Random thoughts… I prefer soft-touch switches, either for use with a relay bypass system, or even a Boss style electrical switching scheme. But that’s my personal preference. To those who prefer mechanical 3PDT I say cool! (Unless they don’t post gut shots of their builds, that’s not cool!
)
One thing on the 30k cycles for typical mechanical 3PDT switches though. I believe that number isn’t a guarantee so much as a MTBF, mean time between failure. Meaning, there’s probably actually a distribution of time-to-failure, probably bell-curve or similar. So just as a wild guess, I’m thinking most probably have a 20-40k typical cycle lifetime. But also the manufacturer probably assumes ideal soldering conditions, and anything short of that likely further reduces lifetime.
And for what it’s worth, go over to diystompboxes and read RG Keen’s numerous posts on the reliability - or lack thereof - of mechanical 3PDT switches. Besides his generous contributions to the DIY community, he’s actually worked directly in the commercial pedal business. If I remember correctly, he basically said something to the effect of virtually all failures were due to mechanical switches.
My other way of looking at it is similar to the soldering thing, in that when the manufacturer establishes reliability numbers, I presume they assume operation under ideal circumstances. I.e. you press directly down, with just the right amount of pressure, in a smooth, controlled manner. But in reality, we’re smashing these things with our feet, possibly during the heat of a gig. Inconsistent pressure, inconsistent angle of attack, etc. If you’re gigging regularly, that also means operation under diverse environmental conditions.
Maybe the switches are tested under more adverse conditions… but usually stuff tested like that is marketed as “ruggedized” or “military grade” or whatever.
Another way I think about it, simplicity versus complexity: I think there’s an argument that the 3PDT switch is more complex, because it’s combining two functions into one device: the “user interface” and the circuit switching. Whereas relay based systems are arguably less complex, because of separation of concerns. I’m a Unix command line guy, I like single tools that do exactly one job but do it well.
I think that “probability of failure increases with the square of number of components” rule of thumb has to be used in context. I mean, if you take it literally, consider a modern CPU with 10s of billions of transistors - how are they as reliable as they are in the face of that rule?
In software development, we also prefer simplicity. It’s generally less error prone, and easier to debug when there is an issue. But eventually, you bundle that complexity up into a library. Now, if you write a short, simple program that uses a complex (but long lived, time tested) library, is the program itself simple or complex?
I think the analog to the hardware world is the integrated circuit. You could make a simple line buffer with a single transistor or use an opamp. Is the transistor better strictly because it’s simpler? In some cases, depending on design goals and other constraints, it could be better… but in other scenarios the opamp would be better, despite the “complexity”.
As such with relay or electrical bypass - if you don’t buy my “separation of concerns” notion, and stand by the idea that the 3PDT is simpler, I still think it’s a stretch to call relay/electrical bypass “complex”. It’s a long-solved problem using commodity components exactly per their design. Nobody’s designing bypass schemes with rare/exotic components, or using parts in an “off-datasheet” manner.
All the above is a big part of why I personally prefer relay or electrical bypass schemes. And I simply prefer soft-touch momentary switches. But I’m certainly not looking down my nose at anyone who takes a different route. My preference is indeed more complex in terms of build time and component sourcing. And despite what I wrote above about 3PDT reliability, I’m quite sure that I’m never going to actually wear out such a switch! I’m not a pro, and even if I was switching my pedals 50x/day, every day, I’ve built so many pedals that I have backups for my backups!
Edit: I might be wrong about my 3PDT MTBF assumptions. Using
this 3PDT Pro switch as an example, the verbiage is "Mechanical Life: 30,000 MAKE-AND-BREAK CYCLES". Not sure if that means MTBF as I assumed, or it implies rated for minimum 30k cycles? Also, the heat resistance is impressive: "We dipped these half way up the plastic casing in our wave solder bed for 15 seconds with no deformation and perfect continuity tests. Lugs stay firmly in place even under extreme heat." I suspect you could in all likelihood generally get quite a bit of mileage out of these switches.
Also, comparing to a momentary SPST (such as typically used by relay or electrical bypass schemes), such as this
Pro-Grade SPST Momentary, it says "50,000 mechanical cycles; 6,000 electrical cycles". Not sure what that means exactly... why would you use it only mechanically, but not electrically? And if the actual reliability is dominated by the lower 6k number, that's disturbingly low!
However, I do think the two-wire momentary SPST is generally easier to replace. It's
almost a wash compared to 3PDT with most PedalPCB boards, especially if you use
the breakout board - then you only have to re-solder 6 wires. Three times the work of a 2-wire SPST, but you'll probably spend more time on setup and cleanup either way. And if you use modular connectors (like JST), and carry a spare, then it's theoretically field-serviceable at the gig.
On the other hand, some DIY effects, I won't mention any names
cough cough Aion cough cough will put actual circuit components on the switch breakout board. Which means you either have to un-solder the actual switch (very painful), or get a whole new board, or try to DIY hack the equivalent circuit components.
It's interesting to me that the reliability for SPST momentary switches isn't much higher than 3PDT (or even significantly lower, if that 6k number is relevant to us). Ignoring the numbers I wrote,
intuitively the momentary SPST should last longer because it's inherently simpler. I mean you could rig up a DIY momentary SPST with a couple paperclips or lengths of wire in a pinch. It's a spring and contacts. I wonder if that 6k number assumes high-voltage switching, where you're likely to have some arcing and therefore fouling of the contacts over time?
