Keithley DMM6500 repair

[drummerdimitri]

I've decided to post a thread as I don't want to spam the DMM6500 thread with my repair issues for simplicity's sake.

My meter's 3.5 A port is no longer working after I had a 220V AC short by falsly connecting the test leads to mains while I had forgotten to switch the leads to the AC V terminals and blue both fuses and apparantly other components within the meter itself 

I've managed to take the unit appart and you can see some pictures of the main board.

Problem is, there is no visible damage and I've tried testing the small resistors which seem to be the current shunts as well as that large diode, the briudge rectifier and the precision shunt facing it and they all seem to be in working order.

I'm stuck now and I'm affraid it could be a non-passive component such as an IC or such.

Any ideas on what else could be the culprit here?

The white wire is the one coming from the port with the damaged circuit just FYI.

The unit is still under warranty but the problem is I need to make low current measurments and with the country on lockdown until further notice, I am not able to post it back and have it repaired.

Also, I would have to pay custom fees once again and with shipping that would come out to be around 300$ 

For that reason, I'd like to have a crack at it and see if luck is on my side.

[Manul]

Try tracing the circuit and consider some creativity.

One idea would be to apply a current on the current input, so you get a small voltage on a shunt. Then start looking where else is this voltage appearing. That can help tracing.

More radical idea is to remove shunt (and destroy calibration!) and apply current limited (like 1mA or so) voltage injection on the points where this shunt was. Efectively making a controled input overload scenario. If you observe the resulting voltage of your limited current injection and go with hot air over the components - you may find a component which will make this voltage go down when sligthly heated. If you are really lucky, this component may be the component which absorbed the real overload.

This is very generalised. I don't know a lot about meters, never had them fail. Anyway, wish you good luck.

[Kleinstein]

Some crude reverse engineering / tracing of the current measurement part would be a first step. Here it can help to look at the plans for a few similar grade meters (e.g. K2000, 34401) to get an idea how the circuit could about look like. At least the higher current part is probably quite similar. The low current part uses an TIA and is thus different. The circuit looks rather obvious.

A next point would be to have a clear picture how the broken meter behaves. So a few readings from the meter and also measurement of the total drop.

Following the signal path is also possibly. If one has a scope,  an AC current e.g. from a low voltage (e.g. 6 V)  transformer may be even better than DC, as it tests the whole range in one run.

Possible culprits could be a relay (e.g. fused or broken contacts), a shunt and maybe the rectifier+large diode, the front/rear switch the initial amplifier (if there is a special one for the currents), a resistor for protection, a MOV and maybe others. Knowing what is still working and what not may help to narrow it down.

[drummerdimitri]

Thanks for the replies,

How do I test the large, white ceramic resistor? What is it anyway it says Fluke 2918 I am assuming it is a current shunt resistor?

How do I trace the fault component by component? Don't have much experience in electronics repair.

[E-Design]

Thanks for the replies,

How do I test the large, white ceramic resistor? What is it anyway it says Fluke 2918 I am assuming it is a current shunt resistor?

How do I trace the fault component by component? Don't have much experience in electronics repair.

The White resistor is the 10Meg divider.. its not part of the amps pathways.

If the unit can be reassembled and powered up:

Put fuses back in and then measure the input resistance AMPS to LO as you click through the ranges to see if all the shunts are in tact.

Also report any measurements for the amps function (with external equipment disconnected)

[drummerdimitri]

Thanks for the replies,

How do I test the large, white ceramic resistor? What is it anyway it says Fluke 2918 I am assuming it is a current shunt resistor?

How do I trace the fault component by component? Don't have much experience in electronics repair.

The White resistor is the 10Meg divider.. its not part of the amps pathways.

If the unit can be reassembled and powered up:

Put fuses back in and then measure the input resistance AMPS to LO as you click through the ranges to see if all the shunts are in tact.

