Powering Nvidia Jetson & Radio, Strange VRM Issues

We’re using a Nvidia Jetson TX1, a radio, an IP camera, an LED ring, and an ethernet switch that all have to be powered via a voltage regulator. We’ve been advised not to run the Jetson and the radio off the same VRM as they may draw too much current and blow the module. However, we’ve noticed some strange issues with wiring. It seems that the Jetson receives power from a VRM just fine but we’re unable to boot it when it doesn’t share the same voltage and current ports as the radio on the VRM upon pressing the power button.

We tried plugging the radio to the 12V 2A port on one VRM and the Jetson to the other VRM on the 12V 2A but the Jetson doesn’t boot, only receives power. We tried all possible permutations of the wiring and arrived at the conclusion that the Jetson ONLY boots when it is on the same VRM as the radio and on the same VRM as the LED ring. When the condition is not met, the radio and LED ring work just fine but the Jetson does not boot. Anyone know why this may be happening?

Did you try swapping the VRMs?

We did, exact same problem. We also used a volt meter and confirmed that 12 volts were going to the board, it just wasn’t booting.

Cool!

If you’re not running the Jetson through the VRM then A) Are you running it through another voltage regulator and B) can you please post some pictures of the setup?

EDIT: The Jetson TX1 is a bit sensitive to minor voltage fluctuations, just FYI.

I’m not currently able to take a picture but we have 2 VRMs on board right now. We tried running only the jetson off of one but it wouldn’t boot. It only seems to boot when it’s plugged into the same VRM as the radio and the led ring.

So maybe it is steadied by an accompanying load.
You could test with a simple resistive load applied as well as the Jetson.

We tried that too, plugged in our ethernet switch instead of our radio and it wouldn’t boot. Somehow it HAS to be the radio. I don’t think its a network issue because we unplugged the ethernet cable.

Know anyone with an oscilliscope you could borrow to plot and compare/contrast the power in both the working configuration and the stand-alone/non-working setup?
If it really is a noisy supply issue you might be able to deal with it using a more suitable conditioned supply, or adding a balancing off-setting load.

What happens if you swap the VRMs? Does the radio and such work on the other VRM? Do you have a bad VRM?