Fluke 5200A repair (?)

[Rax]

I went ahead and did:

• PS adjustments - the 15V was at 15.045V (out of range), adjusted to 15.000V. TP3 on A7 showed 186.70V at that point, compliant.
• Power Amp bias was at .27V - adjusted it to .35V
• Power Amp zero DC was at 222uV to start with, adjusted to 0.000mV on my Fluke 189. I think floating this measurement is likely beneficial. Very finicky adjustment, having the shields on seems critical.

Tomorrow, I may endeavor on adjusting the actual gain and levels.

[Rax]

If anyone's wondering, doing these adjustments didn't change anything discernable in the levels. They still are just about 600ppm high.

There goes my theory!

[tridac]

Neat job on the two resistors, would not know you had been in there. Used metal oxide 1% on mine, good enough. Had considered replacing  the other 2 w comp resistors on the board, but they didn't look overheated, and values were about right. As for the psu electrolytics, replaced all the wire ended types with Vishay 105C as a precaution, but they tested ok as well, off the board. The screw teminal types checked out ok on the cap meter, and they are super expensive, even if you can find them. You can also scope the ripple at the electrolytics, as a check. Mine has to come out of the rack again soon, but too much other stuff on the bench at the mo.

One other thng to check is the reference board at the far left. There is an adjustment that sets the initial dc reference voltage and there are two test points for a voltmeter. Correct voltage value is in the manual, but that could be the cause of the output arror, as it's the dc reference for everything else that follows...

[Rax]

Neat job on the two resistors, would not know you had been in there.

Thank you, tridac!

One other thing to check is the reference board at the far left. There is an adjustment that sets the initial dc reference voltage and there are two test points for a voltmeter. Correct voltage value is in the manual, but that could be the cause of the output arror, as it's the dc reference for everything else that follows...

Will look into this.

I wonder if anyone would have some insight on performing the adjustments. I don't have a 540B - I'll have to rely on a couple of calibrated meters (meaning - long form calibration available at hand) - namely a 8502A and an 8846A. I'll definitely rely mostly on the 8846A, though I have a hunch the 8502A may be a bit more wide bandwidth.

I also have a Prema 6048, HP 34401A and a couple other meters, and the Prema seems very accurate and maybe a bit more digits.

I also have a 931B that passes its "operational tests" but I have no idea of its state of calibration. But maybe as a differential instrument, that's not so critical?...

I guess I'm looking for a sense of which adjustments may be more important (and what specifically they do, for those with good experience with the unit; maybe even recollection of a recent cal adjustment procedure?....). Otherwise, I may try them one at a time (section 4-76 in the Manual - if not available to most, I can snip and post here), and continuously check where the output lands on the characterized meters. Obviously, checking the output at higher frequency is nearly impossible with what I have.

[alm]

I wonder if anyone would have some insight on performing the adjustments. I don't have a 540B - I'll have to rely on a couple of calibrated meters (meaning - long form calibration available at hand) - namely a 8502A and an 8846A. I'll definitely rely mostly on the 8846A, though I have a hunch the 8502A may be a bit more wide bandwidth.

I would only perform adjustments if I think I can do better than its existing state. Depending on how much adjustments interact, it may well be that adjustments need to be made across the bandwidth. So if you could only do a reasonable job until 10 kHz, that might well mean the 10 kHz - 1 MHz part will get worse. Calculate for every instrument you have the absolute uncertainty to national standards based on their last calibration, the calibration uncertainty and the specifications for drift since then (e.g. 90 day or 1 year accuracy), and see how that compares to the specs for the 5200 and the current estimated deviation of 400-600ppm. If not, then how would you know you're not making it less accurate?

Unless your only goal is consistency, not absolute accuracy. Then by all means just twiddle trimmers until the display of whichever meter you like reads all nines or zeroes. Ideally also adjust all but that one meter to match the others.

Unfortunately AC calibration without spending a lot of money is hard. Pretty much everything, including a Fluke 540B, needs calibration across its bandwidth and ranges. There are ratio transformers that can help, but even those have a roll-off at higher frequencies. The traditional procedure involves a set of thermal voltage converters (TVCs). One of them is calibrated across its bandwidth, and that one is used to calibrate the next TVC value up or down, until you are at the highest and lowest values. A very tedious process even if you have all those TVCs and supporting equipment like switches. The modern way would be to throw money at something like a Fluke 792A or 5790A/B and spend a fortune in getting it calibrated. Getting a DMM with good ACV specs and BW is probably the most affordable option. The Fluke 8506 is known for its good AC performance, but rarely cheap. I didn't check how the P6048 compares.

