8 switches

well i can see how my post could be seen as annoying, im designin a device that according to my research has not been created yet and i'd like to design and test it before revealing its purpose.

it's not a question from a book or test. however the design could be seen as an educational tool. i'll be creating a webpage if the design is successful and beneficial to students, i can see it being very beneficial and naturally forcing students to use good form and technique.

it won't take long to construct but the materials used must be very light to keep the instruction realistic and non-fatigueing, and this will be expensive as custom made machined parts will be necessary. can you guess what it is ?

i'll post and graph student's progress and results including my own when this project is complete. hopefully the prototype will immediately eliminate bad habits. any help on the electronics side would be appreciated. im thinking relays or some kind of boolean logic chip :)

i feel bad not writing a more descriptive requirement .. thanks for your input john, ill get back to this post in a few days when i have time and can provide more details

Why are you circumspect about your intended purpose? What is the purpose? Is this a question from a book or test? Or is this for a game of some kind?

Jon

It's difficult for me to suggest an approach. I've no idea if relays are required, at all, from the description. Maybe, or maybe not. What is being driven? How must the resulting switch function be attached? What are the voltage levels that the switch function must handle? How precisely must be all of the timing of observing the input switches and developing the control output? Must the output be held for some duration even if a second switch is then operated? I don't know what is acceptable and what isn't to you. Just way, way too vague.

Sometimes, this is helped a great deal by knowing the application. If it is something within our reasoning and experience, we can fill in a lot of the gaps by common sense and don't have to beg for information. But perhaps, if you won't talk about it directly, we can play a little game and find out indirectly what you'd consider and not consider by just shooting in the dark, blindly tossing darts so to speak, until you clue us all in a tiny bit at a time. A kind of 20 questions?

I'm good for one go at it.

For me, the simple approach would be to use a 16-pin PIC, scan the input sources and debounce them in software, test them as per your spec, and drive _something_ from one of the pins on the basis of that result. But then, I program such things a lot.

There's my tossed dart. You tell me where it hit.

Jon

Start out by making a truth table of what you want.

Well, you probably could use 9 relays, one being the master normally closed relay for the master circuit, and the other eight normally open operated by the switches. Connect the eight NO relay conatacts to a power supply and a resistor, and parallel the outputs to coil of the master relay. Size the resistors such that current thorugh only one isn't enough to operate the master relay, but current through two or more in parallel is sufficient to operate the master coil, opening the master circuit.

How about an analog approach. Tie each switch to a positive supply, and put an equal resistor in series with each. Parallel all resistor switch pairs and connect to zero volts through a 9th resistor equal to about 70% of one of the other 8. Something like 8 of 10k resistors and 1 of 6.8k. Then use a window comparator to sense the voltage across the 9th (lower value) resistor is higher than what you get with one switch closed, but less than what you get with 2 or more switches closed. The reference voltages for this comparator could be a divider made of 3 resistors across the same supply that powers the switches with series resistors. Something like supply to 10k to 3k to 2k to ground.

A good dual comparator for this job might be an LM393.

Then you can energize a relay or whatever with the window comparator output.

Total parts count 12 resistors and a dual comparator, and what ever you need to drive the relay.

Sure! :

Vcc Vcc | | | [4K7] [0.499K] | | +----[1M]-+ | | | 74HC86 +---+---|+\\ | +-----+ | | >--+-------+--|A1 | | +-|-/ | | Y1|--+ [2.49K] | Vcc +--|--|B1 | | | | | | | | | | | | [4K7] | +--|A2 | | | | | | | | Y2|--+ Vcc | +-|---[1M]-+ +--|--|B2 | | | | | | | | | | | | [1000] +---+-|-|+\\ | | +--|A3 | | | | | | >---+---+ | | Y3|--+ +-----//----+------|-----+-|-/ +--|--|B3 | | | | | LM393 | | | | | [1000] [1000] | | +--|A4 | | | | [2.00K] | | Y4|--+ D Q1 D Q8 | +---- |B4 | | IN1>---G 2N7000 +--G 2N7000| +-----+ | S | S | | | | | | +-----------------------------+ | | | | | GND>----+-------|---+------+ | | | VCC VCC IN2>---TO Q2-G | | | | | | [1N4001] COIL- -| IN3>---TO Q3-G | [100R] | | O->| ---TO Q4-G | | | | O K1 | | C | | IN5>---TO Q5-G | +-----B | +-----COM | E | IN6>---TO Q6-G | | +--------NC | GND IN7>---TO Q7-G | | IN8>------------+

