Our team competed at the Arkansas Regional and we ran practice matches all Thursday, no problems at all. There was no score keeping during these matches and we did good. During our first qualification match we ran great until the last 15 seconds we lost power to the C-Rio, and all Comm. and Code. During the next match our battery fell out. And in every other qualification match we would run, then our C-Rio would reboot. Sometimes it would happen when we took hard hits, others we would just drive forward and then reboot. So first we replaced the C-Rio, same problem, all of the wiring was fine, had multiple inspectors confirm that, Then we replaced the PD board, took it to the practice field and it worked fine, we ran at full speed into the low goal multiple times, it was just fine. Then we went onto the field on saturday and same problem happened. We replaced the main breaker, same problem. So then we got picked for elims. and during the time in-between our match, we ran on the practice field with our alliance members at least 20 minutes, just fine. We went out on the field, drove a little bit, and re booted. same thing happened next match. Any ideas what could be the problem?
Although I don’t have any solution or hypothesis for your problem, I did notice something in one of our matches last weekend. The FMS messed up the hot goals (As seen here).
None were hot for the first ~5 seconds, then the left was hot for the last ~5 seconds. But right as the bell rings for teleop to start, the right turns yellow, indicating that it is hot. If you look to the left of the frame, you can tell (watch a few times) that the same thing happened for the blue alliance.
The FMS system this year could definitely use improvement, however it would not cause your robot to reboot. You have a wiring or power distribution problem that you still haven’t discovered. They can be tricky!
I never saw much of that during our regional, and the inspectors were absolutely great working tirelessly to diagnose the problem, it just seems weird because we replaced the all of the electronics and had the same problem and it was even worse in eliminations even though we didn’t take a lot of hits? and we had just been running great? I’m just unsure, not upset, I just want to know what we need to fix.
While the C-Rio is rebooting, it isn’t the only thing, in several matches the d-link did as well, and it seemed like we would randomly just shut down and not work, and under diagnostics the robot light would be green, not the bridge. I’m just wondering if there could be something in our code having a bad reaction with FMS? I would completely agree it is a power or wiring issue if we hadn’t been running perfectly fine any time other than when connected to the field.
here is a link to one of the matches we played in. We are the robot with the huge orange net, at around 1:55 in the video, we lost all communication.
Did you work with the CSA (Control System Advisor aka “orange hat”) at your event? They are experts in robot troubleshooting.
From your description it does sound like a power/wiring problem, possibly a connection to the radio or a short to the frame.
Take a look at the Driver station log viewer for the matches where the problem occurred. Details of how to do this can be found at the WPI screensteps pages
It might also be a good idea to cross post this in the Technical Forums.
^+1 to what Mike said above.
We had a similar issue in one match. Afterwards we found an intermittent short to our frame that had gone undetected during inspection. We fixed that, and never saw the problem again.
Yes we did work with him, he was unsure after we replaced everything, we looked at the logs, definite signs of the C-Rio rebooting, and he also saw the D-Link reboot. The thing that I am confused about is how we can run just fine multiple times when we are off the field, but never work when we are on? I do not know how to cross post? Are you saying to post the same thing in a new thread under technical Forums? Thanks for the feedback guys! I don’t mean to try to argue or anything, I just want to figure this out.
Do you want to post your logs from the matches where you disconnected?
I suspect an intermittent short somewhere in your system. We had a similar problem today with our practice robot, and that was the issue.
We also had a similar problem last year and could not figure it out. We found several things we thought were causing it, but it just kept happening. I’d be very interested to hear your solution if you find it.
In my experience some wiring issues only become a problem in the hardest of hits. These tend to happen in the heat of competition and rarely on the practice field.
You could post your original question in a new thread in the Tech forums and link to this one. It’s very likely the underlying issue is an intermittent & tricky problem with the robot so I’d suggest a different title to the thread.
I don’t have access to them now, but I will definitely post them as soon as I do. In no way am I trying to blame the field, we built the robot, the results are more than likely our fault, I’m just very confused and concerned for for other teams if they run into a problem like this. At first it seemed a power issue, then we made sure everything was good, then we just kept fixing things but it just kept happening, the CSA checked the logs every time showing the same thing. I will try to post those sometime this week.
I made a new thread, and again, thanks for the feedback. http://www.chiefdelphi.com/forums/showthread.php?threadid=127623
When you next get to unbag that comp bot…Look at the Radio Power Converter (there are 4 wires…1 yellow, 2 black center ground wires, and 1 red).
Are the 2 center grounds possibly crossed over? The black nearest the red wire should be paired w/ the red. The black nearest the yellow wire should be paired w/ the yellow…And the yellow should be connected to the radio power + side (black w/ white stripe wire to the radio).
We had no problems whatsoever w/ that exact type of rebooting issue (you seem to be alluding to), w/ our competition bot, then we put that bot into the bag happy as can be after much practice on the 18th of Feb…That 1 was wired correctly.
Started working w/ our practice bot later, since we had been waiting for a ball P.U. gearbox to arrive, and lo and behold after that was in and running right, we had the same exact “once in a while connection loss issue” that you speak of when testing the practice bot"…Shoot, dry shoot, load a ball (no problems, except a very intermittent loss of connection that got worse, not better). We zip tied the radio power connector, changed radios, checked for a frame grounding issue, we thought we did everything…Then, we noticed a real error in wiring (one of the students wired the radio power converter incorrectly, as the 2 center grounds were crossed over).
