Author Topic: Fluke 8845A (like 8846A) 6.5 digit Multimeter showing "overload" at all times  (Read 15211 times)

0 Members and 1 Guest are viewing this topic.

Offline t-17Topic starter

  • Contributor
  • Posts: 24
  • Country: de
Thanks TiN, you're absolutely right.  ;D
My bad. I guess I was just blown away after several weeks of troubleshooting when I finally started making progress.

Talking about progress: I have just solved the last problem and (apparently) fixed frequency and period measurements. I was wrong with the suspicion that U16 would have something to do with the frequency measurement. Rather, it seems to be used for autoranging when in ACV mode. The IC I was looking for was U32 (also an LM393). I found it after close inspection of the PCB. I figured that for frequency/period measurements one would want the AC signal normalized to a certain amplitude range first. So the signal should be tapped off somewhere, where this is the case. And I knew from the ACV repair part that the signal has its largest amplitude, and is also within limits at all times, at the output of U20. And in fact, the output of U20 would feed U18, the True RMS converter, and *something else* (the trace splits into two on the TOP side of the board, and one of them mysteriously vanishes into a via). It took me quite a while though, to find U32 sitting next to the power connector of the analog PCB part. I thought it would rather have something to do with power rail regulation. I could not be more wrong :o

Now, the multimeter also has frequency and period measurements back, BUT: When measuring period or frequency, the flash symbol appears in the display, which usually means "potentially harmful voltage at the input terminals". I guess it may be that there is still something wrong somewhere. Can somebody with a working unit confirm that under normal conditions the flash symbol does not show (when measuring frequency)? Apart from that, all basic measurement functions are back. But I'll perform a complete test, testing every single function right through the user manual to be sure.

I have made a list of ICs and their function, in case somebody wants to repair this beast. Note that the entire AC circuitry is under a metal can and cannot be seen easily as long as it is not removed.

DesignatorIC CodeFunction
U5LF356AC input buffer, only used for higher ranges (at least 750V range, maybe 100V as well-- not checked)
U12LF356AC input buffer, input and output are at the same level (no amplification), only used for lower ranges
U14AD825AC input buffer, input and output are at the same level (no amplification)
U16LM393AC amplitude comparator, 99% positive that it is used for ACV autoranging. Compares against +0.5V and -0.5V.
U18AD637True RMS AC -> DC converter, used for True RMS ACV/ACI measurements
U20AD8510Input amplifier/ buffer for AC, frequency and period measurements.
U36LM393Comparator for frequency and period measurements
U37DG444Analog multiplexer used for DCV range switching
U49AD8510Main integrator opamp of the ADC
« Last Edit: October 07, 2016, 02:53:55 pm by t-17 »
 
The following users thanked this post: TiN, Mickle T., rstoer, voltsandjolts, coromonadalix, Muxr

Offline hbrown

  • Newbie
  • Posts: 3
  • Country: us
Just wanted to say thanks! Your investigative work and good reporting saved me a great deal of time. I received an 8846A for free because all the AC ranges (ACV, ACI, frequency) showed overload. I swapped U18 and U20 and its good as new. Just need to verify calibration now.
 
The following users thanked this post: t-17

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5795
  • Country: ca
@ tin  i would absolutely not use your method in this case, it may stress the pcb pads, the hot air removal is the best and fastest non damaging solution for me, i work a lot on pcb's   and hot air is the best.

Tried on defectives pcb the solder blob on one side and lift the other side ....  you have to be fast and precise ...
 

Offline TiN

  • Super Contributor
  • ***
  • Posts: 4543
  • Country: ua
    • xDevs.com
If that works for you, then already have your solution. Just keep in mind that stress from heat to precision resistor network nearby can cause permanent shift/drift and non-reversible stability issues after using hot air...
YouTube | Metrology IRC Chat room | Let's share T&M documentation? Upload! No upload limits for firmwares, photos, files.
 
The following users thanked this post: t-17, analogRF

Offline coromonadalix

  • Super Contributor
  • ***
  • Posts: 5795
  • Country: ca
kapton can be of some use, but in this case i would provide an heat shield deflector around it.

normally   when pcb's thru holes are installed and soldered like we do in my company, the boards pas thru an pcb oven for at least 30 seconds to 1  minute,  and even precision part have some heat tolerances to do so.

