Who's got short-shorts? Whose flag is at half-mast? The world of path-craft.

[Mark Hammer]

A recent thread on a broken Ross compressor got me to thinking.  What is the universe of "shorts" and "opens" in stompboxes?

By this I mean a complete exhaustive (or as close to it as we can get) listing of the many ways that electronic paths that should not be connected GET connected and electronic paths that should be intact end up NOT being intact.  If we get this right, we can provide a list that would be suitable for troubleshooting, based on how one produces a pedal.  In other words, these are not things that are "wrong" with a cap or transistor pinout, etc., but rather things you didn't planon doing when you put it together, but they are preventing the pedal from functioning nonetheless.

For example:

• On several occasions I have had PCBs made from thin-gauge phenolic bend enough to fracture traces.  If the board is small enough to be stiff, you may not have a problem, but if it is big and rectangular enough to be bendable you may run the risk of breaking a trace without realizing it.
• Press and peel is a wonderful advance in PCB-making.  It IS possible, however, to lay down enough toner, and apply enough heat, that the toner splatters outwards and creates unintentional bridges between traces and pads.  Sometimes these can look like intentional connections as you're making the PCB.  You need to ask yourself, before you etch "Why would this pad/trace be connected THERE if the other end of it goes THERE?".  Then scratch away any superfluous toner that might result in a solder bridge being permanently etched. 
• It also happens that the bridge isn't necessarily implicit in the toner pattern (as laid down) but the board can look fine so you take it out of the etchant prematurely.  Always have a copy of the PCB pattern handy when etching for visual comparison.
• Sometimes you end up needing to stick a d-s-g FET into a PCB where a d-g-s is expected, or an e-c-b bipolar where an e-b-c was expected, or vice versa.  The device will work if you give the legs a little twist and get the right pins to the right pads, but the extra leg length and twisting can sometimes result in pins touching eachother unintentionally.  A good practice is to slide a little bit of wire insulation over the leg before you solder it to the board.
• Many people have experienced the misery of it-worked-before-I-stuck-it-in-the-chassis.  Why does simple installation into a box stop a working board from working?  Sometimes shorts are created in a variety of ways that we can't see once the lid of the chassis is screwed in.  For example, as you tighten the nut on a pot or toggle or jack, sometimes the tightening will rotate the component juuuuust enough that it touches something else.  Jack lugs can touch posts on the side of the chassis.  Pot lugs can touch the cover or other lugs of adjacent jacks of switches or the side of the chassis itself.  Vertically-installed resistors can sometimes short out against panel-mounted controls when you jam the board in the box.  PCBs that juuuuuuust fit the available space can have traces along the edge that short out against the side of the box.  And so on.
• Sure the IC socket looks properly installed from the copper side, but is the IC sitting in it making contact between ALL the IC pins and ALL the socket pins?  Sometimes when we go to insert an IC into a socket, an IC pin can get bent underneath the chip such that not all the pins are inserted.  Alternatively, sometimes all the pins DO go in the socket but they don't make perfect contact.  That could be because of oxidation on the IC pins or socket pins, or because the IC pins are not "springy" enough to firmly grip the IC pin.  In the world of electrons, that 10 micron gap between pin and contact is like the Grand Canyon.  In other words, all visual indications are that contact is being made, but really it isn't.  Certainly one way to avoid that is to use the more expensive machined sockets instead of the cheap and more widely available spring-leaf type, and if the circuit and your soldering technique can support, skip the socket and solder directly to the board.  I've also found that scraping the surface of the IC pins with a utility knife until you see some shine isn't such a bad idea, and have also recently found that a commercial substance called "Stabilant" can be applied in tiny droplets to the pins and bridge that pesky 10 micron gap.

Well, that's an idea of what goes here in the list.  Feel free to contribute your own experiences and solutions to add to that starting group.  If we do it right, this could be a sticky or form part of the FAQ.

[Mark Hammer]

What, NOBODY has any tales to tell of shorts or open circuits and how they go there?

[jlullo]

mark,
one of the problems that has happened to me a few times is filed under "It worked before i stuffed it in the box".  i've noticed that some of my crappy solder joints will get a little wiggly on my where i've soldered wire.... specifically on the input or output!  I'll be going crazy trying to find what the deal is, and then i'll wiggle the input wire, and Voila!