Also report any measurements for the amps function (with external equipment disconnected)

I was able to put the unit back together and measured the resistance between the LO and the front AMPS terminals and here's what I got for the ranges from smallest to largest: 5.3612 k\$\Omega\$, 1.0337 k\$\Omega\$, 130.45\$\Omega\$, 12.39\$\Omega\$, 1.34\$\Omega\$, and 0.4\$\Omega\$ for the 1 and 3A ranges (limited accuracy due to 2 wire measurment (Fluke 289).

The amps reading without any input are 6.9 uA, 69.16 uA, 0.690 mA, Overflow A, 70.12 mA, 0.7A, 0.7A for the respective ranges from lowest to largest.

Also, when should I hear the relay clicking? When it switches between which of those ranges?

[E-Design]

As far as relay clicks,

1 & 3 A are on a relay and 100mA is on a relay. So going into or out of those ranges, you should hear the relay.

Example: 10mA switch to 100mA = CLICK
Example: 1A to 3A = NO CLICK (same relay)
Example 3A to 100mA  = CLICK
Example 100uA to 10mA = NO CLICK

Play around with it, that can let you know the relays are still working (seems likely to me) - 220V won't get from contacts to coil and the fuses hopefully opened before contact damage due to arcing (if there wasnt very high inductance on your wiring path)

Your shunt values indicate you've got some leakage current going on -- likely from a damaged semiconductor (or shunt resistors)  We can find out what it is (if you're willing to follow along)
Remove U42 - its a quad analog switch (DG444) and is quite likely damaged / leaky.
Remove U52 - its an amplifier that could also have damaged / leaky input.
With U42 removed, bridge with solder pins 2 and 3 (of the U42 footprint left behind) or short them with a small piece of wire... this effectively forces only the 10uA range to be enabled (provided your relays are working)

With those 2 semiconductors removed, you now have only the bridge rectifier and the shunts on the path.

Now select 10uA range and measure input impedance (AMPS to LO) with your 289 and report.

One more thing, Next time you have the board out of the chassis -  look at the bottom side of the board  where the bridge rectifier pins are soldered, if you see any component on bottom side bridging pins, it might also be leaky (this was an optional component, not on all boards and I dont know what you have so just check visually) - its a tiny component. if its there, it might also be leaky

[drummerdimitri]

As far as relay clicks,

1 & 3 A are on a relay and 100mA is on a relay. So going into or out of those ranges, you should hear the relay.

Example: 10mA switch to 100mA = CLICK
Example: 1A to 3A = NO CLICK (same relay)
Example 3A to 100mA  = CLICK
Example 100uA to 10mA = NO CLICK

Play around with it, that can let you know the relays are still working (seems likely to me) - 220V won't get from contacts to coil and the fuses hopefully opened before contact damage due to arcing (if there wasnt very high inductance on your wiring path)

Your shunt values indicate you've got some leakage current going on -- likely from a damaged semiconductor (or shunt resistors)  We can find out what it is (if you're willing to follow along)
Remove U42 - its a quad analog switch (DG444) and is quite likely damaged / leaky.
Remove U52 - its an amplifier that could also have damaged / leaky input.
With U42 removed, bridge with solder pins 2 and 3 (of the U42 footprint left behind) or short them with a small piece of wire... this effectively forces only the 10uA range to be enabled (provided your relays are working)

With those 2 semiconductors removed, you now have only the bridge rectifier and the shunts on the path.

Now select 10uA range and measure input impedance (AMPS to LO) with your 289 and report.

One more thing, Next time you have the board out of the chassis -  look at the bottom side of the board  where the bridge rectifier pins are soldered, if you see any component on bottom side bridging pins, it might also be leaky (this was an optional component, not on all boards and I dont know what you have so just check visually) - its a tiny component. if its there, it might also be leaky

I have done as you asked and the resistance between LO and AMPS is now 190.16 k .

Also, I checked the bottom and there doesnt seem to be any small component near the bridge rectifier so I assume this is a later revision of that board?

I have attached a picture either way for your reference.

[E-Design]

Ok that seems too high of a total value.

Try remove Q70 -then measure the total again. (leaving that 10uA solder bridge that connects the series stack of the shunts)
At that point you can also try measure individual shunts

R224 ~ 0.1 Ohms
R225 through R229 are in parallel and ~ 0.91 Ohms
R230 and R231 in series ~ 9.1 Ohms
R233 ~ 90.9 Ohms
R234 ~ 909 Ohms
R235 ~ 9.09 kOhms

See if those are looking correct.

