Breaking News
User Rank
Author
I'm sorry...
7/12/2013 10:18:35 AM
NO RATINGS
I must admit to feeling bad ever since I posted this, because I have to confess to being deliberately misleading.

The reason I used quotes when talking about "input" A and "input" B and "output" Y is that the whole concept of inputs and outputs is a bit nebulous (or perhaps we should say "subjective") when it comes to working with relays.

There, you see, I'm not being too mean... I'm giving you a clue (well, a hint of a sniff of a clue LOL)

User Rank
Author
Claude Shannon explains it all to you
7/12/2013 2:48:53 PM
NO RATINGS
You can get the functionality of the relay with a 2:1 bus switch such as the Fairchild Semi NC7SB3257.  Like the relay, you get bidirectional flow between the common terminal Y and the NC/NO terminals A and B.

I remember playing with relay boxes in high school.  Lots of noise, loads of fun.  I took another look at relay logic when I studied VLSI design, since pass transistors let you do things like the controlling a single light from any number of DPDT switches instead of a bunch of TTL gates.  If anyone wants to cheat, the theory behind relay logic is contained in Claude Shannon's 1937 Master's Thesis A Symbolic Analysis of Relay and Switching Circuits, also published in a 1938 issue of Transactions of the AIEE.

User Rank
Author
Re: Claude Shannon explains it all to you
7/12/2013 3:07:30 PM
NO RATINGS
@betajet: If anyone wants to cheat...

The reasers of my columns aren't like that -- these are the people who scoff ("Ha!") at instruction manuals and prefer to work it out for themselves (no matter HOW long it takes :-)

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
Re: Claude Shannon explains it all to you/XOR Gate with 2 relays
7/15/2013 11:55:24 AM
NO RATINGS
@Weatherbee: Therefore you can realize an XOR gate with a single relay.

I'm going to have to think about this for a bit. One thing I've discovered in my antique relay logic book is that my simple relays wre just th ebeginning -- they also had relays with multiple coils and all sorts of other clever things...

User Rank
Author
Re: I'm sorry...
7/12/2013 4:56:05 PM
I find this relay interesting in the mechanical implemantion.

The relay contacts and arm tension are adjustable. By

taking things apart you can get into the mindset of the

item's designer. My early electronics career was in repair.

I would not trade the design style inspection experience

I got from that era. You see many good and bad practices that

you can improve upon or avoid.

So I say HackAway :-) -Lee Studley

User Rank
Author
Re: I'm sorry...
7/12/2013 5:03:33 PM
@studleylee: I find this relay interesting in the mechanical implemantion.

When I was a student I'm afraid to say that I was more interested in the electronics and I paid little attention to the mechanical side of things.

As the years have gone by, I have developed a much greater appreciation for the way in which things are put together -- and I love looking at a well-made assembly like the one shown in this relay...

User Rank
Author
Balance Relay
8/6/2013 2:34:03 PM
NO RATINGS
the relay I see is a balance relay. The 20 mA current is probably correct because the telephone company uses this as a max current spec on their lines.  Balance relays were used in burgalar alarms for those who could afford the telephone company leased line charge. you had a battery usually 24 or 48 volts subrtract the relay voltage of 5 volts  the rest of the voltage is dropped across the telephone line and series "burden" resistors placed at each end of the telephone line at the central office and the "bank".

Any disturbance in the resistance of the loop causes a change in current and when that change becomes large enough the alarm is tripped.  as long as the burden resistors at each end are large compared to the telephone line resistance normal changes due to climate conditions will not trip the circuit.  Anyone seeking to circumvent the circuit must keep the current constant in the loop so cutting  or shorting the wires is out of the question.  The A and B contacts are adjustable to set sensitivity and the spring is adjustable to set loop current balance point.

User Rank
Author
FYI
7/12/2013 5:21:36 PM
NO RATINGS
Harry Porter's Relay Computer

Awfully new relays, but we don't all have a Mock Electronics handy :-)

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
7/12/2013 6:01:54 PM
NO RATINGS
+(supply)                 -(return)

|---[ ]----[ ]----( )----|

|     A       B      K1    |

|---[ ]---- . . .           |

|    K1  = A & B          |

|---[/]---- . . .           |

|    K1 = /(A & B)       |

|                              |

|---[ ]------------( )---|

|     A    |           K1   |

|---[ ]---|                 |

|     B                       |

|---[ ]---- . . .           |

|    K1   = A + B        |

|---[/]---- . . .           |

|    K1   = /(A + B)    |

Contacts in series implement the "and" function.

