Reasons for/against circuit integrated with 3PDT PCB?

[SmoothAction]

Hey friends.

I'm working on a simple electra circuit in Eagle. I am wondering if implementing a 3PDT bypass stomp switch onto the main PCB would be as detrimental as I think... as in after four months of gigging it fails due to shock and absorption of abuse. Wiring is my weak point for sure, i'm looking to minimize as much as possible while not sacrificing dependability. I've seen quite a few builds with mounted 3PDTs and I think it's a stellar idea in theory...

But in practice?

Anyone is welcome to comment and chat, thanks.

[rousejeremy]

I personally don't like the idea of stomping on something connected to a PCB. If that nut comes loose it's just a matter of time until you have a problem.

[rousejeremy]

I mean connected to the PCB of the entire circuit. I don't see an issue with the separate 3PDT boards besides fitting a battery in there.

[R O Tiree]

I've never had a nut come loose on a 3PDT. I usually incorporate the switch onto the PCB, because it holds the board absolutely rock-steady and it cuts down on the rat's nest of wires a lot. No need for stand-offs, so no drilling or gluing for those at all. No chance of a short circuit to the enclosure.

All that said, I mill slots for the switch lugs. I've got a desktop milling machine with X/Y/Z wheels, so I do BNC instead of CNC (Brain Numerical Control ). This ensures zero insertion force when I push the switch into its slots - make sure it's perpendicular in both axes and then solder. Solid as a rock.

[Mark Hammer]

Stompswitches are meant to take as much use and abuse as pots and toggles; indeed more, since they are footswitches.  I don't see why they should be subject to concern about mechanical reliability or risk to the board itself.

having said that, what you DO see in many commercial builds is the stompswitch mounted to a separate board, from which a ribbon connector can be run to the main board.

[SmoothAction]

Yes Jeremy, I was thinking of that exactly. I see pedal companies doing this. I see guitarpcb.com has this style now too. There are people who etch their own boards with a 3pdt as part of the main board. Maybe I'll do a couple boards with and without... why not yea?

[SmoothAction]

@Mr. Hammer

I didn't think of doing something like a separate 3pdt board with a convenient ribbon connection, that's a great idea. Thanks. I feel "iffy" about terminal slots, I'd probably want to use soldered wire leads instead. Much like the new BYOC board/footswitch combo. Pretty cool, good idea. Thanks.

[gjcamann]

I've designed a small board that has my footswitch, power filtering and LED resistor on it. I plan to use that in all my pedals going forward, i've also put a small stripboard on it, so for boosts I can do it all on the same board, or cut off the strip board and wire to a separate board.

I've got a eagle library with the Alpha PCB footswitch sold by smallbear if you want it. I feel the most reliable part of my pedal will be that footswitch board, because the rest of my pedal looks like a bunch of duct tape and bailing wire. 

[smallbearelec]

Here is how I have implemented it in the Bare Box #1:

https://www.smallbearelec.com/Projects/UrsaMinor/Fig45.JPG

I am pretty sure that the idea will work for most boards. The Molex items and pre-terminated leads are available from my stock.

[SmoothAction]

^ "Snap-in" style terminals would be pretty cool. I hadn't seen that, thanks. Looks pretty secure.

[jkokura]

Footswitches are almost always the first thing, the major thing to fail. By far, the majority repairs I've done have been overwhelmingly switch repairs. I don't know why, but switches seem to fail, or develop pops, or all sorts of things.

Have you every tried to remove a stomp switch from a PCB? It's not fun work, and usually you end up ruining the switch, the PCB, or both. Putting the switch on it's own PCB is no big deal. Will save some time in assembly, and doesn't cost much to make a switch like that, so replacing it isn't really that big a deal. Sure, throw a spot on there for a resistor to ground on the output even, so you can mitigate pops, or a CLR resistor/trimpot. No biggy.

But I think it should be connected by ribbon cable to the rest of the circuit. Replacing a switch means replacing a whole PCB. I won't buy/use PCBs that are setup like that either. Not saying it wouldn't sell, you shouldn't use it, but for long term use/care of your pedal, I choose no.

Jacob

[Arcane Analog]

Connecting a footswitch - which is stomped on by definition - to the main PCB makes no sense. Why would you want to transfer all of the stress from the footswitch into the PCB? The vast majority of repairs I do are because the part was mounted to the PCB and normal use ruined the solder joint or cracked the PCB. In fact, I just finished repairing three pedals for two different people this evening. All problems with PCB mounted external controls.

If you like durability PCB mounted pots and switches are a bad idea.

[SmoothAction]

Good info ^ Thanks Jk and Arcane. I had a gut feeling it wasn't a good idea, but see it often enough to inquire I suppose. I went ahead and just did the PCB without mounting the 3pdt onboard. I'm just gonna have to be more patient while wiring 

[Arcane Analog]

For the record I hate repairing pedals - especially those with PCB mounted components. As jkokura pointed out removing some board-mounted components is a nightmare. I only fixed the pedals this evening because two are owned by a friend and the other is a touring musician who need their pedal fixed fast for a show.

