Setting things on fire (Giant Hogweed problems)

[Timko]

I'm a disc golfer.  When I was younger, I played quite a few tournaments, and practiced non stop.  I always enjoyed playing with people that were just beginning because I think it's a fun game to get hooked on.  I would often tell people "if you haven't lost a disc, then you haven't played enough golf yet."  It's sort of like pedal building.  If you haven't faced some issue yet, it's because you haven't built enough pedals.  So last night I faced a first:  burning out a resistor.

I am working on my second Giant Hogweed build.  I had the board populated, and was ready to begin testing using my testing rig.  I noticed when I hooked the power and ground up that there was something hurting one of my fingers.  I thought for a moment I shocked myself (I did on a preamp build), but quickly dismissed that since there's not a lot of voltage flowing.  When I ran my test, I got no sound.  I looked at the board, and noticed the resistor right after the 9V connection (R42 on the picture below) had turned a dark brown in the middle.  I put my finger on it, and realized that what hurt my finger.



I grabbed the Hogweed that's on my board, and tested it.  The resistor wasn't super hot.  So I began to replace things on my new build.  I replaced R42 and C28 (the big 220u cap), but that didn't make a difference.  It was getting late, so I didn't have time to continuity test every component in the power section, but I'll do that this evening. 

One of the reasons I picked up this hobby was to learn more about electronics.  From my Google research, it seems likely that my resistor got hot because the current going through it isn't being routed anywhere, or at least isn't being routed to enough places.  Does it seem like I'm on the right track?  Unfortunately, I'm concerned about leaving the power hooked up for an extended period of time and burning up something else, so I haven't used my multimeter to probe around much.

[alanp]

The resistor is burning out because it's being forced to draw more current by something in series than the (probably) 1/4W of power that it is rated for.

Check that the diode isn't the wrong way around, the resistors under R42 are correct, etc.

[Timko]

I checked the diode already against my working build and it's oriented correctly.  I'll check the other power supply resistors tonight.

[Timko]

I did some continuity tests around the power section, and I found a major problem.  The 9v and Gnd pads are connected somehow.  I went ahead and did some more testing on all the components in that part, and here's what I found.  Below I drew a line over all the pads that appear to be connected:



I double checked the diode again, and ensured that the diode has the same orientation as my working Hogweed (and the same orientation as the PCB).  I replaced the one that was on the board, and verified it worked before soldering it in.  Now that it's soldered in, there's no voltage across it (I assume a result of my short circuit). 

I want to first make sure that focusing on the power section should be my first move.  It seems like the issue is in here.  It seems that I should remove D6 and validate whether or not the pads without the component aren't connected.  If they are, I should remove C28 and try the same thing.  I wanted to ask for some opinions because I've already changed the components once and I am worried about burning out the parts (I should have measured all of this the first time I took off the parts.  Live and learn.)

[Timko]

I need to measure again, but it seems like the + end of C27 isn't connected after looking at the schematic compared to the troubleshooting picture.

[Timko]

I took all the components in the power section off, and measured the continuity of just the pads. The good news is that the pad that connect R42 to 9V is the only power connection I can find.  The bad news is that both pads of the diode (D6) show as connected.  Same for both pads of C28.

I'm looking for any bridges over C28 (since those are close together), but I don't see any.  There's also nothing bridging D6.

[Timko]

Here's my current state.  I've marked the pads that are connected to one and other in the power section with red circles.



The more and more I look at the board picture in the build doc, the more I think it has to do with C28.  That's the only place where the pads are close enough to have a bridge.  Unless I should be looking elsewhere in the circuit?

[drolo]

it does sound like there must be a bridge somewhere.
In case it helps here a re 2 pics highlighting the ground and positive power rails:

[Timko]

Thanks David!  I have a concert tonight (Radiohead!), but I'll be making use of this tomorrow evening.

[Timko]

I did more investigation last night for a couple of hours.  I got my new multimeter in the mail yesterday, which really helped.  My old one was going totally bonkers and not really letting me know continuity.

I've removed all of the power section components, and after probing around the rails that David posted above, I couldn't find a place where the rails seemed to be bridge by some careless solder.  I looked at the 2 places that seem like likely culprits: 

* C28, which has it's negative leg on the gnd rail and it's positive leg on the positive rail.
* R40/R41, which has one leg connected to power, while the other resistor is connected to gnd.

I've also looked for places where the rails come close to components connected to the other rail. I wanted to ensure that no careless solder had gotten in one of those holes that allows a trace to go from the front of the board to the back (I forgot the technical name).

I'm starting to remove components connected to either rail in any fashion at this point because I'm not sure of a good next step.  If anyone has any troubleshooting tips with something like this, I would love to hear suggestions.

For reference, here's the current state of my board, with a number of components removed.

[drolo]

this here looks suspicious. Might not be the culprit but could you check if there is maybe a bridge there?

[Timko]

this here looks suspicious. Might not be the culprit but could you check if there is maybe a bridge there?

Agreed.  That's the first thing I thought when I saw it.

If the pads go like this

1 2
   3

1 - 9V connection
2 - R42
3 - Anode side of D6

After a continuity test, 1 connects to 2.  1 does not connect to 3.  2 does not connect to 3.  So I believe this is not the cause of the issue.

[drolo]

No indeed ... also now that I'm looking at it, it would not have explained the short you have, only would have shorted R42, which would have not been the problem you have


I'm really sorry, it's hard to find something like this looking at a picture :-(


I have looked closely at the whole back of the PCB and my mind starts seeing shorts everywhere ... where there probably aren't ...


I have attached a pic of another suspicious spot on C28 another suspicious spot (although since you supected it you might already have had a closer look...)


Other than that I am attaching a pic of the 2 images with the 2 rails superimposed in case that might be helpful

[Timko]

Thanks for your assistance.  I appreciate trying to look at a picture to figure out what someone did wrong .

The squared pads you show are C28 and were one of the first areas I investigated, since D6 and C28 are the only 2 components that I can find that connect R42 to Gnd.  Those pads indeed show connection, but I've tried to run the soldering iron between the 2 pads to remove any solder bridges.  I also used an x-acto blade to cut anything that would be connecting those pads.  However, I will go home tonight and ensure that they do look all right, and may attach some new pictures of those areas.

[Timko]

Also, thanks for the superimposed lines!  I'll keep messing around with it .