Teensy 3.6 died mysteriously x 2

Status
Not open for further replies.

CraigF

Well-known member
I have two dead Teensy 3.6 boards that died without obvious cause.

Each one had an audio adapter soldered in place (piggyback) and a ground wire to my oscilloscope, but no other connections to the Teensy other than the USB port. Each worked in the evening, was unplugged from USB overnight, and simply failed to light up when plugged in the next morning. In both dead boards the resistance between the VDD network and ground is about 53 Ohms, which causes the board to draw enough power to shut down the resettable fuse.

Static? USB surge? Excessive heat while soldering, with delayed failure? In both cases the audio board headphone output was connected to a mixer, and there is a significant ground loop problem in this lab -- but I can't see how that would fry anything on the Teensy.

Anybody have any thoughts?

I'll start pulling components off and testing them, but without a component placement guide it's more complicated than I'd wish.


-- Craig
 
Hi Craig. The audio boards headphone output has its common terminal connected to the SGTL5000s HPVGND pin- which is an audio virtual ground and is normally = 3.3V /2. Presumably your mixer's ground is at the same potential as your computers ground, and if you are powering the Teensy/Audio board via USB from your computer, then there is a real bad ground loop issue. If you power the Teensy with a USB charger, then it's ground would be floating, and the above issue would not apply.
Don't know if this scenario fits your situation, but if it does, this may be the cause of the problems.
Note that the line in/ outs are referenced to the actual Teensy GND terminal.
 
Hi Craig. The audio boards headphone output has its common terminal connected to the SGTL5000s HPVGND pin- which is an audio virtual ground and is normally = 3.3V /2. Presumably your mixer's ground is at the same potential as your computers ground, and if you are powering the Teensy/Audio board via USB from your computer, then there is a real bad ground loop issue. If you power the Teensy with a USB charger, then it's ground would be floating, and the above issue would not apply.
Don't know if this scenario fits your situation, but if it does, this may be the cause of the problems.
Note that the line in/ outs are referenced to the actual Teensy GND terminal.

That seems like quite a good suggestion. I do use USB data isolation and a medical grade isolation charger to feed this project, BUT... I also had a chassis ground wire from my oscilloscope to Teensy GND, and the scope ground is very close to the PC ground. Of course, the audio boards remain functional, while the Teensy boards are dead, so I still need to figure out who gets the pain when ground currents do the mambo :)

The final project uses headphones, which is why I was testing with that signal path. Guess I'll have to sacrifice another Teensy or two in an effort to prove/disprove this, since the potential for an end-user to plug the headphone output into an amp or mixer is unavoidable. The requirement to simulate that kind of environment is actually why instrumented ground loops exist on this bench in the first place :)

Thanks!
 
I had the exact same issue using a teensy 3.2 and a teensy audio sheild. Im not sure how i did it but some plugging in the audio jack fried the teensy and audio board. It happened twice, i then just created my own custom board which seemed to correct the issue :)
 
Hi Craig. Yes, it would likely have been the scope ground going to the teensy/ audio board GND and the mixer ground, going to the SGTL5000s virtual ground that I would suspect. In your project, does the audio boards physical headphone jack have to be the user connection, or can you use the following:
Plug a cable into the jack, place a 220 uf capacitor in series with the common lead, and then connect to another socket. This will interrupt the DC path between the conflicting grounds, but let the audio thru ok.
The fact that the audio board makes it though ok and the teensy doesn't is odd. It also might be some higher voltage static charge causing the problem. If so, the capacitor solution above probably wouldn't work.
Regards
 
@ 1101010. The VGND referee to in this thread is only used for the headphone common terminal, which is what the original poster was using. In the case of your SLIC, it has a 60k input impedance, which should allow you to use one of the LINE OUT pins, which are referenced to GND. THe SGTL5000 codec on the audio board has only a coarse volume control for the line output- whereas the headphone output has a very fine output volume control, but I doubt that would affect your application too much. That said, I don't use those SLIC modules myself, so I can't know exactly what your needs might be in that regard.
Cheers
 
Thanks @bmillier! I work on this project off and on and it has been off lately, so i wasn't thinking about the difference between LINE OUT and headphone out when i read the above and got scared. :)
Coarse volume output is perfectly fine. This plays audio files from an SD card through an old telephone in response to numbers dialed. I had been gutting the telephone and hooking the teensy directly to the rotary dial line and the handset audio, (with power fed through the hook so it's off when hung up). I try to do it in such a way that the phone could be recovered later, but it still kind of bothers me to meddle with the insides of these beautiful old things, so hoping i can use the SLIC to do the same thing while keeping the phone intact.
 
Status
Not open for further replies.
Back
Top