[solved] Wavetek / Datron 1281 8.5 digit DMM repair - Error 2433 (RMS 1V Offset)

[Krampmeier]

So I finally got a Wavetek 1281 DMM for my home lab, which I have been drooling over for long time. The meter is in nice condition, but is has been sitting on a shelf for more than 10 years, so the SRAM battery was at 0 V and all NVRAM values were lost. So, after replacing the battery I had to clear the NVRAM and do a preliminary external calibration using a Wavetek 4808. That seemed to work quite well. However, the characterization of the internal references (which can be triggered on leaving the CAL menu) failed with Error 2433 (RMS Converter mean 1V Offset). As a consequence, self-cal is not possible. The full self-test gives the same error.
According to the service manual, this error is a "Digital calculation: Man or RMS +1V and -1V offset", and the limits are +/- 100ppm of FS.
Luckily, this DMM allows to set the self-test pathways manually. As far as I understand, the values used for the test are as follows:

Code: [Select]

P060 +1V in front of RMS converter 	-0.992567 V
P063 +1V after RMS converter 		+0.993058 V

P062 -1V in front of RMS converter	+0.992597 V
P064: -1V after RMS converter 		+0.993095 V

I read these values manually using the "Pathway facility". My guess is that the "offsets" are just the errors introduced by the RMS converters, so that would be

Code: [Select]

+1V offset = P063 + P060 = 0.993058 V - 0.992567 V = 491 µV
-1V offset = P064 - P062 = 0.993095 V - 0.992597 V = 498 µV

The mean value of these is 494.5 µV, clearly outside of 100 ppV of 100 ppm of the 1 V range FS value. Looks like the RMS converter or recitifier does add an offset to the voltage.

Does anyone here have experience with this error, or an idea how to troubleshoot it? The schematic shows the label "OFFSET NULL" next to the trimmer R215 on the AC assembly - should I dare messing around with that one? I don't want to screwdriver this nice meter to death...

I have already ordered all the electronlytic caps from TiN's repair log, but I don't expect that the recapping will solve this issue.

Thanks in advance for all ideas!

[branadic]

Let me forward you the following info:

"Sounds like you need to have it calibrated, its pointless to run self-cal after NVRAM data lost, so full service calibration is required.
Datron 1281 is not like 3458 which can be calibrated by 10V, 10kohm and short, so you'll need to run all service calibrations with 4808 using proper cables."

-branadic-

[Krampmeier]

Hi, Branadic,

yes, I am aware of that, so I did a full "external calibration" (adjustment) of the meter with the Wavetek 4808 after clearing the NVRAM. Also did the DAC calibration (which is fully automatic and did not generate any errors).
The only thing I left out are the spot frequencies for AC, which are not important for me right now, and the DAC calibration for the plotter output. I am planning to do the external calibration again after replacing al the electrolytic caps, and I'll use a Fluke 5720 next time. However, the error I am getting now does not seem to related to the accuracy of the meter alignment.

Regarding the self-calibration:
Right after the external calibration (adjustment), when the meter is at maximum accuracy, it can do the characterization of the internal references. That process generates another set of correction values which is later used during the self-cal, making it possible to eliminate drifts in the meter. It seems to be a similar to the HP 3458A's ACAL.
As the characterization of the internal references seems to be preceeded by a full self-test, which failed with Error 2433, this does not work for my meter now, and I'd like to bring it back into full working order.

The user manual and service manual can be found here: http://ftb.ko4bb.com/getsimple/index.php?id=manuals&dir=06_Misc_Test_Equipment/Datron/Datron_1281

[Krampmeier]

I tried doing some adjustments on the AC board, but it looks like it is not as easy as I thought.

The description of error 2433 is: "Digital Calculation: Mean of RMS +1V offset and -1V offset".
I am not sure on how to interpret this. The offsets are never measured directly, but only

• the output voltage of the selfcal multiplier (reference value),
• the output of the AC Preamp, and
• the output of the RMS converter.

The first thing I assumed was that the difference of the voltages at the Preamp output and the RMS converter output would be called the "offset" there, and it somehow makes sense to me that the RMS converter's output voltage should be the same as it's input voltage.
Based on this hypothesis, I tried eliminating the offset and gain errors which are caused by the rectifier and RMS converter circuits. I order to do that I measured the following "pathways":

1) P060 Preamp output voltage, at rectifier input, with +1 VDC input
2) P063 RMS converter output voltage, with +1V input to preamp

3) P062 Preamp output voltage, at rectifier input, with -1 VDC input
4) P064 RMS converter output voltage, with -1V input to preamp

Using the trimmers R215 (Rectifier Offset Null) and R336 (RMS converter output gain) I was able to trim P063 and P064 so that they matched the magnitude of the corresponding input voltages (P060 and P062 from the Preamp output) pretty well, with about 20 µV error (about 10 ppm FS). The values are not perfectly stable, and they vary a lot when the meter is tilted to the side, but they are clearly within 100 ppm FS.

