order of 20 to 30. While that
works for me, it's bascially a no-go if I want to get the
resistor calibrated. The two other resistors do not show such
a settling. Repsectively, if they do, it's below the uncertainties
of my measurement setup. Even using NPLC 50 with DELAY 0.5 instead
of APER 1 does not appreciably change the measured value.
I have read the whole thread.. sorry that I respond very late as I
was locked out due to a wrong address.
No worries. I'm greatful for all help!
There are several pitfalls in your setup and with the basic
principles of Offset Compensation methods in general for 10k and
100k Ohm range resistance measurements.
There are a few guys who already have explained some of these
effects on the eevblog forum, I cite two threads here:
I'm aware of these issues and read quite a bit of the discussions
on the eevblog forums. I also tried to quantify the influence of
the differnt switches.
Here are several topics I’d like to explain briefly.
- emf voltages and cables
The Offset Compensation method (of 3458A or other long range DMMs) is intended to remove all emfs.
Therefore, it’s not necessary at all to use low emf cables, jacks, aor resistors. Manganine has higher emfs compared to others, so no problem.
The discussion here is misleading concerning the root cause of your noise and instability problems
Yes. I switched to AWG28 PTFE wires and saw no effect whatsoever.
So, neither the EMF nor the dielectric absorption was an issue.
- PVC cables have too low an isolation resistance. For 10kOhm, these introduce errors on the order of ppms when you drill plus and minus, but 10ppms when measuring 100kOhm.
Intersting, I haven't thought about that. Though, as noted above,
the use of PTFE insulated wires didn't make a difference for me.
- The 3458A, as well as other DMMs are quite susceptible to external noise in your lab.
Therefore, you MUST use drilled 4W cables, which implies using PTFE.
These MUST have a shield around, which is connected to GUARD of the DMM (not grounded), and best be connected to the case of your standard resistor.
While the room I'm in is not the quietest, the noise level I see
in my measurements are compatible with what the HP3458 datasheet
says. So I don't think I have an issue there. And on top of that,
I'm averaging over minutes to hours for each "reading". So most
of that noise should average out below the other uncertainties.
- You need to do your measurements with the 3458A at a very
constant room temperature. My basement lab is constant to +/-
0.2°C over a week or so. Therefore I can perform DCV measurements
stable to 0.1ppm over 24h before I have to redo an ACAL. I did not
yet try such Ohm stability measurements, but this mode is
definitely much less stable.
Yes, that's one of the big problems I have. My "lab" is a spare
office nobody is using at the moment, in an 1990s building with
lots of glass and open spaces but no air conditioning. On top of
that I have multiple instruments running, using a total of probably
1kW in a small, unventilated room. That's why I have such high
temperatures. There is no basement room that I could use.
But I'm actually glad that I got even a separate room and could
cordon it off. I'm at an institute of theoretical computer science
after all. I'm one of the few people there, for whom a computer is
not just a black box on a piece of paper. Heck, it took me months
to get it into the heads of people that nobody should enter that room
not even the cleaning lady, because I can see that in my measurements.
To combat the temperature issue a bit (and for a handful other
reasons) I run a full ACAL every 3h. I see quite a bit of temperature
coefficient on the relevant values for resistance measurement, but
everything within what I would expect.
- The OCOMP switches the measuring current on and off, as
explained in those two threads. This leads to polarization
currents, or dielectric absorption, every time when when the
current is switched. That leads to a lowering of the measurement
by at least 1ppm for 10k and 5ppm for 100k, depending on the
cables / isolation used.
Even OCOMP off and AZERO on lead to the same effect here. And
not just 1ppm, but >10ppm for the 10kΩ. Independent of the cables
used.
Therefore, a delay of 1 sec for 10k and
5sec for 100k is necessary. This has to be tested. Increase the
delay from zero, until the measured value does not increase any
more.
1k and lower, as well as 1M and higher should be made with
small delay of 30msec, as no such effect arises. These are
universal parameters, also valid for all other DMMs with a similar
OCOMP function.
The optimal setting of the 3458A can be found in the links. It’s basically APER 1, DELAY 1 in your case.
usage of NPLC 50, instead is counter productive, as this only increases the measurement time unnecessarily.
At least for the 1kΩ and the 100Ω resistors, the difference between
NPLC 50, DELAY 0.5 and APER 1, DELAY 1 is below measurement
uncertainty with multiple hours of averaging. The additional time
needed for NPLC 50 with DELAY 0.5 (about 30%) is well compensated
by the 6dB lower noise. For the 10kΩ resistor.. well..
- use statistics, limited to 16 samples for one data point, see links.
The mean value will meet the transfer uncertainty of 0.1ppm and stability ~ Standard Deviation of 0.2ppm.\
I do apply statistics, not in the 3458, but outside. This
way I can more accurately evaluate things and compensate for
thermal effects and other things that would otherwise get hiden
in the statistics.
