DIY Quality Control

[amptramp]

Be careful what you wish for:

http://src.alionscience.com/pdf/MIL-HDBK-454.pdf

A little light reading.  A lot of items are not applicable to all cases, but it is a bit of light reading.  All 194 pages.

i want Do it yourself, not nuke them yourself

Look beyond the obvious: this military handbook is used by manufacturers of medical equipment, aircraft and any equipment where durability and reliability is paramount.

Leave the nuclear option to me.

(Hey, maybe that should be my signature!)

[Krallum]

I place my pedals on only the softest
down

[amptramp]

Traynor had the right idea - amplifiers and effects units do not live in the MIL-SPEC world of environmental extremes as much as they live (and die) in an environment of abuse.  They used to throw their amplifiers off the roof of their factory as a qualification test - if it passed, the design was qualified to go into production.  If not, back to the drawing board.  Military equipment is usually qualification tested for the following environments as a minimum with other tests being conducted as applicable that are specific to the program:

1. Operating temperature range (a Fuzz Face is a pathetic test subject for this)
2. Storage temperature range (larger than operating temp - think of leaving the equipment in a van in cold or heat)
3. Sinusoidal vibration (the unit sits on a voice coil driven by a sinusiodal generator.  The voice coils are usually 36" diameter and driven by 5 KW amplifiers with a response down to 5 Hz)
4. Random vibration (the unit sits on a voice coil driven by a random noise generator, either analog or digital in all three axes)
5. Shock (pulses in each direction on three axes making a minimum of six pulses)
6. Water ingress
7. Humidity
8. Salt spray
9. Fungus growth (a lot of gasket materials are nutrient to fungus)

and tests are continued for the following interference-related items:

1. Conducted emissions
2. Conducted susceptibility
3. Radiated emissions
4. Radiated susceptibility
5. Power impulse spilkes
6. Reverse battery power

This is in addition to testing to ensure the unit performs the function it is intended to.  If you follow the discipline of designing for these things, even if you do not test for them, the design should be adequate and failures would be due to workmanship only.

[petemoore]

  Philosophy...do everything you can to prevent it, but count on failure. This is where duplicity out-reliables single unit performance.
  As painful as it is to think of 'reliability' this way, at some point it inevitably becomes...inevitable, the spare carried is the spare you wouldn't otherwise be able to attain easily or at all, at gig-night.

[PRR]

> They used to throw their amplifiers off the roof of their factory

No. Only the second floor.

The mil-spec is a nice write-up of what to look for. But it omits practical examples because it must cover "everything".

BTW: this thread should be "RELIABILITY". Not "quality control".

Quality control means if your contract has allowance for 1% duds, you have exactly ONE dud (no more and no less) in every box of 100.

Reliability means it works in a foxhole (or on a stage).

The #1 failure used to be solder joints. Too-too-too many builders are satisfied with re-flowing bad joints in hope of getting lucky. That's bad workmanship, and just won't cut it on a room-size computer (yeah I been around) with millions of solder joints. Even with just a dozen solder joints, EVERY one has to be right FOREVER. Each joint has to be righteously-right all the way to the base metal. Not globbered-over.

Next failure in portable gear is wires breaking off. There is a LOT of vibration from a long wire on a bad road and this concentrates on the ends. You can bend copper once, a hundred times, but keep bending and it WILL break off.

These days, sadly, there's a lot of bad parts. Transistors and resistors are perfect, but jacks are terrible and most pots are hardly worth screwing. Some jacks don't grip right from the start. Others tarnish quickly.

There's also loose jacks and pots, and jacks/pots with such cheezy hardware that they can't be firmly tightened without stripping (even breaking). If you are serious about your work, don't buy from manufacturers who are not serious about their work.

Builder mistakes ought to show up at Quality Control, but some mistakes "work" for a while before failing. Electrolytic caps can _sometimes_ be put in reverse polarity and work for a while. 22K where should be 2.2K may "work", but not-quite-right, and may quit working on a hot bad day. (The resistor works but whatever it is biasing may be "working on the edge" rather than near the middle of its happy-zone.)

[Kitarist]

Hm interesting thread

[jefe]

BTW: this thread should be "RELIABILITY". Not "quality control".

Quality control means if your contract has allowance for 1% duds, you have exactly ONE dud (no more and no less) in every box of 100.

Reliability means it works in a foxhole (or on a stage).

Agree... I wasn't going to say anything, but I'm kinda glad you did.

This thread is really about designing a reliable product. Quality control is more about making sure the finished product meets the original design.

I'll stop there, lol...

[diemilchmann]

board mounted pots/ good switches/ appropriate soldering iron time/foam/zip ties always are great for me!

[svstee]

Big switches. Even good quality subminis and minis go out WAY before any of my 25A rated massive switches that are usually around 10x25 mm. Downsides are that they are big, heavy, ugly and some of them need all the heat my Weller station can put out to solder wires to. Never had one break on me yet. Is the hassle and wasted space worth it? It is for me, but many people don't think so.

