Fluke 87V displaying dashes only

[Multiple Cheese Slices]

This is in fact really strange about the meter having so many faults. Unfortunately I have no idea how this came to be. Who knows, maybe it was lightning. Who knows 


This time I have remembered to probe relative to the circuit ground (2.5V ref output).
Actually, ADIN is at -2.8mV when in Vac mode and not changing at all, regardless of the range. This holds true for DC as well, which AFAIK has no problems with ranging/autoranging with the difference of ADIN being at -0.7mV in DC mode, +4mV for ohms etc.
This does not seem to differ based on shorted or open inputs.


Now about U7 - AD737J, the RMS-to-DC converter, (note that all measurements from here on are taken relative to battery (-), since that seems to be the ground of this IC).   it is getting a stable 5V power supply, as it should,  so that can be ruled out.


I've taken some measurements of the output(pin 6) and Cc(pin1) (not probing pin2, since it is tied directly to 2V5(reference output)) at different AC inputs on the probes, here are the results:

probes state:                                    |     Output pin state:               |     Cc pin state:
-----------------------------------------------|---------------------------------------|-------------------
probes shorted                                |     2.5V (steady)                     |     1.43V (steady)
not hooked up to voltage source     |     2.5V (steady)                     |     1.425V  (steady)
18Vac                                              |     2.225V (steady)                 |     1.97-2.17 (somewhat fluctuating, but by just a few mV)
32Vac                                              |     2.1-2.5V (fluctuating)         |     1.2-1.8V (fluctuating)
230Vac                                            |     2.1-2.5V (fluctuating)         |     1.2-1.28V (fluctuating)

All voltages of the output and Cc pins are measured in DC mode (with my other DMM). When measuring them (with my other DMM(stating this just to avoid confusion)) in AC mode, they all return 0V.

It seems that there is corelation between fluctuating DC voltages on the input/output and the overload condition.
Also to me it looks like there are issues before the RMS converter as well - on the CC pin - the voltage fluctuations on it seem to corelate to the output(pin6).

Honestly speaking, I'm not sure how to read what input voltages should produce what output voltages and what they would "mean", so I'm not sure if the 18Vac translating into 5V occurs here or before/after. Here is a link in case it would help:
https://www.analog.com/media/en/technical-documentation/data-sheets/AD737.pdf


Input(pin2)/COMMON measure 2.511Vdc (as it should be), and the PDN is 3V (as it should be(i think, judging by the datasheet))

[Manul]

Now about U7 - AD737J, the RMS-to-DC converter, (note that all measurements from here on are taken relative to battery (-), since that seems to be the ground of this IC).

I'm not promoting violence, but someone needs to beat you with a shovel to make you forget that battery negative It is not the ground of that IC. In fact, nowhere in the circuit it is a ground. There is power supply section: battery and linear voltage regulators, you may use battery negative as a reference when talking about them if you wish. But after the regulators, circuit ground is clearly drawn as +2.5V reference output (that triangle symbol) and everything else in the schematic uses that. All the signal paths, everything. Voltages in the schematic written relative to that, voltages in the text are written relative to that. Espeacially when we talk about analog signal paths where op amps and other stuff are using that +2.5/-2.5V split supply with virtual ground in the middle acting as common, it is really unnecessary mental stressing and unintuitive to read numbers relative to -2.5V and not the ground.

So with no input ADIN in VAC mode is always -2.8mV at all ranges? Seems quite normal, that is very close to zero, just a small offset. And what the meter shows in those different ranges with no input? Near zero, overloading or what?

[shakalnokturn]

On most handheld DMM's the easy to access "internal" ground (and thus sensible reference for measurements when troubleshooting) is the COM banana socket.

[Multiple Cheese Slices]

I'm sorry about the battery negative, I'll borrow a shovel from a friend and try to fix the issue myself first.  (No, but really, I will 100% remember from now on. I got kinda confused; The COM pin kind of slipped my mind since it had no description in the pin funciton descriptions other than "common".  This is the actual signal ground, yes?)


When the DMM is switched to the Vac mode, it starts off showing 56mV and then drops down.
When it settles and is no longer changing, The 87V is showing:
  1.6mV AC in the auto/manual 600mV range, 
  0.016V in 6V range,
  0.16V in  the 60V range,
  1.7V in the 600V range,
  and 2V in the 1kV range.

Strange thing is that I cannot get it to display 0 no matter if the probes are shorted. I wonder if that's just miscalibration or part of the problem.

On most handheld DMM's the easy to access "internal" ground (and thus sensible reference for measurements when troubleshooting) is the COM banana socket.

Ah, cool trick! There is in fact direct continuity between the COM socket and the circuit ground in the 87V. Thanks for that! 