Typical for all inputs, a SPST NO switch closing to Vcc:

Vcc--O--> | | O | +--->to Qn gate | [10K] | gnd

With no switch or with more than one swith closed, the circuit between the common and the normally closed contacts of K1 will be open. With any single switch closed, the relay will make the contact between COM and NO.

Need a circuit description?

--
John Fields
Professional Circuit Designer

these are all awesome ideas ! thanks everyone ..

which one these designs would be

1) most compact 2) the lightest 3) the cheapest to manufacture in small (50) quantities ?

can someone advise on a better design ? ==================================== separate has 'a rat' in it

John's is an excellent suggestion. You can't get much smaller than 1 easy-to-find IC and 12 resistors.

That doesn't include the output transistor and the relay/diode, of course.

Matt: The other neat thing about John's suggestion is that, if you decide you need more than 8 switches, all you have to add is one resistor for each new switch.

Mine.

--- Looking over John Popelish's response(s), I see that the MOSFET switches I was using are unnecessary and the circuit can look like this:

+5V | [1000] | +----+---+-------+-------+-------+-------+-------+-------+-->Eo | | | | | | | | | [1000] | [1000] [1000] [1000] [1000] [1000] [1000] [1000] | | | | | | | | | O| | O| O| O| O| O| O| O| O| | O| O| O| O| O| O| O| |S1 | |S2 |S3 |S4 |S5 |S6 |S7 |S8 +----|---+-------+-------+-------+-------+-------+-------+ | | GND | | | +--+ | Vcc | Vcc | | | | | [4K7] [0.499K] | | | +-|--[1M]-+ | | | | 74HC86 +---+-|-|+\\ | +-----+ | | | >--+-------+--|A1 | | +-|-/ | | Y1|--+ [2.49K] | Vcc +--|--|B1 | | | | | | | | | | | | [4K7] | +--|A2 | | | | | | | | Y2|--+ | +-|---[1M]-+ +--|--|B2 | | | | | | | | | | | +---+-|-|+\\ | | +--|A3 | | | | | >---+---+ | | Y3|--+ | +-|-/ +--|--|B3 | | | LM393 | | | | | | | +--|A4 | | [2.00K] | | Y4|--+ | +---- |B4 | | GND +-----+ | | +---------------------+ | | | +V +V | [1N4001] [COIL]- -| [100R] | | O-> |Eo | [1000]R2 | GND

and Eo will be:

E1 R2 5V * 1000R Eo = --------- = --------------- = 2.5V R1 + R2 1000R + 1000R

Now, since that voltage will be on the inverting input of the comparator with 4.5V on its non-inverting input, (and will be less positive than 4.5V) that comparator's output will go high.

However, since the other comparator's non-inverting input is at

2.0V, its output will stay low

Next, if we push two buttons, the switch divider will look like this:

+5V |R1 [1000] | +-------+---->Eo |R2 |R3 [1000] [1000] | | +-------+ | GND

and because R2 and R3 are in parallel, their total resistance will be:

R2R3 1000R * 1000R Rt = ------- = --------------- = 500 ohms R2+R3 1000R + 1000R

So, the divider now looks like this:

+5V E1 |R1 [1000] | +----->Eo |R2 [500] | GND

and Eo becomes:

E1R2 5v * 500R Eo = ------- = -------------- ~ 1.67V R1+R2 1000R + 500R

Now, since 1.67V is lower than the reference voltages on either of the non-inverting inputs of the comparators, both of their outputs will go high.