If you use much more samples, medium term instabilities kick
in, which have nothing to do with your test problem.
These are temperature and timely drifts of your DMM and of your
DUT. Think of the Allan Deviation statistics, what it is about.
Yes, I'm well aware of these.
- Characterize your 3458A for Time and T.C. stability of its
resistance mode.
The 40k resistor of my 3458A has < 0.3ppm/K T.C., and is stable
to a few ppm over several years. I already have a VHP101 inside,
maybe your old instrument is not as stable.\
This I would like to do, but unfortunately, I lack the equipment
to do so. I would need to be able to put the 3458 into a temperature
chamber, while keeping a resistor standard in a different camber.
This is equiment I can't get here, where there is no budget for
anything beyond pen and paper and glorified typewriters.
Anyhow, the 3458A is no reference standard, therefore, you
should make long term stability measurements in relation to
another external reference (Volt, Ohm)
That's how I started. I wanted to do some precision measurements,
for which I needed some standards. So I got some cheap resistance
standards off ebay. To see whether they are not some duds and drift
too much themselves, I wanted to use my 3458, that I know has been
treated well and shouldn't drift too much, as "the standard" and
see whether the resistance standards are at least within what's
plausible.
Well, the results so far are, that the 1kΩ and the 100Ω are
probably ok and the 10kΩ has some weird effect that I can't
pin down.
Attila Kinali
--
The driving force behind research is the question: "Why?"
There are things we don't understand and things we always
wonder about. And that's why we do research.
-- Kobayashi Makoto
Hello Attila,
I have understood all of your description, which ruled out a lot of possible sources of error, but what really disturbs me is this passage:
The 10kΩ resistor takes ages to settle. I've been using APER 1, DELAY 1, ZERO ON, OCOMP ON and varied DELAY from 1 to 10.
Between DELAY 1 and DELAY 10 is almost 10ppm of difference. To get below 0.1ppm I would need a DELAY value in the order of 20 to 30.
While that works for me, it's bascially a no-go if I want to get the resistor calibrated.
The two other resistors do not show such a settling.
For 10k, and shielded PTFE cables, APER 1 and DELAY 1..2 is all what you need to get stable and precise readings.
If you still see deviations (in which direction?) when using even longer delays, then either your 3458A has a problem, or your resistor, i.e. the lacquer of the wire wound resistor, which I think is the construction of such Reichsanstalt type resistors, shows an odd polarization effect, or a heating effect, or..
Have you got a picture of your resistor, or of its interior?
To rule out a problem with your 3458A, have you ever measured another stable 10k resistor this way?
You only speak of “The two other resistors”, but I assume you mean the 100 and 1k ones?
What happens, if you just switch OCOMP off, to get rid of any switching / polarization effects, and simply measure the 4W resistance with NPLC 10, Delay 30 msec?
Do you see a constant increase or decrease of the resistance, until it settles?
I don’t want to be obtrusive, but the fastest way to identify the problem, would be to send me your resistor for a try out, because my setup, instrument and environment guarantees that I would only see weird effects from the DUT.
Frank
Well, meanwhile I think your 3458A might have a bug in its 10k range, if 1k and 100 were measured stable in their corresponding ranges.
Then the HW, e.g. the current source of the 3458A might be affected.
How does you 3458A measure the 1k in its 10k range, btw.?
I also read through the whole Service Note 12-B, which describes a lot of Firmware bugs between your 4.6 and the latest, 8 or 9.
I found one which is related to OCOMP and Delay > 0.85sec, or so, and another one which deals with slow implicit reading of the buffer, or so, where the 3458A delivers only 1 valid reading. I could not match those directly with your problem, but anyhow there might be another hidden error which might explain.
At RWTH Aachen in 1989, I used one of the first 3458As, also measuring 10k and 100k reference resistors with OCOMP.
I never used the DELAY, as I was not aware of these polarization problems yet, so I did not encounter a behavior as yours… and usually I find every failure on new instruments. Like I did then: That was the calibration constant issue of the PERIOD function, which was resolved very late with Rev. 8.
Frank
frank.stellmach@freenet.de writes:
I also read through the whole Service Note 12-B, which describes
a lot of Firmware bugs between your 4.6 and the latest, 8 or 9.
…and that list is not even comprehensive, it only admits to the (known) user-visible trouble.
Once you disassble the different releases, it becomes very clear that
4.6 must have had a trouble childhood, and I suspect the out-of-family
".6" suffix also testifies to that.
You /really/ should upgrade your 3458 to version 8 or 9.
(The difference between 8 and 9 is a single numeric value of no significance.)
--
Poul-Henning Kamp | UNIX since Zilog Zeus 3.20
phk@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.