So what happend with our camera?

[Rombus]

So as an electrical guy, I'm trying to pin down some issue with our camera after the end of our season. I cant seem to figure out what was going on, and the programming team wasn't to sure either. So hopefully someone can give me a nugget of info about what happened.

The setup:
Camera at the top of our turret.
Wiring run was 5 feet, About a foot from the end there is a DB25 connector where the 2 servo runs, 1 TTL run, and 2 limit switches run though for quick disconnect.
2 runs of cable: 1 was TTL and both servos, the second was power. Each run was heatshrinked with clear heatshirnk ran though the alum frame of our turret.
All wires had continuity and were solid runs with unmeasurable resistance.

The problems:
1. Intermittent connection, sometimes the camera would work, sometimes it wouldn't. Wiring seemed fine, TTL chip was hot glued to RC, and the Power and data cables were hot glued into the case our camera was at and worked fine for a few matches then mysteriously stopped working. I couldn't track down the issue, but the programming team swore their code was fine, and when it did work, it worked fine.
2. During disable, the camera would want to go to 0, 0 rather than 127,127, it would jitter there, however during normal operation, camera would recenter to 127,127 fine.
3. during programming, the camera would jitter like above.
4. This happened on both a 2006 RC and a 2004 RC, on the 2006 RC, it seemed that when the camera worked and tracked, it would not jitter, but when it was not going to work, it would jitter. on the 2004 RC the jitter seemed almost random.

So what do you guys think? Is it a wiring issue? Code? or something else? Ill try to provide whatever answers and data i can if asked. Thanks for your help!

[Keith Watson]

Originally Posted by Rombus

1. Intermittent connection, sometimes the camera would work, sometimes it wouldn't. Wiring seemed fine, TTL chip was hot glued to RC, and the Power and data cables were hot glued into the case our camera was at and worked fine for a few matches then mysteriously stopped working. I couldn't track down the issue, but the programming team swore their code was fine, and when it did work, it worked fine.

If the robot is running and the only thing changing is the camera working and not working it sounds like an intermittent connection. In our case the tilt servo would stop working intermittently so it was obvious which wiring to check. We found it (we thought) by tapping on the various connectors. We cut out the bad connector and just soldered the wires together. But then tapping on the new connection still caused intermittent problems so the wire obviously had an internal break.

Originally Posted by Rombus

2. During disable, the camera would want to go to 0, 0 rather than 127,127, it would jitter there, however during normal operation, camera would recenter to 127,127 fine.

When going from normal operation to disable the PWM outputs get turned off so the camera servos should simply stop. If that's not happening I don't know what might be going on. We tested and found our camera servos could not be driven to full throw without causing binding which makes the servos buzz.

Originally Posted by Rombus

3. during programming, the camera would jitter like above.

During programming, power on, or anything which changed modes, we would get a single servo movement which is typical of when a PWM signal is turned off and on.

Originally Posted by Rombus

4. This happened on both a 2006 RC and a 2004 RC, on the 2006 RC, it seemed that when the camera worked and tracked, it would not jitter, but when it was not going to work, it would jitter. on the 2004 RC the jitter seemed almost random.

So what do you guys think? Is it a wiring issue? Code? or something else? Ill try to provide whatever answers and data i can if asked. Thanks for your help!

Was your camera "not working" simply the servos would stop moving or was it the camera data would stop working (shown on the terminal)?

[Rombus]

sorry for not being clear!

When i say "Camera not working" I mean no data being recived according to our programming teams code and what it displayed in terminal. The programming team was able to force the tracking code to attempt to search without data so we knew our servos were physically and eletrically fine.

I did check every connection every chance i got, but to no avail. I even went so far as to slightly bending the pins of the TTL port to give us a secure and good conducting connection. Internal break is possible, but im not positive, ill probably end up running my DMM on each wire again and see what i can find next time i get up to the school.

According to the documentation i read, setting disable sets all the pwm outputs to 127 (Nutral to the victors). But i could be mistaken, has anyone tried running a Servo with just +5v and ground connected? I was thinking it might be interferance but now im curious, I think i have some spair pinheaders around here and some servos, so ill give that a go tommrow.

Im curious about your single servo movment. What was the other servo doing? Do you happen to know what PWM you had them connected to? This sounds very similar to what was going on with ours, i think 1 servo was acting how you described during programming, and both servos were acting that way during disabled.

[Keith Watson]

Originally Posted by Rombus

sorry for not being clear!

It sounded like you might have encountered multiple problems so I wanted to be clear.

