Powering Teensy 3

[PaulStoffregen]

I got the Traco part. I've replaced the LM317 on that breadboard with the Traco. I'm just running 9 volts from a lab bench power supply to the input pin on the Traco, and it's output directly to the 3.3V pin on all 3 Teensys.

So far, I've been unable to recreate the problem. I've power cycled a couple dozen times. All 3 boards have started up every time.

[tunahammer]

Ouch. Maybe it's my bench supply that can't handle this or something...
http://www.amazon.com/Mastech-HY3005...ywords=mastech
I'll try some things, including laying out the teensys more simply like you are and unplugging everything else from the supply (though nothing is being powered by it but this). It is comforting somewhat that they are working fine for you though.

[PaulStoffregen]

I just tried a few more experiments, using a fresh 9 volt battery instead of my lab bench power supply.

One time out of about 50 tries I got only 2 of the 3 boards to start up.

I added a 10 ohm resistor in series with the battery, but couldn't reproduce the problem again.

Then I tried a 33 ohm resistor, and a couple times I got 2 of 3 boards to start.

I tried a inductor in series, which I think is 470 uH, but I'm honestly not sure if it's specs (it was in a draw from a very old project). I got 2 of 3 starting up more often with the inductor. Still, all 3 start up most of the time, but the problem does become more frequent.

Sorry, I did this all very quickly without taking careful counts of written notes, but hopefully it helps a bit. It seems like the combination of Traco being fed with series impedance does create some sort of problem with Teensy3 startup. I can't spend much more time on this right now, but eventually I do want to get out my scope and investigate this problem more throughly.....

[tunahammer]

Very nice observation, it is odd that I was having problems with it on my breadboard though previously.

I am currently running on my breadboard with a much more simply layout and did about 30 tests and had 100% startup rate! I will take those parts off my PCB for now and see if they allow for better functionality.

Thank you!!

[PaulStoffregen]

Yeah, whatever's going on here is very strange.

I'm glad it's working for you now. Eventually, I really do want to get to the bottom of this. Not sure if I'll ever get the time, but I really want to....

[PaulStoffregen]

I know this is a very old thread... but this evening I was cleaning up my workbench and found the breadboard with this little Traco power supply.

About a week ago, I built a little shelf to hold the scope and other gear right above the workspace, which makes quick tests much more convenient. So I did a quick test.

I connected a 100 ohm resistor in series with my power supply, which was set to 8 volts. The idea is to emulate a weak power supply, since we know from before the issues were related to the power source.

I also connected a 10 uF capacitor between the Traco input and ground. When the power is connected, that 10 uF capacitor will charge up slowly because of the 100 ohm resistor.

Here's how it looks.


(click for larger)

The yellow is the input power, so the rising edge is where I connected the wire. The green is the voltage at the input of the Traco, and the blue is the Traco's output. The yellow and green are 5V/div and blue is 2V/div, so all look about the same height on these screenshots.

In this test, Teensy 3.0 starts up just fine.

But I was able to reproduce the original problem by removing the 10 uF capacitor. Here's what I see on the scope:



The Teensy 3.0 behaves strangely with this sawtooth-like waveform for its power. It definitely does not run correctly.

In hindsight, adding capacitance at the Traco's input might have solved these problems. But the Traco datasheet implies that's not necessary until you have more than 32 volts input in the "application" section, and even suggests it has internal capacitance in the input specs section. Obviously it's not enough if the power source has series impedance.

[PaulStoffregen]

I posted a followup at Adafruit, requesting they mention input capacitance. Hopefully if will save anyone else a lot of frustration if they run into this same problem.

http://forums.adafruit.com/viewtopic.php?f=19&t=46257

[PaulStoffregen]

Just one more quick followup. Adafruit is revising their product description. Hopefully in the future, people using this Traco power supply will see that recommendation and be spared these difficult problems.