If we make a truth table for all possible combinations of switch activations it will look like this, where 'n' is the number of switches pressed and 'OUT1' and 'OUT2' are the logical output states of the comparators:

n OUT1 OUT2

-----|------|------ 0 0 0 1 1 0 2 1 1 3 1 1 4 1 1 5 1 1 6 1 1 7 1 1 8 1 1

Now, since the OP only wants an output when one buton is pressed, and the only time the outputs of the comparators are different is when one button is pressed, we can use an EXCLUSIVE OR to decode that state. Additionally, since a 74HC86 contains four EXORs, we can run them in parallel and use them to drive a moderate-current bipolar relay driver. (Or anything else, with proper modifications)

-- John Fields Professional Circuit Designer

--
Changing some polarities and wire-ORing the outputs gets rid of the
EXOR, and the relay contact designations were wrong, but are fixed
below:

 +5V
  |
[1000]
  |
  +----+---+-------+-------+-------+-------+-------+-------+-->Eo
  |    |   |       |       |       |       |       |       |  
[1000] | [1000]  [1000]  [1000]  [1000]  [1000]  [1000]  [1000]
  |    |   |       |       |       |       |       |       |
  O|   |   O|      O|      O|      O|      O|      O|      O|  
  O|   |   O|      O|      O|      O|      O|      O|      O|       
  |S1  |   |S2     |S3     |S4     |S5     |S6     |S7     |S8
  +----|---+-------+-------+-------+-------+-------+-------+ 
  |    |
 GND   |
       |
       |
       +--+
          |
          |       
          |       
   VCC    |      
    |     |      VCC 
  [510R]  |       |
    |     |     [1K]           
    +-----|-|+\\   |      +V       +V 
    |     | |  >--+       |K       |      
    |     +-|-/   |   [1N4001]  [COIL]- -|       
    |     |       |       |        | O-> |NC 
  [2.4K]  |       |       +--------+ |   | 
    |     |       |       |          |   O K1     
    |     +-|+\\   |       D          |   |  
    |       |  >--+-----G            |   +--------->COM 
    +-------|-/           S          | 
    |      LM393          |          +------------->NO
   [2K]                  GND 
    |
   GND

I thank you for converting my rambling description to a schematic. I didn't have time to do it, myself, last night.

I think I like it with the switches pulling up, better, because it puts both sensing levels within the common mode range for lower power supply voltages (all the way down to 3). And if you use a PNP relay driver, it can be driven in series with the common pull up resistor on the outputs, so a normally open contact can be used, or the active high collector might be the actual output, itself. Can't say for sure, because we don't know what the intended load is.

Forget that part about the PNP driver eliminating the NC contact. The signal in in the "window" when neiither comparator is pulling down, so an NPN or N channel driver will be on in th window.

John Fields wrote: (snip)

(snip)

I think you will get the largest possible voltage difference between 1 switch closed and 2 switches closed if R1=1/(sqrt(2))*R2. That is why I recommended 10k at each switch and 6.8K as the common series resistor. I chose higher values than you did, because I was imagining (for no particular reason) the whole thing running from a 9 volt battery.

You don't have to reveal the end purpose of a design, just to tell us what's expected to be switched. You want a one-and-only- one gate. I seem to remember this from the Don Lancaster TTL cookbook.

Ew. A one-and-only-one gate. All inputs low, output = low. One input high, all others low = output high. Two or more inputs high, output = low.

Sounds like a _really_ exclusive-or gate! ;-P

I think I'll use that as an exercise on these new tools that this latest client turned me on to. ;-) With a micro, it's trivial - count bits. With combinatorial logic - heck, in Verilog or VHDL it's the equivalent of a "switch" statement.

Oh My God. We've come full-circle.

Thanks! Rich

You've just designed a Rube Goldberg doorbell. ;-P

Cheers! Rich