Stomp 3pdt´s and dpdt´s..I´ve just had enough!!

[R.G.]

RG, why cd4053?, i know is a 3 spdt sitches in a chip, but what´s the difference?
easier layout?. i´ll give it a try!!!  anyway, my questions about popping still apply to 4053, aren´t they?

From June 2000: http://geofex.com/Article_Folders/cd4053/cd4053.htm

The 4053, like all switches can be made to pop or not pop. CMOS switches need to be biased nearly at the middle of their DC voltage supply to let the insides cancel any offset that would cause clicks. The article shows how.

As to why the 4053 - you did ask for making a single module. The CD4053 in combination with your CMOS flipflop makes for a simple, two chip solution that you can operate by only connecting up the parts you need, and it goes all on a remote module easier.

[dschwartz]

oh, thanks for thje article RG!!!..i´ve read it before, but i didnt find it last night..
there´s everything i need to know..

[aron]

OK, what momentary switch is the one to use with all this? Which one is tough enough to last and not fail or break?

[R.G.]

Clever guy! 

Yes, that's the next problem. We use a mechanical actuator we have custom made, and put an ALPS tactile dome switch under it with a block of plastic foam between them. The actuator can't ever actually touch the switch, only transfer pressure through the plastic foam. ALPS rates the switch for 10 million operations. I don't actually expect every one of them to last for ten million, but then I don't ever expect a real-world pedal to get switched a million times either.

Like I said, we were sick of replacing stomp switches. We tried to do a slam-dunk solution. The up side is that after we paid for the casting die for the actuator and got the quantities up, they're as cheap as footswitches. They should be - they're mechanically simpler.

[aron]

>Yes, that's the next problem. We use a mechanical actuator we have custom made

Hehehehe I knew there was some break in logic here. The first thing I thought of was.... ok, we are replacing a mechanical with a bunch of electronics.... driven by a mechanical switch!
However, I did think of something..... what's the likelihood of a momentary DPDT failing on both lugs? I mean, one maybe, not two right?

In any case, for the DIYer, I wonder if it will be a lower cost alternative? I will check.

[earthtonesaudio]

I seem to recall some GEOFEX article or other about a mechanical actuator for a tactile switch using a threaded rod and some other stuff...
and an article about a footswitch "pancake." 

[R.G.]

>Yes, that's the next problem. We use a mechanical actuator we have custom made
Hehehehe I knew there was some break in logic here. The first thing I thought of was.... ok, we are replacing a mechanical with a bunch of electronics.... driven by a mechanical switch!

No, adding more mechanics is not the answer, you're right. I also seem to remember Small Bear having a mechanical spring actuator. As for the reliability of the electronics, once they are working, they are far, far less likely to fail than the mechanics, many orders of magnitudes.

However, I did think of something..... what's the likelihood of a momentary DPDT failing on both lugs? I mean, one maybe, not two right?

The probability of a total of two things happening, given that they are independently probable, is the product of their probabilities. So if the probability of one contact failing is 1 in 10,000, and you have two contacts, then the probability that both fail is 1 in 100,000,000.

Since both contacts of a DPDT momentary were made at the same time, in the same factory, same equipment, same people, same yada yada yada, the issue of independent probabilities gets very strained. I think if one contact fails in a DPST momentary, that it's just a matter of a short time until the other one fails too. But it would be longer than only one. I've used redundant contacts for reliability before in stuff I've designed. But I don't know how to test the actual probabilities without a milllions-of-cycles test machine.

In any case, for the DIYer, I wonder if it will be a lower cost alternative? I will check.

Good plan. Go find out the actual costs.

I've had a few courses in fault analysis and design for reliability, but I'm by no means an expert; only about the minimum needed for a competent designer. The most reliable class of purely mechanical device I know of is a power generating turbine for an electric utility. These machines are expected to be put on line and keep turning at the same speed for over 20 years before the first failure requires them to be stopped.

The study of design for reliability can be fascinating if you get over the initial statistics math hump.

[dschwartz]

for a momentary spst switch..you can use your imagination....even a clip and a nail will do!!!

[Processaurus]

From the experience of designing coin-op games for atari in the 70s, my dad had recommended this debouncing and button configuration, as bulletproof.  Only hitch is you need a spdt momentary switch.  But it was the way those manufacturers went, arcade games probably have a rougher life than stompboxes, with drunks beating on them day in and day out...

Any switch bounce only reinforces the state it's in, and there is no lag.

fig. 3 here
http://www.elexp.com/t_bounc.htm

I recently put a danelectro pedal in a hammond box to make room for some mods, and kept the CMOS switching (momentary switch clocking a 4013 driving a 4053), just like the GEOFEX article on electronic switching), I must say the soft momentary stompswitch that small bear electric sells is nicer feeling than regular stomp switches.

Ooo, with a SPDT momentary switch and 4013 you could use one half of the 4013 as a similar set/reset as the cross-coupled NAND gates, and then the other as a regular alternating flip flop to switch a 4053:


Debouncing and flip flop done with 3 parts!

[R.G.]

Yes, the S-R flipflop and SPDT switch is a classic, and is, as your father says, bulletproof. About the only detractors possible with it are that it needs two throws on the mechanical switch and you need rather more CMOS logic. Looked at as a parts count, you have only the switch, the 14 pin DIP, and the resistors.