Contacts in parallel implement the "or" function.

When relays are energized they close their normally-open contacts (and open their normally-closed contacts).

So, what does this rung do?

|---[/]---------( )---|

/K1          K1

User Rank
Author
7/12/2013 6:44:45 PM
NO RATINGS
So, what does this rung do?

|---[/]---------( )---|

/K1          K1

"Buzz, buzz", as Hamlet would say?

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

LOL

User Rank
Author
7/12/2013 7:16:08 PM
NO RATINGS
You can do a very nice XOR using just two of those Mock Relays.  First a ladder diagram:

`|---[ A]----[/B]--+--( )----|`

`|---[/A]----[ B]--+`

We get conduction if A & !B or !A & B.  Next, we use the fact that the relay can conduct in either direction through the NO and NC connections and use one set of SPDT contacts for both A and !A (connect COM to the terminal on the left) and one set of SPDT contacts for B and !B (connect COM to the relay coil on the right).  Then you just wire the NO/NC contacts of the two relays to each other.  You can swap the NO/NC wires to get XNOR.

This is the same switching you use for an electric light that's controlled from two locations.  If you want more than two locations, you need DPDT switches for the middle switches which require pairs of the Mock Relays.  This gives you the odd or even parity function, depending on how you hook it up.

Gates?  Pfui.  Relay networks are a lot more fun.

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
Relay logic
7/12/2013 8:55:06 PM
NO RATINGS
When I first got my training part of it covered Strowger telephone exchanges.  Which are nothing more than huge collections of relay logic.  They used the old GPO 3000 type relays - probably 2 or 3 times as large as the one you show here Max.  They had specialist tools like tension gauges and what we called "goalposts" - two small pins sticking up from a thin flat bar that were used for precisely bending and adjusting the tension on the contact springs.  Lots of fun, and noisy too when you had a few thousand of them in an exchange.

More recently I have collected some latching relays from some chucked out equipment.  You put 12V on the coil one way and it sets the contacts one way (even when you remove the 12V).  You put the 12V the other way and it pushes the contacts the other way.   I'm not sure what I can do with them, I recall a solar panel charge controller which used one - hence no constant drain on the battery when it was charging, and no diode voltage drop - but I haven't found the diagram for it yet.

I also have some small relays with no less than 8 changeover contacts.  Problem is the coils are 48V.....

Relays in many ways are great things.  If you don't abuse the contacts they last nearly for ever and they are almost totally immune to spikes and surges.    For simple switching they are hard to beat.

BTW if anyone is interested in or can use any of the above, I'll gladly part with them - I work on the principle that if I can see a use for something I'll grab it, and if anyone else can put it to use they can have it.  Works well, except I need a bigger shed :-)

User Rank
Author
Re: Relay logic
7/16/2013 6:23:10 AM
NO RATINGS
Like David I did my early training with relay logic but mine was associated with the signalling control of London Underground trains. We had some amazing relays, everyone designed to do a specfic task in control logic, my favourite were the glass encased Vane relays, easily big enough to use as a gold fish tank.

This is a good link to the kit i used to work with

http://www.wbsframe.mste.co.uk/public/Relays.html

User Rank
Author
Re: Relay logic
7/16/2013 9:09:07 AM
NO RATINGS
1 saves
Lovely stuff, Crusty.

User Rank
Author
Re: Relay logic
7/18/2013 9:00:47 AM
NO RATINGS
@Crusty: I wish I could have seen (and heard) those relay-based cabinets working.

User Rank
Author
Re: Relay logic
7/16/2013 8:52:12 PM
NO RATINGS
Those exchanges are what i thought of when I read the article.  I got to look into one once on a stint with Telstra during summer holidays and listening to the racks of relays handling calls (decadic dialling, terminations) was like a strange kind of music.  I imagine they've all been replace with digital technology now - wonder what was the fate of all those relays?

User Rank
Author
Re: Relay logic
7/18/2013 9:05:34 AM
NO RATINGS
@SteveD_Aus :...wonder what was the fate of all those relays?

Sad to sad, I bet the yended up as scrap -- I know I'm finding it exceedingly difficult to find more copies of the Kurman 223C34 shown in my blog above. I really want to lay my hands on a couple of hundred at least.

