Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 12 of 12

Thread: T4 killing board

  1. #1
    Junior Member
    Join Date
    Feb 2020
    Posts
    17

    T4 killing board

    I built this little board to control a little robot with a T4. It's mainly a puzzle of various pre-made modules.

    It has one big switching voltage regulator to provide power to 12 tiny servos, 5.5V. That voltage (should) never get in contact with the Teensy.

    It has one small switching voltage regulator that provides 5V to VIN to the Teensy and feeds a LDO to make 3.3V that powers a NRF24 radio module and an MPU6050 module. I have cut VIN from the USB port power.

    The thing works fine for a few minutes but eventually kills the teensy. The first Teensy it killed ended up with a shorted 3V rail, stone dead. And the 2nd it killed had the USB stop working. That teensy still works and it is possible to make it load the built in blink program. It still runs perfectly fine, as seen in the attached picture.

    I have measured all I can with the partially working Teensy in place. Scoped all signals. Hooked up an old Arduino UNO to it to test the radio and MPU (those modules are 5V compatible). Everything looks fine and works as intended. No 5V signals I can see reach the Teensy.

    The PWM pins for the servos are connected straight into the Teensy, I should add some series resistors to be safe. But all testing have been done without servos so far. Other I/O are 4 pins with 10k pull up resistors (to teesy 3.3V) for attaching some pressure sensitive resistors.


    Anyone have ideas for what might be wrong? I am out of ideas, and soon out of cash to buy new Teensies
    Attached Thumbnails Attached Thumbnails Click image for larger version. 

Name:	TeensyEater.jpg 
Views:	30 
Size:	151.3 KB 
ID:	18982  

  2. #2
    Senior Member+ KurtE's Avatar
    Join Date
    Jan 2014
    Posts
    6,708
    Sorry to hear about the problems.

    Not sure what you are saying about the 2nd Teensy? You say you lose the USB, but then you say you can reprogram it, which implies the USB is still functional?

    Note: unlike the UNO which has a separate USB controller, the USB on the T4 (like the Arduino Leonardo ) is all done by the main chip. So if your program you are running crashes in a bad way you can easily lose USB. But you should be able to recover it a few different ways. Most of the time you can simply recover by reprogramming it. If the program fails to download you can often simply press the program button on the teensy to get it to take the new program. If that does not work I will often times hold down the program button when I plug it in, and then try to download a program... Or I hold down the program button for something like 15 seconds, until an LED turns on the T4 and then release, which should reprogram the T4 with the default blink program.

    Other things to checkout. Sometimes with setups like this people think their T4 is dead when they try to run it off of a battery, when actually their code has something like: while (!Serial) ;
    in it, so the code is hanging waiting for the USB to be plugged in.

    There have also at times been issues with booting up with some external Voltage Regulators. Sometimes their rise time cases issues. You might do a search on which switching regulator you are using and see if anyone has had issues.

    Good luck

  3. #3
    Junior Member
    Join Date
    Feb 2020
    Posts
    17
    The 2nd teensy with broken USB was "reprogrammed" using the 15s button hold you mention, it is blinking. When in programming mode or running that blink program there is no reaction from the USB host when plugged in, even if I reconnect VUSB to VIN to power from USB. It is quite strange that the USB port broke.

    I scoped the 5V line when powering up, a very nice linear rise to 5V in about 2ms.
    Click image for larger version. 

Name:	5vrail.jpg 
Views:	8 
Size:	151.5 KB 
ID:	18989

    Got some new boards arriving on monday. I will hook one up very carefully, one pin at a time, powering from the bench PSU with a very conservative current limiter.
    Last edited by scaraman; 02-07-2020 at 08:45 PM.

  4. #4
    Junior Member
    Join Date
    Feb 2020
    Posts
    17
    Two more dead Teensies, only one left . At least I am supporting the project

    The last one died when I connected the USB port while the Teensy was powered on, running on battery. I use SPI so the LED is constantly flickering, when I inserted the USB port it went black and the board was dead. No more 3.3V.

    This is consistent with when the others have died, they have always died when I was back at my desk poking with the electronics or reprogramming them. And one has a broken USB port but works otherwise.

    But it makes very little sense.

    Click image for larger version. 

Name:	corpses.jpg 
Views:	8 
Size:	235.9 KB 
ID:	19015

  5. #5
    Member
    Join Date
    Sep 2017
    Location
    Italia
    Posts
    87
    It is very hard to see something in the crystal ball...
    What I understand from your description it says you are feeding 5V into the Teensy via the I/O lines. This will surely fry it.

    Anyway, the PWM signals of servos shall never be connected directly to 3.3V logic. Use a line driver for this. Most servos have pull-up that leak 5.5V into PWM line. Also, the line driver would prevent transients from leaking into the Teensy.

    Could you post a schematic of your board?

  6. #6
    Junior Member
    Join Date
    Feb 2020
    Posts
    17
    It turns out it was the servos sending back their supply voltage to the PWM pins! I was totally sure that one of my boards broke with no servos attached. I was wrong.

    The (very low quality) cables to the Tower Pro servos had worn out over time and were shorting at the servo end.

    For this last Teensy I had added some 330ohm series resistors on the PWM signals, enough protection to save the Teensy.
    I have started making some schematics, I will most definitely add in some proper line drivers for the next version

    Project back on track. Time to scale Mt Floortile

    Click image for larger version. 

Name:	MtFloortile.jpg 
Views:	35 
Size:	346.0 KB 
ID:	19016

  7. #7
    Member
    Join Date
    Sep 2017
    Location
    Italia
    Posts
    87
    Good luck!

  8. #8
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    21,870
    Oh, that's painful losing Teensy 4.0s to badly wired Servo motors!

    But your robot looks pretty awesome. Looking forward to any videos or other info you're willing to share!

  9. #9
    Junior Member
    Join Date
    Feb 2020
    Posts
    17
    Got some videos of it walking/climbing/stumbling around.

    https://youtu.be/a22FzKibE3c

    I could write something about it and post in the Blog section.

  10. #10
    Senior Member+ defragster's Avatar
    Join Date
    Feb 2015
    Posts
    11,571
    Cool. Persistent! What's the umbilical blob hanging down?

    More pics and details would be nice to see what and how it does that.

  11. #11
    Senior Member+ mjs513's Avatar
    Join Date
    Jul 2014
    Location
    New York
    Posts
    5,152
    Agree with @defragster - very cool! Almost makes me want to build one

    Definitely more pics and details would be nice. Putting it in the blog would be great

    Its a lot cooler!!!!!! than the openDog project: https://www.youtube.com/watch?v=SgJcj2Gd0Rc. Posted the link because I thought it might be of interest. Think he uses T3.6s

  12. #12
    Junior Member
    Join Date
    Feb 2020
    Posts
    17
    That umbilical blob is a LiPo pack watcher. It beeps loudly when a cell is too low. I already broke one pack due to discharging it too much, my speciality is breaking hardware

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •