ABY box - 3 states with one switch?

[ethrbunny]

(I think this is a FAQ but I don't know what to search on - redirects are appreciated)

I want to build a switch for a few devices. In an ideal world such a build would have 3 jacks and one switch. AY, BY and ABY (make sense?). Id love to have it just be stomp, stomp, stomp to toggle between the three. LEDs to show who is talking to who. Any pointers on how to get started?
Even vague datasheet references would be appreciated. Im certainly not asking for someone to design this for me. I just don't know where to look first.

Thx.

[R.G.]

No problem - if you are willing to power the box and use some logic to run relays to switch. Parts location problem verging on intractible if you insist on passive.

Sequencing switches to do multiple states in a round robbing on successive activations *do* exist. Finding one with the right number of poles and throws could be daunting. Maybe not impossible, just a matter of haunting surplus places and maybe some luck.

If you will accept the power penalty and need for logic, sure - you want a CD4017 ten state counter chip, a couple of buffer/glue chips, and then two EA2 relays ($2.60 each, Mouser).

There may be other solutions, but those come to mind.

[Outlaws]

I realize the nice thing about DIY is that it will do exactly what you want it to do...but I still have to ask...

Why wouldn't you just use two footswitches so one controls either A or B, and the other controls if it is going to both or not?  Then you can have LEDs and even if it was on A when you step on Y, you can see that when you step on Y again, it will go back to A or you could step on A/B so that the next time you step on Y it will go to only B.

Just seems limiting to have one switch.

But to each his own.   :wink:

[ethrbunny]

Outlaws - thats what I was originally (and probably ultimately) going to do. I was just trying to make it more of a learning experience. Im really trying to grapple with 'logic' in chip form. I know this sort of this is v doable...the trick is making it happen in my brain and on a board.

RG - I absolutely intend to power this. Relays? thats more 'physical' than I was imagining. Cool. Would there be some way to 'transistorize' the switching? I guess Im back to using a relay to move power again. Hmm.. this is proving to be more interesting than I had thought. Nice!

[ethrbunny]

Found a cool IC - the CD4066 "Quad bilateral switch". Looks like something could be rigged using that and a CD4017 counter.

How about this:

non-latching stomp switch -> pulse to 4017 -> signal(s) to 4066 -> opens 9V lines to 2n5088(s) which act as buffered lines for instruments.

Is this a total rube-goldberg? Do I need a mouse and a candle?

EDIT: doing a search on the 4066 turns up *lots* of posts.
EDIT ^2: found a reference to a 'tone god' article - which points me to geofex. LMFAO.

[bigjonny]

How about this:

non-latching stomp switch -> pulse to 4017 -> signal(s) to 4066 -> opens 9V lines to 2n5088(s) which act as buffered lines for instruments.

I think the point behind R.G.'s using of relays was to AVOID having to buffer anything.  Am I right?

[R.G.]

Yeah, I was after that true bypass kinda mojo. Nothing touches the signal but heavy metal.

OK. Look here: http://focus.ti.com/lit/ds/symlink/cd4017b.pdf
The CD4017 is a decade counter with ten decoded outputs. You put power and ground on the noted pins, tie clock inhibit to ground, and then it advances the count once each time the clock goes high. The outputs go high in sequence: 0, 1,2,3,4...9,0,1... and you have one output per count, not binary encoded stuff.

Here's the neato trick with this thing. You can make it count any amount less than ten by tying an output back to the reset pin. When reset is pulsed high, the chip is set to the "0" state, with the "0" output high. So if you want, in this case, three states, you use 0, 1, and 2 for your outputs and tie the "3" output to reset. Now it counts 0,1,2,0,1,2... because the nanosecond when "3" goes high, it flips the chip back to "0" and the propagation time through the chip keeps the "3" output high long enough for the internals to reset themselves to "0". It can count any modulo less than ten.

There is one quirk: the initial state is not guaranteed, so if you don't use an OR gate and explicitly reset the chip at power up, it could come up in one of the unused outputs, 4...9. If that happens, you would have to punch the "change" button up to six times to get it back to 0. That can be fixed with a batch of diodes if it annoys you.

So you use the three states to drive your relays: f'rinstance, "0" is output A, "1" is output B, and "2" is A+B/Y. Now all's you gotta do is arrange some glue logic around the thing to make a clean clock out of your switch closure and to drive the relays. I think I could do that with a hex inverter and one quad two input NOR package, total of three 14 pin packages, each about $0.30 to $0.50.

I know this is not fair to you, but I have to mention it for the experts and pros that watch me: A 12C508 PIC in a single 8 pin package costing $1.75 would also do the job all by itself, no external components except Rs and Cs and perhaps relay driver transistors.

[ethrbunny]

Im not overly concerned about true bypass. I have bigger issues for certain. Im more interested in using the 4066 though I would like to come up with a way to use LEDs (or somesuch) to indicate which channel or pathway was enabled. Maybe something involving the two unused channels in the 4066? Dunno. Driving to Seattle with the family today. This will give me something to ponder.

[erio fraga]

Take a look in
http://members.shaw.ca/roma/switching-2.html

Maybe something there may help you ( be carefull, I believe there are some mistakes in the indicated conections between 4017 and 4066 ).

Erio

[R.G.]

If you're willing to live with the limitations of the 4066, it gets easy. A 4066 can be thought of as four independent switches that turn on when you put a logic high (Vdd) on the control pin for that switch and turn off when you put a logic low (ground) on the control pin.

So: you want to conduct one or both of two signals into an output. You buffer the inputs; call them A and B. A goes to two of the 4066 switches, and B goes to the two others. The outputs of the 4066's are ganged up together and form the output, to be capacitor isolated for DC and so on.

Let's call the switches A goes to A1 and A2. A1 is the A signal only switch, A2 is the "both" switch. Likewise B1 is the B only switch, B2 is the both switch. From your 4017, you hook up output 0 to the control pin of A1, output 1 to the control pin of B1, and output 2 to both A2 and B2 control pins. That's all of the necessary signals and you now have a round-robbin A/B/Y signal switcher. For the LEDs, you will probably need either a transistor per output to pull the LEDs, or more usefully a CMOS hex inverter to drive the three LEDs and use the remaining 3 inverter gates to clean up the momentary switch to make the clock to the 4017.

Two CMOS packages then.

[ethrbunny]

If you're willing to live with the limitations of the 4066, it gets easy.

What limitations are you referring to?

Take a look in http://members.shaw.ca/roma/switching-2.html

This is a cool site! There are several nice projects in there. Thanks for the pointer!

[R.G.]

What limitations are you referring to?

Primarily that it is a non-mechanical switch and that you have to do some tinkering to keep it from popping. It has some voltage limitations, too, but probably you won't run into that.

[ethrbunny]

RG - from reading the datasheet I was under the impression that it was limited to +- 7 volts. Ive seen schematics where it was used to switch 9v+ though. I figure its worth a try either way.

[ethrbunny]

Riddle me this: when I read about 'audio mixers' they always buffer the signal. When a box like this was used to blend the two signals- would it need to be buffered also? to keep from 'leaning too hard' on each channel? Or is it sufficient to have caps at the input?