More on Keithley 2000 AC Volts errors

[Mark Rivers <RIVERS@cars3.uchicago.edu>]

From:	CARS3::RIVERS       'Mark Rivers' 22-APR-1999 17:49:32.88
To:	 CIGOY_DALE@KEITHLEY.COM
CC:	RIVERS
Subj:	More on Keithley 2000 problems


    Hi Dale,

Thanks for your help on the workaround on the Keithley 2000 problems.  This
has fixed the really gross errors which we were having.  However, even with
your suggestion we are seeing errors at the 5 mV level on the shorted input. 
Here is another test program I wrote.  It does the following:
    - Close channel 1
    - Wait for delay time
    - Read channel 1
    - Close channel 2
    - Wait for delay time
    - Read channel 2
    - Wait for delay time
    - Read channel 2 second time
    - Wait for delay time
    - Read channel 2 third time

pro test_keithley1, delay
    send_string, 'rout:close (@1)'
    wait, delay
    v1 = read_value()
    wait, delay
    send_string, 'rout:close (@2)'
    wait, delay
    v2 = read_value()
    wait, delay
    v2a = read_value()
    wait, delay
    v2b = read_value()
    print, 'Delay = ', delay
    print, 'Channel 1 = ', v1
    print, 'Channel 2 (first read) = ', v2
    print, 'Channel 2 (second read) =', v2a
    print, 'Channel 2 (third read) =', v2b
end

I set the function generator to output ~100 mV AC sine wave output at 60 Hz,
putting this signal into channel 1.  I shorted the inputs to channel 2 as
before.  Here are the results in mV as I vary the delay time from .001 second
to 1 second.

Delay time  Channel 1  Channel 2    Channel 2     Channel 2
                      (first read) (second read) (third read)
.001          106.5       16.1        .15           .07
.01           108.1       21.8        .03           .04
.05           110.2        7.0        .40           .15
.10           110.9        4.9        .12           .05
.20           110.9        0.87       .31           .08
.30           110.9        0.42       .13           .05
.40           110.9        0.35       .03           .05
.50           110.9        0.35       .04           .05
.60           110.9        0.42       .05           .05
.70           110.9        0.19       .05           .06
.80           110.9        0.26       .05           .06
.90           110.9        0.13       .05           .05
1.0           110.9        0.10       .06           .06

These results show that even with delays as long as 1 second the Keithley 2000
is not meeting its AC specifications.  Under these conditions (100 mV scale,
60 Hz input) Appendix A, page A-3 in the manual states that the accuracy
should be +-(.05% of the reading, +30 microvolts).  The third read of channel
2 meets these specifications, and the second read meets the specs with delays
of 0.4 seconds or greater. Clearly there is an unacceptable amount of
cross-talk from channel 1 on the first read of channel 2.


                                   Sincerely,

                                   Mark Rivers