Right. I guess I can at least prevent them from continuing to run in Driver Mode. Just still not a fix for mid-autonomous malfunctions.
Maybe an accelerometer is in order…
Hi James,
Sorry to hear about your bad experience with the programming skills. I’m not sure this has any bearing on the topic, but we built a brand new robot for Colorado States’ tournament, it had a total of 6 IME’s on it, and we never experienced any issues with lost control. (although we didn’t qualify for worlds) I was pretty nervous after our experiences early in the year, but in the end it has worked out o.k. for us. The one somewhat relevant point is that the last IME in the chain is at the top of a vertical lift system The wires stretch to about 36 inches. no where near the floor.
I hope you better success in the future, and I hope your team gets to worlds.
Cheers Kb
I have one team qualified for worlds.
The programming skills run was sort of frustrating, there’s a random aspect to every run, you never know quite how the game objects will react and this year is hard (which is a good thing), however, we had two good runs out of six tries (the beginning was the tricky part) but then saw encoder resets in both of them. The software had safety timeouts and the robot could partially recover, it was using line detection, two sonars and a gyro in addition to three IMEs, but the 30 or so points we lost was the difference between making the top 30. It was a long shot anyway and we used a different design to most robots this year that really needed more development.
After that experience I managed to reproduce the issue with my own demo robot which has five IMEs, the last in the chain is also high off the ground. I’ve put numerous hours into understanding this problem over the last week and have been sending test data to CMU. We cannot solve the root cause with software but it can be made to be more tolerant and my recent tests with a forthcoming beta over the weekend show an order of magnitude improvement in frequency of IME resets (ie. 10x less often). I think the new beta in conjunction with software I’m working on to flywheel over any residual errors would have made the previous resets I encountered unnoticeable. Anyway it’s still a work in progress and it’s up to CMU as to when they roll out the next version.
Talking to some DuPont friends here in Delaware, they suggested Mylar. It has wonderful electrical resistive properties and is pretty darn thin.
So if we glued some mylar film to the underside of the IME circuit board and adhered it to the underside using some non conductive electronics glue? You would have to poke a hole for the sensor and be careful you don’t glue over the sensor.
Will the distance be too short and get mylar all jammed in the gears?
Will the sensor hole give a static discharge jumping point still?
Would snipping down any longer leads also help? But could that break the IME if you snip too much and break the solder connection?
Congrats on the qualification, and thanks for your diligence in investigation.
I am looking forward to a new release as well, but I’d rather see it delayed to ensure the stability of the new features is up to par, than have it rushed out, it is a complex program. I have been running beta releases of one sort or another for several months which have worked for what I’m working on, but my focus has been admittedly fairly narrow. I really appreciate the support from CMU.
Cheers Kb
These types of mods seem like they would not be legal for tournament play. I would wait for an official fix, and stick to using anti-static spray on the field for now especially if that works.
Cheers Kb
Well yeah. But they are inside the motor case, who would know?! 
Seriously we’re looking for a real fix for the long term. Practice last night showed they can drive for a while, then bump the wall, and whammo! Dropped! So anecdotal evidence continues to pile up. (using v3.54 RobotC with no anti-static spray on the field in a nice dry forced air electric/propane combo heated room with no humidifier. Arm has a 6" polycarb scoop rubbing against the floor pretty much the whole time.)
We will not be opening up those motors until we know the robot is done for the season. Too much hassle! We don’t have any more I2C IME’s laying around and the Vex site says they are still out of stock (seems to be a popular item even with so much reported bugginess).