Fluke 189 multimeter repair advice needed

[sw2022]

Hi all

I’ve not posted here before though thought I’d try here for some advice about fixing my Fluke 189 meter. I’ve read as many threads here and elsewhere as I can find on the thing but haven’t sorted the problem sadly. I’ve had it since new and it’s been a great tool, recently it started giving leads errors (common I know) and the 10A current range was not working, power on went intermittent too. I sorted the power with new battery connector hardware. I got brief function after that with the meter all working so I think most of the board will be good.  I also had noticed during the battery clip repair that one of the 10A fuse clips was loose, thought that was the rest of the problem. I see it’s normally both glued down and soldered at the two small tabs. Didn’t seem to solder well at the tabs despite cleanup, so I cleaned up and tinned the whole pad below it , plus the back of the clip and attached it well with solder as pictured. I’ve put up pics of the whole board as well as that repair in case it helps. All seemed good, I’ve got continuity from clip to other components on the board - the black ?? component below it.  Repair doesn’t seem to have damaged anything.

I reassembled and frustratingly still get a leads error. The meter boots up fine with no leads attached when on a current range so I guess it *thinks* it’s got current test leads connected even when it has none at all. My intuition says the problem has nothing to do with the original loose fuse clip and repair.

I’ve cleaned the input sockets they look fine. Good continuity to the PCB.  I see there are split connectors in them which are used on the current inputs only to detect banana plug connection. Electrical rather than the IR detection seen on more modern models. One half of the banana input goes to  two separate brown 1 Meg resistors, one for mA and one for A. Resistors seem good. Those resistors connect to what I guess is the base of an SOT transistor , shown on the near edge of the board in the middle in pic 4, one transistor for each input, continuity good. From there I have no clue how the input detection is working. PNP or NPN transistors? Any chance a transistor has failed? 

I wish I had a schematic for this board but none seem to be out there.

Overall the board looks clean with no obviously damaged components or joints.

Any help with info that might help repairing this appreciated, it’s a great meter and I think it should be repairable!  I’m not seeing any professional services fixing older Flukes here in the UK though again if I can’t fix it I’d try that route if it were available.

Steve

[daisizhou]

You have to clean the inside of the socket I painted in the picture.(It is recommended to use 97% anhydrous alcohol for cleaning)
After washing, put it in the sun to dry.The inside of the jack must be clean and free of moisture.

[daisizhou]

Another problem is here, between the 2 pins must be clean.If you use alcohol immersion cleaning, it is recommended to dry operation.
Details are key to maintenance, 187/189 are leak sensitive.

[sw2022]

Hi , thanks. I cleaned the insides of the banana sockets with 100 pc isopropyl alcohol. I dried them with paper and left things to dry out for days, no change. As I said there is easy and good continuity to the PCB from each half of the shells in the sockets. There is also no continuity between the two halves of the socket on the mA or A sockets. So I think the sockets are clean and good,

Steve

[daisizhou]

It is recommended that you use 97% alcohol for soaking.Especially the 4 socket locations.It is best to use a syringe to flush the bottom of the socket while soaking.Clean up the dirt inside the crevice.Dry at low temperature after all cleaning

I don't quite understand what is the main problem you are having now.But if you clean it with an alcohol swab, there will be residue under the socket

[sw2022]

Hi

Thanks. The board itself has not been soaked as it was very clean. I did wipe the area of the fuse holder down with alcohol on paper after soldering.  The sockets were swabbed out fairly vigorously using a sequence of paper plugs held in small forceps, they were soaked in alcohol, so no dirt is left.

Steve

[shakalnokturn]

Do not use your meter to measure itself.

If this is any help: On my (working) 189 powered off and depending on the measurement polarity, 10A socket reads 4.8 or 2.87M , 400mA socket 7.1 or 6.05M .
So if you measure more look for the trouble elsewhere if you measure less it could need more cleaning.

[rsjsouza]

Interesting; my 189 powered off reads about the same regardless of polarity.
mA: 10kΩ (9,97kΩ when inverted)
A: 0,19Ω

In diode mode, it reads the same regardless of polarity.
mA: 1,65V
A: 0V (short)

This was tested with three DMMs (Fluke 87V, Keysight U1282A and Uni-T UT61E)
It is probably dependent on the ohmmeter current of the DMM.

[shakalnokturn]

I realised my previous input maybe lacked an important detail: The measurements are between the 2 contacts on the "split" sockets, the ones that allow "leads" detection not between A/mA sockets and COM socket.

[mqsaharan]

I’ve had it since new and it’s been a great tool, recently it started giving leads errors (common I know) and the 10A current range was not working, power on went intermittent too.

Hi Steve,

The lead error is usually due to the current leakage between two halves of the current input jack(s). That is why everybody is asking you again and again to carefully check and clean these.

I sorted the power with new battery connector hardware.

Did the battery clips break inside the battery compartment?
I'm just curious.

I reassembled and frustratingly still get a leads error. The meter boots up fine with no leads attached when on a current range so I guess it *thinks* it’s got current test leads connected even when it has none at all. My intuition says the problem has nothing to do with the original loose fuse clip and repair.

It does sound like there is still some current leakage between the two halves of one or both of the current sockets.

I’ve cleaned the input sockets they look fine. Good continuity to the PCB.  I see there are split connectors in them which are used on the current inputs only to detect banana plug connection. Electrical rather than the IR detection seen on more modern models. One half of the banana input goes to  two separate brown 1 Meg resistors, one for mA and one for A. Resistors seem good. Those resistors connect to what I guess is the base of an SOT transistor , shown on the near edge of the board in the middle in pic 4, one transistor for each input, continuity good. From there I have no clue how the input detection is working. PNP or NPN transistors? Any chance a transistor has failed? 

If you think the sockets are clean, the 1 Meg resistors are all right, then you need to check these two transistors if one or both have failed. One leg of these transistors is attached to each of 1Meg resistors. Second leg of these two transistors is attached to ground. And third leg of these two transistors is connected together and it goes to the pin 56 (if I've counted it correctly) of the front end chip Fluke 669918, the one under the shield.

There isn't much that can go wrong here. The only components for the detection paths are sockets, 1Meg resistors, the two transistors and then the front end chip itself. You can check these carefully and thoroughly by yourself and hopefully you'll resolve the problem.

Hope it helps.

[sw2022]

Hi

That’s all really helpful. You’ve explained the detection circuit well there.  I’m going to check the resistance between the two halves of the socket contacts this evening and will report. I will clean the sockets once again regardless of resistance results and give more drying time, I do appreciate this is often the issue from what’s been said here. Resistance will be checked on my bench multimeter of course. Thanks once again to all for the advice so far!

Steve

[sw2022]

Battery clips had just corroded by the way. Think low level leakage from a cell may have been the issue. No evidence of anything getting in the meter though

Steve

[bdunham7]

You might want to just replace the jack assembly.  I don't know if they'll come back off backorder, but I got mine from Newark/Farnell.

FLUKE 89-4-8002

https://www.newark.com/fluke/fluke-89-4-8002/input-receptacle-assembly/dp/06H3110

[sw2022]

So ... I’ve measured the resistance across the split socket, best way seemed to be to press the meter probes on the cleaned tips of the jack sockets pins. See my pics. I get 1.31M accross the 10A socket and 1.38M for the 400mA socket. That’s of course with the board out of the meter case as pictured. Seems lower than what was suggested though almost identical for the two, obviously there’s an input circuit in the loop here though. Certainly nowhere near a short...