Unfortunately, the self-test still shows error 2433.

My next idea was that an error caused by the preamp might also be taken into account for this test, and in fact I found that the preamp does seem to have a small gain error:

P061: -0.994,285V (input of preamp)
P062: +0.992,590V (output of preamp)

P059: +0.994.250V (input of preamp)
P060: -0.992,583V (output of preamp)

However, I did not find a trimmer for the DC gain of the preamp, and the block diagram in the manual sugests that a chopper amp should eliminate all DC errors here. Many parts of the preamplifier circuit still look pretty obscure to me though...

I feel like I have reached a dead end here. I don't know what the failed test is actually doing and I have not enough knowledge or instructions to adjust all the trimmers on the AC board correctly.

It would be great if someone here had an idea what might be going on in this meter!

[Kleinstein]

There are 2 trimmers in the preamplifier that should effect the offset. This is R111 for the bias current and R130 for the more direct offset.  There seems to be a little offset, but nowhere near the 100 ppm.
I also don't see a trimmer for the DC gain. So this is likely done by selected resistors and maybe parallel ones not shown in the plan.
The amplifier is indeed quite complicated. It is mainly an inverting amplifier with resistors setting the gain and the capacitive compensation part with extra adjustment, especially for the 1 V and 100 mV range.

The preamplifier part looks OK, though not perfect.
I would more guess the test is comparing the result of +1 V and -1 V  measured through the RMS converter.
The part in question could be the rectifier part, that is also not that simple. The offset in question could be the offset of the rectifier (R209).

Just for a check one could test an external AC signal to follow the signal path in case there is major fault and not just slight drift.

[Krampmeier]

Hello Kleinstein, many thanks for your help!
Rotating R209 did not change the output voltage of the RMS converter at all - I only saw a temporary effect while I was turning the trimmer (with the DMM in pathway test mode and with shorted input). Because of that, I set the trimmer back to the original position. Also, the service manual says "DC_DC turnover balances alternate Half-Cycles" - so that should be the purpose of R209.
I fed about 1 VRMS into the meter, with the range set to 1 VAC, and measured the output of the log amp with the oscilloscope (AC coupled). The half cycles look pretty balanced to me (image attached).

[Edit]: I finally understood the manual's explanation of the rectifier: The negative half-waves of the input voltage are multiplied by -2, while the positive half-waves are just buffered. Then both signals are added before they go into the log amp, which results in a precise full-wave rectification. The "DC_DC turnover" sets the amplification of the negative half-waves only - so it affects the readings from pathway P063, but not P064. That given, I am not exactly sure how to align the whole thing properly, in conjunction with all the other trimmers...[/Edit]

I trimmed R215 (Offset Null) and R336 (Output Gain) so that the output of the RMS converter (as displayed by the meter itself) matched the voltage from the reference multiplier (not the voltage after the preamp). I had to turn the gain pot to the mechanical limit in order to come close to enough gain. This does not feel right, and the error reported by the self-test is still the same (2433).

I am now getting:
P059: +0.994,274 V (input of preamp)
P063: +0.994,248 V (output of RMS)

P061: -0.994,304 V (input of preamp)
P064: +0.994,273 V (output of RMS)

That's about 15 ppm of the FS value off and I really wonder why the meter is not happy with it.
In ACV mode, the readings do correspond to the AC voltage which I supply into the terminals. Of cause, after all the screwdrivering, they are not precise any more - but in the right ballpark.
 

[Kleinstein]

The picture from the scope is normally not very sensitive to see whether one gets the same gain. In this respect the measurements with an DC input should be better, though it only test 1 point of the nonlinear RMS circuit. The signal still looks OK (reasonable waveform for both sides).

The scope trace still shows quite some interference signals near the signal maxima. This may indicate some problem (e.g. with the log amp), unless it's from the signal source.

The measured values look good and well matched.  It is no surprise that the gain of the AC preamplifier is not a pefecct -1. It is more a surprize how close it is to -1.
A small error in the amplifier gain should be corrected in the software and thus no trimmers needed.
Similar there should be no need for the RMS output to directly follow the input ref signal - a different gain should be OK (the preamp gain can be different for the different ranges).

It is more a surprise to have a trimmer for the output gain of the RMS part.

[Krampmeier]

Thank you for following up with me, Kleinstein!
I agree the scope is not suitable to find differences in the µVs, and there is a lot of noise on the signal. The noise seems to be radiated or condicted EMI, not from the signal generator, which is a Rhode&Schwarz SRM RC-Generator. I get the noise on the scope even when I just connec two wires which are shorted at the far end. For this application, I do not worry about that noise too much.

The DMM self-test which fails should be just a math function, using the static values acquired before. And I really don't see what is wrong and it is still obscure which values are actually used (even though we made some educated guesses).

