Keithley DMM7510, SMU 2450 & 2460 problems

[HighVoltage]

The new Keithley instruments are really great, when they work alright. But ...

My DMM7510, the SMU 2450 and 2460 all suffer from time to time of several problems.

1. "Blue Screen of Death"
2. Not responsive to any button

I have contacted Keithley / Tektronix Germany but was not offered any help.
There is a hope that this will be fixed with any future FW update but who knows.
I was told by Keithley Germany, that this is a single incident, that only I have these problems and Keithley is not aware that others have these problems too.

The latest FW Version: 1.6.1a is installed
Release Date: Aug 23 2016 - 12:00 am

I have heard from other people here in the eevblog forum, that they have similar issues / problems.
May be we can help Keithley by compiling all problems in this thread and hope they can use it to fix everything in a new FW update.

So, if you have encountered any issues with your new Keithley instrument as well, please post them here.

Blue Screen Error Codes (so far detected)

- PC=0003D660
- PC=0003D661

[dr.diesel]

FW=1.6.1a on both my 7510 and 2450.

I've seen both issues you list above with #2 being very common.  I can easily trigger #2 by navigating the graph tabs via the instrument's web interface.

We should add a line item for maxwell3e10's noise discovery:

https://www.eevblog.com/forum/metrology/dmm-adc-noise-comparison-testing-project/msg1081347/#msg1081347

During a quasi non-responsive time I managed to catch this before it fully locked up:

 

[dr.diesel]

Posted a message on the Keithley forums, linking to here.

https://forum.tek.com/viewtopic.php?f=363&t=139083#p281815

[maxwell3e10]

1. Confirm blue screen with latest firmware

2. When using trigger programming from front panel to arrange for timed measurements, the meter stops triggering after some random number of samples.
 
3. Not a bug, but when saving buffer to USB, the meter does not check for existence of a file with same name, instead, it just always asks to confirm saving file.

4. Most directly dealing with performance, the autozero does not effectively remove flicker noise on DCV 10V scale for continuous measurements on time scales of 1-50 sec, as illustrated in figures below. As the result on 10V DCV scale the noise performance of DMM7510 is significantly worse than Keysight 34465A.

[_Wim_]

Wow, I would not have expected this from Keithley. I realize that every brand can release a software with some issues left, but then I would expect them to at least acknowledge these and assure the customer they are working on a fix for the next firmware release. Especially because it seems it not very difficult to trigger the fault…

Still, I think their gear looks really fantastic, and would love to have your setup (with our without problems). But for the moment I do to little development to warrant that kind of cost…

[HighVoltage]

Thanks for adding to my list of problems.

Lets hope, it is all only software related, which I think it is.
My 2450 SMU never showed a blue screen until one of the the last FW updates, so this would indicate only a software issue.
But the noise discovered by maxwell3e10 might be HW related?

I will make a similar noise test comparison between the 7510 and the 34470A
Will be interesting, to get the same results

[plesa]

On some instrumets it takes Keithley years to fix issues:
2600B series 3.3
2400 series 33
And some of them still remains.
Do yo know how to reproduce instrument hag up issue or how to prevent their occurrence?
For instance on K2400 is bug with disabling display, but on 6517B series is already fixed:)
It must be quite painful for Keithley and other manufacturers that people are sharing issues on this forum.

[carl_lab]

It must be quite painful for Keithley and other manufacturers that people are sharing issues on this forum.

It should be painful.
They can fix the problem, and everything is all right...

[plesa]

It must be quite painful for Keithley and other manufacturers that people are sharing issues on this forum.

It should be painful.
They can fix the problem, and everything is all right...

I do not think they do not want to fix it but I expect that software team is shared across platforms and they are limited by resources for maintenance.
You can speed it up by describing steps to reproduce and connecting fix to big deal:)

[HighVoltage]

Do yo know how to reproduce instrument hag up issue or how to prevent their occurrence?

Reproduce, partially yes:
The none response can be easily triggered through the web interface and it required a cold / hard restart
The "Blue Screen of Death" happens out of the blue once in a while, but mostly when swiping from the statistics screen to the graph screen.

Prevent:
No.
It is possible to use the DMM7510 for hours and without any problems.
Suddenly it crashes.
After a reboot, it is very well possible that this crash will happen right away again of it can take hours again.
Worst case, one stores a lot of good data and wants to save a screen shot and right then it crashes.
 

[Kleinstein]

Thanks for adding to my list of problems.