Jonathan

[sfr]

The worst offender for myself for a long time was junky Radio Shack 1/4" jacks - tightening the nut often caused the solder tabs on the inside to rotate (particularly if a nut came loose over time and was attempting to tighten it from the outside only - a stupid practice, but one that I've been guilty of) and short against the enclosure, other offboard components, the board itself or even other tabs and contacts on the jack.  (The newer open jacks my R.S. now carries, while still not the quality of the switchcraft I prefer, seem better in regards to this problem)   I almost exclusively use Switchcraft jacks, with lock-washers.

Whilst still "my fault", it's really the *only* problem I've ever found where poor components where partially to blame.

Given the number of issues that can arise from trying to cram more stuff in an enclosure than can fit comfortably, as well as easy  to make goofs in offboard wiring (as well as the fact that I don't always know until the build is complete whether my new pedal actually deserves a place in the my chain) I have made a habit of simply wiring in, out, power and ground wires to the board, and using a breadboard or clips to attach them to a set of jacks that I already have wired here, to make sure that a new pedal functions as it should *before* I box it up.  That way if there's a problem to begin with, I've already taken off-board wiring and shorts against the enclosure out of the equation.  And when I do box it up, if something fails then, I can probably narrow down the list of possible faults. 

Until I recently purchased a nice pair of wire strippers, I occasionally had problems with nicking the conductor on wire when I stripped it, which would eventually (sometimes well after a build was boxed up and used) break right where I stripped the wire.  Often times not completely, or the wire still rested in the general area, resulting in intermittant contact.  (Pedals that cut out as they get bumped or jostled) 

Burned insulation (from soldering inside of a fairly "rats nest" box) is also another problem I've seen - allows wires to short out against the enclosure or other wires, again, sometimes it's an intermittant fault.

[R.G.]

I started to type some things in here a couple of times, but stopped. The more I thought about it, it's like the song, the Gambler - every hand's a winner and every hand's a loser.

Every single part in a circuit, including all wires and traces is potentially open or shorted in some way. I guess a list of subtle, sneaky ways is good to put down.

Ones I have known and hated:
- a thread of un-etched copper so thin it cannot be seen except with x10 or greater magnification
- solder balls dripped off an iron and rolling around inside a metal box almost invisibly ("It only quits when I shake it like this...")
- wires broken inside the insulation so they look OK from the outside
- a carbon comp resistor cracked through the body
- polyester caps self-heal from overvoltages and currents; mostly, anyway
- the bonding wires inside an IC would lose contact when the chip got hot and the wire expanded
- A tent of solder hiding an open solder joint
- a trail of carbonized dust across a tube socket looks like a short to a tube grid
- a MOSFET input is even higher impedance than a tube grid - humidity in dust can "short" these if they aren't pulled to some voltage with a resistor or guarded with a guard ring

[darron]

something stupid that gets me sometimes... when you are working on something in the enclosures, the input jack is NOT on the left any longer! it worked fine out of the enclosure and bypass still works :S

[sfr]

I think the problem with these types of lists is that a lot of folks come here with a non-working build and want us to tell them what's wrong with it, and can end up with other folks listing every possible permutation of what can easily go wrong.  While it's good to realize what can be easy faults to check for, I think it's more important to drill into peoples head how to identify what parts of the circuit are broken or shorted so they can narrow down where to start trouble shooting first, rather than going through every possible scenario until it's fixed. 

I learned this hard way a few years ago with a Small Clone - no chorus.  Blindly rewiring did nothing.  I had fixed previous build problems (mostly in the switch wiring) by using a multimeter and identifying where my output from the board was not making the connection to the output jack, but this was different.

I first had to learn how the circuit worked (understanding the concept of blending a clean and shifted signal, and an LFO to control that shifted signals rate) and then figure out what stopped (the shifted signal, or the LFO, or the clean signal) - in this case, I was getting a functional LFO, that was making it to the control of the shifted signal, and I was getting a shifted signal out, but it wasn't making it to the output to be blended.  After nailing this, I was able to use the audio probe to figure out that my problem lie somewhere around the depth pot, and the multimeter to find the connection in question that was grounding out.  (Turned out to be an extra bit of wire was making continuity between the pot lug and the grounding casing of the pot.)