When i remove smt parts, it take under 10 secs to remove them ..... sometimes even under 5 secs, heathing only the part with the right nozzle and nothing around it with or without kapton protection move or shift ... it really depend of your technique



 

Online analogRF

  • Frequent Contributor
  • **
  • Posts: 955
  • Country: ca
Thanks TiN, you're absolutely right.  ;D
My bad. I guess I was just blown away after several weeks of troubleshooting when I finally started making progress.

Talking about progress: I have just solved the last problem and (apparently) fixed frequency and period measurements. I was wrong with the suspicion that U16 would have something to do with the frequency measurement. Rather, it seems to be used for autoranging when in ACV mode. The IC I was looking for was U32 (also an LM393). I found it after close inspection of the PCB. I figured that for frequency/period measurements one would want the AC signal normalized to a certain amplitude range first. So the signal should be tapped off somewhere, where this is the case. And I knew from the ACV repair part that the signal has its largest amplitude, and is also within limits at all times, at the output of U20. And in fact, the output of U20 would feed U18, the True RMS converter, and *something else* (the trace splits into two on the TOP side of the board, and one of them mysteriously vanishes into a via). It took me quite a while though, to find U32 sitting next to the power connector of the analog PCB part. I thought it would rather have something to do with power rail regulation. I could not be more wrong :o

Now, the multimeter also has frequency and period measurements back, BUT: When measuring period or frequency, the flash symbol appears in the display, which usually means "potentially harmful voltage at the input terminals". I guess it may be that there is still something wrong somewhere. Can somebody with a working unit confirm that under normal conditions the flash symbol does not show (when measuring frequency)? Apart from that, all basic measurement functions are back. But I'll perform a complete test, testing every single function right through the user manual to be sure.

I have made a list of ICs and their function, in case somebody wants to repair this beast. Note that the entire AC circuitry is under a metal can and cannot be seen easily as long as it is not removed.

DesignatorIC CodeFunction
U5LF356AC input buffer, only used for higher ranges (at least 750V range, maybe 100V as well-- not checked)
U12LF356AC input buffer, input and output are at the same level (no amplification), only used for lower ranges
U14AD825AC input buffer, input and output are at the same level (no amplification)
U16LM393AC amplitude comparator, 99% positive that it is used for ACV autoranging. Compares against +0.5V and -0.5V.
U18AD637True RMS AC -> DC converter, used for True RMS ACV/ACI measurements
U20AD8510Input amplifier/ buffer for AC, frequency and period measurements.
U36LM393Comparator for frequency and period measurements
U37DG444Analog multiplexer used for DCV range switching
U49AD8510Main integrator opamp of the ADC

I know this is a very old thread but did you find anything else wrong with the unit with regard to the flashing symbol you mentioned?
 

Offline t-17Topic starter

  • Contributor
  • Posts: 24
  • Country: de
I know this is a very old thread but did you find anything else wrong with the unit with regard to the flashing symbol you mentioned?
First of all, sorry for the late reply. I just want to confirm that even after all these years, the unit is still operating fine. So besides the flash symbol, there is nothing wrong with it. Also, I think I've seen it appearing in brand new units as well. So it does not seem to be a malfunction, but rather a firmware issue (perhaps only with older units like mine)?
 
The following users thanked this post: coromonadalix, analogRF

Offline Miti

  • Super Contributor
  • ***
  • Posts: 1320
  • Country: ca
So besides the flash symbol, there is nothing wrong with it. Also, I think I've seen it appearing in brand new units as well. So it does not seem to be a malfunction, but rather a firmware issue (perhaps only with older units like mine)?

It shows up on mine as well so I don't think it is an issue but rather a feature. When I've got it it had an inguard/outguard comm issue and the culprit was an IR LED. I replaced it with one from an old DVD remote control and it still works after 4 - 5 years.
Fear does not stop death, it stops life.
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
Hi all,

I have a Fluke 8845a that have developed some issues.

DC voltage and current work all fine. This issues are with anything AC.

Frequency measurement:
- Works for "easy" waveforms like square wave it measure correct.
- For a little bit more difficult to measure waveforms like Sine need to be above 7khz for lower voltages (below 1V) to measure correct. Above 1V, after the relay kick into higher voltage range it do not measure correct.
- When measure wrong in frequency mode, typically 20% too high and numbers jumping around. But can be much worse than that.