[Manul]

Well, the offset is not big, seems around 0.2-0.3%. Probably calibration, because 2.5V reference was changed. Anyway, it is quite close to zero with no input, but when actual AC volatge is measured it goes way off, yes? Except 600mV range?

About that RMS converter. Theoretically it should output negative RMS of input AC voltage. So if you measure AC input voltage of lets say 100mV (RMS), then output should be around -100mV DC.

By the way, you can make a voltage divider to use with 18V AC you have. Maybe potentiometer. You can produce different voltage to see how it reacts. Like 100mV or 500mV, 3V and so on.

[Multiple Cheese Slices]

Actually, now that I've checked using the potentiometer voltage divider, the 600mV range looks pretty rekt aswell.
Here are some measurements that I have taken using both my (known to be working) multimeter and the fluke.
(Yes yes, the shovel has helped; this is realtive to circuit ground/GND banana plug)

Probes input:	F87V display:	RMS AD737J input	AD737J DC output
430mVac	6.8mVac	0.22Vac	-4.6mVdc
880mVac	31mVac	0.44Vac	-19.3mVdc
1550mVac	360mVac	0.77Vac	-188mVdc
1840mVac	566mVac	0.93Vac	-283mVdc
2500mVac	Overload	0.1 to 1Vac	-100 to -300mVdc
3333mVac	Overload	0.1 to 1Vac	-50 to -380mVdc

A mistake I have done earlier was measuring the RMS-DC converter input as DC. (oops), Here it is in AC now. It was around -1.3Vdc all the time, so I figured something was wrong.

About that RMS converter. Theoretically it should output negative RMS of input AC voltage. So if you measure AC input voltage of lets say 100mV (RMS), then output should be around -100mV DC.

So, the outputs are waaay off from the inputs. Does this indicate a broken AD737J, or is it possible that the issue lies elsewhere?

I don't know how to check whether the AD737J RMS input voltages are proper, though. The input circuitry seems pretty complicated to my noob eyes.


Well, as always - what do you think?

[Manul]

AD737 input voltage looks exactly proportional to meter input (divided by 2, which seems reasonable). So probably AC AMP, gain (range) selection (U1) works ok and multiplexer (U12) is ok. Now the AD737 output is correlated, but proportion is weird and not linear. So either AD737 failed or its output is overloaded by something. Output goes to U12 B0. Output overloading seems unlikely, but possible I guess. If you could isolate it from B0 and it still makes same wrong output, then it must be damaged. Or just change it and hope for the best.

Good news is that RMS converter output looks to be sampled correctly by the meter, because in 600mV range input volatge is divided by 2, so 600mV AC becomes 300mV AC, which should produce -300mV DC and that -300mV DC seems to produce 600mV on the meter display (I'm using your measurement that -283mVdc RMS converter output produced 566mVac reading on the display). So it seems all good after the RMS converter.

[Multiple Cheese Slices]

I've taken some more measurements after disconnecting the output pin from the rest of the circuit. The input/output isn't all that different than when the output signal was going somewhere, and considering that everything before and after the RMS converter seems fine, this must be it.

Vac input	RMS input pin	DC output pin
1550mVac	777mVac	-204mVac
888mVac	444mVac	-14.4mVdc
446mVac	238mVac	-4mVdc

As I've said, not much difference between connected/disconnected output, as you can see from this and the previous table.

I've looked around nearby electronic part sellers, but none see to have the AD737J in stock. The delivery will take a little longer this time, hopefully it's the last one!
Thank you yet again for your support. I will update you with news as soon as I have any!

[Manul]

DigiKey has it, but you will pay some for shiping, also part itself is not the cheapest. Hope it will not be wasted. I mean, this repair is getting ridiculous. When I think about it, I can hardly comprehend how it all might have happened. Does not look like damage from meter input overload. Reverse battery connection should be clamped by diode CR3. Definetly very interesting case...

[Multiple Cheese Slices]

I have no idea how this happened as well. Unfortunately we'll likely never know; I have received it in this state, and only because it was in this state too. Nobody would want to get rid of a working one, aye?
But yeah, it is strange how so many things seem to have failed. It is possible that someone has attempted to repair it earlier since some screws were missing, but I really don't know more than that.

A thought came to my head. What if by any chance it got damaged by the weird jumping/fluctuating input? The input is always like this when the overload condition is occurring.
And so, I went to investigate - I was planning on supplying the meter with an overloading voltage and probing the AC signal before/after the U1 & the multiplexer and maybe some other stuff depending on what I would find happening with the U1/multiplexer. However now all I was able to find was stable (not fluctuating like with the overload conditions) ac voltage on the RMS converter input. After some testing, it turns out that this only happens when the output pin from the RMS converter is connected to the circuit.