I should explain that my approach to debouncing contacts is a result of optimizing in a different direction, as is so often the case in design. I got hooked early on using SPST switches to sense motion, and you can't use the S-R FF with a single contact switch. I also branched off into many-switch applications where the cost of a single switch is an issue, and forces different considerations. In that case, the R-C network with the switch shorting the cap can be made to work well if you are willing to do some measurements and math. The advantage there is that you can use a CMOS inverter with schmitt trigger inputs and get six debounced switch contacts per IC package. You get more debounced contacts out of a single package of logic and the per-contact cost is lower. A single contact costs you two Rs, one C, one SPST and 1/6 of a 14 pin DIP.

One design I keep in my mental file of "Hey, I did good on that one!" designs is the one where I was able to get a switch contact, debouncing network, and status indicating LED all operated by **one** pin of a microcontroller, and make the internal programming do eight of these in parallel in the same switch sensing loop. The switches all acted independently, and could be changed from momentary action to alternate-action by telling another pin on the uC which they were to be. But that was a specialized design.

[dschwartz]

I´m thinking of using a switch like this one:


do you think it will take the job proudly?

[R.G.]

I´m thinking of using a switch like this one:


do you think it will take the job proudly?

It will do it proudly, for the remainder of its unfortunately short life. That switch won't last very long being stomped upon.

[dschwartz]

ohh, bummer...

[aron]

The footswitches we use are tough. I think something could be made, Steve had something long ago.

[R.G.]

The footswitches we use are tough. I think something could be made, Steve had something long ago.

You mean this? http://www.smallbearelec.com/Detail.bok?no=32
Yeah, it'll work fine. I also did an actuator article on making one up from electric lamp parts from Home Depot.

Making a sturdy, reliable footswitch actuator is not all that simple. Making it cheap is even harder. When I was looking for switches, I ran into a company that specializes in vandalism resistant switches. They were machined from solid stainless steel. the bushing was round and about 2" diameter, and it had sloped sides you could not get a pipe wrench to hold on. The switch button barely protruded at all, and was a small section of a sphere, only protruding about 1/8" or less out of the bushing. The button was about 1" diameter. In the center of the switch was an indicator. It was solid LEXAN, about 1" deep, and was actually a light pipe going to the LED back inside the switch. The mechanical travel was limited to about 0.03", and the inside was shielded and vented to allow stuff like sand to fall through. The connection to the actual switch was a spring. These were intended for use in urban war zones where if there's nothing else to do you tear the place down. They fascinated me until I found they cost $200 each. 

[Hanglow]

I think most guitarists think they can literally stomp on their stomp box - it should be called a tap box for the health of the switches

I don't think that'll catch on though 

[R.G.]

I think most guitarists think they can literally stomp on their stomp box - it should be called a tap box for the health of the switches

I don't think that'll catch on though 

We do get the occasional box back in from our first version of enclosures in what I call "flat-topped" condition. If you get a good run and jump at a box where all the knobs and switches are exposed on the top of the pedal, and come down at a slight angle with your steel-tapped boots, you can break the actuators/shafts off the switch/controls, all at the same time. 

[Processaurus]

Question I can't figure out without trying it, would the Geofex 4053 switching make a pop if the pullup resistors to V1/2 were omitted on the in and out jacks (R5 and R6 on my schematic)?  It would be getting identical DC bias from back through the switch from both throws, so would it ever dip below V1/2 enough to pop as it was switched?  A mechanical switch, yes, from the time between switch contacts, but here, I know analog switches can operate up into radio frequencies, and the typical ~180ns delay from address to signal output (tPHL on the datasheet) would be the same for turning off and on, so it seems like the switching coordination between the two poles would be quite precise?  And the switch wouldn't appreciably be either break-before-make or vice versa?

Here's the 4053 datasheet:
http://www.fairchildsemi.com/ds/CD%2FCD4052BC.pdf

[R.G.]

Question I can't figure out without trying it, would the Geofex 4053 switching make a pop if the pullup resistors to V1/2 were omitted on the in and out jacks (R5 and R6 on my schematic)?  It would be getting identical DC bias from back through the switch from both throws, so would it ever dip below V1/2 enough to pop as it was switched?  A mechanical switch, yes, from the time between switch contacts, but here, I know analog switches can operate up into radio frequencies, and the typical ~180ns delay from address to signal output (tPHL on the datasheet) would be the same for turning off and on, so it seems like the switching coordination between the two poles would be quite precise?  And the switch wouldn't appreciably be either break-before-make or vice versa?

I don't know. I went through a lot of effort getting them not to pop a few years back, and once I got a solution, I stayed with it. Periodically I've speculated on whether I could ditch a pullup here and there based on what the other sides of the switches were doing. When I got down to it, the cost was never more than  a couple of $0.01 resistors, so I determined that for me, it was cheap enough to simply keep using them.

But someone ought to try it. 

[edvard]

Have you built this yet?
A long time ago I made a layout for something just like this, and used a simple one-transistor debouncing buffer I got from a RatShack engineers mini-notebook (remember those?)

When I get home I'll look it up, and here's a great read on everything you'll ever want to know about debouncing:
http://www.ganssle.com/debouncing.pdf