Page 1 of 1

Machinists are programmers, but...

PostPosted: Wed Mar 19, 2014 1:51 pm
by Doug Coulter
Not the way you probably thought. I came across this article on Ars Technica. The old computers used for fire control on battleships were pure mechanical, and there's some darn clever work in them.
The article was good, but the movie made by the navy explaining how the functions worked was better - if you have most of an hour to kill.
Article: http://arstechnica.com/information-tech ... the-waves/

http://www.youtube.com/watch?v=s1i-dnAH9Y4


Wonder if anyone here has (or does) use any of these tricks?

I do know that when I worked for DEC in the very early 70's, that the Naval Academy still used analog computers (electronic) though they had moved to "transistorized" vs the older tube versions by then. (one of our members, Joe Sousa, runs the Philbrick archives) They just do some things better than digital, despite the snark comment on Ars that they could be replaced with an ardiuno. Not hardly, if you know your stuff, it's not even close.
Particularly in speed. To divide by a fixed ratio - two resistors, faster than any digital implementation, and no quantization at all - essentially infinite bitness. And an amplifier (which can have electrically controlled gain) = multiplier, again, can be VERY fast.

Tech is perhaps looking back at this using memristors as the coefficients in neural nets - more processing power/weight/power for NASA, it's been done already.

Re: Machinists are programmers, but...

PostPosted: Wed Mar 19, 2014 8:37 pm
by Jerry
The SR-71 used a analog computer to control the engine. It was a specially lobed spherical cam setup. I cant find the info on it but it was really neat. It was eventually replaced with electronics.

Re: Machinists are programmers, but...

PostPosted: Wed Mar 19, 2014 9:24 pm
by johnf
Doug
Yes the British had something similar called the predictor (an analog computer) that took all sorts of inputs and aimed the guns.
My first antenna rotator used a torque multiplier that was used on the output of one of these to move a bofors gun.
The predictor output went to a selsyn and another selsyn was connected to the torque multiplier on the gun, so oz/in of torque got translated into yard / tonnes of torque to swing the gun around very quickly.
Even the torque multiplier was devilishly clever using a 0.5hp three phase motor geared down to deliver the required torque.
I had a length of 1.5" steam pipe up the centre of my mast with the multiplier connected at the bottom. a Mate ( I say that between clenched teeth) asked what the needle on the shaft of the selsyn was for in my ham shack and before I could stop him he gave it a flick.
My moon bounce array for 2 meters tried to rotate sub second. All that happened ( the antennas only moved a bit) was the steam pipe got two twists in it then broke.
I wish I had never sold that torque multiplier system

Re: Machinists are programmers, but...

PostPosted: Thu Mar 20, 2014 2:33 am
by Nick Strait
if your into oldschool mechanics id recommend looking into the Linotype machine while at it


also maybe the most epic automaton ive ever seen

Re: Machinists are programmers, but...

PostPosted: Thu Mar 20, 2014 4:54 am
by chrismb
In regards machines that mechanically incorporate a series of programmed steps, have you seen Jean-Joseph Merlin's swan, made in 1772/3:


Re: Machinists are programmers, but...

PostPosted: Thu Mar 20, 2014 12:18 pm
by Nick Strait
jean-joseph merlin ^_^ that guy was one of the original oddballs. he also invented the roller skates so he could roll around the gallerys dressed as a bar maid. he also built a motor of some kind that was powered by the shrinking and growing of mercury..