ITSY BITSY 32U4 5V POWER FAILURE
Moderators: adafruit_support_bill, adafruit
Please be positive and constructive with your questions and comments.
- Kenya7c6
- Posts: 3
- Joined: Mon Aug 02, 2021 9:58 am
ITSY BITSY 32U4 5V POWER FAILURE
I have purchased 25 Adafruit Itsy Bitsy 32u4 5V boards for use in a relatively large production run of devices I have made. The Adafruit boards are part of a bespoke PCB I designed which provides power to an ESC/BLDC as well as serial comms between the ESC and 32u4. The 32u4 board is powered via a dedicated power supply (Traco TSR0.5-2465) with appropriate input and output decoupling capacitors. I have had two Adafruit boards presumably draw far too much current and burn up their power traces and also the power trace which goes between the power supply and Adafruit board. Has anyone else seen an issue like this? Essentially 10% of the boards I put into working rotation have failed, and I am trying to figure out if this is an issue with some part of my design or an issue with something on the Adafruit Itsy Bitsy 32u4 5V board (e.g. a dead-short-failed linear voltage regulator on the Adafruit board). To burn and lift a trace of this width (0.025" and 2 oz. thickness) would require over 3 amps of sustained current. One of the boards failed like this almost immediately on first power up, and the other failed after several weeks in service.
- Attachments
-
- IMG_004.JPG (269 KiB) Viewed 206 times
-
- IMG_006.JPG (345.5 KiB) Viewed 206 times
-
- IMG_005.JPG (321.5 KiB) Viewed 206 times
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ITSY BITSY 32U4 5V POWER FAILURE
Yeah, that’s definitely cooked the traces.
The fact that so much power went into the board invites a follow-up question: where did that power leave the board?
My first guess is that it came through the diode next to the burned trace. That suggests the possibility of a strong ground loop.. for whatever reason, the power supply’s GND terminal had a DC connection to some other voltage shared by a connection to the USB cable.
Current capable of doing that to a heavy trace would have vaporised parts of any IC it went through, leaving a trail of dimples, bulges, and outright holes on the package of any chip in the current path.
The fact that so much power went into the board invites a follow-up question: where did that power leave the board?
My first guess is that it came through the diode next to the burned trace. That suggests the possibility of a strong ground loop.. for whatever reason, the power supply’s GND terminal had a DC connection to some other voltage shared by a connection to the USB cable.
Current capable of doing that to a heavy trace would have vaporised parts of any IC it went through, leaving a trail of dimples, bulges, and outright holes on the package of any chip in the current path.
- Kenya7c6
- Posts: 3
- Joined: Mon Aug 02, 2021 9:58 am
Re: ITSY BITSY 32U4 5V POWER FAILURE
Thank you adafruit_support_mike for your reply. You bring up a good point that the current had to leave the board somehow and must have done so via a strong path otherwise we would see burning there too. Strangely, I soldered the split and lifted traces back together and the board powers up and works fine... so now I'm even more confused. Is it possible that an errant piece of metallic debris had fallen in such a way as to cause this short in both cases? Potentially. I guess time will tell and I will have to wait and see if this happens again on other units.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ITSY BITSY 32U4 5V POWER FAILURE
Hmm.. that’s pretty strong evidence against the idea of reversed power polarity.Kenya7c6 wrote:Strangely, I soldered the split and lifted traces back together and the board powers up and works fine.
It’s possible, but seems unlikely. The old saying is that one event is random, two are a coincidence, and three are a pattern.Kenya7c6 wrote: Is it possible that an errant piece of metallic debris had fallen in such a way as to cause this short in both cases?
Sometimes it takes a pretty big coincidence to get two though, and this seems like one of those cases. Any foreign object that could make such a connection would have to be more conductive than the traces that burned, otherwise it would fail first. It would probably leave traces of arcing at both ends of its connection, too.
One possibility is a soldering iron with a grounded tip. Those are like a short to GND that erase their own tracks. The circuit has to be connected to power for a short to occur, but I’ve been guilty of that myself.
Working from the evidence of damage though, do you see any similar damage on the back of the ItsyBitsy, or on the PCB beneath it? The exit damage might be there.
You can also search from the other end by identifying any power sources capable of delivering that much current, locating their connections to the board, and inspecting the connections from there to the damage. The same amount if current had to flow through the entire path, so any traces the same size as the ones above should also show damage. Any traces narrower than the ones above should have burned first. If there’s no visible damage, it suggests a path with lots of copper like power and ground planes.
- Kenya7c6
- Posts: 3
- Joined: Mon Aug 02, 2021 9:58 am
Re: ITSY BITSY 32U4 5V POWER FAILURE
Thanks again adafruit_support_mike for responding, even though this is likely not a problem with the Adafruit product itself. You bring up some very good points there and when it's possible I will have another look at the board in question and see if I can find any more clues.
- adafruit_support_mike
- Posts: 67485
- Joined: Thu Feb 11, 2010 2:51 pm
Re: ITSY BITSY 32U4 5V POWER FAILURE
No worries.. debugging is basically puzzle-solving, and the evidence for this one is practically a locked-room detective story: the damage is obvious and undeniable, but the path of the energy that did the damage seems to be incomplete. That’s kind of challenge is hard to ignore. ;-)
Please be positive and constructive with your questions and comments.