[volt-nuts] 3458A Calibration Education

[J. L. Trantham]

I have successfully modified my A5 board (03458-66505) by removing the EPROM
(U110), the two DALLAS 1230Y NVRAM's (U121 and U122), and the DALLAS 1220Y
NVRAM (the 'CAL RAM', U132) and installing sockets at all 4 positions.  In
the process, I lost the data in the 'CAL RAM', the one chip I really wanted
to archive.  The chip seemed to read differently each time I read it using
my BP Microsystems BP-1600 Universal Programmer and was, clearly, not what I
would have expected.  The two DALLAS 1230Y chips read consistently and I was
able to archive their data with their data in a form I would have expected.

 

I received new DALLAS 1220AD chips (Mouser did not have the DS1220Y),
installed one, and turned on the meter.  It presented me with a series of
messages including ACAL REQUIRED, SCAL REQUIRED, SECURE REQUIRED, and ALL
CAL REQUIRED.  It passed the SELF TEST and, otherwise, seemed to work
although it did not seem to read my voltage standard accurately (high by
about 0.15 volts).  The same thing was seen when I used the old DS1220Y.

 

After installing the new DS1220AD, I went through the calibration procedure
using my 'home' standards (homemade 14 ga. copper wire 'low thermal short',
731B, General Radio 1432-P Resistor, and 3325A level generator) after which
it seems to be completely up and running as advertised with no 'messages' at
turn on or 'RESET'.

 

Before I undertook the effort to modify the A5 board, I did the 'CALNUM?'
command and it returned '1'.  Now, when I do the 'CALNUM?', I get '13'.  Of
note, there are 13 separate steps in the calibration protocol.

 

Also, I archived the DS1220AD chip after the calibration was completed, used
the data to program the DS1220Y chip, installed the DS1220Y chip and it
seems to be working just fine.

 

Therefore, I have several questions:

 

1.	What should the 'CALNUM?' command return after a first calibration
with HP/Agilent (in other words, meter just delivered new)?  Did my meter
have a problem and someone just did the DC calibration and therefore it
would return a '1'?  I can't understand how I would go through the
calibration procedure once and then get '13' in reply to 'CALNUM?' except
for the fact that there are 13 separate steps in the calibration process.
Is there some sort of 'hardware' calibration that HP/Agilent goes through
that completely calibrates the meter yet only results in a '1' in response
to 'CALNUM?'?
2.	Is it possible to 'erase' the 'CAL RAM' by shorting certain pins
together as a result of unsoldering the chip or sitting the chip, pins down,
on an aluminum surface?  I still can't understand how I could seemingly
harvest the two DS1230Y chips without a problem and then lose the data in
the 1220Y chip.  The date code is from 1997 and, clearly, the battery could
be a problem.
3.	Are there any 'programs' installed in U121 and U122 from the
factory?  Or are all these just user entered programs for different
measurement protocols?  The point being that if I replace the chips, should
I 'program' the new chips with the data archived from the old chips or just
insert the new chips?

 

I have learned a lot over the past several weeks and I really appreciate
everyone's input and education.

 

Clearly, one thing I still need to do is to establish a mechanism to
'communicate and archive' via HPIB.  Had I solved that problem first, I
would have been able to preserve the calibration constants.

 

Thanks again and any thoughts and/or insight are welcome.

 

Joe