AC Voltage:
- before this issue happened it was very accurate. Now using DDMcheck, it shows 0.1 V low.

AC current 400mA port:
- No current on terminals and it still shows 5 mA
- At around 15 mA is start to show correct numbers.

AC current 10A port:
- I don not have any strong AC current source enough to try, but with noting connected, it 0.5A.

Any suggestion from the expert panel here where to start looking for the error?




 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
There is a shared AC/DC converter for both the current and voltage readings.  So if zero current results in a non zero reading, I would expect a similar thing with AC voltage - it this happening ?

Testing is usually easier with the voltage ranges.

The frequency reading usually only uses a simple comparator for the input. So low slope signals are always a bit tricky to measure.
To high a reading could be some ringing or interference causing multiple counts in some cases, especially with a low slope, so the input is longer near the sensitive range.

There is a chance the input amplifier may be oscillating for some reason. This could explain trouble with the frequency measurement and the non zero reading. Such an oscillation may be even visible to the input in a lower voltage (e.g. 1 V) AC range.
Another suspect would be hum on the supply voltages.   So as usual: check the supplies.
 

Offline t-17Topic starter

  • Contributor
  • Posts: 24
  • Country: de
I agree with Kleinstein. First check all voltages, then inspect anything related to the AC path. Check the component list I've made. It could also be a problem with one of the many opamps, probably early in the AC signal path since anything AC related is having issues. Perhaps several parts have issues.

I'd check all signals around these parts for stability (there should be no oscillations), then also check if voltages around the opamps are okay and behave as one would expect.

 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
There is a shared AC/DC converter for both the current and voltage readings.  So if zero current results in a non zero reading, I would expect a similar thing with AC voltage - it this happening ?

Testing is usually easier with the voltage ranges.

The frequency reading usually only uses a simple comparator for the input. So low slope signals are always a bit tricky to measure.
To high a reading could be some ringing or interference causing multiple counts in some cases, especially with a low slope, so the input is longer near the sensitive range.

There is a chance the input amplifier may be oscillating for some reason. This could explain trouble with the frequency measurement and the non zero reading. Such an oscillation may be even visible to the input in a lower voltage (e.g. 1 V) AC range.
Another suspect would be hum on the supply voltages.   So as usual: check the supplies.

Thanks for comments!

AC Voltage was also knocked out of calibration when this fault occurred. Not as bad as current though. With shorted probes it show 15 mV, with 5V test reference at 100 Hz it shows 4.88V and with open leads it shows 0.28 V.

So yes, AC measurement for both current and volt is wrong. And frequency. But DCV DCI and resistance is well within specs.

Before this incident the meter had no problem picking if the frequency signals in sine etc. My fluke 87V also can easily do it. So also something that got impacted with this error.

The error is also dependent on what range. So jump around when switching range. Particularly when the relay is clicking for new range.

 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
I agree with Kleinstein. First check all voltages, then inspect anything related to the AC path. Check the component list I've made. It could also be a problem with one of the many opamps, probably early in the AC signal path since anything AC related is having issues. Perhaps several parts have issues.

I'd check all signals around these parts for stability (there should be no oscillations), then also check if voltages around the opamps are okay and behave as one would expect.

Thanks. Will do so.

Any though why different ranges show different error with same input signal?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
The AC input takes some time (e.g. 1-5 seconds ?) to settling when switching range - that is normal with the analog RMS-DC converter.
For testing it would be better to use manual ranging - autoranging only add complications not needed, as it may not be clear which range is used and there may be constant up/down jumping.

Different ranges use different amplifcation setting at the front end an can thus be effected different. Especially the higher voltage (e.g. > 1 V) ranges will have some "divider", which often is a different input path.
The voltage and open input case is of limited use, as it can easily pic up some hum.
DCV and DVI have a different input parts. The problem is likely in the AC input amplifier or RMS to DC converter. After the RMS converter the signal usially joins the DC path somewhere.

15 mV with a short is about as bad as the current readings.
The deviation seems to be combination of some offset/background and a scale factor that is a little low. So different range may be effected differently.

