analogue pong circuit I found on the internet

[ratstar]

https://kaput.retroarchive.org/pong.html

Heres a video of the guy running his machine.

This is really cool, could some of you guys finish one of these?

As I slowly get more finished with my work, I might have an analogue pong circuit of my own as a side effect halfway through as soon as I work out my actual project (in my wierdo style ) but not quite ready yet.
He finished it in less than a week he said, doesn't look like too many components, why I like analogue style. The part reduction is great.

Pong's collision detections are a few subtractions/comparisons, but you have to carry out the response after the detections, which makes the rest of the circuit I'm guessing.

 

[bertus]

Hello,

Here is an other "analog" pong circuit:
http://www.pong-story.com/eaus0576.htm

Bertus

 

[ratstar]

Hello,

Here is an other "analog" pong circuit:
http://www.pong-story.com/eaus0576.htm

Bertus

Thats pretty cool only 13 ic's!! Pong is great to learn electronics from, I'm definitely going to make one, Ill go for a record low amount of discrete components, then i'll get onto slightly more complicated physics for a walking robot. (But hopefully not too much more.)

A single comparitor works for a 3d floor as well as a 2d wall (To detect collision.), and a robot just has to push himself along the floor with his legs, and thats the main thing, hopefully I can get that to a record low amount of components too. Need to simplify things rigorously! =)

 

[Nanren888]

Thought your background was AI?
Not sure those guys ever obsess about simplifying things. thought they just threw larger processors at thtings.

 

[ratstar]

Heres a bit of work of mine->

Theres alot of bugs there, the new version will be way better.
It has little hiccups every now again because the body of the robot has bugs in its hinges and the physics isn't perfect either.

But the code is suprisingly small!!! All it is, is a physics frame update of a chain of legs that push against the floor. (It suits a differential type circuit.) and it balances against gravity and walks along through iterating it a huge amount of times. The only thing is it could suck up any amount of iterations to get it better at tackling the environment. (It would work against a height map ground as well as a flat plane, but the flat plane is easier.)

So if I get the instructions and put them into a physical machine, (TTL probably would work fine.) Then oscillate it fast enough, if it were a megahert, Id get (Given another trick added to not waste as much search space per time step.) about a second in front. It doesnt seem like much a second, but it is, it probably could handle hurdles if the search was a second.

And thats about as fast as the gpu can do it (its actually a bit faster than the gpu), so why bother, but if I ever got it to a gigahert then it would be a dramatic improvement.

Id love to see this little dude walking on an oscilloscope.

 

[Martaine2005]

Very impressive YouTube channel.
But c’mon dude, we know how you did the video feedback trick. It’s so simple.
You used an oscillator and capacitors in series with the video signal. You can’t fool us.

Martin

 

[Harald Kapp]

@Martaine2005 : you forgot the spark gap that's required for amplification of the tiny motion into long moving legs

 

[ratstar]

I admit it, I dont much try on presentation, I go for the more honest lazy messy house style. hehe.

Im actually running short stroke high speed ESA pumps for the drone, and the actual "muscle" is hydraulic.

And his brain is going to run like a high-speed pong machine on steroids. =)

That little magic stick I made, is made out of milk powder. hehe Cause I cant think of how to make something hard enough for the bot to be made from.

check this out: https://www.instructables.com/Make-Potato-Plastic!/