Continual Joystick Disconnect During Match

[java4first]

Hi -

We are having a really weird problem that hopefully someone can shed some light on ...

Thanks!
Stu


In a nutshell, during competitions - in a given match, the two joysticks will go 'disabled' at least once, and usually several times during the match teleoperated mode (they may be doing it in autonomous mode as well, but we obviously wouldn't be aware of it), halting active control of the robot and requiring the 'F1' rescan to bring them back to life. We have no problems at all when running tethered in the pits, or when running wirelessly during build season or during out of the bag practice sessions back home. Needless to say, it makes competing hard (and frustrating for the drive team).

More details:

- We are using Lenovo laptops (refurbished if that matters) running Windows 7 with patches current to the end of Build Season. We have no anti-virus or firewall software, and all three variants of the Windows firewall are disabled. We have a power management configuration that is set to never turn anything off when the laptop is powered on.

- We just competed at the Pioneer Valley District Event, and were not starting the FRC Driver Station program until we were connected to the field (which was a change from our normal procedure of coming ready to go - but that seemed to be causing problems for the field and we were asked to change).

- Because of the problems, we started doing a connectivity check of the two controllers after starting the Driver Station before each match started, and they were always green and flashed when you hit a button.

- We have an Ethernet coupler normally, but tried plugging directly into the laptop

- We are using two Logitech Dual Action gamepad controllers. At our last competition, we went out and bought two brand new ones to swap in, and it made no difference. We just bought to Xbox controllers to try at our next competition and see if that makes a difference.

- During the last match, we switched the tab to the one that shows the USB controller plugins - and when it fails, they go from green (connected / present) to grey ... until you run the F1 scan, when they return and go green again.

- We are running one of the Axis cameras at the default configuration for resolution and compression. We tried upping the compression and it didn't make a difference.

- If we look at the log file after the match, the data rate is down at the bottom of the graph, so I don't think it's a problem with communication bandwidth to the robot (although from what I have read in the FMS documentation, I don't see how the connection of the controllers could be affected by that anyway).

- We do notice what seems to be a correlation in spikes on lost packets at the times of the disconnects ... although I'm not sure whether it is related to the problem or caused by the F1 rescan process.

- The software on the robot isn't doing anything special or different from previous years in terms of just using the standard WPILib Java classes to instantiate and poll the controllers for stuff.

- There doesn't seem to be much correlation as to when it occurs, although in the match where we watched the USB tab, it seemed to happen right when the transition from autonomous to teleop occurred.

- It seems unlikely that it's related to the student driver & operator physical actions, as we never have problems at home and they are just as active with the controller motion in their hands there as at a competition.

If there is anything else I can look at or provide for information that would help troubleshoot this, let me know ... but we are really at a loss trying to figure out what the problem is. We do plan to bring several different laptops to our next competition and see if we can change that variable.

[otherguy]

We had a similar issue at our event last weekend Springfield ma, ne district event).
Almost every match we had on the field our operator and driver joystick would swap while in teleoperated mode. The drive team reported having seen something indicating that the joysticks were unplugged in match. When the sticks were recognized as being plugged in again, they would occasionally (often) be swapped. As in the stick which was previously used to move the drivetrain was now affecting the lifting mechanism and intake (operator joystick). And the stick which was previously used for the operator was now controlling the drivetrain.

We are using two Logitech F310 gamepads. Our DS are running fresh installs of Windows 7, w/ latest chipset and Logitech drivers installed.

We attempted using 3 different laptops over the course of the event, one of which was used by another team (4048) without issue (AFAIK) during eliminations.

We have never seen this kind of behavior in the pits/practice field/or during at home tests.

I had presumed this was caused by ESD, as there was n abnormally high amount of static electricity buildup due to the floor coverings used at the event.

We attempted to aid static discharge during matches by having the DS laptop plugged in during the match, this didn't seem to make a difference.

I have not had a chance to go through all our log files yet, but they are available here for reference if anyone finds that helpful: https://drive.google.com/folderview?...DQ&usp=sharing

Logs from the event would be on 3/5 - 3/7, I've inclouded logs from days prior as a baseline comparison in the event that is useful information.

My only though going forward was to try to use two different kinds of joysticks, one F310 and one xbox360 controller, in hopes that if we do lose connection to our sticks in match, they will at least come back up on the right port in the DS.

[Mark McLeod]

Is the laptop plugged in at the Driver Station, or on battery?
USB has a separate power saving setting that can cut power to the usb ports while on battery.
You can (and should) disable that feature through the Device Manager -> USB controllers properties.

Also avoid flashy game controllers with flashy LEDS, as they drain more power and force a usb power cutback all the sooner.

Connect each controller directly to a laptop usb port and not through a usb hub. A hub concentrates the power requirements and will more readily drive can force a usb power cutback.

For the joysticks switching positions, do you double-click on them the lock them in position? You can tell if they are locked because they get an underline.

[otherguy]

I have not tried the double click feature, we will give that a go.

Do you know how the DS software uniquely identifies the sticks? They are the same make/model.

[Greg McKaskle]

Here is what I remember.

DirectInput is used for the nonXBox joysticks. If the joysticks are of different types, I believe that will be the primary distinction. For identical types, the USB port number will be used.

For XBox controllers that are plugged in when the DS boots, they are port based. But if the DS is running and the joysticks are plugged in, the enumeration is based on insertion order. This makes things very confusing at times.

Greg McKaskle

[otherguy]

I was hoping you would respond. Thanks!

Sounds like using two different models of joysticks is the most fool proof route.

[dougwilliams]

We had the same issue during some practice matches this last weekend. We had a new (this season) laptop and xbox controllers.

We believe we found the root cause when we the inner part of the USB connector fell out entirely. We will try and get this replaced by Dell under warranty.

One of our off season projects is now to find a more robust USB solution for competition play. We are considering soldering some controllers or cut-off USB extension cables to the innards of the USB connectors directly.

Does anyone else have a more robust USB connection solution?

[Kingofl337]

Just an update.

We uninstalled the Thinkpad power saving software, changed the USB power settings to always on and switched to wired XBox controllers. We had zero drop outs at UNH this past weekend.

[BitTwiddler]

Originally Posted by Kingofl337

Just an update.

We uninstalled the Thinkpad power saving software, changed the USB power settings to always on and switched to wired XBox controllers. We had zero drop outs at UNH this past weekend.

Team 1726 is having this problem also. We are using a 3 year old Lenovo Thinkpad with Windows 7 for the DS. We use two different models of Logitech gamepads (one is a dual action and the other is similar but with colored buttons). One or both will drop out together during a match but come back after an F1 rescan. We use LabView on the robot.

Sounds like the Lenovos running Windows 7 is a common configuration judging by what I read in this forum.
How did you uninstall the Thinkpad power saving software? I don't have access to the machine right now.

[MrForbes]

I'll see if I can find the laptops in the back of the Suburban