Without knowing the circuit configuration it is a bit tricky to localize the error more.
After checking the supplies one could look at the input to the AD637 - ideally with a scope, to see if the problem is more with the amplifier or the converter itself and maybe the connected filter.
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
The AC input takes some time (e.g. 1-5 seconds ?) to settling when switching range - that is normal with the analog RMS-DC converter.
For testing it would be better to use manual ranging - autoranging only add complications not needed, as it may not be clear which range is used and there may be constant up/down jumping.

Different ranges use different amplifcation setting at the front end an can thus be effected different. Especially the higher voltage (e.g. > 1 V) ranges will have some "divider", which often is a different input path.
The voltage and open input case is of limited use, as it can easily pic up some hum.
DCV and DVI have a different input parts. The problem is likely in the AC input amplifier or RMS to DC converter. After the RMS converter the signal usially joins the DC path somewhere.

15 mV with a short is about as bad as the current readings.
The deviation seems to be combination of some offset/background and a scale factor that is a little low. So different range may be effected differently.

Without knowing the circuit configuration it is a bit tricky to localize the error more.
After checking the supplies one could look at the input to the AD637 - ideally with a scope, to see if the problem is more with the amplifier or the converter itself and maybe the connected filter.

Hi, thanks for comments!

Driving voltages looks stable and nice. Will still swap all electrolytes.

Yes, AC input path ICs or the RMS to DC converter clearly the suspicious items. I put an order of all chips in should have them here in a couple of days.
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
With the AC part one can usually pin-point the defect better than just a curde swap all caps and chips. When changing OPs one may need to do a new adjustment of the frequency flatness. The AD637 is a relatively expensive chip. One can usually at least check at the input of the AD637, if the signal there is OK. This would tell if the problem is with the amplifier or the AD637 and maybe behind this.

It is not so much about the cost of the chips, but the indirect changes and possible stress to the PCB, that one should test more to narrow down the defect more to individual OPs / CMOS switches.

Another way for checking would be to apply a test signal AC signal and then look at the output of the OPs in the area in the different ranges. This would help to know which chip is used in which range. This can help to get a crude idea of the circuit and signal paths.
If some OP is oscillating the defect could be as simple as a cold solder joint / broken (litterally mechanical) small capacitor.
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
With the AC part one can usually pin-point the defect better than just a curde swap all caps and chips. When changing OPs one may need to do a new adjustment of the frequency flatness. The AD637 is a relatively expensive chip. One can usually at least check at the input of the AD637, if the signal there is OK. This would tell if the problem is with the amplifier or the AD637 and maybe behind this.

It is not so much about the cost of the chips, but the indirect changes and possible stress to the PCB, that one should test more to narrow down the defect more to individual OPs / CMOS switches.

Another way for checking would be to apply a test signal AC signal and then look at the output of the OPs in the area in the different ranges. This would help to know which chip is used in which range. This can help to get a crude idea of the circuit and signal paths.
If some OP is oscillating the defect could be as simple as a cold solder joint / broken (litterally mechanical) small capacitor.

Thanks for more comments! Much appreciated! I bought most chips, just in case, so i don't end up figuring out i need something I don't have when i am repairing. The AD637, as you wrote, the most expensive.  Around 30 USD form Mouser. Then around 20 USD altogether for one or two of the other potential suspects. Then a new set of electrolytes as those that are now in the meter have been around for 15 years or so.

So with the extremely good mapping of purpose of chips in this thread, I am quite hopeful to pinpoint the problem and swap out the correct chip only.

Again, thanks for comments. Will be out traveling for some time, so may first get to dig into this in some time. Will report back on this, and if I do any other findings.
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
Hi repair wizards!

I now been looking more into my 8845a.

I discovered that the negative 15 V supply line is negative 7 V and the mosfet controlling is much hotter than the 15 V positive mosfet. All other voltages seems correct.

This makes me suspect that there is some short circuit or some other issue that draw way too much current from negative 15 V.

I have checked negative 15V in input pins on OP amps etc that use negative 15 to confirm that problem is everywhere where negative 15V is used, which it is.

I checked if any component hot, but none except the voltage regulators and power mosfets.

if it would be possible to measure current this could be a way to trace the problem, but since everything soldered on a PCB it is very difficult to measure currents.

Or could it be the regulation circuit for the mosfet that have issues? A LM358M is used. I cant find any reference voltage from any zenerdiod thought. Are there any OP amp + mosfet voltage regulators that do not use any zenerdiod (or similar as voltage regerence)?

