Author Options:

Diode Calculator will not work Answered

I've been working on a diode only (well, plus switches for input [numbers, yet there's only one pad! not one for each digit], resistors for various things, and LEDs for the display) calculator that could add two two digit numbers (xx+xx=xxx) for a while, putting countless hours into it.

I have deemed it impossible.

There's so many problems with a mass diode resistor network that is just won't work. As some and gates conduct then some other gate (like it's suppose to) conducts, and eventually causes a voltage divder (with the pull up/ pull down resistors) and the output gets wacky and there's no wway of amplifying the signal.

The diode adder is dead


You can do this very easily if you incorporate transistors into your design. Use them to create "and", "or", & "not" gates. That is what you need to do, then work on combining those gates to build half-adder and full adder circuits. Then hook up input and output, and you're done! In the old days, this was called DTL or diode-transistor logic. Worked then. Works now.

the whole point of the thing was to show that you can create a smart machine without any "smart parts" such as transistors and other switching items.

but the TOD explanation clearly indicates that only diodes and transistors were the only allowed semiconductors

Well transistors are sort of diode pairs. But it's really old news that you can't cascade diode logic gates, so I thought I'd "help" with my suggestion. That's settled engineering. Either that or you make a 10000 input diode ROM. Up to you.

Post circuit. If someone here can find the problem and make it work, then you'll feel better because your concept was good. If no one can fix it, then you'll feel better since not only were you right about it being impossible but there's no one here so much smarter than you they made it work. Win - win! Post circuit now. You'll be sorry if you don't.

well its several pages of work... And my scanner wings scan it good since I made it on graph paper... If you are wondering why I did it on paper its because I can do it faster on paper and there's a ton of changes on every page Maybe last night I was just getting frusterated... I'll work more on it today with a new design.

If you are wondering why I did it on paper...

The absolute last thing you have to do is explain to an old(er) geek like me the advantages of using paper.


I'm curious how you went about designing it. It sounds like a really cool project, but I'm not sure a look up table is the best way to implement an adder. A binary adder should be a bit simpler and have a lower component count. I know you can build an adder from and/or gates so it should be possible make one out of just diodes. I have some experience in digital logic design and would be happy to take a look at your design and help figure it out. Good luck

Defiantly an interesting problem. Not being able to cascade logic gates makes things a lot more difficult. Have you tried making the resistors different, (e.g. 10k,1k) with higher value on the second gate so that the voltage measured across the second resistor is above the threshold? I haven't tested it yet, but it seems promising. I'm not sure if it would scale up or create any other problems. I hope this helps.

Yes, It still creates a voltage divider, but because the voltage is proportional to (R1/(R1+R2)), making the resistor that the output is measured across much larger than the resistor in the AND gate should make the high level output greater than the logic high threshold. I tested it with some LEDs (I couldn't find signal diodes) and it seemed to work. I'll do an analysis of the circuit and try to figure out if it works, and if it can be scaled up.

oh And if this works I'm expecting it to have a couple thousand diodes, it'll be fun to solder the first time since no breadboard in the world could do it I'll probably use a cub design. I'll make a copper frame of a cube, and make a smaller one inside it. Thisll be power and ground. Inside it'll be a matrix of diodes and vertical and horizontal copper bars

scratch that, more than a 100000 diodes unless I find a different system

one last thing trcnically it wouldn't be called an adder, more of a memory bank. It couldn't add 12+72, is would more so remember that 12+72=84

The outcome would be the same, though