Keithley 2700 schematic diagrams

[Taviman]

Greetings,

I have recently purchased a Keithley 2700 and it appears to have the input section broken.
DCV has an offset of about 5mV with leads shorted.
DCI shows 2.2A without input.
Ohms show( almost) dead short in 1Ohm and about .19 Ohm in 1M and above ranges

Can anyone help me with advice and/or schematics. I read some threads here but I don't have yet a solid path on how to approach this.
Power supply rails are ok, and the input  FET's seem to be correctly switched in respective modes. I spent a few hours today in an attempt to find faulty components but it is hard without schematic.

[TizianoHV]

As a starting point I would expect many similarities to keithley 2000 multimeter.
On Xdevs.com (https://xdevs.com/doc/Keithley/2000/) you can find schematics (https://xdevs.com/doc/Keithley/2000/K2000.pdf) and repair manuals for it.

[Taviman]

Thanks! I didn't think about this 
Will check and report here my progress

[TizianoHV]

*By the way, if you want to run the built in self test you have to enter the "secret" mode by pressing the "OPEN" button wile powering up (keep it pressed for a few secs...).
Afterwards, by pressing "SHIFT", "DIGITS"("TEST"), you'll find the hidden self test.
Good luck

[Taviman]

Amazing tip. I have error codes in 300's, 400's, 500's and 600's
I don't have the cards so I have 700's codes also.

So... I needs lots of coffee

[Taviman]

Well, thanks to a previous repair thread, I started with checking the U115 LM339D, but I found that to be operating normally.
Now, you will have to keep in mind I am average at best in electronics, but found that difficult projects are the ones that educate me most.

Finding myself in a dead end, I decided to check the analog switching states, as per service manual, and found the K102 to be stuck in reset condition.
Afaiu, that is responsible for switching in the AC voltage section so I am not holding my hopes high, but still. That relay is controlled by U116, a ULN3003L, a Darlington Array. But I couldn't check it's operation with the multimeter, until I realized there must be a short pulse that latches the relay in a new state.

So I pull out my cheap Hantek DSO, and here it was: a nice rising edge on the input that triggers a short low output that latches the K101 in set or reset state. Trouble is the SETK2 and Reset K2 inputs are not operated when I switch from 1V to 10V in ACV. So there must be something else.

I followed through to the diagrams of the K2000 (thanks @TizianoHV for suggesting) and reached to U109, a MC14094BD 8-Stage Shift/Store Register. If one or more of these 8-StageShift/Store Registers (U121, U134, U109, U106, U130) are broken that would explain a lot of the defects I am seeing (basically all modes are bananas, only DCV exhibits a constant offset. ACA is showing 2A and ohms are shorted). Of course, in the process I am trying to understand the whole picture on how all the switches interact with the current path, and what is doing what.

Now I am looking for advice on how to reasonably check these shift registers, so I don't have to pull them all out.

[Kleinstein]

If the actual shift register is not working all following stages would be bad. So except for the last bit(s) it would effect multiple control lines.

If available on the scope one could use an SPI decoder function to read the data send to the 4094 chips. This could be directly from the µC/FPGA but also between individual chips down the chain.
Even with simple DSO the data between the 4094 chips (e.g. pin2, pins 9/10) are interesting. So one could see if there are actually data comming in or not. The strobe signal (pin 1 of 4094) could be used as a trigger.

[Taviman]

Yeah, I don't have decode on my Hantek. I was actually considering this possibility, that the following shift registers are silenced since the malfunctions go beyond that specific stage "business" . Thanks for reinforcing my hunch.
I also have a logic level analyzer, I'll try to see if that one shines light upon the situation.

"Not working" would mean dead silent or bad bit timing or else?

[Kleinstein]

Not working would normally mean stuck to a fixed value. The 4094 shift registers are a relative insensitive type (especially if the 2nd output for chaining is used). So the signal quality is usually not an issue.
SPI is not that complicated and for not so many bits also just the scope picture (2 channels or more) could be OK too as the data are binary anyway.

[Taviman]

Well, I finally managed to repair it and the dang thing is awesome.
Upon checking the relay states I found K102 not working (reset state in ACV). That relay is controlled by U116, which in turn is controlled by U109, a MC14094BD. The damn thing was hot so I thought replacing it was going to solve the situation. It's supply rail was at 1.8V.
Now, Keithley in it's wisdom chose to separate the 5V rails between AC and DC with  a precision 50 ohm resistor. Anyone can clarify why?

However, after I removed U109 it I checked the local 5V rail and it was still 1.8V. Additionally,  a precision opamp in AC section (x10 gain), the LF357M was hot. Great, because I couldn't source that part anywhere around here. I continued checking with the schematic and found that it's output was shorted by U105, a JFET analog switch. Interestingly enough, this switch needs both +/- 15 V and +5V. If it wasn't for that 50ohm separation resistor I guess my poor LF357 would have been fried...
Bingo!
When I removed U105, and reinstalled the U109 I got 5V on the local rail and all the functions except AC related ones were restored. I was looking at the OVR.FLW message in ohms and couldn't process it's meaning

Anyway, one week later I received the missing DG211 and now everything works.  I am in awe at it's speed and precision. Cal seems intact for my purposes. It can measure f up to 1MHz!

Thanks again TizianoHV for the hint on the schematics, it would have been impossible without it. I can confirm they are identical to my K2700, at least in the parts where I looked.
Thanks Kleinstein for taking time to answer to a noob.