Originally Posted by Rombus

When i say "Camera not working" I mean no data being recived according to our programming teams code and what it displayed in terminal. The programming team was able to force the tracking code to attempt to search without data so we knew our servos were physically and eletrically fine.

We unfortunately had this problem all the time. We finally determined it only happened when the robot was powered up while in autonomous mode and the reset button was pressed. The camera would not initialize when that was done. It will be a while before I again have access to our robot and can run some more tests in an attempt to isolate the problem.

A few other people have reported a similar problems. Some sound like our problem and others sound like an intermittent problem where the camera just stops working. If you look at the 8.2 Battery Voltage Bug thread in the programming forum people reported they changed their code to check if the camera stopped sending packets, and if so to reinitialize the camera.

Originally Posted by Rombus

According to the documentation i read, setting disable sets all the pwm outputs to 127 (Nutral to the victors). But i could be mistaken, has anyone tried running a Servo with just +5v and ground connected? I was thinking it might be interferance but now im curious, I think i have some spair pinheaders around here and some servos, so ill give that a go tommrow.

If the pwm outputs were set to 127 then you could not move the camera servos by hand to another position without them fighting you.

Originally Posted by Rombus

Im curious about your single servo movment. What was the other servo doing? Do you happen to know what PWM you had them connected to? This sounds very similar to what was going on with ours, i think 1 servo was acting how you described during programming, and both servos were acting that way during disabled.

Our pan and tilt servos were on pwm1 and pwm2 as used in the default camera code. During camera search mode if our tilt servo stopped working the pan servo would continue to sweep back and forth.

[KenWittlief]

Im curious about your single servo movment....

When first powered up its normal for servos to twich for a second. If thats what was meant by a 'single servo movement' then this is not unusual.

If he meant that a single servo would move (and continue to move) then something was not right. The servos are suppose to be disabled (127'd) when the robot is disabled.

[Keith Watson]

Originally Posted by KenWittlief

When first powered up its normal for servos to twich for a second.

Yes. There is a big difference between the single twitch when a servo is powered up and the continuous jittering/twitching when the servo is driven until something binds.

Originally Posted by KenWittlief

The servos are suppose to be disabled (127'd) when the robot is disabled.

Think about it, not all devices connected to pwm outputs use 127 as a value for "off". So if disable mode sets values to 127 some devices, including drive motors, would still be operating. Not good.

[Alan Anderson]

Originally Posted by Rombus

2. During disable, the camera would want to go to 0, 0 rather than 127,127, it would jitter there, however during normal operation, camera would recenter to 127,127 fine.

Ours did that too. We chalked it up to crosstalk from the RS-232 communication to the servo PWM inputs, as we have all the signals running through a multi-conductor cable up to the camera assembly. It was fun watching the camera "nod" to the beat of the music before a match.

Today we replaced the last bit of cable to the serial input of the camera in an attempt to track down an intermittent communication problem. It fixed the problem, and it also seems to have stopped the servo jittering at zero during disable.

[Keith Watson]

Originally Posted by Alan Anderson

We chalked it up to crosstalk...

Good find. Long runs of servo wire, or any unshielded straight wire, are susceptible to crosstalk. Twisting the wire helps, as does separating them as Alan points out.

[Rombus]

Well, i thought it might have been crosstalk, but i a fresh pwm wire (2 connected together to make the length) and it still didnet get data. The servos signals wrapped around the ground would cut the crosstalk correct?

[Michael Hill]

Originally Posted by Rombus

Well, i thought it might have been crosstalk, but i a fresh pwm wire (2 connected together to make the length) and it still didnet get data. The servos signals wrapped around the ground would cut the crosstalk correct?

I would check for continuity. Just because something is new doesn't mean it will work. You may have something wrong with the internal pwm pins (inside the body) that may not be letting a signal transfer from 1 pwm cable to the other.

[Greg Marra]

Our CMUCam has an annoying tendancy to freeze up and stop tracking/communicating with the RC. The whole unit is freezing, since the 'tracking' status LED on the camera locks itself on or off and no longer responds to the green light.

We still aren't sure exactly what is causing it, but I think we're going to re-initialize the camera if it stops talking to us. I'm not sure if that will fix it, since it seems like it loses communication. It's quite frustrating, overall.

[Keith Watson]

Originally Posted by Rombus

The servos signals wrapped around the ground would cut the crosstalk correct?

Yes, just like "twisted pair wire.

[Rombus]

Originally Posted by Keith Watson

Yes, just like "twisted pair wire.

i should have just used my orginal idea of cat5 twisted pair to begin with

Thanks for the help everyone, Unfortunitally i havent had a chance to run by the school and go over eveything with a fine tooth comb. When i do ill report back with my findings