Whilst I think it's good to keep in mind the things that can cause problems, like the bits that have been mentioned in this thread, I think it's more important to focus on identifying the general problem, and locating the portion of the circuit it is in so you can "drill down" as it were - I'm afraid newer folks will see threads like this and just start "shotgunning" at everything that might be a problem.  I know I've been victim to this methodology when I started, and I've seen more than one thread here recently that makes me think others work the same way.  Worse, I've seen these shotgunning approaches recommended to people who come here asking for help - (And I'm sure I've been guilty of recommending the same thing) before we start "fixing" things, we should be trying to identify the problem first.

I hope this post doesn't make me sound haughty - I have no illusions about being one of the experts.  Maybe everything I've said is wrong, this is just how I see it.  Again, I'm not saying the direction of this thread is bad, I just worry about how it can be perceived; particularly knowing how I would have perceived it a few years ago. 

[blanik]

you also forgot the dreaded: "the gold bar on my cheap resistors look yellow or the red looks orange so i put a 150 ohm oinstead of a 150 k..." 

happened to me... 

[newbie builder]

you also forgot the dreaded: "the gold bar on my cheap resistors look yellow or the red looks orange so i put a 150 ohm oinstead of a 150 k..." 

happened to me... 

That one happened to me once- i stayed up until 4 in the morning trying to figure out why I had almost NO volume coming out of my green ringer- I un-soldered and then re-did the entire board 3 times...only to finally realize that I had replaced a 6.2k resistor with a 620K (or something along those lines), and that I had in fact done it corectly each time, but with the wrong part. Put in the right resistor and it fired right up

[StephenGiles]

There's always "We wear short shorts" by Freddie & the Dreamers!

[brett]

Arggh.  I had a distortion pedal that had gone bad, but I checked it sooooo many times and couldn't work out what could be wrong.  The power supply cable had gone microphonic.
My most annoying and expensive stompbox mistake was this:  My amp went totally distorted, even at low volumes.  I checked everything and even changed the tubes (pre-amp AND power amp).  The wire to the LED in my fuzz pedal had broken and the fuzz had been on all the time even though the LED was off. 

[MikeH]

A little embarassing:  I wired up a pedal once (late at night after a couple of beers) and it worked great at first, then it became intermittent.  I figured it had to be the offboard wiring, since it worked fine before I boxed it.  So after messing around with that for a day, I finally realized the last wire I soldered to the pcb, the 9V+ wire, was not actually soldered.  I just stuck it through the hole and left it there.  D'oh!

[Mark Hammer]

I understand sfr's sentiments completely.  I was hoping, however, for this thread to be ONLY about continuity errors, rather than about ANY impediment to a functioning circuit.  Misreading a resistor value is certainly a common problem and worthy of considering, but again, I was hoping to have this be only about:

...things being connected that shouldn't be....

and

...things that ought to be connected but aren't.

Now that's not exactly as narrow as "how does this control on this pedal work", but it is relatively focussed.  I have to apologize here for the stupid thread heading.  Stephen caught the song reference, but I suspect only my classmates from grade 9 caught the flag reference.  "Flag at half mast" was our code-phrase to let someone politely know that their fly was open without giving away their faux pas in public.  In recent times, there have been far too many reasons to have actual flags at half mast, so I imagine that phrase caused some confusion and bewilderment.

Work is kind of hairy these days, but perhaps I'll be able to come back to this threasd once it has run its course, and distill all the things mentioned into a list for insertion into the FAQ.  In the meantime, carry on.

[Sir H C]

I hate the IC sockets that have PCB traces under them.  I almost always get shorts there.

In high impedance paths, solder flux plus moisture can create a path.  Or hide the short.  Clean that stuff off!

The wires through the PCB that are longer than the standoffs are high.

The large blob of solder that masks that the two lines that are now joined were just really close together.

The cracked copper on a stressed PCB.  Works fine most of the time until that connection gets slightly more open...

[Paul Perry (Frostwave)]

I don't know whether a short INSIDE a component counts - but they happen! (bypass caps - sometimes they explode, which is OK because, then they are out of the way.. but they can fail as a short sometimes.
But most shorts I see are on stripboard... i recommend checking each trace cut with a meter (suprising how easy it is for an invisible thread of copper to hold on!) and checking for shorts between each parallel trace after soldering.
Most bizarre open circuit I ever had was a machined pin socket, with no metal collar inside pin 14! plugged the chip in... and the logic 'kinda" worked, because the +5 that should have come from pin 14, was being supplied via other pins. Now that was a nasty one... since then, I try to put the meter on the leg of the chip, not the leg of the socket!

