Analog: Integrator, Capacitor Accuracy (publ. 2025-03-31)

I added an additional integrator, bring the computer up to four integrators.

I did some measurements of my feedback capacitors. I use two 500 nF capacitors in parallel, for each integrator, the target being 1uF. My measurements (averaging three readings each) showed some capacitors as high as 545 nF, and some as low as 497 nF. Some pairs were, added together, totally as high as 1047 nF whereas some were as low as 1030 nF.

In the end, what I did was mix and match the capacitors so that the high capacitors were paired with low capacitors. In the end, I was seeing an average of around 1010 nF per pair, with a deviation of not more than 1.3% from the average. I think this is good enough. It is important to remember that our feedback capacitors don't necessarily need to be very close to any arbitrary target, so long as they are all close to each other.

I must note, however, that that I was using a cheap TC1 multi-function tester for the readings. Something odd about it is if I take repeated readings of a capacitor without touching the leads, each reading is 10 nF higher than the previous reading. But in the end, all the feedbacks capacitor pairs were giving around 1010 nF on the first reading, so at least the capacitor values seem to be close to each to other.

Recently I was able to borrow, through ILL, a copy of Analog Computing, 2nd Ed., by Dr. Bernd Ulmann. This is a very interesting book which I have been reading at home, and I have hardly been able to put it down. I also have been able to communicate with Dr. Ulmann over e-mail, regarding several questions I had about the Analog Thing and analog computers more generally.

Copyright

CC BY-SA 4.0 Deed