When we removed the power converter (holding it w/ wires away from you, wires top, from the left), the order should have been yellow /black-black /red…(Make sure they are paired just like that).
Wired correctly now (;-0), we have never lost connection again no matter what we hit, or do, after many rough hours of practice. (It was simply the 2 blacks crossed over…The black nearest the red was w/ the yellow…and the black nearest the yellow was w/ the red. Those 2 grounds are not created equal by any means on that power converter). Wasn’t easy to notice as that converter was in a spot way down under and very hard to see unless you were looking very closely.
The power connector to the d-Link radios are also very famous for intermittent loose connections and subject to wiggling in the socket easily…tape it, zip tie it, etc. (but of course, you cannot solder it…You need to get that radio configured at the events at the kiosk, as you do already know).
Please post back if you find that is the problem please (It may not be, but ours was…and both act the same). And GOOD LUCK in tracking down an electrical gremlin, as they are not always easy.
Edited: BTW, that practice bot electrical board was the 1st 1 wired…(It was the practice board of course).
Let’s start from the bottom with this:
Please locate a multimeter with a Min/Max function, put it on your robot setup to measure minimum voltage DC and attach it to your battery lugs.
Then drive your robot around for 3 minutes (longer than a match) and see what the lowest voltage recorded is. Please post that here.
Then drive your robot around gently bumping stuff and see what the lowest recorded voltage is (similar to a situation with a field robot collision) and post that lowest recorded voltage here.
Generally a multimeter records voltages in Min/Max about 1,000 times a second. That’s not bad considering the most (not all) critical electronic loads on the robot have additional capacitance which will ride through short drops.
Now once you achieve this, let’s keep in mind that at high current you might have wiring somewhere that’s too high in resistance (poor connections or an intermittent short). To test this put that same multimeter on the power inputs to these critical electronic devices (cRIO, radio, radio DC/DC converter, digital side car) again set to minimum voltage DC and repeat the tests above. Post that data.
With this in hand you are not depending on any aspect of your FIRST robot to measure the power supply critical to your robot. In fairness usually the driver’s station charts will show voltage drops - but just in case there’s something not right there this process I’ve outlined will show such problems as well.
Also I may have missed something but which programming language did your team use for your robot? C++, Java or LabView? Does your robot use vision?
Keep in mind that the field system does have some very real timing differences from other private wireless and wired robot situations. It’s possible to cause yourself some headaches. This year I’ve seen problems with putting the Java camera get instance in the robot init section (this is a perfect example it only was an issue on the real field). I’ve seen problems with LabView and one team’s attempt to use the standard compressor and pressure switch cause their robot to reboot consistently because LabView kept saying that relay 1 was an invalid index (clearly that’s not an invalid index and strangely if you edit the block and set that directly it’s fine for a bit and then the problem returns). Also with C++ you can easily shoot yourself in the foot (not so much a problem with C++ but just the nature having that much power it can blow up).
So I advise you even if you provide the voltages I request above to take your code and features apart starting at the most simple and adding on till you reach the full compliment. In this manner if one thing is causing headaches you can isolate which one(s) require more evaluation.
As always techhelpbb has a good set of steps to follow.
He asked
Please locate a multimeter with a Min/Max function, put it on your robot setup to measure minimum voltage DC and attach it to your battery lugs.
Putting on my psychic problem hat (*) I’m wondering if there is a correlation between your battery falling off the robot issue and the power cable and attachment to the Power Distribution Board. I’m thinking that the match was not the first time the battery has come off the robot.
So I’d like you to Please locate a multimeter with a Min/Max function, put it on your robot setup to measure minimum voltage DC and attach it to your Power Distribution lugs. That way you will be able to see if there is a problem with your main power feed.
Report back with the stuff that’s been asked in the prior threads, that should exactly pinpoint the problem.
Good luck!
(*)Yes, I said psychic hat. I know the difference since I have one of each. :rolleyes:
To echo above. Check your software. When on the field several tasks on the CRIO interact with the FMS that pretty much sit idle when on a practice field.
If you have some sort of memory/buffer overrun or bad pointer in C++ that can cause that type of behavior.
Software will not cause your radio to reboot. This definitely sounds like a power problem.
We had a similar problem this year. During driver practice we would be able to run the robot for about 3-4 minutes. After that the cRIO would randomly reboot or we would pop the 120 amp main fuse. We checked if the frame was grounded or if any metal shavings were found in the PD board and found nothing. This problem continued multiple times afterwards but we still haven’t diagnosed the problem because the next day was bag and tag day.
However after searching chief delphi for answers we found this: http://www.chiefdelphi.com/forums/showthread.php?t=91733
We use a Banebots 775 to tilt our shooter. It is directly mounted to our aluminum frame. Since the motor only shorts out as different spots. This can explains why you don’t see any electrical problems when we test for conductivity to the frame. We currently think that this may be our problem.
If you use a Banebot I would disconnect power to it and see if you experience the same problem.