Restoration glory of Keithley 2001 DMM

[Hydron]

Also, forgot to ask.  Is there a Keithley logo out there to replace the faceplate one that is damaged?

https://www.eevblog.com/forum/testgear/keithley-old-logo-in-good-quality-someone/?all

Sorry, couldn't resist :p

[mapleLC]

https://www.eevblog.com/forum/testgear/keithley-old-logo-in-good-quality-someone/?all

Sorry, couldn't resist :p

LOL, in that case, is there a way to adjust the screen brightness?  Mine is a little dim.  I saw that heating method out there but that scares me a bit to try.

[maxwell3e10]

I looked at temperature stability and noise of two K2001 and two other 6-1/2 meters that I have on hand. These are very basic measurements relying on temperature variations in a poorly air conditioned room. The voltage source is an LTZ1000 HP3458 A9 board, not using any output buffer and sitting in a cardboard box.

Day 1

Day 2


There are several differences between the two plots. First, I moved the black internal cover from K2001 B to K2001 A on the second day (unfortunately I have only one). It makes a big difference in the noise. Second, I switched HP3457 from using 30 V range to using the 3 V range with a 1/4 divider based on LT5400-5. It reduces the noise quite a bit but makes the temperature drift a bit worse. Third, in the second plot I show the actual raw voltages of each meter instead of deviation from average. It's not bad that all 4 meters are within about 10 ppm of each other after 2-3 decades without calibration.

The temperature drift of the two K2001 is very similar to each other, roughly 1 ppm/K. In general, they are noisier than other 6.5 - 7.5 meters. Yokogawa 7263 uses SL161525 reference and seems to drift more with temperature than other meters.

[Kleinstein]

The K2001 really looks noisy. The 3457 in the 30 V range is already rather noisy as it uses an 1:100 divider at the input.
The 2001 has a 20 V range and a 0.5 amplifier gain - so a little divider, but not that much.

The thermal cover has quite some effect: some 30 µV peak to peak from thermal effects is quite a bit. It looks like there must be something quite sensitive or a really nasty variable temperature. There are quite a few points where the trouble can happen: the protection (2 MOSFET stage in series), the amplifier gain (resistors) and the ADC itself. The K2001 uses a rather frequent correction of the ADC / amplifier gain. So variations in the resistors for the gain don't show up in the long time TC, but they may show up in the shorter time noise.

[fma]

Do you think it is possible to replace the transformer+regulators with a switching power supply?

[Kleinstein]

The transformer need rather good isolation and shielding. It is possible to replace it with a different supply, but it is a demanding task to build one. This would likely be a 2 step process: 230 V AC to some 5 to 12 V DC (maybe to supply the digital boad and display) and than an extra low noise  isolated DCDC with a custom tranformer with shields and good isolation for the analog part.

[fma]

Ok, thanks!

[w.v.s.]

Hello there!
This morning, I finally registered to this forum. I've enjoyed reading it (especially this thread and the one about the seek thermal camera) for years and benefited several times. I've even bought a broken K2001 (electrolytic spill, burnt R100 and transformer, too and more) quite a while ago, because I liked this thread and wanted to share the experience  . Yesterday I did some tracing under C117, today I've read your question and saw a chance for a small first contribution. I hope you're still interested and have gimp installed to read my file and can benefit from it. It was never intended for publishing and it may contain errors. It's mainly derived from photos published by xdev and ones I did myself. Use it at your own risk.

[w.v.s.]

Hello,

when I should share the way of my 2001 on this platform: Should I make this great thread even longer or better open a separate one?

[chekhov]

There are maybe a dozen shorter topics dedicated to each similar repair.
In some cases shorter topics help to find information better (exact info about repair), however beefy topic likely to contain all information you may need.

@w.v.s., are you going to fix transformer yourself ? If so, could you please count turns in your transformer (I want to compare).

[w.v.s.]

Hello chekhov,

regarding my transformer, I bought a replacement via ebay quite a while ago. I think the seller was from china. It is not a exact match to the original one (even a bit larger). But the transformation ratios appear to be plausible. Secondary side of my original transformer seems pretty fried. I did not dig too deep into it, but it might be hard to count.

[w.v.s.]

The seller from Shenzhen was selling 6 of them ("for Keithley 2001 Multimeter Replacement Transformer Without terminal") in 2019. So if someone from this forum bought one, you could share your experience 

[w.v.s.]

Here is what the original transformer seems to have done to the digital board at the end of the transformers life. However, the good news is that the digital board seems to be working after cleaning and resoldering of the EPROM sockets.

[w.v.s.]

Before I remove some burned PCB under R123 and R250 with a dental drill, I check what the vias in this area are about. I was a bit surprised, that some Zener diodes and transistors under the switch were missing in my device. Before, I did not find further signs of attempted repair in that unit. I took a while to find those, because the components are hidden under the power switch.

[w.v.s.]

Hi,

there was some interest in the dimensions of the case of a K2001. I took some pictures and hope they help. Let me know, if I should provide additional information. By the way, in the picture "top" you can see the transformer, which I've purchased from china and mentioned before.

[Qw3rtzuiop]

I was lucky to get a good price on a used 2001m on ebay.
To be safe I replaced all the electrolytics. It was probably not necessary because the desoldered caps still measured fine.

The fan was replaced with a quiet Noctua NF-A4x10 and a rubber gasket.

Since the DMM just had the standard memory option i wanted to upgrade the memory.
I read this old idea from Zucca and designed a small pcb to get the MEM2 option without buying an expensive Dallas NVRAM.

I did some homework and indeed it was not a clever question...

First of all we need to find a 32 pin parallel FRAM, like this one FM28V100:
http://www.cypress.com/?docID=50822

Second it is not 1:1 pin compatible, so we need to build an adapter board.

