Keithley 199 uncal

[Spike]

Hi all,

I visited a local garage sale yesterday and picked up three Keithley 199 multimeters.  Of the three, one appears to operate perfectly. While none of them have probably seen the inside of a cal lab in forever, it appears pretty spot on compared to my 34410a (which is in cal).

The second one displays UNCAL on power-up, and then after pressing any button it resumes normal operation.  According to the manual this is expected behavior if there is a problem reading from the EEPROM in which calibration constants are stored.  Searching here and elsewhere I found a couple of threads that indicate this is a not too uncommon problem, and a bad PROM is likely.  I have a few questions:

1.  Can anyone tell me which PROM I need?.  The manual shows a part of 2816a and I was able to see that the chip in the second unit was a Xicor X2816AP, but there's no indication of the speed, and there appear to be a few choices on eBay (the only place I've found these).  I don't know if it will matter for this case but want to make sure.  This appears to also be made by Samsung or Intel, is there a preferred manufacturer?

2.  If it can't read the calibration constants, what values does it use?  After resuming operation it appears to function pretty well.  I tested DC voltage, current and resistance and while maybe not bang-on, it agrees pretty well with my Agilent meter.

3.  Is there a way to save calibration without actually calibrating it before I attempt to replace the PROM?  I simply don't have the equipment or ability to calibrate it and I'm afraid I'll make it worse.  As it stands now it is at least usable after the annoyance of the UNCAL message on startup.

As for the third meter, it also exhibits the UNCAL issue, but beyond that it does not seem to function.  It displays overload for most ranges and functions.  It's unlikely that I would ever purchase a needed piece of equipment that is for parts or repair, simply because my skill level is not adequate.  In this case, though, it might be a fun and educational experience to try to fix it so I will attempt that at some point.


Thanks in advance for any information you can provide.

[Spike]

Okay, to answer some of my own questions.  I finally found a datasheet for the EEPROM and it seems that the one with no speed designation is 300ns but, as it turns out, it doesn't matter anyway.  It would appear that there's nothing wrong with the EEPROM or the meters.

I ran a faux calibration on the 3V range of the non-functioning unit using what I have available, at the end of which it saved the data without complaint.  I was then able to make measurements on this range and it appears to work flawlessly.  I powered the unit off and then back on and no more UNCAL message and the 3V range continues to work.  So it would appear that all this unit needs is a calibration and I have three working units.

Unfortunately I'm not sure of the wisdom of dumping a bunch of money on calibration for these.  Does anyone know of a place to get these calibrated in the S.F. Bay Area that won't charge an arm and a leg?  I don't need data or traceability, just for it to be reasonably calibrated.

One other question that I do have: does anyone who knows these meters know whether losing cal data like this is common?  It seems that two of the meters have some sort of corruption of the data on the EEPROM, but one just happened to continue to function and the other not.  I haven't tried a faux cal of the other UNCAL but working unit because I don't want to screw up the calibration, such as it is.  I'm just hoping it's the same issue.

Thanks

[Smith]

Calibration data is lost after some years (probably 20-25 years. I got mine for free as most ranges didn't work. Turned out cal values where gone. After calibration all worked fine.

[Spike]

So much for that 100 year data retention spec in the datasheet.

[macboy]

For hobby purposes, your in-cal 34401A is quite sufficient to use as a standard against which to calibrate these meters. Luckily, the Keithely 199 can be calibrated per-function and per-range; there is no need to adjust everything in one go, just the ones you want to do. So you can use whatever good stable sources you have available, measure them with the 34401A, then set that same value into the K199. This is exactly what I did for my two K199 using three K2001 meters as the reference. Your success with this will depend on the stability of your sources and your technique (not introducing temperature differences to connections, etc.). Of course the best you can do is a match to your 34401A, so any error in it is reflected in the 199 too.

Note that the tempco of the K199 is not so good, so do the cal at the same room temperature that you will use them in. Also, let everything warm up well (>1 hour) before calibration and again before use. (the 199 uses a simple low tempco zener reference compared to the heated buried zener reference of the 34401A).

[Spike]

Yeah, I'll be satisfied with that level of calibration and that was in the back of my mind.  It'll be a bit of a pain for me to generate all the signals needed, though.  Low voltage DC shouldn't be a problem and I can muddle through with what I have for resistance and probably current, but the highest DC voltage range will be out of reach for me and AC as well.   It's not that I have great need for this, but it would be nice to just do it all and be done with it.