My regards,
YO4OCM - Tavi

Edit: Yes, all the fault codes disappeared, except for the 700 ones, since I don't have the cards.
I admit, I bought this one for the speed and because it was cheap since broken. What options do I have to connect to a PC? I don't have experience with data logging or ever needed such but I want to learn

[TizianoHV]

Well!!

Yes, these keithleys have the best and fastest auto-ranging.

If you aim for remote programming you can use RS232 or GPIB:

RS232 is the cheap way: you can use a RS232toUSB dongle (like this https://www.amazon.com/Adapter-Chipset%EF%BC%8CDB9-Serial-Converter-Windows/dp/B0759HSLP1/ref=sr_1_3?keywords=rs232+to+usb&qid=1684876051&sprefix=rs232%2Caps%2C231&sr=8-3) but I found RS232 unreliable (In particular at higher speeds) and slow.

GPIB is a bit expensive but you can use one cable for multiple instruments, I bought a used "GPIB-USB-HS" (National Instruments) for 100€ on ebay and it has been extremely reliable .
_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Are you using Windows? Here what I did (a noob):
After choosing between the two GPIB/RS232 you'll need to install the drivers and VISA needed to comunicate with the instruments.
Windows will automatically install the RS232-USB drivers. But VISA and GPIB drivers aren't that simple, since I'm using a NI gpib-usb converter I've installed NI VISA(edit: https://www.ni.com/it-it/support/downloads/drivers/download.ni-visa.html#480875 - The last version doesn't install GPIB drivers. After installing NIVISA open NI package manager and install NI488). Alternatively you can use Keysight IO Libraries Suite (https://www.keysight.com/us/en/lib/software-detail/computer-software/io-libraries-suite-downloads-2175637.html), it's the same thing but will install drivers for keysight converters (82357B USB/GPIB...), at the time it wasn't working with windows 11.

Anyway, these will install all drivers and will offer some convenient features, for example they permit you to record all the commands you are sending/receiving from the instruments (really useful for debugging).

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

Automation:
After installing all drivers and VISA you can start automating:
I'm using python with pyVISA, here a really good "Rapid Start" tutorial https://www.youtube.com/watch?v=W5Brxiwnp5g.
Python is pretty easy and you can find a lot of tutorials, you can use matplotlib to plot your data, here an example using similar configuration and meter -> https://www.youtube.com/watch?v=TLUTCDbt52I&t=66s.

See photo: you can be fancy and build an interface or leave it on command line (like I did).


Obviously there are many other ways, but this was quite convenient.
_______________________________________________________________________

Keithley over remote:  not the fastest meter, getting maximum speed is quite trikey and it can't be done continuously (since you'll need to use the internal buffer).

[Taviman]

Wow, that really helps!
Yes, I'm using W10. I was looking for a simple explanation but I was only coming across official links that aren't' that helpful, understandably so.

Grazzie mille!

[Taviman]

I understand now the pain with RS232... Commands are not recognized on instrument end, error 800... I'll see about that GPIB

[TizianoHV]

Yeah, with a desktop was almost ok but with my laptop no.
Older Keithleys develop an issue with the RS232 chip, there are some articles on EEVblog (Keithley2000 - https://www.eevblog.com/forum/testgear/keithley-2000-rs-232-wtf/msg1561996/#msg1561996).

Did you get at least one answer (query   *IDN?    on NIMAX VISA interactive control)?

If yes I found out that after getting an error (800, 802...) to get it working again you had to first disconnect the RS232-USB dongle from the usb port, then close the terminal, wait a few seconds and reconnect it. Otherwise it would keep throwing errors .

Someone built an arduino GPIB-USB converters but I don't know if they work with VISA and they can't work with multiple instruments.

[coromonadalix]

you have some usb-gpib interfaces too,  some works with visa tools / packages and dont cost a fortune

i have 2x from  xypho, works very well with R&S visa drivers / small footprint compared to others

and yes the  *IDN?    is always a good start

[Taviman]

Did you get at least one answer (query   *IDN?    on NIMAX VISA interactive control)?

Yes it did, but not consistently. Like you said, if I connect with Max, then for sure Py cannot connect.
if I connect first with py, then I have to restart my laptop to connect with NI-Max. A pain, like you said.
 

[TizianoHV]

That could be an issue with the RS232-USB converter driver. Do you have an other converter (different brand) or pc?

Luckily for you I needed to reinstall NI VISA, after reinstalling it 3 times without getting it to work I noticed that the latest version of NIVISA does not automatically install drivers for GPIB (You have to manually install NI488 via NI package manager). I corrected my  original post.
(This shouldn't solve your RS232 issues).

If it finally works don't touch it! 

They just sold a few agilent 82357B on ebay.
With windows 10 you can use keysight io. I liked it before upgrading to W11 and buying a converter from NI.

[Taviman]

That could be an issue with the RS232-USB converter driver.

True, that might be the problem, I'll see about another one.
Anyway, all those GPIB-USB on Ebay are clones, if I understood well. Right now I went through re-arranging my desk/lab so I have to postpone a bit other investments

[SanderMTI]

Gentlemen,

I too was looking for schematics for a 2700 for the purpose of adding a trigger output which I can use to trigger my other meters.
I also have a 2000, so now I have them for that model.
Thank you.

On the subject of remote reading/ control:
I use a Prologix USB-GPIB adapter with  EZGPIB and the Datalogger program I wrote for it
No drivers or complicated stuff needed to talk to GPIB and RS232 meters from the same software, output is a .CSV file.
PM me if interested, I'll send you a copy.

Sander