Thank you for posting these videos. I love your explanations. I am still amazed that memory was stored as pulses in a tube (and in wires like I have seen in other videos). Fascinating.
Wonderful project and all can be congratulated on collaboration to stitch the EDSAC together. As a 64 yo retired electronics engineer here in Australia who worked on valve equipment I really appreciate the issues that valves bring to a project like this. Love the current day technology being used and deployed in resolving some limitations of 1940’s technology. So important to preserve and record for history sake something only our generation understand. This will stand the test of time albeit limited to the eventual lack of spare parts, I believe 20 years was mentioned in one of the videos al, of which having just discovered your You Tube channel I am in the process of watch all relating to the EDSAC project. Is there a completed project video? Seems from around 2018 your excellent update video ceased? Have I missed something? Be very interested in where the project is at? What a wonderful project created by like minded electronic veterans who value what has been in the past.
Thanks for your appreciation. The EDSAC project is still in progress. Please take a look at our RU-vid channel, there is a playlist for EDSAC. The latest video as March 2020 shows work on the delay line storage: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-9BA4AyvlKnM.html
6:06 "It has worked in the past with recent changes we've made it has stopped working and without it working main control can't fetch instructions and data." Sounds like what happens to me when I change code without committing the previous working version, thinking I can just use the undo history to revert. It never works out that way.
You can see why, as soon as we could, we went for parallel architectures instead. ... project manager asking that perennial project manager's question "What's that?.. An hour's work?.. two hour's work?" I find it kinda amusing how the old war-time, national-service thinking still touched on EDSAC... how there are no "instructions", instead they're "orders" ;) ... and... what's that fascinating wooden doofer with the two big dials on it in the corner of the meeting room?????
You are going to have trouble synchronizing your clock to the delay memory, the delay and the clock are going to change with temperature and age. What if you used one of the memory delay tanks as a sync pulse and then a phase locked loop to multiply that pulse by 576?
That's an interesting idea that hasn't been though of so far AFAIK. At last week's progress meeting we were talking about getting better stability in the main clock by using positive or negative TC caps. Trouble is, it always ends up as a discussion about authenticity - the attempt to not change things from the original circuits unless there is no other way, so I doubt that your PLL method would be allowed. I have suggested synchronising the main clock from injection of pulses from a programmable oscillator, and I guess that we could measure temperature and tweak the oscillator frequency to track the requirements of the tanks. That way, the main clock would still be the original circuit with just the addition of a small capacitor for injection of sync pulses.
We have a very few original parts of EDSAC Mk.1 at the National Museum of Computing, Bletchley Park, UK. Unfortunately, the great majority of EDSAC Mk.1 was scrapped when its use was superseded by EDSAC Mk.2 The Science Museum in London also has one EDSAC Mk.1 chassis on display.