Steve

[shakalnokturn]

If you've been through mqsaharan's advice and found nothing wrong it's probably time to open the link to the split sockets, either by lifting the 1M resistors if they allow it, or by removing the socket assembly (which is what I'd do).
The point is being able to measure the resistance of the split sockets separately from the rest.
If the sockets alone still have a measurable resistance they need more cleaning. (Don't forget what daisizhou mentioned: I think he meant that the bottom of the sockets is not necessarily water-proof so you may end up with the conductive residue washed down in an invisible spot between the PCB and socket assembly that's why it's better to remove the assembly. Without heating too long...)
If they don't then the problem must be on the PCB, In that case I'd remove the shielding and double check for battery damage.

[sw2022]

Thanks for that, I thought maybe the circuitry was playing a part. I’ll treble check everything then consider removal of the sockets,

Steve

[daisizhou]

Moving sockets is a dangerous move.Because you don't have the same solder material as the original,In addition, if the temperature cannot be accurately controlled during the soldering process, it is easy to cause the PCB to deteriorate (overheat) and lead to conduction.(Solder refers to the formulation of the solder, as different solders result in different contact resistances)

So I suggest you soak all sockets in 99% alcohol,Then take it out to dry.

I think this is the safest operation

[daisizhou]

(Sorry, this is the best picture I can get,I don't have a dedicated image capture camera)
The socket is through and if you wash the top it will leak to the bottom,Because you usually insert and unplug the plug, the metal powder will remain inside.If you wash it will flush the residue to the bottom (PCB touches here).lead to switch on.

So I suggest you soak the 4 sockets in a container of 99% alcohol.And use a syringe to draw alcohol solution, spray the top and bottom of the socket.Wash the inside of the gap multiple times

[sw2022]

Thank you very much for that advice. I will try the soaking rather than going straight to socket removal then. Your experience is appreciated,

Steve

[sw2022]

I just want to say a big thank you for pushing me to clean this meter properly, both you and the other contributors here were so helpful, it’s really appreciated. I now understand the issues with high impedance boards a lot better and how easy it is for contamination to become a problem. I both soaked and sprayed the bottom of this board for multiple days using isopropyl alcohol. I also cleaned every place I could get to again. I then dried the board out using warm air. I left if for a couple of weeks to fully dry. Guess what, when reassembled it works perfectly, no leads errors. The advice here has given me my expensive meter back! 😀😀.   Just need a new display window and two new battery case lock buttons in a perfect world now, though it’s pretty great as it is really!

Thanks all

Steve

[Shock]

Wouldn't bother with the replacing the display window, it looks quite usable still. Just use some plastic polish on it and keep it away from the surface cracks. Here is how my 87V turned out.

[sw2022]

Thanks. I had some metal polish to hand and that did not remove the large scratch, think it’s pretty deep, I will try some better polish though, your display looks good,

Steve

[shakalnokturn]

For the deep scratches, ones you can catch your nail in, I plane the area with multiple passes of a cutter blade at right angle to the plastic.
After that and on regular scratches  I use the Novus scratch remover products bought after reading some discussion on the subject on the forum, definitely practice on a less expensive DMM tough.

[daisizhou]

very good. :-+Prevents water or liquids from entering the 4 sockets for future use.I've even found that both sunny and rainy days can affect the data

[sw2022]

Thanks, I may not go as far as using a blade as I can live with the damage, I will try different polishes though. I wondered if a screen protector might make the damage less obvious too

Steve

[BILLPOD]

My 189 still has the protective film on the screen that came on it when I bought it new in 2007.  I don't think it is an 'official' screen protector, but it has worked well.  I buy cheap cell phone screen protectors and cut them down for use on all my other meters.  Good job getting your 189 working again.

[sw2022]

Thanks, it made me very happy 😀. I’d used that meter for years even if I temporarily forgot how expensive it was to buy. Part of you thinks all through the process that it’s going to be scrap, people here inspired me to keep the faith!

Steve

[daisizhou]

fluke 189 Note that there are lithium batteries inside, which are easy to leak.
While 189 can store data, I prefer fluke 187

[sw2022]

Wonder if it’s possible to replace the lithium battery with new

[Shock]

It's not a battery, it's a super cap. A known issue and source of contamination due to physically leaking. It's well documented on the forums here in several threads.

[alm]

What I've done in the past, though this was on flat display lenses, was to sand them with wet-and-dry sandpaper from coarse enough to remove the scratch down to the finest grit I could buy, and then polish until its smooth again.

[sw2022]

Sounds possible I might try that, a bit scary in that it will dull the plastic and will definitely need the right polish to finish I guess

[alm]

Yes, you basically go to fully opaque but no big scratches down to polished and transparent again. Search this forum, I'm pretty sure people have written about this and recommended specific plastic polishes. Quite possibly by robrenz.

[sw2022]

Will do thanks!

[rsjsouza]

Great job in restoring one more fine piece of equipment!

I would be very careful with the polishing process: it is easy to make mistakes which, if anything goes wrong, will feel much worse now the meter is working. I would practice and practice on anything else before even trying something on the 189. The user Robrenz mentioned by alm is an incredible machinist with gobs of practice on his hands.

[sw2022]

Comments noted. Still thinking on whether to do anything more at all. I used the meter in anger yesterday and was very happy with it!

Steve

[alm]

I agree that you should consider if you want to fix this. For what it's worth, I certainly don't have the experience Robrenz has, and I did this without any problems. But I can see a potential problem polishing close to the edge of the frame.

User modemhead used to repair a lot of multimeters. You could search this forum and his website if he came up with a simpler solution.

[sw2022]

Thanks. His pics of a 189 overhaul were really helpful. Especially when a couple of switch contacts dropped loose, I sorted the orientation from there,

Steve

[Trader]

So ... I’ve measured the resistance across the split socket, best way seemed to be to press the meter probes on the cleaned tips of the jack sockets pins. See my pics. I get 1.31M accross the 10A socket and 1.38M for the 400mA socket. That’s of course with the board out of the meter case as pictured. Seems lower than what was suggested though almost identical for the two, obviously there’s an input circuit in the loop here though. Certainly nowhere near a short...

Steve

Hi sw2022, congrats you solve that problem.  I have the same issue, when you measured 1.31M (on A) and 1.38M (on mA), I measured 1.35M and 1.25M (oscillating between 1.2M and 1.3M).

It's possible for you to check again the same places?

I removed the jack, and even without the input jacks and cleaning the PCB, the LEADS beep doesn't stop.

If I blow warm air, the LEADS warning stops, but returns after a while.  Weird.

[sw2022]

Hi,

The meter still works well, the resistance readings across the two socket shells are 400mA socket 2.97M, 10A socket 2.43M. These are higher than yours. I pre cleaned the sockets physically using a sharp probe then pipe cleaner socked in alcohol. I then soaked my board for days in 100 pc isopropyl alcohol, with also lots of spray rinses with clean alcohol after the soak. I sprayed out the sockets also using a syringe. I then dried it both with hot air then leaving it in a cupboard for days. It’s interesting that function returns to your board when you warm it, are you sure there is no damp? Was the isopropyl alcohol 100 pc, I guess any water is not good,

Steve

[Trader]

My F189 was stored in a humid environment for several years.

I kept hot air (150F) on it for 10 hours, I guess the DMM is super dry and cleaned, I also removed the input jacks. Is weird that up to now the only solution for this problem is "cleaning", I wonder if some components (diode, resistance, etc) could be defective.

After warming, it works and the measures are correct, when the temperature returns to normal, starts beeping.

So, you soaked the Entire Board in 100% alcohol?

Thanks.

[sw2022]

I soaked approximately the bottom half of the board. I left the sockets in place and stood the thing in a plastic container, leaving the display end out.

Steve

[ewtotem]

It is somewhat common for the 189 and 187 to have the 4 post assembly break. The only solution is to replace it (Fluke 4611374)