A little logic question

[TomWaterman]

Alright guys,

I'm musing over a possible application that may require some logic, it may be pretty trivial but I've not dealt with logic before.

Basically I am looking at a way to implement the following:

A stereo, balanced 3 source input select.

• A DPDT relay each side for assigning each source to a buss, (possibly separate left/right PCBs to save on 4PDT relays and improve x-talk).
  A non-latching push button for each input.
  3 L.E.Ds for indication of selected input.

Essentially I am looking for a way to have all selections toggle so that only one of the three can be active at one time.
Selecting one will turn the other off etc etc

I was wondering if any logic gurus have an idea of what kind of circuit would be needed.
I require a non-latching button so that it doesn't need to be released when another input is selected. The LED provides the indication and both that and the relay will need to latch on until another button is pressed.

I'm guessing some sort of flip-flop is needed.

Any thoughts - just looking for a point in the right direction, not asking for anyone to design this one for me.....I'll be raiding Horowitz+Hill in the morning.

Thanks for any help.
Cheers Tom

 

[PRR]

If you can train your hand to twist: a 4-pole 3-throw rotary switch is a ready-made $2 solution.

There are ganged interlocked pushputton switches, or used-to-be. I had heaps of them. Remember the old car radios? Similar but with up to 12 contacts per switch, enough for floating stereo, lights, and even to cut you over to Musak when no button was down. However a quick search of DigiKey and Mouser doesn't find them available today. If you find the right surplus dealer you can probably get a crate for $1 a button, or $20 for one 4-gang (easy to cut to 3-gang), but that could be a hunt.

Yeah, a 3-bit memory (or 2-bit plus binary decoding) is the "obvious" way. But you actually need two layers of flip-flops (to force the old selection to drop-out while rememebering the new selection). There should be a more elegant approach.....

Someone here, I think "mcs", has a website of volume and selector switches implemented semi-digitally.

Rather than H+H, I'd look to Don Lancaster's (RTL,) TTL, and CMOS Cookbooks. (Kids: ask your daddy who Lancaster is.)

 

[PRR]

> I'm guessing some sort of flip-flop is needed.

Ya know.... it seems absurd.... but you can do it with 10,000 flipflops and be done with it before you can wrap your head around a 3-flop solution.

Get a BASIC Stamp©. Little microcomputer on a board. 8 I/O pins. Program 3 for input, 3 for output. A few lines of BASIC code read the buttons and set the outputs. $34 BASIC Stamp.

The Stamp outputs won't drive a multi-contact relay directly. But small relays just need a resistor and transistor.

That LED is the least of your troubles. If you can clack a relay, you can hang a resistor and LED across the relay coil.

 

[bcarso]

I have somewhere a reasonable bus-connected latch-cancel circuit, actually a few of them, somewhere (boy do those go back), but PRR is absolutely right. And you will have the added cachet of saying it's microprocessor controlled.

 

[TomWaterman]

Hey thanks guys,

You know I was originally going for rotary switch :green:, then I wanted to use a PIC and do a whole bunch of things like relay attenuators etc....

Got a bit overwhelmed and thought about just using some logic for the input switching, dropping the rest of the microcontroller circuit. I was under the impression (wrongly I guess?) that logic would be simpler and easier to implement??

Ganged switches like the 1176 would be cool but I can't find many, and I have a particular style/look in mind.

One reason I wanted to avoid a toggle was so that selections could be as close to instantaneous as possible. If all three sources are active but I want to A/B between 1 and 3, I didn't want to have to click past 2 and get a short blast of whatever that is.....I was trying to maximise the memory of the ear.

I read somewhere that its very quick, hence an idea for a quick switching solution.

Another reason is, that I have a nice switching arrangement with two banks in series so that a source can be assigned to either monitor or process buss. A monitored source cuts its feed to the process buss to avoid parallel loading. Relays are nice because I can do all of this in close proximity with small PCB tracks. Switches would mean running P2P wire all over the case....

I've seen Mikkels designs - very cool. I shall look again. I didn't want to just copy him though....

PRR, thanks for the tip on Don Lancaster! :thumb:

So do you guys have any tips about board layout and such for the microprocessor thing? I am worried about polluting the analogue signals with crud.

Thanks again
-Tom

 

[Svart]

although a switch is clearly the easiest and simplest answer, but maybe he just wants to build logic circuits? I would choose a small set of gates and a few tactile switches over a rotary switch too if it were only for 3 functions.

3 NAND gates, some pull down resistors, an array of 6 switching diodes,( which you can get in a single package too) and 3 tactile switches will get this done easily and cheaply, probably less than a few $ and much smaller than a switch.

feed the outputs of the gate through a diode back to a common trace that then splits through diodes to the inverting input of each gate. if any of those gates go high they will pull the inverting i of the other gates high shutting them off or keeping them off. Control the gates by pulling the non-inverting input high or low. add pulldowns to taste.. simmer for 20 minutes and voila!

just a quick idea though..

 

[TomWaterman]

Hey thanks svart, that sounds kinda simple and interesting....

I will look into it, I preume the gates feed a transistor driving the relays?

Cheers Tom

 

[Svart]

Yep, but some gates have totem pole outputs(actually most i think..) and some are MOSFET output that can even drive a smallish relay directly.

try the TC4469 for logic and drive. Kind of expensive for logic, but has MOSFET outputs that will drive over one amp so you can leave out the actual switching transistor/FET.

TC4469COE-ND @ digikey for 2.23(SOIC)

 

[TomWaterman]

Thanks svart, I'll peruse the data sheets this weekend!

Nice one.

-Tom

 

[FredForssell]

You could also try the 74C922 keyboard decoder IC.

 

[TomWaterman]

Hey Fred!

Hope you're alright?
Thanks for the suggestion, I'll look into it. This for the monitoring controller I emailed you about a while back.

Tom :guinness:

 

[Svart]

hmm just thought of something.. what I thought might work earlier with the diode arrays might not work due to the global pullup arrangement.. you'll have to look at the logic and what you need, also try the 4467 and 4468 ICs for different gate types.

:thumb:

 

[gyraf]

There is an old Philips channel-selector IC, made for 6-ch television sets - maybe named SAA- or SAB-something or the like - that will do what you want. Try looking that up. We use this for our speaker-select switching..

Jakob E.

 

[mcs]

[quote author="TomWaterman"]I've seen Mikkels designs - very cool. I shall look again. I didn't want to just copy him though....[/quote]
Have you seen the Control1 and Control2 here?

They could control both your input selector and attenuator...

If you need special features, you could just ask or program the microcontrollers yourself.

Best regards,

Mikkel C. Simonsen

 

[Svart]

looks like you have the pick of the litter Tom!


:guinness:

 

[TomWaterman]

Nice one guys! Lots to think about.

Thanks for all of your help and the offer Mikkel. You're kits seem to have moved foward quite a lot since I last looked at them.

Gotta dash now but I'll be back.

Cheers and beers as they say!
-Tom