How to test final product?

[diemilchmann]

Let's say i finish a pedal. Something not too special, random distortion or phaser or something. What would be some good steps to take to make sure this pedals will almost never break, e.g. power shorts, switch failure, dropping it etc?

Thanks,
Julian

[R.G.]

Wow. Talk about bringing back memories. I read this and it was 1986 all over again. Japanese factories were kicking the collective posteriors of USA factories by producing stuff that worked when you opened the box, and kept on working. In retrospect, it seems a little silly to even contemplate why the factories that *weren't* producing stuff that worked first time, every time, hadn't already had their butts kicked in the market, but I guess there's a time for everything. 

Let me give you a few high points of the lectures I had to sit through over and over and over on quality and reliability.

1. You can't realistically "test in" quality or reliability. You have to design them in from the start. Otherwise, the cost of repairing the failures gets out of hand very, very quickly.
2. "Quality" does not mean high performance or a peak of achievement. Quality means -*conformance to requirements*-. The corollary is that if you can't write down what the requirements are, you can't design to them, and you can't achieve quality. This idea was the thing that ended up producing the ISO9000 family of standards for manufacturing quality.
3. There are many design engineers who specialize in design for reliability and spend a whole professional career on it. To do a good job of designing for reliability, even modest reliability, you have to first know how things break and then design those possibilities out. Knowing the likely failure rates and modes of your components - which varies from maker to maker on each component! - is kind of a first step on how to design things that work and keep working. Bathtub curves, infant mortality, accelerated aging, mechanical stability, shock and vibration testing... ARGHH~!!!

I started typing in a bunch of stuff, but this is such a deep subject that I get frustrated at what to type in. The only short and sweet way to say this is - do it like pedals you've seen before. If anything flexes or moves around, it will break eventually. Design for electrical reliability is another very deep subject all on its own. If it gets hot, or voltage or current stressed, it will fail eventually. But "stressed" is different for each component; that's what the maximums on datasheets are for.

You're asking good questions, but the answer are not small, they're huge.

[Taylor]

RG's right of course, but it's also useful to give a quick mention of things we think are most important:

-stomp switches and audio and power jacks get the most abuse from musicians. These should therefore be wired offboard, so stress doesn't crack the PCB.
-3PDT stomps are themselves a weak link. Design them out.
-depending on your philosophical perspective on "dumb-proofing", design a pedal that can be connected to wrong power supplies and not die. IMO Aphex wins here: their pedals will not just take wrong power supplies, they'll actually run off of a very wide range of voltages, DC or AC and regardless of the polarity of the plug.

Those things cover the majority of broken pedals I've seen in my life.

[CynicalMan]

Everything RG and Taylor said is right, but I think for DIY pedals the main cause of failure is poor soldering technique. I'm not pointing the finger at you specifically or at anyone, but a lot of the DIY equipment failures I've seen, my own included, have been because people just don't have enough experience with soldering to be able to consistently solder a reliable joint or to tell when a joint isn't reliable. So, if you deal with that and eliminate moving/flexing parts, your builds will be as reliable or more reliable than a lot of commercial effects.

Another thing to keep in mind is serviceability. If you use short, efficient, and organized wire runs, well laid-out pedals, and good records, you'll be able to get things up and running much faster when they break.

And, a pet peeve of mine, don't use positive ground! A pedal that can be broken by daisy-chaining it is terribly unrelaible. If it's important, use this: http://www.geofex.com/circuits/+9_to_-9.htm

[Processaurus]

I try to imagine everything that could go wrong, going wrong.  Stomping on the controls, stomping on the patch cables where they go into the jacks, having heavy stuff stacked on top of it, getting dropped down some stairs, having a full beer spilled on top of it, wrong adapter, being taken apart by someone else who won't be super careful of the wiring.  Kind of like good parents imagine and address all the dangers for little kids around the house.  DIYing stuff is a great way to learn design for reliability, because you're the one who directly benefits from the work put in.

[petemoore]

  So 'get rid of anything that fails' ?
  That's everything though !
  Starting with the battery, next...who's next ?
  No-one part wants to tell you that, they're all trying.
  Electromechanical, then electrolytical [or vice versa].
  Electromechanical can sometimes be made to be near fail-proof, capacitors...they may require replacement.
  The frequency decreases markedly once we're past pots, switches and capacitors, and have been lucky enough to take for granted jacks/plugs/cables/wires all are intact and functional.
  We're left looking at transistors and resistors box and phelonic or other board material. If there is rohs solder in there...then we may have to look at that also.
  Transistors can be damaged with electricity of the wrong type, resistors...pretty darn resistive, and the box will last for quite some time, I vote either the aluminum or the phelonic will survive the longest in their present form.

[Hides-His-Eyes]

Never underestimate the power of idiots. What damage could an idiot do to your pedal? Prevent it.

Could an idiot unscrew the stomp switch and then stamp on it, snapping the board? Fix it. Could an idiot lose the rear cover and snag a wire on something? Fix it. The pedal will be stamped on many times; could an idiot physically stamp on your toggle switches? fix it. Use big knobs to guard your pots from kicking damage.

Most component failures we encounter will be electromechanical or due to incorrect power supplies.

One thing I really don't get; most power supplies provide 9.6v. A Schottky diode drops 0.3V at reasonable current draw. It costs very little. It takes up little space. And yet people still seem to use suicide diodes for polarity protection! Weird.

[petemoore]

Most component failures we encounter will be electromechanical or due to incorrect power supplies.
  Using a typical DC jack ensures competitively high failure rate [stiff competition here, /many or most pedals feature a DC jack] an Excellent source for catastrophic PS effects.
  Get rid of the electromechanical components which would probably be the #2 failure rate thing...that you can sometimes do something about, once the DC jack feature convention has been transcended.
  I should say the #2 failure rated failure-prone-components...that you may do something about...are electromechanical [switches pots connectors]...unless a way is figured out to not use electrolytics.