[drummerdimitri]

Ok that seems too high of a total value.

Try remove Q70 -then measure the total again. (leaving that 10uA solder bridge that connects the series stack of the shunts)
At that point you can also try measure individual shunts

R224 ~ 0.1 Ohms
R225 through R229 are in parallel and ~ 0.91 Ohms
R230 and R231 in series ~ 9.1 Ohms
R233 ~ 90.9 Ohms
R234 ~ 909 Ohms
R235 ~ 9.09 kOhms

See if those are looking correct.

It's taking me way too long to find these parts on a board with hundred of components.

Could you please indicate where they are located? It would make the task a lot easier 

I was unable to find Q70 so please assist with this thanks.

[Kleinstein]

As far as I can see Q70 is the TO92 part right of the red relays, right next to the resistors in question.

[coromonadalix]

Ok that seems too high of a total value.

Try remove Q70 -then measure the total again. (leaving that 10uA solder bridge that connects the series stack of the shunts)
At that point you can also try measure individual shunts

R224 ~ 0.1 Ohms
R225 through R229 are in parallel and ~ 0.91 Ohms
R230 and R231 in series ~ 9.1 Ohms
R233 ~ 90.9 Ohms
R234 ~ 909 Ohms
R235 ~ 9.09 kOhms

See if those are looking correct.

It's taking me way too long to find these parts on a board with hundred of components.

Could you please indicate where they are located? It would make the task a lot easier 

I was unable to find Q70 so please assist with this thanks.

Seriously ??   it took me  2 secs from your photos to see them ?  check near the 3 orange relays, you have an to92 transistor standing up near an pcb hole ??
on your 20200407_180632.jpg  image. 

And yes as Kleinstein wrote  its Q70   i've put an red circle around your parts.

[drummerdimitri]

Ok that seems too high of a total value.

Try remove Q70 -then measure the total again. (leaving that 10uA solder bridge that connects the series stack of the shunts)
At that point you can also try measure individual shunts

R224 ~ 0.1 Ohms
R225 through R229 are in parallel and ~ 0.91 Ohms
R230 and R231 in series ~ 9.1 Ohms
R233 ~ 90.9 Ohms
R234 ~ 909 Ohms
R235 ~ 9.09 kOhms

See if those are looking correct.

It's taking me way too long to find these parts on a board with hundred of components.

Could you please indicate where they are located? It would make the task a lot easier 

I was unable to find Q70 so please assist with this thanks.

Seriously ??   it took me  2 secs from your photos to see them ?  check near the 3 orange relays, you have an to92 transistor standing up near an pcb hole ??
on your 20200407_180632.jpg  image. 

And yes as Kleinstein wrote  its Q70   i've put an red circle around your parts.

I was looking for a component that looks completely different like other QXX parts on the board that's why it took me so long and couldnt find it.

Thanks for the help.

[drummerdimitri]

Ok that seems too high of a total value.

Try remove Q70 -then measure the total again. (leaving that 10uA solder bridge that connects the series stack of the shunts)
At that point you can also try measure individual shunts

R224 ~ 0.1 Ohms
R225 through R229 are in parallel and ~ 0.91 Ohms
R230 and R231 in series ~ 9.1 Ohms
R233 ~ 90.9 Ohms
R234 ~ 909 Ohms
R235 ~ 9.09 kOhms

See if those are looking correct.

I have checked all the above shunts and their values are correct.

With the Q70 removed, I am no longer able to get a proper resistance reading between the LO and the AMPS ports.

I am getting values in the Mega and Kilo  ranges and even negative ones that change when the leads are switched around.

[E-Design]

You are still measuring with the DG444 pin 2-3 shorted and power up and select 10uA range ?
I believe the configuration you have now should let you see the whole sum of shunt values - unless there is still a damaged part.
You had said the diode bridge was ok right? - thats about all left in there.

