Robot loosing communication when switching directions

[Tom Line]

All we know at this point is that you're losing communication.

This could be due to a physical problem like loose ethernet connections or power connections. It could also be an electrical problem like low battery voltage (or plugged in to the wrong port). You are using the transformer, correct?

[dudefise]

It's a problem with power distibution.The load spike caused by reversing the motors is causing the cRio or radio to lose power... use the protected 12 and 24 volt boost supplies on the bottom of the PD board.

[cgmv123]

A little more unlikely explanation, but still worth considering, is that when you reverse the directions, you are causing a spike (literally a spark jumping a gap) inside the motor which can mess with you Wi-Fi signals. You can sometimes see this with the Fisher-Price and other unsealed motors. It's unlikely, but it's still worth considering. Does the Wi-Fi re-lock after some time?

[Al Skierkiewicz]

This is exactly the result when the Crio chassis is electrically connected to the frame and a motor or other wiring also is connected with the frame. Typical of certain Banebots motors shipped after the beginning of build in 2011, was a case short. Even if you were sure you had adequate insulation under the Crio, check it again with a meter. When you reverse motors under this condition, the power going to the Crio is cut in half by the short. The result is a Crio reboot.

[Mr. Lim]

I second both James' and Al's suggestions, and have some tips to offer to help troubleshoot:

1) Al's suggestion: If it IS a cRIO reboot, you'll lose comm for about 1 minute, then it will come back. The time duration of your comm loss is a good indicator here. If you are running any Banebot 775s on your robot this is likely your problem, and you can check by taking a meter, disconnecting the motor from the speed controller, and measuring resistance between the positive terminal on the motor and the case of the motor. You will want to rotate the motor slightly as your measure. If at any point you see <100k ohm resistance, you have a defective motor which is causing the voltage to dip on your entire electrical bus. There are a few fixes for this that have been posted on CD. Try searching for "case short".

2) James' suggestion: As soon as you lose comm, have two people immediately check the status lights on robot. One person for the cRIO, the other for the radio. Both have telltale blinking-light sequences which are indicative that they've been reset recently. If only the radio looks like it's been reset, then James' suggestion is your best bet. Generally with radio-only resets, comm is re-established faster than 1 minute.

[Phalanx]

I was working with / trying to help these guys yesterday at Brunswick Eruption. There were numerous problems being identified and we were trouble shooting them one at a time. This is as far as were able to get.

1) The ground wasn't firmly attached to the stud on the PD. Fixed
2) One Gear Box was loose and wobbling. Fixed
3) We measured the power drop between the studs on the PD with a multi-meter when they changed directions it went from 12volts to 8volts and lower.
We recommended fully charged batteries.
4) I didn't check if the CRIO rebooted, but the DLINK rebooted every time.
5) They also have the D-Link powered by a 20AMP circuit like James posted previously.
6) Lastly, I'm wondering if direct driving Mechanum Wheels using 4 - CIMPLE Boxes wasn't enough of a mechanical reduction and hence cause too much draw.

[Alan Anderson]

Originally Posted by Phalanx

5) They also have the D-Link powered by a 20AMP circuit like James posted previously.

Fixing that should take care of the communication dropout immediately. Then they can start working on the other electrical issues (battery charge/condition, connection reliability, etc.).

[Ether]

Originally Posted by Phalanx

Lastly, I'm wondering if direct driving Mechanum Wheels using 4 - CIMPLE Boxes wasn't enough of a mechanical reduction and hence cause too much draw.

Yes. With 6" wheels, at least 1.8x more reduction would be appropriate. For 8" wheels, make that 2.4x additional reduction.

[Greg McKaskle]

If it IS a cRIO reboot, you'll lose comm for about 1 minute,

Can you time this to see what the actual number is? And if it is close to one minute, can you give language details?

Greg McKaskle

[Al Skierkiewicz]

The loose negative power supply lead and the improper wiring for the radio are enough of a smoking gun for me. The battery will regularly drop to eight volts in a complete direction reversal. All motors are going through excessive current draw so the voltage drop across the internal resistance of the battery and the wiring and breaker feeding the PD is enough to cause this in four motor drives. The three terminal regulator is only speced to maintain output voltage down to about seven volts input. It is for this reason, the voltage regulator must be wired to the 12 volt radio output of the PD. That output will maintain voltage down to about 4.5 volts on the battery input.

[kinganu123]

I apologize for the lack of communication on my end. I actually had one of the members post this up for me and I never found the actual thread.
But yes, thanks to Mr. Reffler, we did do some troubleshooting, as he said.
Now, there ARE 3 Bane Bot Motors on the system, but I have no idea when they were ordered, but they were definitely before the 2011 season.
We are currently off of school for the week, but once we get in, I'll switch the radio over to the dedicated port and check our motors for the case short.
I'll update you guys once I can do the few tests and possible fix, now that I know where this thread is.
Thank You All Very Much For Your Help!
UPDATE
Now that I think about it, the Bane Bots Motors are the unlikely cause of the problem, because they aren't running when the radio resets itself. It's the CIM Motors that are the most probable cause.
@Ether, could you please explain to me what you mean by a 2.4X reduction? We are using 8" Mecs and I believe the ratio is 1:3 or some absurdly low number.
UPDATE
btw, to answer Mr. McKaskle's question, we do get connected back to the robot, but after a good 5 minutes. It's definitely a radio reset. Furthermore, we are using Java and running cRIO image version 28

[Mark McLeod]

Originally Posted by kinganu123

btw, to answer Mr. McKaskle's question, we do get connected back to the robot, but after a good 5 minutes. It's definitely a radio reset. Furthermore, we are using Java and running cRIO image version 28

If your radio is taking 5 minutes to reconnect to the field, then it could be that the manual switch on the back of the DLink is in the Auto position rather than Bridge.

[Joe Ross]

Have you rewired the robot since competition, or were you able to pass inspection twice with the radio wired to a PD breaker?

[Ether]

Originally Posted by kinganu123

@Ether, could you please explain to me what you mean by a 2.4X reduction? We are using 8" Mecs and I believe the ratio is 1:3 or some absurdly low number.

I said 2.4x additional reduction.

You said in an earlier post "I'm wondering if direct driving Mechanum Wheels using 4 - CIMple Boxes wasn't enough of a mechanical reduction and hence cause too much draw".

If you are direct-driving 8" wheels with nothing but a CIMPLE box for gear reduction, your motors are overloaded. CIMPLE box provides 4.67:1 reduction. You need more gear reduction. Preferably at least 2.4x more gear reduction. So, 2.4x4.67 = 11.2 total reduction from CIM shaft to wheel.

[EricVanWyk]

Originally Posted by Ether

I said 2.4x additional reduction...

I think this should be treated as a separate issue, possibly in a separate thread. Incorrect motor reductions should be unable to cause the root issue, provided the electrical system is fine. They can definitely aggravate the issue, but they won't be the root cause. Even if the problem appears to go away after changing the reduction, it will still be lingering in the background (and harder to debug when it returns).

As many other have already said, move the radio to the dedicated radio power supply. Also, add a few things to your standard pre-flight checklist - electrical connector security, chassis fault, motor chassis fault.