I know I could opt for modern versions ... but that wouldn't be quite the same (sad face)

User Rank
Author
Re: Relay logic
7/19/2013 5:45:29 AM
NO RATINGS
@Max....I have possibly 30 or 40 relays, with DPDT (ie 2 changeovers) BUT they have 48 volt coils.  They are not as old looking as yours above, standard looking relay with a transparent case.   If they'd be of interest I'll root them out and give you more details.  I might be able to find a 48V power supply as well (oh, hang on, that'd be for our 240V AC mains.... :-(  )

@SteveDAus...most of those nice relays are probably slowl leaching heavy metals into our water table, in a landfill somewhere....

Strowger exhanges were lovely to listen to,  The guys who worked in them could pick up sticky relays and malfunctioning selectors just by the sound.    When I lived in a very small town in Zimbabwe about 11 years ago, we still had a 1000 line Strowger exhange for our town.  I told the techs I'd trained on that stuff and they let me look around.  Nostalgia par excellence....  you had to remember to put ATDP commands in your modem strings (for pulse dialling) or nothing happened...

User Rank
Author
Re: Relay logic
7/19/2013 12:14:56 PM
NO RATINGS
@David: BUT they have 48 volt coils.  They are not as old looking as yours above, standard looking relay with a transparent case.

Thansk so much for the offer David, but given the choice I prefer to work with 5V relays (or 12V at a push).

Brian LaGrave found a mention of someone who had 26 more lik emine, but they wanted \$100 a piece for them!!!

I think I'm going to have to go for modern relays in plastic cases -- but I will mount my old one in the center
.

Let's not make any decisions yet -- look for my blogs here on EE Times and also my Saturday Morning blogs on All Programmable Planet for the next 10 days or so whil eI think this out more carefully.

User Rank
Author
Re: Relay logic
7/19/2013 7:53:35 PM
NO RATINGS
@Max "I prefer to work with 5V relays (or 12V at a push)."

Don't we all.....I think I will chuck them eventually, no use to man or beast.

You're certainly ambitious....I was thinking about how to do (say) a full adder with relays....or maybe a 4-digit counter....but it rapidly gets bigger than Ben Hur ( almost literally....)  thank god for TTL and CMOS....

User Rank
Author
Re: Relay logic
7/22/2013 9:06:53 AM
NO RATINGS
@David: I was thinking about how to do (say) a full adder with relays....or maybe a 4-digit counter....

I must admit that I'm starting to rein myself in -- I don;t want to spend years doing this -- I just want to get a relay "thing" that's doing something vaguely useful and clicking away while doing it -- I'll be asking for suggestions in a future blog -- one think might be generating the value of Pi to an infinite number of digits (well, a large number) -- maybe outputting the values on a Teletype?

User Rank
Author
Re: Relay logic
8/29/2013 12:17:45 PM
NO RATINGS
If you really want to do logic with relays, you'll be happier making mux-based logic. You can think of a SPDT relay as a mux with the coil as the control input, the common as the output and each throw as an input. From that you can build all your common logic gates. I was writing about that a few months ago: http://www.drdobbs.com/embedded-systems/making-contacts/240152060

This is much denser than just using parallel and series switching circuits.

As for the adder circuit in relays, that was my April 1st project this year: http://www.youtube.com/watch?v=3n6WVXx4MV0

It does work.

User Rank
Author
Re: Relay logic
8/29/2013 12:22:32 PM
NO RATINGS
@wd5gnr: If you really want to do logic with relays, you'll be happier making mux-based logic.

Thank you so much for sharing this -- I have been thinking along these lines a little, but it's difficult when you are on your own. I will be reading your articles as soon as I get a free moment.

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: 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).

User Rank
Author
Re: Relay logic
7/18/2013 3:12:38 PM
NO RATINGS
Max

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).

Althoug not a D-Type we used to make a latching relay as follows: Let's presume a 24VDC coil and I'll polarize it, just to simplify the explanation. You will also need a 2 pole relay normally open, or changeover contacts- pole one is for general use. Connect the coil positive to 24Vdc and the negative end is what's used to activate the relay when connected to 0V. Connect C2 (common of pole 2) to 0V and NO2 (not laughing gas)  to the negative of the coil. When the coil is activated, the closed contact connects 0V to the coil so that even if the external activation is removed the relay remains energised.

The only way to de-activate the relay is to somehow open circuit the power loop to the coil.

