🚀 go-pugleaf

Bruce Dunn wonders if my outside temperarute effect could be due to line 
voltage variations.  Don't think so Bruce.  There was a table in the early
version of the Como paper that got cut out of the final version due to the 
space limitations.  It listed some of the control measurements taken over
a three day period of operation, I repeat some of them below:

Variable                                   Mean           Sigma
+5 Volt Reference                         5.0008        .00022
ADC Reference                             4.5299        .000081
+15 Supply                               14.9590        .00072
High Quality Ground                       0.00049       .00012
Cell Current                             -0.5557        .0011
Current Sum                               0.0017        .00068

In general while running I watch 60 odd such variables.  Any that look 
suspicious I can plot, and I store the last 3 1/2 hours so that the recent
history of anything funny can be brought up.  I usually plot about 10, which
is about the limit of colors that can be distinguished.  

The ADC reference has been through three layers of regulator before it is 
used.  The last two in a temperature controlled environment.  So there are 
lots of checks.  But never enough.  The Mark II version suffered from some 
cleverness on my part that is too complicated to explain, but probably was the
source of the reported heat pulses.  This was partially fixed in the Mark II/
V, and will be fully fixed in the Mark III.

So this remains a tough business.  Simple open cells, i.e. Mills, have large 
error bars.  Complex designs like my kludge can have complex interactions and 
so tale long and careful study.

Tom Droege