[tridac]

You already said that the output is off by 600ppm, from memory, so what confidence do you have about the accuaracy of that ?  If you have several meters, different vendors,  that agree with each other within a few microvolts, it's reasonable to assume that they are all close to accurate, but leaving out any outliers, take an average of all those meters ot get a reasonable result. Not everyone has the luxury of a calibrated hp3458, so those without cal lab facilities have to rely on the best they can do at the time. As for the null voltmeter, that sort of kit was pre high input Z dvms, which make loading errors ignorable in most practical cases, especially when tuning  for a null. Just use a dvm as the null voltmeter. A good 7.5 digit dvm like the 3457 will resolve microvolts in 10 volts, much better than any of the early Fluke analog null voltmeters. Measure the reference anyway, as the output impedance at those pins is low and unlikey to be affected by a dvm load. At least you will have an idea if it's close to what it should be...

[alm]

Not everyone has the luxury of a calibrated hp3458, so those without cal lab facilities have to rely on the best they can do at the time.

That's very true, but sometimes the best you can do may be to not touch any adjustment.

As for the null voltmeter, that sort of kit was pre high input Z dvms, which make loading errors ignorable in most practical cases, especially when tuning  for a null. Just use a dvm as the null voltmeter. A good 7.5 digit dvm like the 3457 will resolve microvolts in 10 volts, much better than any of the early Fluke analog null voltmeters.

I'm not sure where this null voltmeter came from. I didn't see it in the F5200 suggested calibration equipment list. It used to be common to use a Kelvin-Varley divider + null meter for precise voltage measurements before long scale DMMs became available. You can substitute a good DMM here, though keep in mind that the F720A has a linearity specification of 0.1ppm of input, so while operating at ratios between say 0.1-1.0, it will likely be more linear than your DMM if that's important.

As for replacing a null meter by a DMM. Fluke has a white paper about it where they try to sell their DMM over null meters. There are two caveats to be aware of. One, adjusting a trimmer for null on a DMM is extremely tedious. I have done it trying to adjust a Fluke 720A, and because of the sign flip it's so much harder than with an analog meter movement. Instead of adjusting in a fluid movement, it's making small adjustments and waiting for the display to settle. Even 'analog' bar graphs are useless around 0, or if you're working on uV levels. The second consideration is bias current. Depending on the impedance of the two nodes the null meter is connected to, bias current may be a big problem. See for example this thread on replacing a null meter with a nanovoltmeter or DMM and the linked post of Dr. Frank describing his experiences with using a 3458A as null meter with a Fluke 752A Hamon divider and the errors this introduces.

[Rax]

I'm not sure where this null voltmeter came from. I didn't see it in the F5200 suggested calibration equipment list.

From the Service Manual, Section 4-30 Output Accuracy Test - Voltage Ranges, but it has nothing to do with DC nulling. The 931B (which is an exclusively AC differential voltmeter, and which I have, as I mentioned prior) is listed as an alternate instrument to the 540B.

DC nulling the power amp - I have no issues doing it with a regular DMM; this just balances the amp to have no DC out, as far as I can tell, so no reason to employ a KVD or anything anywhere near that level. Unless I'm missing something, but I've nulled many instruments before and it's a very cursory operation.

If you review my prior posts, the PS adjustments, power amp bias and power amp DC zero were all done a couple of days ago. And as I said after that, this regards Section 4-76. Let's please focus on that. And thank you very much tridac and alm for thinking on this with me.

[Rax]

To the inquiries on what I have available to dial this in - I attach a spreadsheet with the equipment at hand that I used to gauge how well the 5200A is doing, and some measurements.

Namely:

• An 8502A that was relatively recently adjusted and then "Z540-calibrated" (unfortunately, no uncertainties provided) in 02/2023.
• An 8846A that was calibrated by Fluke at the end of August. Unfortunately, they also replaced all the guts, so the internal reference is probably a little green for "certainty" (is this a pun?...). Not sure if it impacts the ACV accuracy, but it's likely my best reference to assess the accuracy of the 5200A. I enclose its 17025 long form cal data for ACV.
• A Prema 6048 which I have a fairly good level of trust for.
• A recently acquired 34401A, which I have no strong reasons to trust, but no reason to distrust either. I look at it as a fourth opinion that can reinforce the first three, and doesn't get disqualification rights over anything.

[Rax]

For the 8502A, the "calibration result" is what the cal house said it'd tell on standard levels the lab would apply to it.

[Rax]

On my measurements spreadsheet, I recommend focusing on the mid range of the Hz spectrum (100Hz to maybe 20kHz) as, obviously, that's where a meter like those used would have something useful to say.