[R O Tiree]

I concede that the switch is a point of failure internally, but there should be no forces transmitted through to the board... you press the switch and the force is dissipated into the enclosure via the external nut/washer and no further. The "blue box" containing all the internal workings is simply hanging off the inside face of the enclosure. If you have designed/drilled/milled the slots/holes matrix in the board accurately, then there is zero insertion force when you install the switch, so no (read very minimal) stress on the board. Since the fixing of the board to the enclosure is also just in one plane and one point, there is no stress due to expansion/contraction. The only force on the lugs when you press on the switch is due to the spring/lever action internally. Milling slots for the lugs makes for a much more stable solder joint compared with holes (although holes are much easier to do). Once again, I've never had a board/solder joint failure on any of my pedals.

I have only ever had one pedal returned due to a broken switch - basically because the customer had tried to reverse-engineer the circuit and hadn't realised what was connected to what and messed it up in spades. Since the switch was utterly trashed, I simply removed the metal actuator assembly and then de-soldered the lug individually, popped in a new switch and that was that. Even if the switch is not trashed as that one was, you just get a small screwdriver and lever the little metal "fishtails" away from the body, remove the actuator/springs/levers assembly from the top of the switch and then you can remove the lugs one by one as I just described.

Similarly, for board-mounted pots... admittedly, little 9mm ones that mount flush to the board with extremely short pins are a pain and I wouldn't do that, but 16mm ones, with "relatively" flexible legs are fine, as any expansion/contraction stresses are alleviated by the longer legs. Since the pot is trashed, just snip through the legs, de-solder each pin individually and then install a new one. Not really any more hassle than a "wired" pot.

[Arcane Analog]

That is fine an dandy to say that but what happens in the real world is another story. A player's feet do not always hit a switch perfectly. Mounting something that is routinely stepped on to a PCB is not smart.

[candidate]

If you populate your boards using outsourced Chinese labor the best approach is to mount everything to the board.

[GGBB]

I concede that the switch is a point of failure internally, but there should be no forces transmitted through to the board... you press the switch and the force is dissipated into the enclosure via the external nut/washer and no further. The "blue box" containing all the internal workings is simply hanging off the inside face of the enclosure. If you have designed/drilled/milled the slots/holes matrix in the board accurately, then there is zero insertion force when you install the switch, so no (read very minimal) stress on the board. Since the fixing of the board to the enclosure is also just in one plane and one point, there is no stress due to expansion/contraction. The only force on the lugs when you press on the switch is due to the spring/lever action internally. Milling slots for the lugs makes for a much more stable solder joint compared with holes (although holes are much easier to do). Once again, I've never had a board/solder joint failure on any of my pedals.

I have only ever had one pedal returned due to a broken switch - basically because the customer had tried to reverse-engineer the circuit and hadn't realised what was connected to what and messed it up in spades. Since the switch was utterly trashed, I simply removed the metal actuator assembly and then de-soldered the lug individually, popped in a new switch and that was that. Even if the switch is not trashed as that one was, you just get a small screwdriver and lever the little metal "fishtails" away from the body, remove the actuator/springs/levers assembly from the top of the switch and then you can remove the lugs one by one as I just described.

Similarly, for board-mounted pots... admittedly, little 9mm ones that mount flush to the board with extremely short pins are a pain and I wouldn't do that, but 16mm ones, with "relatively" flexible legs are fine, as any expansion/contraction stresses are alleviated by the longer legs. Since the pot is trashed, just snip through the legs, de-solder each pin individually and then install a new one. Not really any more hassle than a "wired" pot.

How many people are going to mill their slots?  It sounds like it is many times better than holes, but no-one (besides you) is going to do it, so the question is really whether standard 3PDT mounting (not milled slots) is reliable.  I think the problem with it is more to do with the overall support of the PCB, not just the 3PDT connection.  If the 3PDT is the only thing supporting the board, any forces and vibrations transferred to the board either by stomping or handling, dropping, kicking etc. are going to be amplified and stress the joints leading to premature failure, especially if the weight of the populated PCB is not perfectly balanced on all sides of the 3PDT connections (which it never is).  The larger/heavier the PCB or more off balance, the worse this gets.  If the PCB is supported elsewhere - ideally at the opposite end - by standoffs or pots for example, the vibrations are dampened and you have a far less problematic situation.  But it's important to make sure there are no new stresses introduced by having multiple support points.  Still, no matter how well you do it, there is always going to be more stress from handling and use on the 3PDT joints when it is directly connected to the PCB than when it is wired (assuming the wiring is stranded and is not constantly flopping around inside the pedal).  So it will always fail sooner, but could still last a very long time.

[Mark Hammer]

Need I remind people that the force required to actuate mechanical switches, and the increase in risk incurred the more contacts have to be made, are but two of the reasons who so many dependable manufacturers stick with e-switching?

The rationale for TB switching has its historical roots in vintage pedals that used SPDT switching.  Pedals that left the input of the circuit connected to the input jack in both effect and bypass mode did not pose THAT much of a problem.....as long as you only used one pedal in your rig.  Connect up two or more, and the loading down of the signal, even when all pedals were ostensibly in "bypass" mode, would be significant.

DPDT switches solved that particular problem, and eventually 3PDTs would address the same problem in a compact way but also allow for mimicing the status LEDs that were part of what accompanied e-switches.  While a convenient solution for DIY-ers and cottage-industry types, much mystique and cachet has been built up about TB, making it seem more critical than it really is.

And sometimes, it's omnipresence has led to it being seen as more faulty than it really is.  It seems I have to repeat this at least once a year:  if switch manufacturers had the same failure or problem rate with their commercial clients that DIY-ers had with those same switches, the manufacturers would have been out of business long ago.  The difference in failure rate tends to stem from inappropriate practices amongst hobbyists, or, to put it in a more positive light, more optimal practices amongst larger-scale manufacturers.

What do they do that we don't?  First, I'd imagine their switches are purchased in sufficient quantity that they arrive with gleaming solder lugs that lap up solder like it's going out of style.  They aren't tarnished such that they fight the solder and seem to require higher heat.  The soldering temperature is like juuussst right such that the joint is made without applying enough heat to reflow the grease inside the switch.  All contacts to be joined are pre-tinned such that joint formation is equal parts soldering and "sweating".

Plenty of commercial companies that do not want ANY servicing to be necessary, rely on 3PDTs soldered directly to boards.  I would agree that having too much of the shaft sticking out of the chassis can result in greater stress applied to the switch, because of the angle that feet apply force.  But if the switch is mount properly, and all that's going on is simple up/down piston motion, they shold live up to their anticipated lifespan, problem-free.

Here is more info about the inner anatomy:

[R.G.]

Connecting a footswitch - which is stomped on by definition - to the main PCB makes no sense. Why would you want to transfer all of the stress from the footswitch into the PCB?
...
If you like durability PCB mounted pots and switches are a bad idea.

No necessarily. It depends on the mechanics of the situation, and where the stresses go. Materials vary in the amount of elastic stress they can withstand without creep or permanent deformation. They also vary in their resistance to fatigue cracking.

Solder has remarkably low resistance to plastic creep, deformation and fatigue cracking. This is mostly due to its melting temp. At room temperature, solder is being hot-worked, much like steel heated to a dull red with a torch. But it has some resistance to creep and failure. If it had none, all parts would eventually fall off all PCBs.

So the trick with any board mounted part is to keep the applied stress on a solder joint below the creep and fatigue points on the solder. This comes on two flavors, both elements of mechanical design. First, design the setup so that a board is fixed in one plane only. No two-way mounting. If all the pots, jacks, etc. mount to the inside of one face of the enclosure, then you do not have the tightening force of the nuts on the jacks, switches, and pots trying to pry the solder joints apart. Obviously, the parts have to have the same mounting height above the PCB, or you can be trying to force solder joints apart by just tightening the nuts down even in the same plane.

This means that you can't put opposing-direction jacks on the same PCB - they pull each others' solder joints apart. And don't put pots on the face and jacks on the side, both pulled to the enclosure with nuts. Don't hard-mount a PCB and use pots that bolt to the enclosure unless you can get pots and standoffs of the same height.

Second, make your holes big enough so that the variation in side-to-side placement of the pots/jacks/switches in PCB holes is smaller than the slop in the holes. This prevents tightening the nuts in bushings from forcing the controls sideways, putting stress on the solder joints.

I speak from experience here. In my day job, I've used PCB mounted pots and switches literally tens of thousands of times. We do have failures, of course - nothing is perfect, or lasts forever! - but the numbers I collect and review on failures is down in the random-component-failure range.

In terms of footswitches, the stresses are higher because the switches are obviously being stomped on. So you have to take extra care to keep the stresses off any PCB the switch mounts to. The simplest way is to float the PCB, leaving it mounted only to the switch and not to other stuff. If you do that, the forces from stomping can act only on the button of the footswitch and the bushing, and do not stress the PCB or the solder joints. No (excess) failures.

This is all very much like the refrain in "The Gambler"; you gotta know when to hold 'em, know when to fold 'em.  It is possible to mount pots, jacks, and switches to a PCB and have this be reliable. But you do have to put in the time to learn what makes it reliable and not. There are lots of ways to get this wrong. The tidal wave of the pedal cottage industry has created a whole lot of people rediscovering the ways to get it wrong.

We're all still learning here.

Mark posted while I was fumble-typing. Right again Mark.