Forum Rule: Always post complete source code & details to reproduce any issue!
Results 1 to 13 of 13

Thread: Teensy 4.0 doesn't power on with less than 4.10V

  1. #1

    Teensy 4.0 doesn't power on with less than 4.10V

    Hi,

    I'm trying to power a Teensy 4.0 using a Lipo battery and an Adafruit Lipoly Backpack https://www.adafruit.com/product/2124

    I couldn't get the teensy to power on, unless I would "jumpstart" the teensy by hooking it up over micro usb.

    After a bit of testing over a PSU I found out that the teensy doesn't start unless it receives 4.1V or more over its VIN pin.

    This is confusing (and unfortunate) since on the teensy's welcome card it states that it runs on 3.6-5.5V.

    Am I doing something wrong here? Or is indeed the teensy not able to run of the 3.8V I'm getting from the Adafruit Lipoly Backpack?

  2. #2
    Senior Member+ defragster's Avatar
    Join Date
    Feb 2015
    Posts
    16,702
    What other testing was done to get 4.1V on VIN? Was that just with USB cable?

    No code provided ...

    What does setup() look like?
    If it has some variation of: while ( !Serial ) this would explain what is posted.

  3. #3
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,375
    I tested a couple Teensy 4.0 boards programmed with LED blink. Both are able to boot up when 3.25 volts is applied to VIN.

  4. #4
    I'm running the basic blink example with nothing but the Teensy on a breadboard and a PSU attached to VIN and GND. also VIN isn't cut or anything, its just a Teensy straight from the package. I'll test one or two more to see if it changes, but I'm pretty sure I got the same behavior on another one. Also no USB cable attached or anything like that.

    EDIT: https://photos.app.goo.gl/DZYDDrbYNUC9Bid37 video recording to show whats going on.

  5. #5
    I just tried it on another Teensy, straight from the package, I could get as low as 4V, but not lower than that.

  6. #6
    I tested a Teensy LC this morning, that works fine, even with 3.6V. Could it be that I simply have a bad batch of Teensys 4.0...?

    In an effort to still try to get my Teensy4.0's to boot in my currrent PCB without too much hassle, I got some voltage regulators. The one that fits the best in my board though sets the current to 5.75V, which is 0.25V above the limit stated on the card. Will I risk damaging the teensy with such a voltage overshoot?


    EDIT: Getting rid of the Adafruit Lipoly and going for a charging circuit with a voltage regulator that outputs 5V..
    Last edited by mick; 07-15-2022 at 11:06 AM.

  7. #7
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,375
    Sorry about the delay. I've been investigating here. Looks like there is indeed a hardware issue with some Teensy 4.0 boards not starting up with less than 4 volts at VUSB / VIN. It's related to the voltage regulator chip. Nearly all Teensy 4.0 and 4.1 have been made with TLV75733P. But this part ran out several months ago. On Teensy 4.0 we've used 2 alternate parts, AP7366 and NCP705E. On Teensy 4.1, we recently used NCV8186A. The startup issue with lower voltage is happening with the boards using AP7366.

    I'm still working to get to the bottom of *why* the lower voltage startup doesn't work. It's not a simple matter like dropout voltage. Startup is getting stuck with a problem at the AP7366 enable pin. I still don't fully understand the problem, and I'm pretty embarrassed this escaped notice during quite a lot of testing I did with all of these alternate parts. Most tests checked whether the parts could deliver enough current (and avoid overheating with only the thermal vias to the internal PCB planes acting as a heatsink) with Teensy overclocked and external loads connected.

  8. #8
    Hi Paul,

    Thank you for your answer. Don't be too hard on yourself, everyone is currently struggling with the shortage and how to deal with it. I have been and for a long time will be a very happy Teensy customer :-)

    If you find out anything interesting please keep us posted.

    Best,
    Mickey

  9. #9
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,375
    I've spent some more time investigating this problem. It all comes down to resistor R4.

    Click image for larger version. 

Name:	r4.png 
Views:	15 
Size:	49.2 KB 
ID:	28976

    If you just remove R4, your Teensy 4.0 will be able to start up with approx 3.2 volts (and the 3.3V power will be around 2.9 to 3.0 volts, just barely enough to run).

    But without R4 the 15 second button press restore process will not work. Nothing bad will happen, it just won't work and your Teensy will reboot right after you release the pushbutton (if within 4 seconds of the quick red LED blink). R4's only purpose is to keep the power turned on during the restore process.