Sorry for abusing the forum as some kind of log book for my amateurish messing around.

I have adjusted the DC_DC turnover so that the signal at the RMS converter output is the same (within 10 µV = 5 ppm FS) for positive and negative 1V input voltage, and the test still failed.

I connected about 4.8 mV RMS to the DMM input and measured the following signals in ACV measurement mode:
CH1: TP215 Rectifier amplifier feedback path, before diodes
CH2: TP214 Rectifier input (from preamp)
CH3: TP216 isolated and inverted negative half-waves

Screenshot attached. One thing that looks a bit strange to me is that CH2 and CH3 seem to have an offset of about 3-4 mV, but I later saw that the signal generator has a small DC offset which might be the reason for that.

With no other ideas, I thought of away to adjust some of the trimmers in a defined way.
1) Offset Null R215: The rectifier's amplifier M201 gets feedback through D210 for positive half-waves and through D209 for negative half-waves at its output. Because of the diode forward voltage, there is a 400 mV voltage "jump" on every zero crossing at the input. I believe that the offset trimmer R215 should be adjusted with 0 V input voltage, so that the output at TP215 starts becoming unstable. I shorted TP214 to GND with the meter in VDC mode and this adjustment seemed to work as expected. I did only a very minor adjustment.
2) With the Offset set, I should be able to tune the DC turnover so that the rectifier (measured through the RMS converter) has the same amplification for the positive and negative half-wave. I used the pathway test functions like before and was able to adjust DC turnover so that the rectifier had identical gain for -1V and +1V. The gain is still about 1.002 from my previous tinkering, but I guess that should be OK (and it compensates the preamp's gain which is about 0.998.
If someone is trying to do the same thing - beware, the RMS converter's gain is quite temperature dependet, the value changes by approx. 100 µV when the meter is tilted by 90 degrees on its side.

At that stage, I got:

Ref mux output P059 +0.994,268 V --> Preamp out P060 -0.992,595 V --> RMS out P063 0.994,185 V
Ref mux output P061 -0.994,299 V --> Preamp out P062 +0.992,590 V --> RMS out P064 0.994,181 V

So, the preamp output and RMS converter output look very symmetrical. However, the preamp output should actually not be as symmetrical, as the Ref mux output is not as symmetrical. I suppose trimmer R111, "Bias Current" at the preamp input could fix that. Ugh, another trimmer to screwdriver into lower orbit...

After messing with R111 I get:
Ref mux output P059 +0.994,263 V --> Preamp out P060 -0.992,582 V --> RMS out P063 0.994,149 V
Ref mux output P061 -0.994,296 V --> Preamp out P062 +0.992,608 V --> RMS out P064 0.994,174 V

With the limited understanding I have of the circuit and the self-test functions, this looks very reasonable to me and I don't see anything wrong with these values.

So, I triggered the internal source calibration again, and got error 2433 again. Bummer 

Yes, I could turn down the gain of the RMS converter again now. But I I think I have been there already before - anyway 

Ref mux output P059 +0.994,258 V --> Preamp out P060 -0.992,581 V --> RMS out P063 +0.992,579 V
Ref mux output P061 -0.994,291 V --> Preamp out P062 +0.992,606 V --> RMS out P064 +0.992,597 V

Ok, one last internal reference characterization for today (it takes about 10 minutes each time)...
Nope, same error again. 
I just don't get it...

[Krampmeier]

It looks like the Datron 1271 uses the same self- test and error codes as the 1281, and finally found a very helpful contribution in another thread:
https://www.eevblog.com/forum/repair/datron-1271-meter-repair-help/msg163433/#msg163433

I just checked the 1VDC range of my meter with a negative input voltage, and sure enough it is totally off (about one percent low, just like the positive ranges were right after clearing the NVRAM). During external calibration I never did the alignment for negative voltages. The need to to that is buried in the text of the user manual, but I had missed it  I also find it surprising, as other meters such as the Keithley 196 or Keithley 2001 don't seem to need separate alignment of positive and negative input voltages...

I transferred the positive 1VDC alignment to the negative 1 VDC alignment, and sure enough, the meter passed the internal source calibration after that. 

Checking the pathway measurements again revealed that the reference voltage has drifted by about 10 ppm FS, and the RMS converter gain has gone down about 100 ppm, probably due to temperature change. I did not see that huge 1% change in any of these values - so it seems like the "RMS 1V Offset" tests does not use the displayed values from the previous tests directly, but applies the correction values in the background before the offset is calculated.

Thanks for thinking along with me, - Branadic, you were right -, and I hope this thread will one day help others who struggle with the same problem!

[branadic]

Nice you found the culprit. I just forwarded the info someone gave me, so I will forward the flowers to that person.

-branadic-

[seebeck]

Krohn-Hite 523