Quote:
Originally Posted by jblay
1676 looked dominant when it played but appeared to have communication issues. Can someone clarify exactly what was going wrong?
|
I can try. We would lose comms, the lights on the radio were solid blue (not flickering), the RSL never missed a beat, and the cRio was up and happy every time an FTA could see the lights.
We changed radios several times, radio power supplies (and all related wiring), ethernet cables, redeployed code, re imaged the cRio, cyt the frame rate on the camera again (to 10) and a whole bunch of other things - basically, anything anyone suggested.
Here's the exact scenario, which is 100% repeatable: We power on, connect to the field, then after ~1 minute, we lose comms. Reboot the cRio, wait another ~1 minute, lose comms again. Reboot again, and comms stay solid forever. We were 100% successful with this once we stumbles upon it mid-dat Friday.
We had at least 5 FTAs look it over, and we couldn't nail it down to anything specific.
I have to call out Pete K especially for his HUGE efforts in helping us troubleshoot this thing. He graciously stayed late Thursday and ran a match for us 3 or 4 times, and graciously agreed to let us run a match with 6 robots Saturday morning (the bots were on carts, not on the field, but we were trying to ensure the wireless was properly loaded). The NI guy, and several others all gave it a great shot, but Pete was our hero for the MAR Region Championship.
From what we learned, we think we'll be able to duplicate the field conditions at home and use a packet sniffer (again) to see what's really going on.
I'm not convinced it is a field problem, but more likely a problem in our bot that makes us more sensitive to something in the field. The fact that we never failed as red but always failed as Blue is suspicious. Especially when several teams also failed as blue.
Now this bug is getting personal, we're going to find it and fix it - but maybe not in time for St Louis.