Lets hope, it is all only software related, which I think it is.
My 2450 SMU never showed a blue screen until one of the the last FW updates, so this would indicate only a software issue.
But the noise discovered by maxwell3e10 might be HW related?
.....

The noise problem gets smaller if triggered readings are used (visible in the curve in the noise comparison thread). So the software can do better, if used in a different mode. If done the normal, simple way the AZ function should take care of the flicker noise on time scale longer than two readings (input and zero) even better - just like it did with the Keysight instruments. There would have to be something really odd in hardware to get this type of noise spectrum with active AZ mode. Even if there is some special Hardware involved the better performance using a different mode shows that at least some of the noise is avoidable.

The much more likely reason it, they tried to be clever and used different (more complicated), non standard way of doing the AZ mode and this just does not work well. It might be possible it worked ok on older very different HW. The curve for the K2002 somewhat suggests that there might be a similar weakness with the K2002 too.

Different from the other SW bugs, this is not a rare glitch happening under special conditions. This is more like a well defined point in the software, where a not so good formula is used. Normally this should be relatively easy to find and fix.

[maxwell3e10]

Has anyone tried to keep track of the code it displays on the blue screen? Perhaps it would be useful to make a list of the codes.

When I program a trigger series from the front panel, there is 100% chance it will lock up soon, although not blue screen.

[dr.diesel]

Vince W from Keithley just responded:

"Thank you for bringing this to our attention. I have forwarded the links on to our product teams to address these issues. We would post any responses on our forum, so the blog participants should watch here."

Direct link:

https://forum.tek.com/viewtopic.php?p=281837&sid=f0617886f132ba13c5af865cee429a5d#p281837

[dr.diesel]

Has anyone tried to keep track of the code it displays on the blue screen? Perhaps it would be useful to make a list of the codes.

I don't see the blue screen near as often as the general lockups, but this is the error I've seen repeatability.

PC= 0003D660

[HighVoltage]

I just had a different blue screen error code on the screen:
PC=0003D661

All error codes will be edited in to the first post.
Any others?

[dr.diesel]

Just caught another one on my 2450.  Then happened via Quickset > Voltmeter > Clicking Yes on the warning box.

PC= 000282DC

I suspect we'll need to mirror this info on the Tekley forums.

[adranp]

Just wondering, have you had the same issues prior to FW 1.6.1a?

[dr.diesel]

Yes, I've had stability issues with all FW versions, however it does "seem" like the most recent is more problematic.

[boborich]

Just to report that have blue screen issues with the recently purchased DMM7510 with the latest FW 1.6.1a.

Code: PC=0003D660

This happens a lot when playing around with the buffer and chart operation.

[plesa]

Just to report that have blue screen issues with the recently purchased DMM7510 with the latest FW 1.6.1a.

Code: PC=0003D660

This happens a lot when playing around with the buffer and chart operation.

Are you going to return this unit?

[adranp]

Did anyone running 1.6.1a tried to downgrade to an older FW release and do some tests?

Thanks.

[dr.diesel]

Further response from Vince at Tekley:

"We are addressing these issues now and applying engineering release control and audit rigor for a new firmware release. We apologize for getting this latest release out with the issues you have found. Each release takes time and this is a difficult time of year to get all resources working at full speed because of the holidays."

Sounds like a new FW revision might be headed out way soon.   

[boborich]

Just to report that have blue screen issues with the recently purchased DMM7510 with the latest FW 1.6.1a.

Code: PC=0003D660

This happens a lot when playing around with the buffer and chart operation.

Are you going to return this unit?

Not sure yet, it is a nice bit of tool and I have printed the 1000 pages of the reference manual

Hopefully keithley will fix these issues with FW releases.

Btw, if I return this, what shall I get? 34470 or 34465? In China, 34470's price is close to DMM7510 and that is the reason why I bought 7510 at the 1st place.

[boborich]

Further response from Vince at Tekley:

"We are addressing these issues now and applying engineering release control and audit rigor for a new firmware release. We apologize for getting this latest release out with the issues you have found. Each release takes time and this is a difficult time of year to get all resources working at full speed because of the holidays."

Sounds like a new FW revision might be headed out way soon.   

I saw that reply too, get my finger crossed, and thanks for putting everything together and pushing Keithley to look into this for us!

[boborich]

Did anyone running 1.6.1a tried to downgrade to an older FW release and do some tests?

Thanks.

I downgraded to 1.5 yesterday and seems to be much better than 1.6. Haven't got any blue screen or lock up so far. Recommend you try that before the new FW come out.