Keithley DMM7510, SMU 2450 & 2460 problems

[HighVoltage]

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.

Good to know.

My DMM7510 came with FW v1.5 installed and had no issues.
This confirms that the problems increased with the FW v1.6

[dr.diesel]

Two updates from Tekley:

When I asked specifically about the AZ bug:

"I am waiting on an analysis from our engineering team in regard to performance/autozero; not sure it's a bug unless I missed something. In the meantime, I'll suggest some ways to reduce noise (flicker was mentioned). Autozero is designed to reset drift from the 0.000 reference, but it comes at a time expense. To reduce noise levels, you can increase NPLC, turn on Line Sync -both under Menu->Settings (under Measure)- and use a moving average filter -see Menu->Calculations (under Measure)->Filter."

And then:

"REGARDING BLUE SCREENS AND LOCKUP CONDITIONS:

We have confirmed blue screen and lockup conditions and are testing a hot fix for v1.6.1a firmware. This hot fix would be available for all touchscreen products. The hot fix will not be a general download because it is not a standard firmware release. We will be making it available to customers who have a critical need and are not able to use the next-rev-down firmware on their instrument. This hot fix will be incorporated into a future standard firmware release.

We encourage you to down-rev your firmware as a first attempt to correct the situation.
To down-rev your firmware from v1.6.1a, you can retrieve the most recent prior rev:
2450 http://www.tek.com/source-measure-units/2450-software-12
2460 http://www.tek.com/source-measure-units/2460-software-1
2461 http://www.tek.com/source-measure-units/2461-software
DMM7510 http://www.tek.com/digital-multimeter/dmm7510-software

If you have a critical need for the hot fix, please send me a private message with your email address for follow-up (click on Private messages [X] in the top right corner of the page, select Compose Message)."

[maxwell3e10]

I switched to earlier version of firmware, but the problem with triggers is still present. Attached is the .tsp set up file, renamed .txt to allow download. It is actually just  a text file anyway.

The triggers are programmed from the front panel. The goal is to make measurements every 100 msec for 100000 samples. The program works, but it has never reached 100000 samples, usually stops somewhere between 5000 and 30000, at a random number.

Perhaps someone can upload this script and try to run it on their meter, to see if its a common problem.

[HighVoltage]

Just to update everyone on this issue.
I have updated my 2450 SMU successfully to the hot fix (interim) FW version 1.6.3b  and the blue screen has not happened so far.
All looks good so far.

Keithley is planning to release a full FW update right after the holidays.

[bson]

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.

Of course, no one sits around and waits for a bug report to appear in their in basket and leaps to it...  Bugs are added to the scheduled and prioritized work of any engineering group and handled as they get to it.  However, HOW bugs are prioritized, and what's considered critical enough to drop other work and fix ASAP is telling of how customer-focused an organization is vs how engineering-focused it is.  (Engineers hate being broadsided like this and are very uncomfortable with context switching, especially when they're neck deep in working on a new product and have it entirely in their mind.  Not fun to have to switch this to some other product they haven't touched in a while and get back into that.  Tends to bias the views when prioritizing.)

As for inadequate information, if there's not enough in the report, such as engineering needs a reproducible test, then they should ask for it - reach out to the customer and ask.  But in a lot of organizations engineers aren't permitted to communicate directly with customers.  This is totally ack basswards - engineers should talk to customers when they need technical information and not have to go through product.  Product should prioritize bugs!  The reverse is anathema to customer response.

[HighVoltage]

Just another update ...

Today I upgraded my 2460 SMU to the the hot fix (interim) FW version 1.6.3b
And I used the instrument all day and had no blue screen.
Really good so far.

[plesa]

Just another update ...

Today I upgraded my 2460 SMU to the the hot fix (interim) FW version 1.6.3b
And I used the instrument all day and had no blue screen.
Really good so far.

Kudos for Keithley 

[maxwell3e10]

Good to hear the BSD seems to be solved. Now they just need to solve the functional problems:

A. Make the meter trigger at uniform timing without crashing. In contrast, Keysight 34465a has two different methods to program uniform sampling.

B. Resolve noise issues. I made a few more plots to show the problems. Of course the suggestion of their engineer to use the moving average filter makes it only more obvious that it doesn't work properly.

[HighVoltage]

Yesterday I updated the last one, my DMM7510
It has been used for many hours and also did not show any Blue Screen anymore.
I think the BSD issues have been solved completely on the 2450, 2460 and 7510.

Thanks to Keithley for taking care of this issue so fast and so well with an interim FW solution.

May be Keithley can fix the remaining problems in a future full FW update soon.

[maxwell3e10]

Has anyone tried to run the script I posted above for uniformly timed measurements?

If its confirmed in other meters, we should break it out into a separate item and post on their support page.

[dr.diesel]

Has anyone tried to run the script I posted above for uniformly timed measurements?

If its confirmed in other meters, we should break it out into a separate item and post on their support page.

I'll get mine running your script today! 

[dr.diesel]

Has anyone tried to run the script I posted above for uniformly timed measurements?

If its confirmed in other meters, we should break it out into a separate item and post on their support page.

I'll get mine running your script today!

Well, mine made it to 34464 readings, then just quit, no errors.   

Running the test 1.6.3b FW.

[maxwell3e10]

Thanks. That is what I was seeing, it stops at a random number without errors, the data are still there. The problem was both in 1.6.1 and 1.5 versions.

[dr.diesel]

Thanks. That is what I was seeing, it stops at a random number without errors, the data are still there. The problem was both in 1.6.1 and 1.5 versions.

I have forwarded the info/script to Keithley for their confirmation.

[maxwell3e10]

Thanks. Who are you talking with?
We still need to press them about the noise issue, but I figure this will be easier to handle first.

[dr.diesel]

Thanks. Who are you talking with?
We still need to press them about the noise issue, but I figure this will be easier to handle first.

PM incoming.

[HighVoltage]

I just ran your script as well

FW  1.6.3b
Stopped at 34464 rdgs

Also interesting:
The symbol on the top line switched from RUN to WAIT but sometimes it just stopped and did nothing for a few seconds and then alternated between RUN and WAIT again.
 

[maxwell3e10]

Ah, interesting, stopped at the same number of points. For me it would usually stop at fewer points and not the same number, but maybe once it was in the 30 thousands.

[donkey77]

Maybe it stops before 34465 as it would then promote Keysight!

[HighVoltage]

Just an update and confirmation:
It has been a few weeks now that I have used this interim FW on all three instruments and I had zero blue screen effects.
That problem seems definitely solved.

Remote control hangup still happens from time to time on the DMM7510.

[dr.diesel]

I can echo the above, zero bluescreens since the interim FW.

The web interface is still unusable, programmed trigger/AZ issue are still present.

[maxwell3e10]

I wonder if we should separate the issues into 3 items and post them individually on their support page, to increase the visibility of the problems.

[Kleinstein]

AFAIK the 3 instruments use a very similar software for the display part and it seems like the the big problem with blue screens is more or less solved.

The not so good AZ implementation is a little different from that. Some of the shown curves in the DMM noise comparison thread even suggest that this problem could be common with even more Keithley instruments (like the 2002). I have not seen noise spectra of Allan deviation curves for the SMUs - so not sure they are also effected.

[HighVoltage]

I wonder if we should separate the issues into 3 items and post them individually on their support page, to increase the visibility of the problems.

I was told that Keithley is working on all issues and it will be addressed in the next official FW update.

[dr.diesel]

I was told that Keithley is working on all issues and it will be addressed in the next official FW update.

Dumping a change/fix for the AZ bug could be interesting.  Although quite small it might pose some problems for critical automated test cells.