My aim is to power the TX2 using a battery for 24 hours, to record some datasets. After reading the forums, the power requirements of the TX2 seem to be 12V and roughly 1A, though the adapter that comes with the developer kit supplies a 19V 4.75A current.
So I got a 12V, 100Ah deep cycle battery. Furthermore, I have connected the battery to a barrel plug (http://uk.farnell.com/lumberg/1634-02/connector-dc-power-plug-4a-12v/dp/1568397) which is connected to the TX2.
The CR5 LED comes on the TX2 , however, when I press the on button, it fails to power on and the CR1 and CR2 LEDs don’t come on.
Any ideas of what I am doing wrong?
- Needs a 19V to start? can use an step up converter in this case
- The barrel plug isn’t the correct specification, but the CR5 LED comes on?
Before trying the TX2, I tried to power the raspberry pi and that worked fine. Measuring the voltage of the battery checked out fine too.
You may need to add a large capacitor close to the barrel connector. 12V should work fine, but I think Jetsons are a bit sensitive to the initial current draw.
Also, not all of the barrel connectors of the 5.5mm outer diameter have the same inner diameter…but you will rarely see any spec listed for “real” inner diameter. This could effect contact resistance. As an example go here, look at the actual inner diameter being stated:
Even if your barrel connector is stated as 5.5mm, it won’t necessarily be the same as the one used in the Jetson’s default power adapter.
I have used 12V to power the Jetson (even a 11.1V LiPo battery) and it’s worked fine.
I did find that one of the 5.5mm barrel connectors I had did not have a good fit on the interior pin, so that’s worth checking.
However, if the red LED comes on, that should indicate contact.
Also, the 1A draw depends a lot on what the Jetson is doing. If running full tilt with all CPUs and GPU, and all of the peripherals on the motherboard working, then you can see > 1A. If you’re mainly sitting idle with a load of 0.1 and not much going on, it will draw much less.
These are all “of course I did that” things, but it bears checking again:
Is the 12V battery actually charged?
Are you using sufficiently low-gauge wires? (20 Gauge should be enough; thin stuff like 28 or 32 probably won’t be.)
Did you actually push the right power-on button?
Did you hold it down enough and wait long enough for power-on?
Is the Jetson module properly seated onto the motherboard?
Has your motherboard or Jetson module been in contact with any kind of metal, that could have caused a short and burned some part?
Good luck on the project, and please let us know whatever you find!
@linuxdev thanks for your reply. What purpose would a capacitor serve?
@snarky thanks for your reply. I have a 25 gauge wire(https://www.amazon.co.uk/gp/product/B017O5TJYG/ref=oh_aui_detailpage_o03_s01?ie=UTF8&psc=1), but it should be able to carry 6A. I will try a thicker wire and run through the rest of your list.
I will try to measure the amps being drawn too and let you know how I get on.
Thanks for your help!
A large capacitor near the connector is for no other reason than to provide a momentary current source for power spikes at the moment of turning on the Jetson. This is something a few people have noticed helps when using other power sources as increasing reliability. I don’t think normal operation is as sensitive to power requirements as is the moment of turning on the Jetson.
I have noticed the Jetson not being happy with large swings in input voltage. It’s OK with a wide variety of input voltages, as long as the voltage only changes slowly.
With 25 gauge wire, it’s possible (not super likely, but not impossible) that the inrush current when turning it on would cause enough voltage drop that it wouldn’t run.
The “6A” rating for 25 gauge wire depends on a lot of assumptions related to amperage, cooling, environment, and so forth. It also depends on how long the wire is.
However, I don’t get where you get the 6A value from. Standar ratings for copper wire at 25 gauge says 2.7A chassis wiring, 0.5A transmission wiring. Resistance will be 0.11 Ohms per meter. If inrush current is 20 A (just momentarily) and you use a meter of wire, then you will drop 2V during the turn-on event, which might be enough to cause it to refuse to turn on.
This is all speculation, though – please let us know what you find, because I’m curious :-)
@linuxdev thanks for your clear explanation, I will keep that in mind.
@snarky I realised the wire was 20 gauge but I was using 20 meters of it. With 0.11 Oh ms resistance per meter, no wonder the TX2 failed to power on.
To solve this, I kept everything else the same and made the wire shorter. The TX2 now happily powers on! :)
Great speculation, thanks for your help.