PDA

View Full Version : Short circuit when using USB and external power?



09-25-2007, 06:46 PM
Hello,

I am powering a robot using a 118 Wh Portable Power Station from Batterygeek.net (max current 5A). The output is set to 12v to power a Via mini-ITX computer I am using as the brain. The 12v terminals are also connected to an adjustable DC-DC transformer that I have set to 7v output. Checking with a volt meter, everything seems to be A-OK: 12v before the transformer and 7v after.

I wanted to connect a Phidgets 4-servo controller to the 7v side of the circuit. Here's what happens:

* I turn on the battery
* I turn on the mini-ITX board and it boots WinXP fine
* I plug the Phidgets servo board's external power jack to the 7v side of my circuit
* I plug two servos into ports 1 and 2 of the Phidgets servo board
* I connect a USB cable to the USB port on the Phidgets servo board

And now here's where the trouble comes:

* The instant I plug the other end of the USB cable to a free port on the mini-ITX, the battery shuts down and kills power to the mini-ITX. I'm guessing the battery's short circuit protection is kicking in for some reason.

I'm pretty sure this is not a current problem since the mini-ITX draws only 1.6A on its own, the battery can handle up to 5A, and the servos are in neutral when this happens.

Not knowing a lot about USB, I suspect that somehow the 5v USB power from the mini-ITX is doing something weird to the rest of the power circuit by way of the Phidgets servo controller.

Any ideas on how I can fix this would be greatly appreciated!

--patrick

Alex
09-26-2007, 08:51 AM
Dave may be able to better shed light on what I'm trying to say below (he's more of the hardware guy), but for now:

What sort of servos are you using?

If you are using higher current servos (pretty much anything above a 645 (http://www.trossenrobotics.com/store/p/3292-Hitec-Servo-HS645.aspx)) and are connecting directly to the USB port on your ITX board and not through an externally powered USB hub, make sure you don't have any servos plugged into port #0. Port #0 on the Phidget 4-motor servo controller receives its power directly from the USB, whereas ports #1 - 3 get their power directly from the external power adapter plugged into the servo controller.

Like you said, there's probably some sort of protection built in and it's preventing your USB ports from overloading. If the mini-ITX only draws 1.6A and you're using higher current servos, then there is a good possibility that the servos are drawing too much power from the Phidgets which in turn draw too much power from the USB, thus tripping tripping protection on the ITX board.


I hope that made sense:)

Dave
09-26-2007, 10:47 AM
If the servos aren't moving, they shouldn't be drawing enough current to trip anything. Have you tried any other USB devices? How about a different cable? Does it still cut out if the dc input to the servo controller isn't plugged in?

09-26-2007, 11:37 AM
Many thanks Alex and Dave for your replies. I did another experiment with no servos attached to the Phidgets servo controller and I got the same result so current drain is not the issue. To answer Dave's questions, I can power other USB devices like webcams (Quickcam 5000) and USB speakers fine using the mini-ITX USB ports. And I can use the same USB cable with the servo controller on my desktop PC and it works fine. The problem does *not* occur when I connect the servo controller DC power jack to a separate 7.2v battery pack (6 AA NiMH). It only occurs when I connect to the 7v output of the DC-DC transformer that is inline with the main 12v battery that in turn powers the mini-ITX.

Something just occurred to me but I don't know enough about electronics to know the answer: When I did the test with the 6-AA battery pack, I connected it to the DC power jack of the servo controller by way of my power bus on the output side of the DC-DC transformer. Surprising to me, the transformer power LED lit up even though the battery pack was on the *output* side of the transformer. This made me wonder if the USB power from the mini-ITX is causing a similar "back flow" through the transformer and thus back to the mini-ITX via the 12v line, thereby creating a loop. Could the problem perhaps be fixed by some kind of diode on the output side of the DC-DC transformer that prevents this kind of backflow?

Thanks!
patrick

Dave
09-27-2007, 01:30 PM
And I can use the same USB cable with the servo controller on my desktop PC and it works fine. The problem does *not* occur when I connect the servo controller DC power jack to a separate 7.2v battery pack (6 AA NiMH). It only occurs when I connect to the 7v output of the DC-DC transformer that is inline with the main 12v battery that in turn powers the mini-ITX.
Sounds like the fault is in the DC-DC transformer. I don't know enough about your power system to say for sure, though.


Surprising to me, the transformer power LED lit up even though the battery pack was on the *output* side of the transformer. This made me wonder if the USB power from the mini-ITX is causing a similar "back flow" through the transformer and thus back to the mini-ITX via the 12v line, thereby creating a loop. Could the problem perhaps be fixed by some kind of diode on the output side of the DC-DC transformer that prevents this kind of backflow?

This shouldn't be the case, because the USB is isolated from the DC input on the Phidget board. Take a good look at the Phidget board. Check for damaged traces and make sure there are no solder blobs bridging pins on the IC's.

fitchett
09-27-2007, 11:19 PM
Patrick,

I recommend trying connecting the 12VDC directly to the PhidgetServo DC Input. This is not great for efficiency, but it would be a good test.

It's not surprising that your DC/DC converter allows current flow from output->input, there is normally a diode that allows that.

Most servos will have a glitch on startup that briefly drives the motor at full power; depending on what servos you are using, there might be a inrush of 2.5 Amps for each servo.

Chester

09-28-2007, 07:38 AM
I have solved the problem: I placed a diode (60V, 5A) in series with the + line on the output side of the DC-DC transformer and everything now works A-OK. So it looks like Dave was right--it was the transformer, in the sense that I guess a better design would have had the diode already in place.

Thanks again for everyone's help!!

--patrick