Amazing find! I really think everybody should check out the rest of his website, it's full of gems. I especially like his Telefunken's Bouncing Ball implementation.[1] Ever since I took a dive into hardware, I've been excited by the possibilities of analog computing.
It's great to see fun and stimulating examples that you (might) be able to reproduce yourself. Also, his Analog Computing Bookshelf photo is very cool.[2]
Anybody else on HN have experience / pointers on where to find more stuff like this w.r.t. analog computing?
Various 20MHz secondhand scopes are available on ebay and suitable for beginners. Delivery is usually expensive as they are heavy, the CRT ones even more so. If you want more bandwidth the cost goes up exponentially.
You can get ones that are cheaper and use a PC interface, but they have other limitations and lack a certain authenticity.
I was just talking about this sort of thing with a friend. I really admire the 'all in' analog (or analogue :-)) mode in which he built it. I would be sorely tempted to do it with a DSP chip and an AD9361 :-).
> It probably doesn’t need to be said that figuring out the Fourier components values for each character, and thus the resistor values of the R.O.M. board, was a particularly tedious and laborious task. Each character was first traced out on to graph paper. As a periodic waveform is a closed function, once the character was manually scribed, the curve had to be closed by a segment of a length commensurate in proportion with the time of the blanking interval to the net 50 uS period. This was achieved by measuring the total length of the visible portion of the characters curve with a length of string. The string was then cut down in length to the correct proportion (six eighths visible to two eighths blanked) and then used as an aide to trace out the segment length closing the curve.
This seems like a fun math problem to solve - given some basic character traces, create the appropriation using the given harmonics.
It's the sum of the sin/cos with coefficients. There are two outputs, one for X and one for Y that draw the shape of the number without "raising the pen".
He describes how he calculated the coefficients by drawing the shapes on paper first.
It's great to see fun and stimulating examples that you (might) be able to reproduce yourself. Also, his Analog Computing Bookshelf photo is very cool.[2]
Anybody else on HN have experience / pointers on where to find more stuff like this w.r.t. analog computing?
[1] http://www.glensstuff.com/ballimkasten/ballimkasten.htm
[2] http://www.glensstuff.com/analogbookshelf/analogbookshelf.ht...