[petemoore]

  The one where the DMM becomes useless is when everything or many non-grounded things are grounded. Then you have...however may ways ground could possible get in there to find the fault in. exhaspaerating because the DMM just tells all about too many grounds, no more. When the close in light and mag-glass eyeing doesn't reveal it, I start lifting connections...trying the DMM again.
  Debugging a FF for half a day, as a way of finding a bad cable.
  Debugging a PS for half a day as a way of finding a bad DMM lead.
  Battery nowhere near the clip...that one got me more times than I can count, habits are hard to break, I used to disconnect it for everything except testing.
  Saying to myself for fifteen minutes, Wth'nF gadzillampS I built was werkin' fine, ..ok and still fine werkin' alone, only to realize [the wire/angle jack plate literally sparked my realization] I can't always just grab a 2jack plate [non insulated] like when connecting two amps to one cabinet...amps saying...hear us buzz: we can see eachothers ground points...way over there. ..just the other day I was surprized when I caught that jack connection...in the back of a 2way reflex cab., for daisy chainings a common metal jackplate is fine, for two amps into 1 cab...not so good.
  Trick is to take every mark on the schematic, and compare that to the board connects/disconnects. If it says 220k and that's not what's there..or if it doesn't show a connection between SP and box ground and it Is there...or even when it doesn't show say the next jack in the chain of the signal path as connected, and the cable doesn't connect..stuff outside the box is as important as everything, because some things  are assumed [as in you need to know, but it's not marked] and required doesn't mean they don't have to be there...
  We had a bad 4x12'' GB loaded cab. sounded like 9/10ths baked woofer...kinda, while taking it apart, strange things happened...the bad speaker made DC connect/lift tones..with no power source..except the other speakers. Pushing on one of the good speakers made no noise until the bad speaker coil connected and clicked. Turns out it was a bad flexwire between the the speaker terminals and the cone, something that surprized me...for a 'new-ish' greenback, DIY cab musta gotta cutter hooked to that wire?...we took a stretch of flexwire off a junk speaker and managed to tack the connection back together.
  Taking stuff for granted can cause serious short and long term hair loss.  "I know that speaker cabinet works"...4 hours later still knowing the unknowable, then finding by proof, that a few seconds of consideration could have '9vclick-diagnosed' the bad second link from the last link in the chain...a speaker wire in a 600w cabinet, off the jack.
  It sometimes takes me alot of extra time to debug a 'professonal's' system, it may be necessary to let them perform the 'decreasing radius circle - circuit jerk debugging method' repeatedly before spiraling down to consider logic, and allow the effective debugging technique..the one that starts at the end of the chain and tests every link in order<<. When there's a bad driver or driver connection, anything/everything seems bad...and with a real 'prophessional', this can go on four many hours, I refuse to not start with the driver...I'm a hardheaded one that way..lol..nope..gotta start at the driver...

[runmikeyrun]

sticking with the continuity thing:

Forgetting to plug your cable back into input/output/guitar/amp jack

alligator clip coming off wire when testing components

cheap battery connector wires breaking off where soldered (i always solder on a short lead of good wire to the ends of my battery snap leads, it's happened countless times)

leads breaking when putting into box

component shorting to box lid when putting the cover on

toggle switch internals broken/shorting

switching ground/tip on jack

forgetting to plug battery/ps back in (ok, not on the list but still...)

those are a few that have buggered me many times.  Now when i have a problem i start with these because about 10% of the time one of these is to blame!

[enigmur]

Make sure the schem or layout you are using is verified.

That one certainly got me....

[Paul Perry (Frostwave)]

    Debugging a PS for half a day as a way of finding a bad DMM lead.
 

I feel your pain, bro. Like having an intermittent in the ground lead on a cro probe... woah, where is that intermittent hum from? it is 200 times bigger than the signal

[Sir H C]

In college we had to build up a digital circuit on those solderless breadboards.  Well one group could not get their circuit to work.  Turns out there was a short to the metal backplane on some of their busses and that put them 3 weeks behind.