Keithley 2002 repair (Help please)

[SelectLOL]

Hello good people of the EEVblog Forum

To me: I live in Switzerland and am a soon to be student that studies electronics engineering. 

To the subject: A while back I bought a defective Keithley 2002 from an auction site and put it aside. Now I have some time in hand and would like get it running again.

I have read through the excellent post https://xdevs.com/fix/kei2002_u2/ and some other posts on the forum that successfully repaired a Keithley 2002 or similar.

To the device: It came with the cap C117 and rectifier CR113 missing. To be seen at 'pic2'. The LCD lights up fine but on whatever you choose, Voltage, Ampere etc. it shows 'overflow', as to be seen on 'pic1'. I have measured the Voltages and they seem as following:

5V -> measured about 4.1V
+15V -> measured about 2.2V
-15V -> measured about 1.1V

The device fails so many tests, that I did not bother to look into those.

After that I carefully removed all the electrolytic caps on the analog board. The reason for that is that, the inside really smelled fishy and I have heard many times that it is good practice. 'pic3' 

To no ones surprise it was a great idea because, some of the caps already started leaking. 'pic4 and 5' 

The caps I removed look fine, besides one 'pic6 and 7' The one on 'pic7' is not near its rated capacitance, it measures some nanofarads. I hope that that is why the Keithley did not work 

I cleaned everything nicely with isopropanol and looked at the traces. None of them see to be damaged severly, as far as I can see no damages have been done. 'pic8'.

I am currently making a basket with all the caps I need (electrolyte and tantal). I will install everything this friday probably and here is where I need your help.

I am suspecting that the device wont work, because that would be too easy woudlnt it be?  If anyone has any other suggestions or places that I should check while waiting for the parts, Id be really greatful to hear them.

Thanks everyone in advance and greetings

SelectLOL


Sorry for the image quality, had to resize alot to be able to post 

[Kleinstein]

Unless one of the old electrolytics has a short, there is likely something else wrong at the +-15 V supply to let the voltage drop that much.
A point to check (in circuit should be OK) would be the rectifiers to see if there is a shorted diode. It very much looks like something is pulling down the -15 V and likely also the +15 V. The low voltage for the 5 V may be from the transformer voltage going down from overload, but also from current flowing from the +5 V towards the +15 V rail.  Some of the CMOS switches may have both a 5 and 15 v supply and may provide a parasitic path.

The damage to the PCB does not look too bad, but it would still need quite some cleaning to be on the safe side. The nasty part is contamination below the SMD chips and resistors.  One may have to remove some, just for better cleaning.

For a first test one should probably leave out the ADC board, and check the supplies.

[Sarvesaa]

That rectifer looks abnormal check that up.

We both are the same catergory. I completed high school and in one month time, I will be studying EE as you do. 

The repair will go alright.

Thank You
With Regards
Sarvesaa

[alm]

I'd try to measure resistance across the plus and minus 15V rails. So +15V to the ground for that rail, and then -15V to ground. I believe the power supply is linear, so if the rails are so low, then either the regulator is open or there's a low resistance across the rail. After removing the caps and cleaning the pcb, but before putting in the new caps, would be the perfect time because large caps make resistance measurements more difficult.

[SelectLOL]

Unless one of the old electrolytics has a short, there is likely something else wrong at the +-15 V supply to let the voltage drop that much.

The old electrolytics dont have any shorts. As far as I can only measure the capacitance with my Multimeter, only the red one is far off. The othner ones seem kinda fine. (I will of course change them).

A point to check (in circuit should be OK) would be the rectifiers to see if there is a shorted diode. It very much looks like something is pulling down the -15 V and likely also the +15 V. The low voltage for the 5 V may be from the transformer voltage going down from overload, but also from current flowing from the +5 V towards the +15 V rail.  Some of the CMOS switches may have both a 5 and 15 v supply and may provide a parasitic path.

When I bought the device it had the CR113 and C117 were missing. The Repair Manual for the 2002 says, that CR113 is the rectifier for the +15V and -15V rails. And C117 has something to do with the Bootstrap supply (about 38V). I have put a screenshot in the Attachments. It should be page 21/148 from the manual (2002_902_01B)

The damage to the PCB does not look too bad, but it would still need quite some cleaning to be on the safe side. The nasty part is contamination below the SMD chips and resistors.  One may have to remove some, just for better cleaning.

For a first test one should probably leave out the ADC board, and check the supplies.

Yeah I havent thought about the underside of the SMD parts you are right, there might be some electrolyte left there. I will give it a good cleanup again and hope for the best, Im to afraid to desolder anything else besides the caps.

Is there anything else that comes up to your mind that might be worth measuring? I have found some testpoints on the board but no place where it says what voltages I should expect there 

[SelectLOL]

That rectifer looks abnormal check that up.

Yeah when I bought it was just missing (screenshot of the functionality on my answer before), very weird, somebody must have had tried something 

We both are the same catergory. I completed high school and in one month time, I will be studying EE as you do. 

The repair will go alright.