User Rank
Author
Re: Relay logic
7/18/2013 3:17:09 PM
NO RATINGS
@Antedeluvian: Although not a D-Type we used to make a latching relay as follows...

That's pretty much what I was thinking -- look for a follow-up blog in a week or so where I intend to follow up with more details...

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: Relays as delay lines
7/15/2013 11:58:25 AM
NO RATINGS
@Ron: I think I am now able to tell this story without some legal beagle chasing me...

Don't look behind you :-)

Thanks for sharing this very interesting story -- I wish we could all get together one day -- sitting round a table quaffing a few cold beers and sharing tales of things we'd seen and done...

User Rank
Author
Why Positive Earth?
7/15/2013 12:03:17 PM
NO RATINGS
@All: In my antique relay book, which was written by folks at AT&T, they are using a positive ground. For example, if you look at the diagram for the coil on a simple relay, the left side of the coil is connected to one side of a switch (could be a mexhanical switch, could be another relay); the other side of the switch is connected to ground.

Meanwhile, the right-hand side of the coil is connected to th enegative terminal on a battery symbol (which they are using to represent a power supply) and the positive battery terminal is connected to ground.

So, my question is -- was there any underlying reason why they should have a positive ground (i.e., the positive battery / supply terminal connected to ground)?

User Rank
Author
Re: Why Positive Earth?
7/15/2013 2:13:12 PM
NO RATINGS
We're talking AT&T or the phone company, back to the days of landlines, the CO battreies were always -48 volts and I believe the reason has to do with there's less corrosion on the ground rod (that actually goes into the soil and conducts real DC current) that way, something about electrochemistry and avoiding corrosion of the electrode, I can't provide details but they're probably not all that important.

User Rank
Author
Re: Why Positive Earth?
7/15/2013 2:23:52 PM
NO RATINGS
@JeffL_2: We're talking AT&T or the phone company, back to the days of landlines, the CO battreies were always -48 volts and I believe the reason has to do with there's less corrosion on the ground rod (that actually goes into the soil and conducts real DC current) that way

That makes a lot of sense -- OK, so now I won't feel bad about using a negative earth in my implementation (because that's the way I think :-)

User Rank
Author
Re: Why Positive Earth?
7/15/2013 2:52:00 PM
NO RATINGS
From Wikipedia: Tip and Ring:
Originally, the potentials on the wires were positive with respect to earth (ground). This is called negative ground, since the negative side of the battery is grounded to earth. Telephone companies discovered that with positive voltage on the copper wires, copper wires experienced corrosion due to electrolysis. Operating in reverse, positive ground (negative voltage on the wires), the copper is protected from corrosion, a process called cathodic protection.

User Rank
Author
Re: Why Positive Earth?
7/15/2013 3:14:01 PM
NO RATINGS
Oh yeah. Well I had the "corrosion" and "polarity" parts OK. Couldn't make the concept work in my head because there's always two wires in a pair, would just shift issue to the other wire, but here they're talking about all the UNPAIRED wires inside the CO, have to protect THEM, the pairs are mostly "customer premises" so if the customer loses service as a result he'll pay for a service call anyway. Gotta think like a telco!

User Rank
Author
Re: Why Positive Earth?
7/18/2013 8:59:19 AM
NO RATINGS
@Betajet: It's funny how we get locked into certyain ways of thinking. Based on my days working with TTL, I prefer to have a +ve supply and ground.

Now I understand why my AT&T books shows positive ground -- and I also understand that there's no reason I shouldn't use a negative ground for my relay-based experiments.

User Rank
Author
Re: Why Positive Earth?
8/1/2013 5:24:30 PM
NO RATINGS
There were two types of open wire telephone lines in the '60's and 70's.Steel wire was used for local service,but long distance required copper wires.

Most poles for long distance had 10 crossarms, with 10 wires on each, and a transposition bracket every so often(can't remember exact spacing intervals).The transpositon brackets prevented noise and induction from the earth's magnetic field.

These wires ran for hundreds of miles,from office to office across the country.

The wires were left up long after they were obsoleted, then someone realized that there were millions of dollars worth of copper on the poles,and started having them removed.If they had stored the wire until today, it would have been a better investment than just about anything else.

The power of hindsight!