I will have access to a Keysight-calibrated 3458A/002 for a weekend at some point in the next few weeks or months, but that's a thing I need to plan the heck out of, and am not sure it'll serve this purpose. I don't think the ACV performance of the 3458A is amazing, particularly at the ends of the frequency spectrum... or is it?

[alm]

So what you need to do to calculate what uncertainty relative to the national standards you could achieve, for the meters that were calibrated, is like described in the GUM sections 1-6, identify all the terms contributing to the uncertainty. Start with the value at the time of calibration:

• If you have data, then say at the time of calibration the value was the measured value plus the offset of their stimulus from nominal. It might be that the calibrator was not sourcing 1.0000 V but 1.0001V. See the first paragraph I wrote here for how to correct for that. The uncertainty at this time was the stated expanded uncertainty, which will likely be a 95% confidence interval, so divide my 1.96 to get the standard error.
• If you don't have the data, you should assume the meter was at that time at its nominal values but with the uncertainties it was adjusted to (1 year? 24h? it might say on the certificate). Ideally with some guard banding, like if they say they adjust if the value is beyond 80% of the tolerance band, then you can assume the value was within 80% of that value. Calculate the absolute value in Volt for a power of ten value near the top of the range (e.g. 1 V) reading by multiplying the value by the "of value" specification and adding the "of range" specification multiplied by the full range (check datasheet, could be 10V, or 12V or 20V for example). This should be treated as a rectangular distribution as described in section 4.3.7 of GUM, so the standard error is \$\sqrt{a^2/3}\$.

This calibration was some time ago, so apply the stability or accuracy specification (90 days, 180 days or 1 year depending on how long it's been any what kind of different specifications the instrument has). Convert to absolute volts (not percentages or ppms) as described above. This should be treated as a rectangular distribution as I described above.

The calibration was at a certain temperature that should be on the certificate, and the 90d etc stability / accuracy specifications will say something like within 5 degrees of the calibration temperature. If the ambient temperature at time of measurement is outside this range, apply the temperature coefficient from the specs multiplied by the number of degrees outside the range. This should again be assumed to be a rectangular distribution.

Then calculate the geometric sum (root of sum of squares) of all these components, and multiply by the coverage factor (k=2 is recommended by GUM). This is an estimate of what kind of uncertainty you might achieve, though the actual uncertainty will be slightly higher because you also need to add the standard error of the mean of the measurement as a factor, although this can be reduced by collecting more samples.

Do this for your 8502A and 8846A for every range and frequency. You could do this in Excel or Jupyter notebook with Python and Pandas. What I did for MM2022 was to make a table with all the calibration certificate data, one with the datasheet specs, and then had a function that looked up the correct value in each and did the calculation. You could do the same in Excel with VLOOKUP etc. Regarding the 8846A with replaced guts, it should still be within the manfucturer's specifications, but might drift more than a couple of years from now.

For every range / frequency I'd pick the meter with the lowest calculated uncertainty. This might well be the same meter for all. As a sanity check, the intervals of meter1_result +/- meter1_uncertainty and meter2_result +/- meter2_uncertainty should be overlapping for every value. The 34401A can vote but especially for the higher frequency AC ranges I wouldn't attribute too much importance, since I believe those are the most frequent to drift out of tolerance. Compare this expanded uncertainty to the accuracy and stability of the 5200A and to the offset that you measure now. The stability of the 5200A is like a lower limit: there's no point in setting the 5200A down to 1 uV/V if the 24h stability is 100 uV/V. The ~400-600 uV/V offset is the higher limit: If you measure the 5200A to be high by 400 uV/V, but the uncertainty of your measurement is 800 uV/V, then it might was well be 400 uV/V low and you wouldn't be able to tell from the measurement. Ideally you aim for the accuracy specifications in the data sheet, but otherwise you'll need to note that the uncertainty is higher than in the data sheet for those specific ranges.

You can do the same math I described above for the 3458A. The 3458A isn't really good enough above 100 kHz or 300 kHz to calibrate the 5200A, but it might be the best you have available. I would first focus on using it to measure the most stable standards you have, probably the 5440A. Depending on how long ago it was calibrated, the uncertainty might be about 4 uV/V for 10 VDC. If you plan any adjustments (like the 5200A) with the 3458A, I would rehease the adjustments with one of your other meters before you get the 3458A and investigate the best settings for the 3458A, since just punching the 'ACV' button may not give you the best results.