[drummerdimitri]

You are still measuring with the DG444 pin 2-3 shorted and power up and select 10uA range ?
I believe the configuration you have now should let you see the whole sum of shunt values - unless there is still a damaged part.
You had said the diode bridge was ok right? - thats about all left in there.

I am measuring with the three components previously mentioned removed and DG444 traces 2 and 3 shorted as can be seen in the picture below.

Not sure why I am still not getting a proper impedance reading from ranges 10 uA to 1 mA. The larger ranges are reading consistently impedance values of 10.083 k , 1.254 and 0.335

[coromonadalix]

Is it me or you habe a small bodge on one pin of u42 (center of the chip where two pcb vias) and a shield trace ??  please  clean the surroundings properly

[TiN]

R236, R594, R593, R386 and shunt R233 look like have liquid damage or corrosion. Those resistors look very nasty to me

R255 have deep scratch across surface, could be broken. These small parts require gentle and careful work, otherwise it's easy to ruin delicate instrument like DMM6500 beyond repair.
Please be careful.

[E-Design]

Check for guard bridges. Put your 289 in continuity mode and one lead on the mask relieved trace you see everywhere (somebody pointed out a bridge on pin 12 of U42). With the other lead, probe around on each of the shunts and some nodes nearby. You can cover alot of them quickly.. nothing to should indicate continuity.

Also, you are measuring ~10k ohm on the 10uA range? If not, check continuity from AMPS input to one terminal of R235. You should be able to get a low resistance to this point (its just wiring, fuses and copper traces) - see what you've got.

[drummerdimitri]

Check for guard bridges. Put your 289 in continuity mode and one lead on the mask relieved trace you see everywhere (somebody pointed out a bridge on pin 12 of U42). With the other lead, probe around on each of the shunts and some nodes nearby. You can cover alot of them quickly.. nothing to should indicate continuity.

Also, you are measuring ~10k ohm on the 10uA range? If not, check continuity from AMPS input to one terminal of R235. You should be able to get a low resistance to this point (its just wiring, fuses and copper traces) - see what you've got.

I checked the board and there are no shorts between any of the shunts/components and the guard bridges except for the components circled in red in the attached picture (right side shorted).

I am getting a resistance reading of -109 k whatever that means and it changes when i reverse the test leads.

Resistance between AMPS port and one terminal of R235 is around 335 m\$\Omega\$.

[coromonadalix]

please clean your job and the ic pads  they dont look clean to me,  it seem to short in some place with the shield trace ... on your last photo      you could have problems without knowing it  ...

[drummerdimitri]

please clean your job and the ic pads  they dont look clean to me,  it seem to short in some place with the shield trace ... on your last photo      you could have problems without knowing it  ...

I just did and it made absolutely no difference 

Also, none of the traces are shorted with the guard bridges.

[E-Design]

Alright, seems you have good connection up to the shunts and previously, you measured each of the shunts individually as correct so lets check from the bottom of shunts back out.

So, measure the 5 parallel (R225 to R229) from the side not connected to R230 measure that node out to LO terminal. It should look low like a low value again (it just contains the 0.1 ohm shunt and low resistance wires, traces.)

See what that shows us.

[drummerdimitri]

Alright, seems you have good connection up to the shunts and previously, you measured each of the shunts individually as correct so lets check from the bottom of shunts back out.

So, measure the 5 parallel (R225 to R229) from the side not connected to R230 measure that node out to LO terminal. It should look low like a low value again (it just contains the 0.1 ohm shunt and low resistance wires, traces.)

See what that shows us.

I measured the resistance from the parallel resistors to the LO input terminal and got a value of 0.123 as expected.

[thm_w]

R236, R594, R593, R386 and shunt R233 look like have liquid damage or corrosion. Those resistors look very nasty to me

R255 have deep scratch across surface, could be broken. These small parts require gentle and careful work, otherwise it's easy to ruin delicate instrument like DMM6500 beyond repair.
Please be careful.

R225 I think you mean.
R75 in the original photo has some "browning" as well.
Also R283 in the follow up photo, is that a ferrite bead? It looks slightly suspicious.