I was assuming there was something wrong with the input section before the AD737 as well (because of the weird inputs), but would that be ruling it out?

P.S. Oh, and one more thing - what's the actual difference between AD737JRZ and AD737JRZ-5? The only thing I can find about it in the datasheet is that the JRZ-5 is "tested with supply voltages of +-2.5V DC". What does this mean? Will AD737JRZ also work?

[Manul]

Input to RMS converter is very unlikely to ever cause damage, because amplitude is naturally limited by supply rails of previous amp stage, so it can jump as much as it wants, but +-2.5V peak will be roughly the limit which is ok for AD737. Also I guess the multiplexer has clamping diodes. In my opinion, serious over voltage will fry the U1 long before RMS converter.

AC input voltage is stable, because meter does not know that there is any voltage at all if RMS converter output is not connected. Better don't do such experiments, especially with high voltages, just wait for new part to arrive. Jumping input was because of range switching or some protection kicking in (not sure, have not studied all schematics).

Service manual has (full?) part list. For example, U7 is AD737JR-5. U8 is LT1790ACS6-2.5 (these have different grades too).

Tested at +-2.5 means that it meets specifications at that supply voltage. Other parts are guaranteed and specified only for higher supply voltages. So yes, you need AD737JR-5.

[srmahaffy]

Hi, are you continuing to troubleshoot your 87V with the dash problem? I just purchased one on eBay with the same issue, that was missing the external selection knob, hoping the related circuit was why the “all dashes” . After opening it up, and doing some troubleshooting the switch circuit looks okay.
Please let me know if your continuing your “dash” repair, and do you have a schematic/ service manual for the 87V. I don’t see that on the fluke website.
Thanks in advance!

[3db]

Hi, are you continuing to troubleshoot your 87V with the dash problem? I just purchased one on eBay with the same issue, that was missing the external selection knob, hoping the related circuit was why the “all dashes” . After opening it up, and doing some troubleshooting the switch circuit looks okay.
Please let me know if your continuing your “dash” repair, and do you have a schematic/ service manual for the 87V. I don’t see that on the fluke website.
Thanks in advance!

Look at page one of this thread Manul posts a link to the manual.
Getting the manual and reading this thread may also help you fix your meter.

Regards
3db

[Multiple Cheese Slices]

Oh boy, it sure has been a journey. 

I am happy to announce that it's over though. But it has been lots of fun!
To get the most important (technical) things out of the way first - I have replaced the AD737JR-5, and now the Vac and A/mAac modes seem to be working fine. That means It's fully functional (though I haven't checked all the function buttons, haha) 

Now, for the less important stuff (you can skip this paragraph if you're not curious about non-technical stuff) you might be wondering what took me so long (this was definitely not planned). Looking at the shipping/component costs, I have decided to look for alternative ways of obtaining that part cheaply. I found out that Analog Devices had a free sample program in which the company would send customers free components and pay the shipping bill. I assumed it wouldn't take more than a few days, a week at maximum to arrive. Oh, how wrong I was... the procedure on Analog's part took around 12 days to get my order confirmed and packed up. Later came the fedex courier which extended the wait up until today. But hey, I can't complain too much - I'm not the one to look a gift horse in the mouth - it was all delivered for free.

I guess I should get the multimeter calibrated now if I want to use it. After the horrors it has seen, I think it would be perfectly in place. There is one metrology laboratory in Poland which claims to calibrate fluke equipment (here's their website if you're wondering: https://anmar-pl.com/)). I'll likely try to get a little more information about them, see if they are legitimate, and if yes, send the 87V to them for calibration. It can't be that expensive, right?

To srmahaffy,

Hi, are you continuing to troubleshoot your 87V with the dash problem?

Well, as you already know, my troubles have ceased for now. Seeing that your 87V has the same symptoms, I recommend checking the thing that have been found to be faulty in my case. Hopefully the culprit of your case is one of them, this would make things much easier.


Anyways, all the DMM modes now fully work. (at least I think so ^_^). It really isn't possible to convey how much I owe you people, especially Manul and Giosif - this really wouldn't be possible without you. I don't know how I can ever repay you for the amount of your dedication to help solve my issues. You are amazing; Thank you.

[srmahaffy]

Hi, Multiple Cheese Slices, I’m so glad you got your meter working! I will start a new post for my repair journey.  eevblog is a great resource! Thank you everyone so for.

[Manul]

Congrats on fixing that thing. I would just say to make sure that you bought the correct "A" grade 2.5V reference before doing calibration. A very strange multi-point failure, but with happy ending.