MM> G'morning Mike,

MM>  MR> "Miles Maxted" bravely wrote to "Mike
Ross" (13 Jul 04  05:09:00)
MM>  MR>  --- on the heady topic of "Re: Relative Humidity"

MM>  MM> radar receiver bits seem to be likely suspects !

MM> (And or proximity fuses for anti-aircraft shells)

Those were a natural development of VT fuzes already used to 
burst bombardment shells above ground for shrapnel effect.

MM>  MR> Silicon junctions have a natural log function just the right thing for
MM>  MR> an analog calculator built up using differential operational circuits.
MM>  MR> These were already described in the 20's and would have been a natural
MM>  MR> in the 30's for extracting logs then adding and subracting voltages to
MM>  MR> derive multiplication, division, squares, roots, and some trig
MM>  MR> functions. It would have been complex but doable.

MM> Mmm ... they developed an analog `differential analyser' at Rugby
MM> using Meccano - the British Boy's Own constructor kit of the day
MM> (and still in business, I discovered t'other day).  There's a
MM> working variation in the Auckland Museum of Transport and
MM> Technology, coincidently...

One of the WW2 AA radar sets included a totally mechanical 
analog predictor computer made of wheels and rotating rods.
The radar fed in several positions for the target and the computer 
derived height, speed and direction, then computed the aiming point 
for AA guns to put their shells where the target was *going to be* 
by the time the shells got all the way up there.
Could give height commands to the fuze setter on guns not using 
proximity fuzes.
It was quite accurate, but needed to be run in a temperature controlled 
oil bath to avoid errors from thermal changes in the size of the 
precision wheels.

MM>  MR> I recall wiring up
MM>  MR> an analog computer in my early tech classes. This used IC opamps and
MM>  MR> wasn't very precise but still quite good when compared to the
MM>  MR> alternative then of a room size digital computer or worse nothing.

MM>  Heh...  my '59 Psych Dept Calculating Room was equipped with
MM> electromechanical Marchant and Munro calculating machines.  These
MM> could hunt down square roots - given half an hour for the
MM> mechanicals to thrash and spasm the desk around the room - great
MM> fun.

We had Facit ones - also quite acrobatic on occasion.

MM> There was a special machine which could accept raw numbers in a
MM> cross-tabulation before throwing a hissy fit that spat out a
MM> correlation coefficient.... only senior and graduate students
MM> got to use it...

MM> Ah me ... took a whole term to derive the principal components of
MM> a 10 item table of correlations,  doing everything by hand and
MM> mechanical adding machine as a fulltime student  ...   and now, on
MM> this old Pentium, I routinely do a much fuller multivariate
MM> analysis of 3 or 400 variables in seconds....

Only after Pentiums had that FP error corrected!



Terry V.
---
 þ MM 1.1 #0367 þ JUDGE SLAMS BID TO DEPORT ORANGE MAN


--- QScan/PCB v1.20a / 01-0462