[armstrom]

it's not commercial level production, but you can go a long way by doing a few basic things, IMO.
1) Don't use solid core wire. It  just breaks far too easily.

2) Even if you use stranded wire avoid bending wires too much. They will break. It doesn't take many bends in the same place to break solid wire, and even stranded wire will become "solid" when you solder it to the board (solder wicks up into the wire making it "solid" near the board, right where it is likely to bend while you're fiddling with your pedal)

3) Don't let your board "float" in the enclosure (see #2 for why). Standoffs are the way to go. As others have said, the adhesive ones are the easiest but you run the risk of the adhesive letting go, but you don't need to deal with screws on the top. You could mount the board to the lid with screw standoffs, but then if you have to remove the lid to access the battery (Assuming the pedal has a battery) then you will be risking bending the wires too much (see #2).

4) As others have said, protect soldered wire connections with heat shrink tubing. This protects against accidental shorts as well as provides a little bit of "strain relief" near the solder connection.

5) Don't "float" individual components off the board (examples of this are caps soldered between pot lugs, current limiting resistors for LEDs are often left "floating" off the board).

Others have given many additional steps you can take but I've found that the above five, combined with good soldering and the use of a proper PCB (ideally one from a board manufacturer with a solder mask, but a good home etch can work just as well if done properly) and good design (reverse polarity protection diodes, PS filtering, and a multitude of other things that can help your circuits survive abusive users) can go a long way toward eliminating headaches and stress when building.

If you want to go above and beyond and build even more professional pedals (again,  not commercial quality, but the "next level" so to speak above "typical" diy techniques)

1) Minimize wire connections when possible, but keep them where they are important. For example, board mounting pots and then using those connections as the primary means of supporting the board is just asking for a failure, in my opinion, so using wires to connect pots is important to me.

2) Only solder one wire per lug on off-board components. This sounds silly, but having two or three wires (I've seen more!) forced through the solder lug on your input jack it a recipe for disaster. Trying to keep all those wires properly connected to the lug and to each other is a difficult task. It also complicates repair work as you're more likely to damage the solder lug trying to remove more wires than were intended to fit in the hole.

3) If you use a 3PDT for true bypass switching with LED then buy/make a PCB to handle all the connections to/from the LED, input output jacks and power jacks. This goes hand-in-hand with the suggestion of using one wire per lug for off-board wiring. Plus, it looks SO much better than a rats nest of wires snaking around the already crowded 9 pins on the switch. I would consider this an acceptable exception to the rule of not using solder connections as the sole mechanical support for a pcb because there are 9 connections between the board and the switch in a very tight area, the solder lugs on the switches we use are much stronger and more rigid than those of typical board mounted components and the switch PCB is likely very small and light.

4) Use a battery drawer/compartment. This adds machining time/costs to the enclosures but it removes any need for the customer to fully open the box.

5) Use wire to board connectors to attach all your offboard wiring. If you need to fiddle with your circuit you can easily unplug ALL the offboard wiring to access the PCB and avoid bending/breaking wires. Seems like overkill, but it makes working in the tight confines SUCH a pleasure when you can simply unplug some wires to get them out of the way. This doesn't have to be difficult. You can buy pre-crimped molex connectors with a matching through-hole PCB connector from places like SparkFun for a buck or two each. Yes, it adds cost to the pedals, but it also adds an extra level of professionalism to your work when someone does pop the hood (in addition to the benefits described above)

6) Properly secure all your offboard components. Loc-tite or locking washers on metal threads (pots, switches) a dab of hot-glue on plastic ones (9V DC jack). And if you really want to be safe, KEEP the little tab on your pots and drill the matching hole in the enclosure. Over time if your pots do work loose it's easy for a pot to rotate causing extra stress on the wires or possible shorts against the enclosure body.

I know, that's a LONG list... do as much or as little of that as you like, I haven't followed all of these myself in my recent builds, and it has cost me in debugging time... So some of these items I'll be adding to my next build, some I won't. Lessons learned and all...

[DougH]

Don't "float" individual components off the board (examples of this are caps soldered between pot lugs

Nothing wrong with that. I do it all the time with no problems.

My rule of thumb is don't float a connection between 2 components in mid air. As long as both connections on a cap are connnected to a stationary point like a pot lug, for example, it's okay. Sometimes if I want to point-to-point wire a couple components like that in an amp, I install a single non-grounded terminal lug to attach the "free" ends of each component to. Works just as good as the other terminal strips I use in the amp, which is just as good as perf, pcb, or pot lugs in a stompbox.

For LED resistors I connect one end to a ground point, the other to a wire that goes to the switch (or vice versa), and the end that is connected to the wire gets heat shrink to stiffen the joint and insulate it. If the resistor lead is long enough I just slip a piece of stripped wire insulation over it and connect it to the switch. No problems.

