Keithley 2100 Calibration Manual Errors??

[essele]

I've been recalibrating a Keithley 2100 after repairing it and have noticed a couple of errors in the cal manual. Has anyone else noticed this??

For all the resistance ranges you set the calibrator to 0R, STANDBY, configure the 2100, then calibrator to OPERATE, then send the offset cal command, then calibrator to STANDBY.
Then you do the actually resistance: calibrator to value, OPERATE, send full scale cal command, then calibrator to STANDBY.

The cal commands include the actual value of the resistance using scientific notation, i.e. 0.9999615E6 for a 1M range potentially.

To quote the manual as an example:

18. Set the 5700A output mode to OPERATE.
19. Adjust the input offset by sending the following SCPI command:
                    CAL:PROT:DC:STEP 1,1E6
20. Wait until the following message is displayed on the Model 2100 display:
Cali OK
21. Set the 5700A output to 1MΩ.
22. Adjust the full scale measurement by sending the following SCPI command:
        CAL:PROT:DC:STEP 2,<ACTUAL CALIBRATOR OUTPUT>
23. Wait until the following message is displayed on the Model 2100 display:
Cali OK

All seems very sensible to me.

This works all the way through the resistance ranges until you get to 100M.

You do the zero offset as normal, but then it says:

31. Set the 5700A to 100MΩ and output to STANDBY.
32. Adjust the full scale measurement by sending the following SCPI command:
CAL:PROT:DC:STEP 2,99.99750
33. Wait until the following message is displayed on the Model 2100 display:
Cali OK

So three issues in my mind:

1. Why would you set a calibrator to 100M and then just put it on STANDBY??
2. Why would you not put the actual calibrator value in, rather than 99.99750
3. Why have we dropped engineering notation ... surely it would assume 99R rather than 99M?

Anyone got any experience of this?

You can do individual ranges quite easily so I'll probably experiment a but, but just wanted to check I wasn't going mad.

[boggis the cat]

I've been recalibrating a Keithley 2100 after repairing it and have noticed a couple of errors in the cal manual. Has anyone else noticed this??

...

1. Why would you set a calibrator to 100M and then just put it on STANDBY??
2. Why would you not put the actual calibrator value in, rather than 99.99750
3. Why have we dropped engineering notation ... surely it would assume 99R rather than 99M?

Anyone got any experience of this?

You can do individual ranges quite easily so I'll probably experiment a but, but just wanted to check I wasn't going mad.

I haven't calibrated this particular model, but it isn't uncommon for documentation to have errors.  My assumption would be that the person who wrote the procedure up was going from notes supplied by the engineer responsible for the process documentation, and the review prior to publication didn't catch these errors.

So my educated guesses:
1. You wouldn't
2. You would
3. The notation should be present (so "100E6" in this case).

It is almost always the case that the process is the same for each range, so you would duplicate the same process for all of the high resistance settings.

An observation for this particular process: most calibration adjustment procedures for long scale DMMs request a short to be applied then step through all of the zero offsets, before moving on to span correction.  It seems unusual to be using the calibrator output as the zero offset, although it isn't likely to introduce significant errors for MOhms ranges.

[essele]

Thanks Boggis ... that matched my thinking as well.

An observation for this particular process: most calibration adjustment procedures for long scale DMMs request a short to be applied then step through all of the zero offsets, before moving on to span correction.  It seems unusual to be using the calibrator output as the zero offset, although it isn't likely to introduce significant errors for MOhms ranges.

The first 2 steps of the overall process are zero range setting using a short for the front and rear terminals respectively, but then each individual setting (at least for DCV and resistance) have a zero setting phase using the calibrator.

The whole process seems a little strange to be honest ... but I know this isn't really a Keithley device!

[dmderev]

Hello, I am trying to follow the calibration manual of K2100 and correct slight difference in some ranges.
Is it possible to do only partial calibration (i.e, zeros and few voltage ranges like 10V and 100V)?
If so, will the calibration on other ranges be taken from previous calibration cycle?
In other words, is it safe to save the partial calibration state with CAL:PROT:SAVE ?
I do not have AC and current calibrator, but would like to at least calibrate some DC and resistance ranges.

Also, what is the meaning of second argument in CAL:PROT:DC:STEP 1,10 ? This is the calibration of zero offset at zero input.
Apparently, the meaning of the second argument here CAL:PROT:DC:STEP 2,10 is the actual voltage on the input, so if the reference is, say, 10.00025V, one has to write: CAL:PROT:DC:STEP 2,10.00025 - correct? 

[dmderev]

In addition to my former question:
It looks that it is possible to do calibration only on few selected modes and ranges. But this calibration coefficients hold only until a power cycle. The CAL:PROT:SAVE saves new calibration date, but does not change the calibration coefficients used in next power cycle.
Is there any secret sauce to store partial calibration data to EEPROM? Any special command or jumper?
Is it possible to read calibration coefficients vis SCPI to check that EEPROM was actually updated?