My "craziest"analog design (1995 or so)

I LOVE the name of your company !

Thanks Bob. I made it up way back, during the communistic dark ages, when thinking of a private company was more of a diagnosis than anything else. Obviously I was putting in also some laugh calling it so, I was quite young (20-s) and way way less knowledgeable than it took to think of a company. But being the stubborn bastard I am I stuck to the name.

Looks cool. What does it do?

It is the acquisition module for an MCA (Multi Channel Analyzer), used for high-resolution gamma spectroscopy. Done analog - the last one I did so before going digital. The current model:

Thanks for sharing. If I understand right the motorized multipole switch is to select different low pass filters - just wondering if you rejected switched capacitor tunable filter chips?

piglet

At the bandwidth of the input signal (rise time starting well below

There was a more interesting twist with the variable gain cells. In my previous (first of the kind) design I had used just relays; here I used JFET pairs, one half as a voltage controlled resistor, the other as temp compensation (how I got to select/make the pairs is another longish story :). It worked reasonably well (though I used relays again in my next design....) but introduced a lot more integral nonlinearity than was acceptable (about 1%, I needed something like 0.025). So what I did was acquire a somewhat longer (say 8.5k) spectrum "raw", nonlinearity bad and converted that live (5-10 times a second) into an 8192 channel spectrum with linearity good (very good actually). Since the linearity error would differ from one FET to another we had to take calibration measurements and store some coefficients into the module's HC11 eeprom... Then the live transformation was done by a 20 MHz 68340, was anything but trivial to make. Mad as it was it worked :-).

Dimiter

....................................................... Dimiter Popoff, TGI