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

Thread: Dying 4.1 teensies

  1. #1
    Junior Member
    Join Date
    Aug 2018
    Posts
    7

    Dying 4.1 teensies

    Hello,

    I have been having trouble making teensy 4.1 to work reliably.
    On 2 separate projects I had teensies dying for no apparent reason. (6 in total)

    On one project we designed a pcb where the teensy reads a couple of pwm signals and retransmit them , a couple of can transceivers and using a couple of serials.
    All signal coming in from a flight controller are 3.3v so no problems there. Teensy is powered by a custom buck converter using TI parts and is basically the exact same design as in the datasheet of the main converter (12V->5V). VIN pads underneath the teensy was cut so that we could use usb SERIAL. Some of the teensies work, others die immediately others fail for no reason. I have been monitoring the 5v rail and I can confirm no spikes or over voltages occur.
    Using a Teensy 3.5 or 3.6 works perfectly.

    Then again I made for personal use a couple of pcbs of the kicksid( dual sid replacement for c64) project that uses 4.1, and then again 2 teensies are gone. Checked for shorts, overvoltages prior connecting teensy but nothing there. Still they failed immediately.

    Most of the teensies wont recover using the 15sec reset recovery method. But after a few days some may recover using the same technique.

    My best guess so far is that Teensy 4.1 can be "picky" on the power supply and it can fail under certain conditions.

    Any insights are welcome.

  2. #2
    Senior Member
    Join Date
    Apr 2014
    Location
    Germany
    Posts
    9,285
    I guess the PWM signals are 5v? Also, c64 uses 5V.
    T4.x are not 5V tollerant. 5V on any of the digital or analog pins kills it.
    That's printed on the card you got with the teensies.

    T3.2 and T3.5 are OK for 5V signals.

  3. #3
    Senior Member
    Join Date
    Jan 2015
    Location
    UK
    Posts
    174
    There might be transient voltage to the Teensy's input and power rail.

    Can you post some photos of your board and the schematic.

  4. #4
    Junior Member
    Join Date
    Aug 2018
    Posts
    7
    Quote Originally Posted by Frank B View Post
    I guess the PWM signals are 5v? Also, c64 uses 5V.
    T4.x are not 5V tollerant. 5V on any of the digital or analog pins kills it.
    That's printed on the card you got with the teensies.

    T3.2 and T3.5 are OK for 5V signals.
    No, PWM signals coming from the flight controller are 3.3v as the rest of the signals. Also tried a teensy 3.6 ,which is not 5v tolerant, and worked but I had to reroute some pins because 3.5/3.6 and 4.1 are not pin compatible.
    5v on the kicksid pcb for c64 is routed to vin of the teensy which accepts 3.6 to 5.5volts.

  5. #5
    Junior Member
    Join Date
    Aug 2018
    Posts
    7
    Quote Originally Posted by skpang View Post
    There might be transient voltage to the Teensy's input and power rail.

    Can you post some photos of your board and the schematic.
    Sorry cannot post schematic for the board because its used in a consumer product. but the power supply is this https://www.ti.com/lit/ds/symlink/tps54560b.pdf pg.24 .

    There is nothing connected to the board that can draw excessive power from the teensy or individual pins. The teensy fails even if nothing is connected. Just power.

  6. #6
    Senior Member
    Join Date
    Apr 2014
    Location
    Germany
    Posts
    9,285
    Ok, then, I think skpang is right.
    How do you translate the levels in c64?

  7. #7
    Junior Member
    Join Date
    Aug 2018
    Posts
    7
    Quote Originally Posted by Frank B View Post
    Ok, then, I think skpang is right.
    How do you translate the levels in c64?
    Sidkick uses 3 74LVC245

  8. #8
    Senior Member
    Join Date
    Apr 2014
    Location
    Germany
    Posts
    9,285
    I've never had a Teensy (and i have many) that died just by powering it.

    Replace that step-down supply, and use a linear vr.

  9. #9
    Senior Member
    Join Date
    Jan 2015
    Location
    UK
    Posts
    174
    Bit difficult to help if we can't see the board and the schematic.

  10. #10
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    25,057
    Have you verified your DC-DC power supply design's startup and shutdown behavior by watching the input and output voltages with an oscilloscope? And with the range of expected input sources (eg, if a 12V battery, when the battery is fresh and when it's low)?

    Especially if the capacitor at the power supply input is not chosen well, it's easy to end up with a power supply that appears to work but it actually starting up with overshoot or oscillation or other terrible issues. Avoiding those problems can be quite a challenge if your input power covers a wide range of voltages and equivalent source impedance.

    Even companies who design power supplies sometimes get this very wrong. In particular there is a little 3-pin DC-DC converter from Traco which Adafruit sells that has caused a lot of problems over the years. Traco's datasheet says extra capacitance at the input is only needed for the higher end of the input voltage range, but in fact when used with most Teensy boards where the load current changes so drastically as the voltage ramps up, they go into oscillation and subject the Teensy to as much as 7 or 8 volt spikes! Adding capacitance at the input solves those problems. I got Adafruit to add a warning on their product page... and initially they incorrectly said to add a capacitor only at the output, which actually makes the problem even worse! In that case, it's the impedance at the input which matters. Getting stable performance from DC-DC switchers can be tricky sometimes.

  11. #11
    Junior Member
    Join Date
    Aug 2018
    Posts
    7
    Thank you all for your replies. I will try a linear power supply on the next design.

  12. #12
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    25,057
    Two other tests you might try...

    Connect your oscilloscope to the 3.3V power and set either normal or single trigger (not auto) and adjust the trigger level to just slightly over 3.3V. If something is going wrong like large currents though the ESD protection diodes on the pins, odds are good that sort of stress would momentarily add noise on top of the 3.3V power.

    With the power to Teensy turned off, measure the 5V and 3.3V power. If some other part of your system is remaining powered and tries to drive Teensy pins, you'll probably see some residual voltage, perhaps even 1 volt or more.

Posting Permissions

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