FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 

Marbles meltdown smoke
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author Marbles meltdown smoke
Just finished Marbles today and had triple checked etc all connections and all the chips are the right way round. I cant imagine what might have lead to this disaster? THe only area i wasnt sure of was the 7 leds at the botttom of the board so made a guess as to their orientation. I put all the cathodes to the left and anodes to the right. The power connector was correct and no problem with other modules on the rack. With my doepfer psu 3 it flashes if i plug something in wrong. I as very methiodical in part placement but something as gone horribly wrong and i dont know where to start. I put a dab of ipa on all the chips and it evapourated really quick on the cpu (anyone know what the operating temp might be? Says max of 85c on the datasheet.)and was okay on the others. Checked the 3.3 volt line and there is a dead short to ground. Working my way round the board looking for potential suspects. I removed the Recom R-78E3.3-0.5 still 3.3v ground short. Also removed the bat54s and lifted the cpu leg near c12 top right hand corner, still dead short between 3.3v and ground. Checked 3.3_A to ground and it was fine, no short. Also j6,j7,j8 and j9 are lit and orange.

I have built Clouds and Elements and they work great, so i think my basic practices arent too bad. But there is always room for error in these compact builds. Hope someone can help shed some light on this. Thanks in advance. Cheers
You will learn a lot debugging that build. Dead Banana

Just breath (not the fumes wink and take one or two days till you feel redy to go for it.

Do you know what chip was smoking?

A 3v3-gnd short can be anywhere. Diodes are not the culprit. I would check the caps and the ics pins. Appling flux and hot air does wonders reflowing the smaller pins. Take it easy and go for it when you feel ready.

Also you can upload a hires photo, so more eyes checking out
HI MapacheRaper, thanks for the feedback. I finsihed the build on monday and have spent the whole week following traces. At one stage i was wondering whether one or more of the 100nf decoupling caps had failed short but couldnt find any evidence for that on the net. KInda burnt out looking lol. I hand soldering everything, as i have no experience with solder paste and dont have an oven, though heard of people using hot plates and have one of those. Most of the passives soldering is blobby, used plenty of flux and fluxed solder and highish iron temps. Heard that blobby can also be too much solder i used 0.5mm, checked out finer stuff on mouser but it was really expensive, so i had to put up with joints i wasnt happy with. The ic joints are quite nice though. I checked behind the pins with the microscope and couldnt see any blobs of solder, though i could have missed some maybe? LOdas of slightly wonky passives, only have a low clearance under my cheap microscope, so hard to manipulate them straight sometimes. Mostly work with my Rolson head magnifier and check with scope for passives. Using Eagle you can select where the 3.3v lines go, so followed that around and apart from lifting more pins on the stm the only thing left that i can see is C2. I had the front panel on when i powered it up ready for flashing, then i saw the smoke, though nothing i can see looks burnt? I did wash the board, to get rid of flux with ipa, it could have been some of that from under the stm? Maybe this washed some solder fragments somewhere under the stm? I dont know lol. very frustrating [/]
could be a number of different things. there are a few ways to go about troubleshooting it. if you have a sensitive enough mulitmeter, you can measure the resistance betwen 3.3V and GND, and it should go down as you get closer to the short. another way to do this is to use a current limited powersupply, attach it to the 3.3V rail, and then measure the voltage along the traces to see where it drops to exactly 0. if you dont have a super sensitive multimeter, you can try feeding that voltage into a VCO and listening for pitch change.

another option is to lift off the DAC and STM32. the short is either a solder bridge, a bad PCB, or a busted IC. this would help narrow it down. also, if this short is under one of the ICs, you wont see it without taking it off. if you inspect all of the components carefully, and see no evidence of damage, the smoke you saw was most likely from a solder bridge or bad PCB.
Aww man, sorry you have to deal with this! SMD is pretty much beyond my old weak eyes and shaky hands so kudos for being able to do this! thumbs up

I can't see anything that looks suspicious other then maybe IC1? What it that, a voltage regulator? Doesn't look soldered well but maybe it's soldered on the other side of the board OK? If it's intended to be on the pictured side of the board, did you flip it to make sure the pins are orientated correctly?

I'm only guessing so ignore if I'm wrong, best of luck! SlayerBadger!
Maybe a dumb question but did you check for shorts prior to populating the pcb?

- Jim
these pcb's are extensively tested, it's unlikely to be a pcb defect
HI johnlrice this ic goes on the other side of the board and i removed it to rule it out as a possible short. Thanks a lot for the interest. Appreciate the feedback. Take it easy
Hi Adam thanks for that i expected as much as other boards i got from them work perfectly, meant no disrespect to them. Cheers
Hi Jimfowler, no i didnt check, wished i had, but there you go, one for next time. Cheers
HI guest, thank you very much for all the suggestions. NOt seen anything specific to Marbles and troubleshooting, thats why i started this thread. Theres some great suggestions in there that i would not have thought of. Will get back to you tomorrow when i have tried them. Would love to test more voltages around the board when connected to my rack but the STM seems to be getting too hot, maybe its already toast, i dont know but dont want to make things worse. As anyone noticed their marbles getting hot, lol theres a euphemism in there somewhere. Take it easy. Cheers!
Don't know if the 3 pin part around D3 should be there or not.

I don't have the schematics but if there are series resistors between the power connector and the circuit check those. They can burn up if there is a short and to much current is drawn.

If potentiometers get too much current they can burn up and you won't be able to see anything. Might try ohming them out, but in circuit you might have to think about what you are seeing since it may not be the real value of the part.

One trick that sometimes works for burnt parts you can't see is just smell the board and that could get you close to the area the part is in.

Good luck.
I have twice run into a short in a Mutable DIY build, and the following suggestions come from those experiences. Thankfully I found the shorts by performing resistance tests before powering up!

Verify you have the correct flavor of BAT54, they come in various configurations and the wrong one can result in a short.

Remove the polarized cap or caps decoupling the shorted 3.3V supply and recheck for short. Its easy for solder paste to wick up under these parts.

You can use your nose to find the burned parts - they keep on stinking after the smoke is gone. Just use a small straw and snort around. Burned parts may have started out fine but became shorts when they failed so you will have to remove them.
Hi Fuzzbass, thats brilliant, was what i hoped for when i posted. The rest of the guys have come up with some great suggestions too. Yeah was beginging to wonder about c2 and the decoupling caps as potential suspects. I removed the bat54s for debugging and still the short, i got it off ebay as Mouser didnt stock them. Now i know Digikey does spreadsheet imports, i will try them out next time, as i have had to find replacement parts as they were obsolete or out of stock from them before. Who knows what damage as been done? I was trying what guest suggested above this morning but ran out of time to finish the whole test but none of the pots seem to be burnt they all test the same. Thought i might have found the lowest resisitance point at c2 but it could be the 100nf around it. Great you nipped the shorts in the bud, im over optimistic with my builds. Will be more thorough for sure in future, great you had direct experience, is reassuring. Thanks for the feedback. Take it easy. You guys are ace!
collumn2016 wrote:
Using Eagle you can select where the 3.3v lines go, so followed that around and apart from lifting more pins on the stm the only thing left that i can see is C2.

The 0402 C1 looks like a good thing to rule out early, especially if you have another to replace it.
Have you got some sort of microscope? There may be a tiny bridge somewhere that you haven't found yet.
one more quick thing: you mentioned that the STM32 was getting warm. typically, if the short is outside of the chip, it would not be getting warm. but, you have the analog 3.3V rail going to it as well, so it was probably being reverse powered by that rail, generating a lot of heat. if you decide to apply a current limited power supply to the digital 3.3V rail, be sure that it does not power up the analog 3.3V rail as well.
HI all. I removed c1,c2 c9, c13, c12 and c16 and there was still a short between 3.3v and ground. Had to bite the bullet and remove the STM and used chipquick, which is a bit messy but worked in the end, my first success with it. Tested for shorts after clean up and the short as gone. Not sure what the sequence of events was here and whether a dodgy Bat54s may have caused the short in the first place and caused a failure to ground in the STM or whether there was a solder bridge out of sight under the STM? Im wondering whether any short on the 3.3v could have damaged the DAC now and whether i should replace that too? Thanks a lot everybody for your help.
probably worth replacing the dac too
Are you using the correct variety of BAT54? It must be the ā€˜Sā€™ type.

Just a thought.
Hi Adam, thanks for the suggestion as i did worry that it may have gone. Take it easy.
HI cnicht yeas i made sure i was using bat54s but you never know what you are getting on ebay, had a big mistake with some leds for my Plaits attempt, which i have problems with, over on storder=asc&start=50. MEasured the normalisation pulses from it (plaits), which is what i think it is there for? Cool to see. Take it easy.
sounds like either a bad chip, or a short underneath you couldnt see. the latter is more probable. you can measure the resistance between the Vss and Vdd pins on the IC to see whether the IC was the culprit. if the IC seems good, you can try reseating it. since youve gotten good at chip removal, it wouldnt be too much of a time sink if it was bad. either way, the DAC shouldnot have been effected, as the supply voltage for it never went above 3.3V.

and congrats on finding the problem.
HI Guest, been waiting for parts before i next commented and been working on getting my Plaits to work. Also received my Tides PCB, which is my next project, hope everything goes well with it. Will keep you all posted after i replace the STM and a few passives in Marbles, i will definitely check for shorts before powering it up this time. Guinness ftw! Cheers
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 1
Powered by phpBB © phpBB Group