Third that FRAM runs with 3.3V and not 5V like the DS1245Y, so a VREG should go on that adapter and pray the 3.3V FRAM will be interpreted from the 2001 as logic "1".
Fourth the FM28V100 has two chip enables pins CE1 and CE2... (not a show stopper but another difference)

It´s not looking to be that easy. Still it is a cool Week End project to work on.

It works fine so far...
https://github.com/Qw3rtzuiop/DS1245Y_FRAM

[macboy]

I tried a cheap clone of the Dallas, and it has worked great for several years. The only real problem with the clone is the strange pin spacing, maybe 2.50mm instead of 2.54mm. Fortunately the 2001 has a leaf type socket, not machined, so it still fits fine.

[w.v.s.]

I tried a cheap clone of the Dallas, and it has worked great for several years. The only real problem with the clone is the strange pin spacing, maybe 2.50mm instead of 2.54mm. Fortunately the 2001 has a leaf type socket, not machined, so it still fits fine.

Do you remember the name of that clone? When my repair is done, the device should get some nonvolatile memory, too.

[BD1QMP]

Thank you， TheSteve,  My k2001 can measure but when I run build in test, it reported the same error codes. 
Followed your methods, I found the failed component was Q519, after replaced the Q519 with a 2N7002, the meter is working now. the G--S of Q519 is shorted.
Maybe the root cause of Q519 failure was the C114 leakage brought some high voltage to U505 Pin13--G of Q519, or when I replaced the U505, some ESD from my soldering cause the G of Q519 broken.

I am pleased to report my Keithley 2001 is now error free and working great!

The failed component was Q547 - it wasn't totally open but but didn't turn on properly so relay K503 never closed.
I didn't remove any of the other IC's  as they all looked very clean and didn't appear to have suffered any contamination.
Just to have it all in the same post the faults I was seeing were:
407.1 Front end
407.2 Front end
407.3 Front end
409.6 /500 Freq Comp(this one may not occur 100% of the time)
410.1 TRMS
411.1 TRMS
411.2 TRMS

When I swapped the caps C116 and C117 they were changed to 50 volts(as previously mentioned in this thread they should never have been spec'd at 35 volts).

Having the schematics made all the difference - I can't believe you guys slogged through this thing without them originally, especially being the PCB layout is not very refined.

I still plan to dump my firmware and calibration eeprom next time I visit a buddy with a programmer and will upload them. I guess I may as well convert it to a MEMS version at the same time.

[Zucca]

I was lucky to get a good price on a used 2001m on ebay.
To be safe I replaced all the electrolytics. It was probably not necessary because the desoldered caps still measured fine.

The fan was replaced with a quiet Noctua NF-A4x10 and a rubber gasket.

Since the DMM just had the standard memory option i wanted to upgrade the memory.
I read this old idea from Zucca and designed a small pcb to get the MEM2 option without buying an expensive Dallas NVRAM.

I did some homework and indeed it was not a clever question...

First of all we need to find a 32 pin parallel FRAM, like this one FM28V100:
http://www.cypress.com/?docID=50822

Second it is not 1:1 pin compatible, so we need to build an adapter board.

Third that FRAM runs with 3.3V and not 5V like the DS1245Y, so a VREG should go on that adapter and pray the 3.3V FRAM will be interpreted from the 2001 as logic "1".
Fourth the FM28V100 has two chip enables pins CE1 and CE2... (not a show stopper but another difference)

It´s not looking to be that easy. Still it is a cool Week End project to work on.

It works fine so far...
https://github.com/Qw3rtzuiop/DS1245Y_FRAM

Oh wow you have all my respect! I will surely get mine as well!
You should organize a group buy...

[__BriKs__]

Hello!

I save a 2001 from trash in my customer. There was a label on the power plus with the mention 'don't power on before repair' and the fuse is missing.
Last calibration was on 2019, I would like to repair it to replace my poor DMM.
Then I open the device and see two things: the power transformer has suffering and there is two dead diodes.

PCB are clean, this units looks like there wasn't very used.

The dead diodes (CR514 and CR112) are relative to bootstrap voltage generation. I don't understand why CR514 was dead with CR512, schematic sugest if CR514 dead then CR511 or CR509 suffer also but not CR512 ?
Anyway the issue is relative to BS+ and BS- generator, these voltage rail seems to be used in input buffer section only. How can I test if there is issue in input buffer section ? Impedance between BS+ and BS- ?
The electrolitic capacitors looks really good but I read this is a recurrent faillure in these device.
Impedance of winding of transformer looks good (I find a picture on the net with impedance of each winding), despite the fact the transformer has been hot.

What repair plan can you suggest me ? In first approach I think about changing dead diodes and also relative electrolitic capacitor. I read all electrolitic capacitor should be changed for a good rework, maybe will do that.
And CR512 is a melf BAV103 component which can be found, but CR112 is a BYD17G which seems hard to found, anywone have a sugestion of alternate buyable part ? I think change all diodes relative to BS generator is good idea.

Thanks for any help anyone could provide

[__BriKs__]

  actually other diodes are dead, seems logic
CR109,110,111 dead...

[__BriKs__]

Howw there is a alectrolitic capacitor 'sure' dead also!

Faillure scheme starts to be understandable...

Good news, I have to fix this BS+/- psu!

So I have to find an alternate part for this BYD17G diodes... First look purpose of these diode is to make a full wave rectifier so will say there is no specific parameters for these component, only voltage/current to take care.

[w.v.s.]

Hello and welcome to the forum!

My notes say "GL1M DIO" for this diode. I don't remember, how I came to this conclusion. Good luck with fixing your device!

[__BriKs__]

  love GLM1 because availlable on Mouser!

I have an order to do for another project, will add some electrolitic caps and diodes