Powering your USB Hub at Competition

Technical Technical Discussion
#1

Even though this thread is discussing the control system, I found this to be important enough for its own thread.

Twice our robot became unresponsive, either right at the start of teleop, or shortly after. Pete the FTA at NJ looked carefully at the robot when it happened both times, and in both cases the symptom was no comms from cRio to the wireless adapter.

We strongly believe the solution was insufficient power from the classmate PC to the USB hub for the joysticks.

The solution was to plug both the red and the black USB connectors into the Classmate, doubling the available power for the hub devices.

We have four devices connected to the hub: a joystick, and gamepad, the E-Stop button and the PSOC. I heard from three teams that they had experienced similar issues, and connecting the hub to the classmate twice ( the cable to the hub is a “Y” cable) fixed it. As it did for us.

I can’t say how USB affected comms on the 'bot, but after we connected the hub this way, the problem went away. YMMV.

Don

#2

Remember that the E-Stop button can be removed for FMS official rounds - so that will aleviate some power consumption.

#3

I thought teams were being encouraged to plug the Cypress directly into one USB hub, and plugging the joysticks and gamepads into the other. As for how this may cause comms issues, since the Cypress device has to be reprogrammed each time it is enumerated. Perhaps power dropouts cause this to happen repeatedly.

Greg McKaskle

#4

Here’s our problem, though sort of off-topic.
How do you recharge your classmate during eliminations when you have to stay on the field the whole time? There is no power to plug into on the driver station area.
This was a major issue for our team as we were scrambling to beg and borrow extra CHARGED batteries.
What happens at CMP?

#5

Would it be correct to assume that the PSOC is the biggest culprit for USB power consumption?

We have two joysticks and an Xbox controller (but do not currently use the PSOC), and have never had any issues running on Classmate battery power (using only the black plug).

#6

read this thread

#7

In these situations on the field that I witnessed, Comms were never actually lost, it just seemed like they were. All statuses were good, but the robot didn’t respond whatsoever to any driver controls. Essentially, the Classmate wasn’t hearing from the joysticks connected through the hub.

At NJ if I was consulted or found out about it, I had teams plug the PSOC directly into the Classmate and just use the hub for low power devices. A team that didn’t use the PSOC, but did have a higher power consumption game controller also had to move that device to a direct Classmate connection.

Using both the red and black USB plugs did not always solve the problem for some, including one of the finalists. They finally went with direct connections for both joysticks. They would have benefitted from just a better redistribution of devices, but under the gun during a match the goal is to just get it working.

#8

Does anyone know if there is a difference in available USB power between when the classmate is plugged in, or just operating off the battery?

#9

There should be no difference, according to the USB standard. In order for USB devices to work on an extremely wide variety of machines (from laptops to desktops, Dell to HP to Apple, etc) the standard dictates exactly how the USB port should act, including how much power is available, the transfer rate, protocol, etc. For any device with a USB output, that standard is going to be followed.

#10

The Targus USB hub became a real problem for us during a practice day with the Fembots. Luckily I had a spare USB hub in my laptop bag and donated it to the cause. When we returned to our classroom, we found a new use for the hub…Target Practice!.

#11

If this solves the problems we were having at the Traverse City competition, this will be very, very appreciated. Out of all of the matches played, we were able to play 1 full match completely, and 2 others for a very brief period. After that it appeared that there was no comm. The field manager helped out a ton, and told us that each time a few seconds into teleop, comm was lost. We had tried multiple wireless adapters and locations for it to sit with no avail. We also have a bunch of joysticks connected to a usb hub only connect to one port on the classmate.

After seeing this thread and mentioning it to one of our mentors, he suggested getting a powered usb hub that connects to a normal ac/dc line, but whip up a device to connect it to something like a 9V battery to power it while up at the match stations. Would this provide more available power for the usb hub then getting a y connector and plugging it into both ports on the classmate?

#12

If you were actually losing communication as indicated by the FMS system, then the issue mentioned in this thread is not your problem. The issue mentioned here is where communication is present, but the robot cannot be controlled using USB devices.

#13

hmmmmm good point. It wouldn’t hurt to do it anyway. It’s weird because we stay through autonomous, but as soon as start to move in teleop, controls cut out and all the victors flash and no comm :confused: . Could this be a programming error? Our team used LabView this year. As a second backup plan, i got nominated to recode our controls in java to test that out. We’re still utterly confused by this problem though.

#14

This is exactly what happened to us in two matches. In one match, we lost communication immediately after switching from auto to teleop. In another match, we were able to drive in teleop for a couple of seconds before losing comms.
We don’t know for sure whether plugging in the second USB cable on the hub fixed the problem because we also switched out the radio and an ethernet cable. However after these modifications, it didn’t happen again.

#15

Huh. Thanks for the input. That may have narrowed down our problem to the power going through the USB hub considering we swapped multiple Ethernet cables and tried 3 different wireless gaming adapters in different locations on the robot.

#16

We had the same error.

We fixed it by moving our radio away from other electronics and metal, uninstalled the DS, reinstalled the DS (from a different image than the origional), and reupdated the DS (from a different image than the origional).

We believe that we had a bad install of the DS since another classmmate (one loaned to us) worked well.

#17

My team has found that on the classmates one usb port gets no power or very little and the other one gets practically all the power ( i could better explain this but im extremely tired right now). i advise test both usb ports and label the one that gets the power and use the usb hub on it

#18

Thanks Mark. Sorry I didn’t have a chance to sneak onto the field and say hello.

We did have the PSOC connected to the right-hand port, and the hub on the left hand port, and we experienced the problem (as described by Ethan above). In the pits, we saw the security light flash orange on occasion, and later saw it flashing orange and green very rapidly. I don;'t know what, if anything, these were telling us.
All at the same time, we did swap the ethernet cable, secure the power plug better, and swap out the WGA with our spare. Our WGA was well-located, away from most metal, out in the clear. (Hey, I’m a ham. I know these things)

So, the conclusion is, we have no idea what fixed it, but both 1089 and 25 gave us that advice, and someone asked the pit announcer to say it as well.

Don

#19

This is triage. Just do whatever works to get back on the field.
If they were dead on the field, I had drivers patch their drive joysticks directly into the Classmate and forget kickers and other controls just to get moving. That usually worked (except when that velcro pulled out that Classmate battery).

I came late when your robot failed on the field and got that yellow warning light, and stood behind you at the practice field after you’d swapped out the WGA and were retesting. You had things well in hand.

I must say your radio placement was indeed excellent and it was very easy to see the status lights when you were on the field.

There was one particular robot that had everything so buried that not a single status light could be seen, and they bolted their cover on so we couldn’t get at anything. They sat dead for a few matches and there wasn’t a thing we could do to help them.

#20

Except for one victor. But that’s how we troubleshoot, too, so I was pretty adamant when we were discussing component placement.