Is my new Fluke 731B defective - frequent 1mV positive spikes every 30 sec?

[KJ-90]

I recently was lucky enough to acquire a Fluke 731B, which seems to have been one of the last ones made, close to 1987. I don't know the unit's cal history ( no stickers present ). The unit also has a new Ni-Cd battery installed, which reads the correct voltage
( about 16V ). However, when I did a quick basic test with a Fluke 287 ( recent calibration ), I found a potentially ominous issue. Although the voltage is initially 10.000 Vdc on that output, there is a 1mV positive spike about every 30 seconds. At first I thought this was normal for a unit that had been in storage a very long time, and the unit needed a long time to stabilize, or the unit was really, really out of cal. However, even after running it for several days, the 1mV positive spike still will not go away, nor has it become less frequent with time. Also, the unit has a negative -1mV spike about every hour. These problems exist both when the unit is run on both battery power and the unit is charging from being plugged in. No AC voltage is found on the 731B's output with that test. The spikes are DC voltage only. Also, when measuring the 731B's voltage, the analog bar graph on the 287 is constantly jittering
slightly upwards, it's never stable.

At first I though it was the switch issue, and I found all the gold plating was gone, so I tried measuring the 10.000V output on the Ref Amp board itself, at TP1 & TP2. The same 1mV spike problem exists, both here and on the front panel terminals. When I checked the 731B service section of the manual, I verified that Q1's collector voltage is correct. According to the manual, the only other potential areas of trouble are that either the LM308 ( U1 ) or- really shudder- the SZA263 ( U2 ) ref amp has become defective.

My only other theory is a defective R11- the 10V output adjustment control is either damaged, or has become noisy over time.

The other problem I have is that currently I don't have access to a 6.5 or 8.5 digit DMM to measure the 731B's voltage outputs more accurately. A Fluke 287 has a basic DC accuracy of only 250 ppm, so the 1mV digit can't be really relied upon. However, I did use the recording feature of the meter to do a basic voltage stability plot every 1sec for 2 hours. It clearly shows an almost constant peaking of 1mV positive, and confirms the once a hour -1mV drop.

Has anyone experienced this fault before, and if so, know what it is? Can it be repaired, or was my unit sold because it had developed a serious instability and isn't worth fixing?

[Edwin G. Pettis]

My 731B is much older than yours, I did do some restoration on it last year.  It was working okay otherwise.  Those 1mV blips are not normal for the circuit but they aren't really a usual failure mode for semiconductors.  There is the possibility of external interference causing the blips, they are very periodic.  There is a 1uF tantalum cap in the LM308 circuit that could possibly be generating the blips but that would be unusual.  Those blips could be generated by something cycling on the power line and also radiating it through the air so that you're seeing it whether or not the 731B is connected to the power line.

You should see nothing more than some mostly higher frequency noise of the output of the 731B, no blips unless you're looking at uV levels.  At this point I would not suspect your SZA263 is malfunctioning nor the LM308A for that matter.   

[wraper]

1mV is one least significant digit for Fluke 287. You cannot rely on that or make any conclusions.

[Edwin G. Pettis]

Quite true, 1mV resolution is somewhat marginal for such a measurement, you really need at least one more digit of resolution.  However, if the readings are otherwise stable between blips, that doesn't sound like the normal ±1 digit bobble inherent in digital meters, but just to be sure I'd try to get that extra digit of resolution just to make sure.  One digit bobble tends to be a-periodic, not periodic, they are random.

[Conrad Hoffman]

You should be able to see 1 mV on a scope. I've got several 731sa and if I see something like that, it's invariably external stuff in the house.

[orin]

I get occasional jumps down on my 731B, but I'm pretty sure it's EMI.

I'd check with the meter on an ordinary 9V alkaline battery and see if you get any jumps with that.  If you do, it's either EMI or the meter itself.

[ManateeMafia]

I agree with Orin and do that first. Start by building up a test voltage near 10v. Perhaps alkaline 9V + AA (10.5v)or similar (Ni-Cd) and do a measure of series opposition but connecting both negative outputs together and put your meter across both positive terminals. You are not looking for an exact voltage but a test for the spikes. This will allow you to use the 500mV range of your meter.

If you have a cheap 10v ref board even better.

[MisterDiodes]

KJ - All good advice here.  A 731b is really pretty stable, and they either work really really well or not at all.  There shouldn't be anything in the 731b to be causing a periodic signal like that.

As Manatee and Orin say let's make sure your meter is behaving first.  The bar graph on the meter doesn't mean a whole lot except you might be hovering between two segments.  A battery substitute on your 287 around the same voltage ballpark is perfect - no need to be exact.

Another suggestion I can add is to take you, your 731b, your test batteries and your 287 out in the backyard or some other different location away from power lines, PC's. Lights etc.  Play with it and see if you still see the spikes.

That "once every 30 seconds" symptom sounds really environmental, or perhaps just something on your meter.  We had something similar happen once on a Vref and tracked it down to a lab freezer with a faulty defroster switch on the other side of a wall.  It was on a thermal cut-out that kept cycling on and off around every 45 or 60 seconds - but made enough periodic EMI to confuse everyone.  The freezer did seem awful icy...

[SL4P]

1mV pfft.

I was working on a site 30 years ago, where unterminated leads would show 3V transients at regular intervals at various times of the day/week.

scroll down














Radar at the international nearby airport sweeping the neighbourhood. Mascot NSW

[TimFox]

I had an annoying problem (with a different system) at a higher frequency:  10 Hz, and line-synchronous (to 60 Hz).
The problem was a soldering station plugged into the same outlet, with the iron idling in the stand.
Apparently, the temperature servo used a zero-crossing switch, and the idle cycle caused it to switch at approximately 100 msec period, which then was synchronized to the line by the zero-crossing switch.
Something else on your power line may be cycling at a much slower rate.

[VintageNut]

I have two 731B units. They will occasionally blip. I think that it is appliance(s) throwing hash on my house wiring. I suspect that the 731B is susceptible to EMI/RFI/etc.

The time between blips is usually hours. When calculating long-term statistics, I think that these blips are not a concern.

For the money, a working 731B is a great value. I have had both of mine calibrated against a bank of 732Bs. I drove them to the cal lab. After a day of settling at the cal lab, they were quite stable.

[HighVoltage]

I just tested one of my 731B for noise yesterday with my Keithley DMM7510,

Settings:
- PLC = 5
- AZERO ON
- FILTER OFF
- 33 Samples taken over 10 seconds
- Y-scale = 1µV/Div

Here are the results: 2.8 uV Pk/Pk