Breaking News
<<   <   Page 2 / 5   >   >>
User Rank
Author
Relay logic
7/13/2013 2:19:54 PM
NO RATINGS
It would make a nice demo to use relays to show combinatorial logic but I think it would be too hard to take it much beyond that. The hard part only starts when you try to use standard relays to make an element capable of toggling ie the basic component of counting, at least so long as you don't cheat and use special mechanics like alternate-action latching relays! I suppose (in theory at least) you could use four relays to implement something like a master-slave flip-flop but I can't recall having seen it done, maybe you could use fewer if you allow some "non-relay" components. Then you'd just have an asynchronous counting stage, to make it synchronous and running on a master clock would be quite a challenge, can you imagine compensating for contact bounce and unequal closure delays? No thanks, if you want to carry it to that extreme I'll watch, you design, thank you very much!

User Rank
Author
Re: Claude Shannon explains it all to you/XOR Gate with 2 relays
7/13/2013 4:13:00 PM
NO RATINGS
Well your observation with the NC7SB3257 was close but not quite right.   Actually, given the description of the relay if the coil operates at the logic voltage then it would be equivalent to a 2:1 multiplexer like you mention *plus* an XOR gate in front of the 2:1 multiplexer control pin.  This is because the relay coil is polarity insensitive, so as long as the inputs to each leg are opposite the magnetic field will move the Y contact from A to B.  Therefore you can realize an XOR gate with a single relay.

C1 = Input 1, C2 = Input 2, Y = Output, A = GND, B = +5V

Note that the above actually indicates that you can get even more then an XOR out of a single relay, since in the above case A and B are statically connected

In comparison to non-tristating gate logic (assuming all contacts must have a valid logic level at all times)

The relay connections C1, C2, A, and B are inputs while Y is an output.   So there are a number of 4-input, 1-output functions that can be realized.   If you allow passives on A, B and Y then there could be some 2 output functions additionally.

User Rank
Author
Mock Relay Truth Table
7/13/2013 5:00:41 PM
NO RATINGS
Some quick C code to generate a truth table for the Mock relay:

#include <stdio.h>
#include <stdlib.h>

int main (int argc, char **argv)
{
unsigned int C1, C2, A, B, Y;

printf ("C1 C2 A  B  | Y\n"
"----------------\n");
for (C1=0;C1<=1;C1++)
for (C2=0;C2<=1;C2++)
for (A=0;A<=1;A++)
for (B=0;B<=1;B++)
{
if (C1 != C2) Y = B;
else Y = A;
printf ("%d  %d  %d  %d  | %d\n", C1, C2, A, B, Y);
}
exit(0);
}

User Rank
Author
Re: Mock Relay Truth Table
7/13/2013 5:03:36 PM
NO RATINGS
And the truth table:

C1 C2 A  B  | Y
----------------
0  0  0  0  | 0
0  0  0  1  | 0
0  0  1  0  | 1
0  0  1  1  | 1
0  1  0  0  | 0
0  1  0  1  | 1
0  1  1  0  | 0
0  1  1  1  | 1
1  0  0  0  | 0
1  0  0  1  | 1
1  0  1  0  | 0
1  0  1  1  | 1
1  1  0  0  | 0
1  1  0  1  | 0
1  1  1  0  | 1
1  1  1  1  | 1

User Rank
Author
Relays as delay lines
7/14/2013 5:24:28 AM
NO RATINGS
Ah relays, delay lines and equivalent circuits. I think I am now able to tell this story without some legal beagle chasing me. Long, long ago I worked on a very simple stock control computer. It used a Sperry drum as the magenetic memory, an arithmetic unit with electronic adders that employed wonderful arithmetic features like end-around-carry for subtraction, but I digress.

The output printer was a Creed Teleprinter. A delay line with a large number of miilisecs delay was urgently needed, it was required for synchronization purposes (I suppose nowadays we would call it clock domain transition) between the computer and the printer. Miles of coax might have done the trick or some big capacitors and inductors. At that time I think IBM had developed a wire relay that was the fastest available. Four in series, the contacts of one driving the coil of the next, provided the required repeatable delay. What appeared on the documentation? A box with the conventional inductor-capacitor delay line schematic with annotation please contact the manufacturer for further details. Why would I (we) do such a thing? Because the sales department had promised delivery tomorrow, when there was still six months of development required, Situation Normal All...etc.

User Rank
Author
Re: FYI
7/15/2013 11:39:41 AM
NO RATINGS
@Betajet: I saw Harry's relay computer on the Internet quite a few years ago and ended up in an email conversation with him. It really is a "tasty" piece of work, even if he did end up using a silicon chip for the memory :-)

User Rank
Author
7/15/2013 11:41:49 AM
NO RATINGS
@betajet: "Buzz, buzz", as Hamlet would say?

LOL

User Rank
Author
7/15/2013 11:45:27 AM
NO RATINGS
@betajet: You can do a very nice XOR using just two of those Mock Relays.

Very interesting -- I was wondering about an XOR, but I was leaving that for later because I've been plunging into my old relay logic book.

The strange thing is how hard I find it to wrap my brain around ladder diagrams -- I never used them myself. At the moment I find it easier to actually think of the relays as a circuit -- but I can see that ladder diagrams would be much more useful -- I need to look for a simple tutorial.

User Rank
Author
7/15/2013 11:48:07 AM
NO RATINGS
@betajet: Speaking of my old relay book -- this was written at a time when Boolean algebra was not widely known, so they sort of make up their own notation, which can make your eyes water.

They also were talking about "sneak paths" but their explanations were really confusing -- then they showed one example and I thought "Ha, so THAT'S what they mean!"

User Rank
Author
Re: Relay logic
7/15/2013 11:52:28 AM
NO RATINGS
@JeffL_2: It would make a nice demo to use relays to show combinatorial logic but I think it would be too hard to take it much beyond that.

If I can lay my hands on a bunch of them, I'd certainly start with just a bunch of regular switches on the left feeding a bunch of relays wired in a combinational manner -- plus have loads of miniature incandecent bulbs scattered around so you can see everything operating.

But I would have to move up to sequential logic. I actually think you could make a D-type latch with a clear input using just two relays (but I'll have to think about thsi some more).

<<   <   Page 2 / 5   >   >>