I see 0V measurements in your spreadsheet. I don't think measuring 0 VAC is of any use since multimeters are usually not specified below 1% or 3% of full scale. I don't think the 5200A performance verification procedure calls for a 0V AC measurement either.

[Rax]

So what you need to do to calculate what uncertainty relative to the national standards you could achieve, for the meters that were calibrated, is like described in the GUM sections 1-6, identify all the terms contributing to the uncertainty.
[...]
I see 0V measurements in your spreadsheet. I don't think measuring 0 VAC is of any use since multimeters are usually not specified below 1% or 3% of full scale. I don't think the 5200A performance verification procedure calls for a 0V AC measurement either.

Thank you very much, this is great, detailed, and very thoughtful information. Very much appreciated!

Not sure where are the 0V measurements. If you mean the value provided by the 5200A, I had it output .001V, .01V, .1V, 1V, 10V, and 10V. The low levels more of curiosity - I do think it makes sense to read useful, conclusive stuff only from the relative high levels 1V, 10V, maybe 100V), and the middle of the frequency spectrum 100Hz to, say, 20kHz). In a couple of cases, I had to output 3V or 7V, because that's what they did during calibration and I wanted to match their stimulus. Same for the frequency, at times I had to do other frequencies than what I started with.

Sometimes the reading would be zero - such as when maybe the frequency would be too high and the level too low etc., or it'd be "OL" if I was not careful enough to stay within the design limits.

But I don't think I applied "zero ACV" in any of the cases saved in the spreadsheet.

[alm]

But I don't think I applied "zero ACV" in any of the cases saved in the spreadsheet.

Ah, I see, due to the column width Excel was truncating 0.001 to "0" instead of "0.00" or some other clue that something is being truncated. Excel is such a wonderful product

[Rax]

But I don't think I applied "zero ACV" in any of the cases saved in the spreadsheet.

Ah, I see, due to the column width Excel was trunctaing 0.001 to "0" instead of "0.00" or some other clue that something is being truncated. Excel is such a wonderful product

Yup... I use it as a google sheet, where all that's better handled. But the columns may need to be expanded here and there to read the values listed. LibreOffice has a little red arrow, I think, where a value is being truncated.

[Rax]

Well, I guess I am finding an (intermittent!) issue with this.

Sometimes, when I switch to, say, a 1.000000V out, I'd get that with a jump up by four units on the second level control (namely, 1.040650V or so). Increasing that control higher by one unit would return as below (for clarity, I am listing all relevant values going up).

• 1.010000V -> 1.050650V
• 1.020000V -> 1.060650V
• 1.030000V -> 1.070650V
• 1.040000V -> 1.040650V
• 1.050000V -> 1.050650V
• 1.060000V -> 1.060650V

and so on.

So this "hopping by four" would stop at the "actual four" and it'd be correct from that point on. This gets replicated exactly on all ranges etc. It seems to be caused by the logic generating that specific digit.

Not sure what can cause this yet, but looking at the schematic next. I did clean the corresponding switch with Deoxit (not very hopeful it'd fix it, due to the symptoms), to no avail.

[tridac]

Ther reference voltage comes from a pwm signal generated by a programmable 1-2-4-8 input bcd counter chain, driven by a 10MHz clock. The pwm is low pass filtered to provide the dc reference voltage.  Sounds like one of the switch to counter chain inputs is at fault. You need the extender board, but just trace the logic levels from switch onward, for the offending digits. If you have good meter, can also measure the reference board outut dc for correctness, as you change the switch settings, which should be completely monotonic. If not that, then  it may be in the sine wave oscillator / rectfier board group...

[ken goodhew]

the purpose of the handle inside the power supply assembly is simply to push the cards into the slots on the motherboard,otherwise it is held out of the way by the bracket in the top cover

[Rax]

Ther reference voltage comes from a pwm signal generated by a programmable 1-2-4-8 input bcd counter chain, driven by a 10MHz clock. The pwm is low pass filtered to provide the dc reference voltage.  Sounds like one of the switch to counter chain inputs is at fault. You need the extender board, but just trace the logic levels from switch onward, for the offending digits. If you have good meter, can also measure the reference board outut dc for correctness, as you change the switch settings, which should be completely monotonic. If not that, then  it may be in the sine wave oscillator / rectfier board group...

Thank you for this input. I've been focusing on other projects, though the 5200A has been very useful to even one calibration check this far (with a calibrated meter characterization). The "+4 on second digit" persists, but is occasional. If I get a chance, though, I'll seek to further investigate. The other roadblock is I don't have an extender board for this, and I didn't look for them extensively, but I imagine they may be pretty hard to find.