[Steve Mavronis]

Don't let your board "float" in the enclosure (see #2 for why). Standoffs are the way to go. As others have said, the adhesive ones are the easiest but you run the risk of the adhesive letting go, but you don't need to deal with screws on the top. You could mount the board to the lid with screw standoffs, but then if you have to remove the lid to access the battery (Assuming the pedal has a battery) then you will be risking bending the wires too much

I'm not a big fan of adhesive pads or foam so I've mounted them using right angle PCB mount pots as a kind of standoff.  I'd also like to look into using plastic card slot holders if they can be fixed to the inside walls of the case, or just sandwich the PCB between plastic standoffs that fit against the top and bottom of the case to keep the PCB secure and snug inside.

[DougH]

I'd also like to look into using plastic card slot holders if they can be fixed to the inside walls of the case,

Don't remember the model # but Steve used to sell an aluminum enclosure with "ribs" along each long side for doing exactly that.

[Steve Mavronis]

I'd also like to look into using plastic card slot holders if they can be fixed to the inside walls of the case,

aluminum enclosure with "ribs" along each long side

Yes but then you are restricted with an even smaller PCB area to work with, unless you want to do multiple board layers inside one pedal.

[stringsthings]

There's also loose jacks and pots, and jacks/pots with such cheezy hardware that they can't be firmly tightened without stripping (even breaking). If you are serious about your work, don't buy from manufacturers who are not serious about their work.

+1

tip #248:  as someone mentioned before, look for the mechanical parts and spend more time/money making sure they're bulletproof ...

example A)  in the venerable Echoplex, the moving parts would most often develop problems over time; no pun intended ... intend ... inten ... inten ... inte ... in ...  

example B)   a cassette recorder would most often develop problems with the machinery that moved the tape ...

example C)   an electric guitar often needs work on the input jack ... tightening ... cleaning ... and the occasional replace ...

[armstrom]

Sorry to dig this up but I thought I would share my efforts to improve the quality of one of my pedals. I recently had a build come back from a friend of mine due to intermittent problems with the effect being nearly silent when not bypassed. I rushed the initial build too much and had some wiring problems. I decided I didn't want to risk having any more problems so I completely redesigned the pedal from the ground up (with the exception of the enclosure).
Here are the results:

Here's a link to the "before" shot (actually, I had started to disassemble the old pedal but kind of slapped it back together to take a before shot)
http://www.mr2-power.com/oldguts.jpg

Here's a list of the changes I made:

• Redesigned PCB for smaller layout and to consolidate all offboard wiring (with the exception of the pots and a small toggle switch) into a single connector. This has 4 pins, Ground, Power, In and Out and is located at the bottom-left of the main board.
• Created a daughterboard to eliminate the nest of wiring associated with 3PDT switching. No jumper wires. The board also handles all of the input/output jack wiring as well as power wiring and holds the LED and resistor. It connects to the main board through the connector I mentioned above.
• There is only one wire per solder lug on any offboard component. If two wires need to connect to the same net the connection is made on the board, not by twisting wires together or cramming two or more wires into the same solder lug.
• Every critical offboard component that might need to be serviced is connected to the main board or daughterboard with molex connectors. I can swap out a scratchy pot or a damaged 2.1mm power jack without any soldering
• The battery snap is connected to the daughterboard and is "strain relieved" by passing through a second set of holes in the PCB before being soldered to the board.
• Mounted PCB to a pair of standoffs that have been epoxied into the enclosure to avoid problems associated with a "floating" board
• Switched to 9mm pots to save space (not really a quality control thing, but it helped with the layout)
• Used a more appropriate 78L05 regulator rather than the huge 7805 I had in there before. Not that there's anything really wrong with a 7805 but it added extra height and increased the risk that it would fail by having its pins bent too often.
• No solid core wire! That was my main issue with the previous build. One of the solid core wires broke somewhere in the insulation due to trying to work around the nest of wires and repeated bending of the wire. That connection then became intermittent and would fail if the box was subjected to shock (like stomping the switch)

Hopefully this helps give people some ideas. Sure, this was probably overkill and you could easily do without the individual connectors for the pots or some of the other things, but it's what I chose to do

-Matt

[StereoKills]

There is a NASA build guide somewhere on the net that shows a lot of their build quality requirements for their electronics (hopefully someone has a link). Every time I start getting an ego about me I'll read that and realize my pedals are not YET ready to be shot into space.

http://workmanship.nasa.gov/lib/insp/2%20books/frameset.html

[jasperoosthoek]

A friend asked me to fix his broken line6 super distortion. The boxed was built like a tank but several surface mount ICs were burnt beyond recognition. For no obvious reason...

[rousejeremy]

Bump. This is a great thread. I started drilling holes for the little tabs on pots and some switches, started using standoffs and hot glue for wires etc. I feel much more secure about my pedals surviving in the hands of the people that buy them as a result. All learned from this thread. Great stuff.