Any suggestions on what to check next?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
One point to check is if the supply is DC or oscillating.

In quite some older circuits they use one supply (e.g. the positive) as a reference for the opposite sign reference.

Most if the OPs would be using something like a +-15 V supply, so this would be the same current for the positive and negative side.
It would however make a difference of the output of an OP is shorted to ground or similar.

Another point to check is the voltage on the input side of the regulator - with a bad fitler cap one could have excessive ripple there and this can make the voltage drop at least on average.

7 V is still not a dead short - a short from a MLCC is nothing so unusual. With extra series resistors this would usually give quite hot resistors, possibly even burnt ones.  This would also cause a much lower voltage for the effectged part (e.g OP).
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
One point to check is if the supply is DC or oscillating.

In quite some older circuits they use one supply (e.g. the positive) as a reference for the opposite sign reference.

Most if the OPs would be using something like a +-15 V supply, so this would be the same current for the positive and negative side.
It would however make a difference of the output of an OP is shorted to ground or similar.

Another point to check is the voltage on the input side of the regulator - with a bad fitler cap one could have excessive ripple there and this can make the voltage drop at least on average.

7 V is still not a dead short - a short from a MLCC is nothing so unusual. With extra series resistors this would usually give quite hot resistors, possibly even burnt ones.  This would also cause a much lower voltage for the effected part (e.g OP).

Awesome! Thanks a lot! Found it! Cause of the -7 volt was a somewhat shorted tantalum capacitor between ground and the output of the power regulating mosfet. Removed it, and now -15V! Replacement on the way. This cap is located just next to the hot power mosfet, so at first i didnt find it, as i didnt want to touch to close to this mosfet as it was burning hot.

Now DC is working properly, AC volt and AC current still off. Will do some tracing of AC signals and see what i find. Maybe just putting back a working tantalum capacitor may help.

Thanks again!!!
 

Offline aronake

  • Regular Contributor
  • *
  • Posts: 183
  • Country: hk
Kleinstein, t-17,

You are true superstars! Thanks to you I now have a fully functional 8845a, and learnt a load about electronics repair! Hats off! And it even measures more accurate now than before.

Showed out that removing the -15V 22uf tantalum capacitor brought DC measurements back to spec. Soldering on a new also brought AC measurements back to spec!

I also have 100 USD worth of AD637, OP amps etc that I bought and never had use for. But they may come to use some other time.

I thought I had checked the power rails before first, before looking into more complicated stuff, but apparently not.

Its a really great feeling to manage to repair something like this!

Again, thanks!!!
 

Offline yildi

  • Newbie
  • Posts: 4
  • Country: de
Hello all together,

I also have a Fluke 8845A that no longer works properly. To be exact, only the 400 mA current input is not working. The 10 A input and all other functions work. I checked the fuses, of course, they are both fine!

Since my unit is still sealed, I have not opened it yet. First I wanted to do some research, but unfortunately, I haven't found anything about this exact problem yet. Has anyone of you ever had such a problem? It would also help me a lot to have a high-resolution picture of the board to be able to estimate how promising a self-repair is. Maybe someone has tips and/or a good picture of the PCB.

Thanks a lot for your help!

EDIT: With "the 400 mA current input is not working" I mean that the displayed value is in the uA range and changing due to noise. However, when connecting a current source the displayed value does not change. In fact, no current from an external current source is flowing into the 8845A.
« Last Edit: November 23, 2021, 09:53:42 pm by yildi »
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
https://xdevs.com/fix/f8846a/

Are all the low current ranges inop or just the actual 400mA range?
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 

Offline yildi

  • Newbie
  • Posts: 4
  • Country: de
None of the manually adjustable ranges of the 400 mA connector work. Thanks for the link, I will check that out!
 

Offline bdunham7

  • Super Contributor
  • ***
  • Posts: 7729
  • Country: us
None of the manually adjustable ranges of the 400 mA connector work. Thanks for the link, I will check that out!

OK, just be aware that there are multiple shunts involved then, and if the fuse is actually good (I'd triple check!) and hasn't been replaced with a wrong one sometime in the past, you hopefully will have a straightforward wiring or connection problem.
A 3.5 digit 4.5 digit 5 digit 5.5 digit 6.5 digit 7.5 digit DMM is good enough for most people.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf