Unwanted beeps from my multimeter, what's causing them?

[elliot42]

I have a Fluke 87 IV, it's a second hand unit from eBay. It was working fine when I first got it but after sitting unused for a few months I pulled it out and it just beeped constantly. The beeps at the moment are consistent, but it has made more erratic random chunks of beeps at times as well.

It's in pretty decent shape physically; it does have corroded battery terminals, but there's no corrosion that's made it to any pads on the PCB. I've put in fresh batteries too.

Video of the problem: http://youtu.be/bvpiZjnImwM

Does anyone know what this means or what could be wrong?

[amspire]

Are the probes genuine Fluke probes?

Try the current ranges with the leads plugged in correctly for the current range you choose. Do the beeps go away? If so, it may the the lead detect switch (or whatever Fluke have) in that socket is shorted, or contaminated.

Richard.

[TerminalJack505]

Have you cleaned the battery terminals?  That may be the whole problem.  Piezo buzzers are actually very sensitive to voltage fluctuations.  It takes just a couple of hundred mV (even from a high impedance source) to make a Piezo element buzz.  You might have ground noise or something similar due to the corrosion.

You probably want to start out with something like baking soda to neutralize the corrosion.  Once that's done maybe used some fine sandpaper.  I'd avoid using steel wool since the little particles like to get into things and short stuff out.

[elliot42]

I haven't properly cleaned them because some of the contacts are actually broken (corroded through at the bend) and the batteries are held tight kind of bodgily. I contacted a local Fluke service centre about replacement parts and didn't much like the price; $55 for all new terminals
I might just have to go with that and see how it goes.

The leads I was using in the video aren't Fluke ones, I'll try again with the Fluke leads too.
If there is a problem with the lead detection wouldn't the display show 'LEADS' in error?

[mikeselectricstuff]

My guess is it thinks one of the  current leads are in - could just be something has fallen into the socket, or a pin has come off - pull it apart & take a look. 

[G7PSK]

Does it have an audio low battery warning. If so the corroded terminals could fool the system into thinking the battery is low even if you put a new one in.

[elliot42]

I can't see anything in the terminals but it is a bit hard to see in there.
I haven't replaced the battery terminals yet but I'm not so sure if I should now, I've had a closer look at the PCB and it looks like there is some corrosion on some pads and tracks. And now I'm having general power issues after messing about inside, probably the battery terminals not contacting the PCB properly...

[elliot42]

I've now tried powering the multimeter directly from a lab power supply, just touching probes to the PCB pads that normally contact the battery terminals in the rear part of the case. I still got the same result, beeping, so that should rule out the batteries/contacts causing the issue.

I don't think I can remove the test lead terminals without desoldering but I've looked inside with a magnifying glass and light and can't see anything more than some dust that could short out the terminals. I've also tested continuity between the 2 PCB pads of each of the terminals. The V and COM both have continuity with and without leads plugged in. The A and mA both have continuity only when leads ore plugged in, otherwise not.

I've also inspected the PCB more closely with a magnifying glass. There was muck on the rear side of the PCB (nearest the batteries) that looked like some residue from something, the front side was all clean.

There is corrosion on the screw and PCB ground via that holds on a shield over the main processor and holds the PCB to the front part of the case (there's one other screw lower down that was fine).
There is corrosion on the left-most pad of a long, thin, vertical white component.
And there is possibly some corrosion on one of the traces leading to the main processor; or it could be something else.
These are circled in the attached image.

What is the white component? (I think it's the thick film resistor network?) Should I clean and re-solder the corroded joint?
Any other ideas?

[TerminalJack505]

Is the "thin, vertical white component" in the picture?  The only white component I see is a ceramic cap.

The corrosion in the picture doesn't seem too bad.  I doubt it's an issue.

[elliot42]

No, sorry, it's on the reverse side of the board. The corroded pin and the line of pins are part of the thin white component. I'm also pretty sure its the precision resistor network.

[amspire]

Does it beep in every current range with the leads plugged into the correct socket for that current range?

[elliot42]

A/mA range, 10A socket = a few beeps but none to start with
uA range, 10A socket = constant beeps and "LEADS" error
A/mA range, 400mA socket = beeped when I applied power but then nothing, when I turned the meter over to look at the screen it started beeping...
uA range, 400mA socket = sporadic beeps...

It seems to change from beeping to nothing when I turn the meter in different ways on other ranges too (ohms, volts).
And the beeping is always at the same even rate, like an alarm.
And I can still take a measurement while it's beeping (I was just measuring the voltage I'm supplying the PCB).
It now seems more and more like something shorted out or something moving and shorting something out when I turn the meter different ways.

[elliot42]

Buuut, I still can't detect any shorts in the current terminals without leads plugged in.

[amspire]

I can't help thinking it is some weird fault in the 10A probe detect circuitry. What, I just don't know the meter well enough to guess.

This looks like a service manual, but no schematic it appears.

http://assets.fluke.com/manuals/87_89iv_umeng0200.pdf

Richard.

[SeanB]

It could be the corrosion has either shorted or opened a track or a via, buried or blind, on the circuit board. Only way is to use a schematic and trace all the pins associated with that for continuity and high resistance shorts to either each other or to power rails. Difficult and long process.

RBR

[elliot42]

A schematic would be useful, but I think I'd have a hard time (impossible, even) tracking down the problem as you've described, SeanB. That's a bit beyond my skill, and patience.

I also don't want to have it serviced because I imagine it would cost more than I actually paid for the meter in the first place to fix it (considering just the replacement battery terminals are $55).  So I guess there's no harm in me going at it with a soldering iron to try and fix that corroded joint and remove the lead terminals in case there's a short somewhere in there.

[pickle9000]

Here is the location for the fluke service manuals (with schematics). They are also available on flukes website but this one is much easier to navigate.

http://www.mods.dk/manual.php?brand=fluke

Check your switches (including the main selector). Clean everything up but avoid solvents for now use a dry clean brush (toothbrush), remove solder from thru holes (in the offending area) and replace with fresh. Check the probe detection switches, should be indicated on the display. Take some measurements of known items and see if they are correct. Write everything down. Solder a couple jumpers in place and remove any clips that are corroded. You need to see what is down there.

Make sure you are looking at the correct revision number for the PCB. Always start at the begining and take nothing for granted. That means check your power supply voltages, and xtal freq before doing anything else.

If you do decide to clean the board be aware that even plain isopropyl alcohol can cause problems if it melts a coating and puts it where it should not be (like on a switch contact). Solvents can also get into pitted solder and when heated it will spatter hot solder in your general direction. Do your soldering first and allow the board to completely dry before testing. Use canned air to blow solvents out from under components. Put the board in a warm place 100-140F for a couple hours to ensure its bone dry.


...mike

[elliot42]

Thanks for the extra info, Mike.
There's nice detailed information, schematics and block diagrams for the original 83,85,87 series. Unfortunately though I can't see any schematics for the 87 IV series.

I'll start off by cleaning / re-soldering the visibly corroded joints and cleaning any dirty parts of the board with a dry brush.
The rotary switch contacts look nice and clean but I'm not sure about the lead terminals so I think I'll try removing them from the board and testing with some wire shorting the pads to simulate an attached test lead.
There are no other physical switches on the board, hopefully some of this will help, otherwise I may just have to write it off as I don't have a schematic to test further.
I'll steer clear of the alcohol for now as it looks pretty clean.

[pickle9000]

Sounds like you are the right track. Even if you don't have the exact schematic check one of the other 87's and see if you are using the same processor. If so you will probably have the same power supply and beeper circuit (part numbers may differ).

I haven't looked at any of the schematics but check to see if the beeper is driven directly or is there a continuity check system involved. In either case you can eliminate the actual beeper as the problem cut the lines that activate it. If the processor is the same don't be afraid to lift or cut a couple legs on the input lines to the processor. If it isn't use the schematics a guide, the input circuitry will be similar if not identical.  That should allow you to see if it's localized to the processor or not. If it is then try getting into the diagnostics for the meter (or do that first). To get into diagnostic or features hold down a button while turning on. You may need to try several buttons. That's pretty typical of most high end gear (but not all). It's like the BIOS settings on a PC. It's broke now so screwing around with it is not a major problem. You may as well learn something from it.   

If you are scoping the crystal be careful of loading, use a 10x probe if you have any problems. If you have a digital scope it should be fine. Frequency counters are great at this.

If you can localize the problem, even to a place you can't repair it's a great exercise, and I for one would give you a thumbs up.

...mike

[sonicj]

if you use a dry brush, it should be of the ESD safe type.

battery terminals & boards almost never spontaneously corrode, especially from a quality manufacturer like fluke. i suspect a battery has leaked its corrosive electrolyte at some point in the meter's past. i would start by neutralizing the electrolyte, then cleaning the board with isopropyl and a brush. if that doesn't fix it, i would then test any suspect traces. my 2¢ fwiw.
-sj

[elliot42]

Thanks for all the tips.
Yes, the corrosion is all over the battery terminals and has eaten through 2 of the sprung terminal ends, breaking them off. So it would have originated from a leaky battery (or batteries) some point in the past. Some must have leaked through onto the PCB, the corroded pad on the PCB is underneath some shielding but it's still pretty close to underneath the battery terminals.