    The startup issue involves loading of the PMIC_ON_REQ signal as the voltage ramps up. 100K worked well with the original TLV75733P voltage regulator. But with AP7366, the regulator's output has a low impedance path to ground much earlier in the startup process.

    If you want both lower voltage powerup and working restore process, you can solder a different resistor in place of R4. We're going to build future Teensy 4.0 & 4.1 batches with 470K. And when/if we need to switch to yet another voltage regulator (they're all unavailable right now and we have enough AP7366 and NCV8186A to last until about October) next time I'll be sure to test the low voltage startup behavior, rather than focusing only on power capacity and heat dissipation.

    Again, really sorry this escaped notice. Hopefully just desoldering or cutting R4 off those boards will let you use them with the batteries you wanted.

  10. #10
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,375
    Final followup on startup voltage. All new Teensy 4.0 and 4.1 now have the 470K resistor, which allows startup down to about 3.2 volts.

  11. #11
    Senior Member
    Join Date
    Oct 2016
    Posts
    103
    Will this problem affect running the 4.0 using a 5V step up regulator? I had a problem using the Pololu 2564 with a Teensy 3.2. I finally had to drop the speed down to 72Mhz. Dropping it to 150Mhz on the 4.0 didn't allow it to boot from battery only.

    Will removing the R4 allow it to boot even with the slow startup of the regulator?

    I have a few 4.0s. How can I determine if I have any that have been fixed or have the working regulator?

  12. #12
    Senior Member PaulStoffregen's Avatar
    Join Date
    Nov 2012
    Posts
    27,375
    All sorts of unrelated problems are possible with low quality DC-DC converters. Especially the very small ones tend to not have enough capacitance at the input and output. Some people (including folks at Adafruit) sometimes assume the answer to power supply problems is adding more capacitance at their output. Often this does help, especially in the scenario of a weak "wall wart" power supply with a long cable. But if the DC-DC converter has stability problems, as many of the small PCB mounted ones do, sometimes piling on more capacitance at the output without also adding a good quality capacitor close to the input can make matters even worse.

    The Teensy 4.0 boards with the newer regulators and 100K are able to work with slow startup power supplies which eventually (usually over many milliseconds) do rise up to 5 volts. Slow rising input power is not a problem for those boards. They will turn on properly as the input voltage crosses about 4.0 to 4.1 volts. Removing or changing R4 won't have any change for handling of slowly rising power, other than the threshold where Teensy 4.0 enables its 3.3V regulator.

    Probably the simplest way to test if your Teensy 4.0 has the original design (which starts up around 3.25 volts) or the new regulator with 100K resistor (4.0 to 4.1V startup) or the updated design with new regulator and 470K resistor (~3.2V startup) is to program your Teensy 4.0 with a LED blink or other easy-to-observe program and then connect it to lab bench power supply. First set it to 5V and make sure everything is working. Then dial it down to 3.6V and cycle the power. Then test again at 3.4V and other levels as you wish.

    Attempting to measure R4 with an ohm meter is tricky. Multimeters in ohms mode transmit a test current through the leads and try to measure the resulting voltage. The specific current used can vary quite a lot with different brands of multimeters, and of course changes depending on the measurement range the it has configured. For higher resistance, the voltage across the resistor will be higher. With my old Fluke 87 multimeter, I can measure R4 at ~100K with all power shut off. But measuring a 470K resistor gives a strange-looking result because the voltage across the resistor becomes enough the much of the multimeter's test current starts flowing through the other parts on the PCB. I usually see it settle around 150K. The slowness of this process is probably due to the current flowing into some of the capacitors. As a general rule, measuring resistors while they're connected to other stuff is tricky, because your ohm meter's test current doesn't necessarily all flow through the resistor. It has plenty of other paths to take on the circuit board.

    So back the original issue with your Pololu 2564 regulator. If it's also having trouble with Teensy 3.2 and Teensy 4.0, that's probably a sign you should try adding capacitors at both input and output. You might also try putting a switch (or just a wire you touch) between the Pololu+capacitors and Teensy, so you can compare the results of turning on the power to the regulator with Teensy connected versus letting the regulator fully start up and then connect Teensy to it.

  13. #13
    Senior Member
    Join Date
    Oct 2016
    Posts
    103
    Thanks for the detailed reply. Unfortunately due to space constraints I need to stick with the small form factor boost regulator. I did have a capacitor on the output but not the input. My assumption was that the battery was a clean DC source so it wasn't needed.

    I'm in the process of setting up a testbed to identify which version of the teensy I have and to allow experimentation with the battery circuit.

Posting Permissions

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