Ahahahah thats amazing, I wish you all the luck and great success in your studies Thank you for checking in 

[SelectLOL]

I'd try to measure resistance across the plus and minus 15V rails. So +15V to the ground for that rail, and then -15V to ground. I believe the power supply is linear, so if the rails are so low, then either the regulator is open or there's a low resistance across the rail. After removing the caps and cleaning the pcb, but before putting in the new caps, would be the perfect time because large caps make resistance measurements more difficult.

Thank you for your suggestion, I have measured between +15V and -15V, currently the larg caps are disconnected as on the pictures from the original post and the analog board is connected to nowhere. The measured resistance is 2.4kΩ, I think thats not to low or to high, what do you think? Anything else that I should measure before continuing?

[Zucca]

When you try to fix a 8.5 DMM it could be a never ending story.
I wish you luck.

Do not be discouraged by the above lines, the opposite.

What you are doing it is one of the best way to learn, and learn a lot.
Go slow and do not cut corners, always document/post here your progress.

Even if you do not bring it back to life, you are not wasting time.

PS: Very interesting stuff here too:
https://www.eevblog.com/forum/repair/keithley-2002-repair-help/

To the repair, a voltage supply issue is promising. Use IPA alcohol or finger to spot high temperature components that maybe are in short....
If nothing is warm, you need to check the components that should generate the +-15V

[alm]

Thank you for your suggestion, I have measured between +15V and -15V, currently the larg caps are disconnected as on the pictures from the original post and the analog board is connected to nowhere. The measured resistance is 2.4kΩ, I think thats not to low or to high, what do you think? Anything else that I should measure before continuing?

2.4kOhm sounds sensible to me. How about measuring just +15V to its ground and just -15V?

[SelectLOL]

When you try to fix a 8.5 DMM it could be a never ending story.
I wish you luck.

Do not be discouraged by the above lines, the opposite.

Thank you very much. My main goal, even above getting the Keithley back to a running state, is to learn something new. So my motivation is always high 

What you are doing it is one of the best way to learn, and learn a lot.
Go slow and do not cut corners, always document/post here your progress.

Even if you do not bring it back to life, you are not wasting time.

I really love the way the forum works and how you can keep people updated and get help from them. I will definitely write down every little step of mine, so that someone, somewhere, with the same problem, might resolve his problems.

PS: Very interesting stuff here too:
https://www.eevblog.com/forum/repair/keithley-2002-repair-help/

To the repair, a voltage supply issue is promising. Use IPA alcohol or finger to spot high temperature components that maybe are in short....
If nothing is warm, you need to check the components that should generate the +-15V

Interesting technique with the IPA alcohol to find out high temperature areas. I know someone, which I hopefully might be able lend me a IR-Temp camera to measure high temp. areas in the near future.

In my last comment I posted a screenshot of the repair manual, which shows me that the rectifier for +-15V is missing. I got it that way when I bought it 

[SelectLOL]

Thank you for your suggestion, I have measured between +15V and -15V, currently the larg caps are disconnected as on the pictures from the original post and the analog board is connected to nowhere. The measured resistance is 2.4kΩ, I think thats not to low or to high, what do you think? Anything else that I should measure before continuing?

2.4kOhm sounds sensible to me. How about measuring just +15V to its ground and just -15V?

So I took those measurements of the bottom side of the Analog Converter Board, because they were nicely labeled.

+15V to VGND -> 183k
+15V to GND -> 2.9k
+15V to DGND -> 38k

-15V to VGND -> 185k
-15V to GND -> 1.7k
-15V to DGND -> 35k

All the measurements were either rising or falling, seems to be connected to some sort of capacitance. But all in all that seems fine, right?

In the attachments there is a picture from where I measured.

VGND 3. from top, GND 5. from top and DGND 2. from bottom

[alm]

All the measurements were either rising or falling, seems to be connected to some sort of capacitance. But all in all that seems fine, right?

Yes, none of the values are low enough to draw the rails down. When you have the caps replaced and can power it up again, my next step would be to look for hot components like Zucca suggests. An oscillating op-amp, for example, can also draw a lot more current than normal.

I would also measure the voltage around the +/- 15V linear regulators, like the input voltage into the regulators and the AC voltage that's going into the bridge rectifier for the +/- 15V rails. The problem with the 15V rails could also be at the supply side, like an open transformer winding, though this appears less likely.

[SelectLOL]

All the measurements were either rising or falling, seems to be connected to some sort of capacitance. But all in all that seems fine, right?

Yes, none of the values are low enough to draw the rails down. When you have the caps replaced and can power it up again, my next step would be to look for hot components like Zucca suggests. An oscillating op-amp, for example, can also draw a lot more current than normal.

I would also measure the voltage around the +/- 15V linear regulators, like the input voltage into the regulators and the AC voltage that's going into the bridge rectifier for the +/- 15V rails. The problem with the 15V rails could also be at the supply side, like an open transformer winding, though this appears less likely.

Thank you, I will be posting my basket from Farnell later on. The parts should arrive tomorrow hopefully, then I can measure out said points and post the results.    

[SelectLOL]

Alright, so this is my final Shoping Basket/Cart. I have selected all (besides one rated for 2000hrs) to be rated for 10000hrs @ 105C. The size of the caps should be the same too. I also added some tantalum caps to the basket, which I might change later on. If I get the meter fully running, I will change all the other caps too, I am just on a tight budget currently. 

For the missing rectifier CR113, I found out that the part is https://ch.farnell.com/vishay/df02m-e3-45/br-ckengleichrichter-diode/dp/1651041?ICID=I-RP-STM7REC-3 but it is currently not available. I will be ordering the SMD variant https://ch.farnell.com/vishay/df02s-e3-77/br-cken-gr-diode-1ph-1a-200v-dfs/dp/2335133 instead. It should be fine, right?

I will be ordering in a few hours if something critical is missing please let me know, thank you.

The Shopping Basket/Cart can be found in the attachments as ZIP, which contains a HTML file from Farnell.

[Kleinstein]

For the rectifier there should be THT versions available. It does not matter if the current rating is a little higher or the voltage rating is different, as longs as it is sufficient (e.g. >=60 V). The round form may also fit.

There is not need to change all the capacitors. Just the electrolytics are some-what critical.

[SelectLOL]

For the rectifier there should be THT versions available. It does not matter if the current rating is a little higher or the voltage rating is different, as longs as it is sufficient (e.g. >=60 V). The round form may also fit.

Alright then, I will choose this one instead > https://ch.farnell.com/multicomp/df04/br-cken-gr-diode-1phase-400v-1a/dp/2748159?MER=sy-me-pd-mi-acce As far as I have understood only the Voltage changes for this one from 200V to 400V so it should be fine.

There is not need to change all the capacitors. Just the electrolytics are some-what critical.

I see, I see, so I will leave the ones I already selected and only focus on the electrolytic ones for future changes, thank you.

I will order now then, so that everything definitely arrives tomorrow 

[Per Hansson]

CR112 looks like it overheated? I would replace it also!

[SelectLOL]

CR112 looks like it overheated? I would replace it also!

Thank you very much for pointing that out, I have found following replacement and have added it to my basket : https://ch.farnell.com/multicomp/w04mg/br-cken-gr-diode-400v-1-5a-wob/dp/1861437?ost=w04m

[manicdoc]

I'd try to measure resistance across the plus and minus 15V rails. So +15V to the ground for that rail, and then -15V to ground. I believe the power supply is linear, so if the rails are so low, then either the regulator is open or there's a low resistance across the rail. After removing the caps and cleaning the pcb, but before putting in the new caps, would be the perfect time because large caps make resistance measurements more difficult.

Agreed. Try as much as possible to get power rail resistive measurements per board in isolation, and write it down, compare and do the power maths...

[SelectLOL]

I'd try to measure resistance across the plus and minus 15V rails. So +15V to the ground for that rail, and then -15V to ground. I believe the power supply is linear, so if the rails are so low, then either the regulator is open or there's a low resistance across the rail. After removing the caps and cleaning the pcb, but before putting in the new caps, would be the perfect time because large caps make resistance measurements more difficult.

Agreed. Try as much as possible to get power rail resistive measurements per board in isolation, and write it down, compare and do the power maths...

Thank you, I will try to measure some more points and document and post them. 

[SelectLOL]

CR112 looks like it overheated? I would replace it also!

Thank you very much for pointing that out, I have found following replacement and have added it to my basket : https://ch.farnell.com/multicomp/w04mg/br-cken-gr-diode-400v-1-5a-wob/dp/1861437?ost=w04m

Amazing, I have soldered CR112 out to see if its intact, looks like it bit the dust. My measurements can be seen in the attachment. Please tell me if I have measured something wrongly. Thank you, luckily I already purchased the replacement 

[free_electron]

many of these machines die due to the capacitors leaking. some are unrepairable. the leaky electrolyte has caused electromigration on inner layers and created hard shorts. your ohmmeter may tell "open" but the moment you apply a voltage that stuff goes short.

i;ve repaired a few and had 2 that were not repairable.

check out the glass diodes close to the capacitors. they sometimes get fried too.

[IanJ]

Nice clean looking unit, including it's PCBs. Looks like you caught it in time with those leaky caps.......it's a whole different ball game if the PSU side of things fails.....trust me!

Ian.

[HighVoltage]

It seems to me that Keithley instruments are much more prone to have leaky electrolytics, compared to other instrument manufacturers. Especially the 2001 and 2002 models.

Why is that?

Compare this to the Agilent 34401A, I never had one bad electrolytic capacitor, not even on the older models.

[SelectLOL]

many of these machines die due to the capacitors leaking. some are unrepairable. the leaky electrolyte has caused electromigration on inner layers and created hard shorts. your ohmmeter may tell "open" but the moment you apply a voltage that stuff goes short.

Ou thats very interesting, so the only way you can actually tell is by applying voltage and checking for abnormal temp. and current drawing? Luckily mine isn't that far gone, that that can happen, right? 

i;ve repaired a few and had 2 that were not repairable.

check out the glass diodes close to the capacitors. they sometimes get fried too.

Oh what an amazing feat! Glad you are helping me out  I have looked into those diodes and they seem to be described as "Avalanche" Diodes and I really had to google first what they even are. I will test them now before soldering up everything together, thank you.