User Rank
Author
Relay logic
7/16/2013 8:04:38 PM
NO RATINGS
In the Telco of the 1960's era,dialtone was 53 volts.When dial tone was broekn by picking up the receiver and dialing,you had talking battery,which was 48  volts.When the phone rang, it was 96 volts pulsating DC.The ringer had a capacitor in seres that blocked the pure DC.

The Central Office had thousands of stepping relays,that stepped one step for every number dialed on the rotary phone.After a standard delay, it moved the to the next relay.This repeated through the whole number sequence.

Most service was party line service of 10 people or more sharing the same line,especially in the rural areas.

About 1970, the FCC mandated a maximum of 4 parties on one line,so a lot of infrstructure had to be upgraded.Open wire was replaced with buried wire,and eventually,2party lines became available,and now private lines are the standard.

Fiber optic has replace most copper now,and the "central office" is now mounted on a single circuit board in the field.

Cards called DSLAM (Digital Subscriber Line Analog Multiplexer) cards convert the analog voice into digital and pump it down a fiber cable.

Any logic that can be performed with gates can be performed with relays, and was done that way first.

Series circuits are AND circuit,Parallel circuits are OR circuit,and holding(interlocking)  contacts are memory bits.

For NAND and NOR simply use the opposite contact.

You can make anything with these.

However, I prefer modern solid state logic.

It is so much eaiser to reprogram than to move wires around physically,as in PLCs.

Much easier to modify any part of the circuit or time delay or addressing.

Addresses had to be hard wired with wire-wrap terminals.What a pain.Especially if you had to change them.

Anyway, enough nostalgia for now.It is time for my toddy and nap.

User Rank
Author
Re: Relay logic
7/18/2013 9:02:45 AM
NO RATINGS
@DrZuhoch: I prefer modern solid state logic.

Well, me too for real work, but there's something about the sound of a relay-based computer chattering away in the background that takes one back in time...

User Rank
Author
Reduction & Relay Computers
7/18/2013 10:55:51 AM
NO RATINGS
I am quite familiar with relay ladder diagrams. My first job with an industrial robot and automation manufacturer used PLCs for control, but also used a lot of relays. I remember discussing this with my boss. I thought it would make more sense to use TTL for a lot of things like safety interlocks. He correctly pointed out that that would require power supplies, and special mounting cabinets. It was always cheaper to use relays (at least in the 70s). Then there was the problem of fault modes. The relays tend to fail in know specific ways. With relays that have been tested/approved for safety interlock applications you can have some assurance that they will work in a safety application.

Another interesting thing about relay logic is that the combinatorial reduction for it is quite different (as others have noted) because unlike TTL or other logic families, there is no input & output. Current can flow in both directions. If you look at old relay designs that used complex circuits with hundreds of relays, they used the formal mathematical reduction techniques to eliminate large expensive relays. The resulting schematic seems impossible to figure out. It is not the simple contacts in series are AND, and in parallel are OR circuits. Kind of like code without comments.

I have an ancient Friden Flexowriter teletype with paper tape reader/punch that it took me weeks to repair when a contact got bent. It has complicated logic to do parity checking and all kinds of other control stuff. Even with schematics, it was very difficult to figure out the 90VDC relay logic.

Then there was the Friden Computyper that the teletype was connected to. It was a completely mechanical computer. It used many hundreds of relays as well as stepping relays, and a mechanical adding machine (the ALU) that used solenoids to press the buttons, and had means of reading out the output digits electrically. This particular machine was used in the 50s & 60s to calculate payroll and print paychecks for the USAF.

User Rank
Author
Hi voltage detector
7/18/2013 3:17:18 PM
NO RATINGS
On several occasions I have seen requests on different forums for a simple way to detect whether mains voltage is present in a system. There are always the suggestions of resistors and caps and diodes and optocouplers in all sorts of configurations.

The simplest way- a relay with a coil rated to the mains voltage. Of course there must be some consideration paid to the wetting current through the contacts, but still...

User Rank
Author
Paperclip Computer
7/18/2013 6:25:51 PM
NO RATINGS
Even more primitive than a relay computer is the educational Paperclip computer, introduced by  the book "How To Build a Working Digital Computer" by Alcosser, Phillips, and Wolk.   See http://archive.org/details/howtobuildaworkingdigitalcomputer_jun67/

User Rank
Author
Re: Paperclip Computer
7/19/2013 12:00:57 PM
NO RATINGS
@grnewell: ...the educational Paperclip computer...

Tell me more!