sluggish robot

[abhijit10425]

Hi,

At our regionals, our Robot suddenly started
acting very sluggish. Even at 100% power it moved like it was
on 60% or less. We checked the obvious :
a. friction - if something was touching the wheels or motors
b. if the connector plugs were properly in
c. changed the bricks

Nothing helped.

Other potential issues I can think of:
a. the connector wires are broken inside - need to change the connector
wires and check
b. one of the B or C motors is dying - but I think this would cause the robot
to swerve to one direction which is not the case.
I guess both could be dying - highly unlikely
Do the Mindstorms motors die often ? Is this common ?

Anything else folks can think of ? It's really bizarre. The day before
everything was running fine, and all of a sudden we have this issue.

Abhijit

[Dave Parker]

First make sure you eliminate programming as a possible problem by writing a simple test program to just drive straight (e.g. Move block straight, 100% power, 10 rotations). Then also try two Motor Unlimited blocks followed by a empty loop to eliminate any Move block sync weirdness if a rotation sensor is failing.

You didn't mention changing batteries or testing the voltage. It is possible for the Lithium rechargeable to die and no longer accept a charge even when near empty (and the LED indicator may show green only, implying full charge). If you only have one rechargeable, try putting in some AAs to see if that wakes it up (although note that AAs are naturally stronger than the rechargeable anyway). If a rechargeable gets into this error-dead state, sometimes it will come back if left plugged in overnight, but sometimes it will be dead permanently (in which case you have a $50 paperweight, or LEGO may offer to replace it for you). You can tell when it is dead permanently when the voltage no longer reads as expected (8.2 volts or so) when it claims it is fully charged (green LED only), even after an overnight charge.

Dave Parker
www.nxtprograms.com

[dnb]

Originally Posted by abhijit10425

At our regionals, our Robot suddenly started acting very sluggish. Even at 100% power it moved like it was on 60% or less.

When you say it was sluggish, did you measure the time it took to accomplish a task? And did you compare that time to earlier measurements?

There are times when one of our robots has seemed slow. In these situations I've used a stopwatch and found that there were actually no differences relative to earlier run times.

[abhijit10425]

@DaveParker:

We've been using AA batteries solely for a while. Somehow parts of our programs seem to vary based on battery power. Easier to monitor and change batteries (and run them down a little) with the AAs.

But I'll try out your other suggestions and see what happens....

@dnb:

I didn't actually time the runs. But my observation was just based on my experience the past couple of years of using the NXT. Also, for this year's patent capture run, we had designed a fast single arc using a Move block that got there really fast. And it was visibly slower and veering off to one side on the tournament day.

Abhijit

[eaglesc]

Originally Posted by abhijit10425

@DaveParker:

We've been using AA batteries solely for a while. Somehow parts of our programs seem to vary based on battery power. Easier to monitor and change batteries (and run them down a little) with the AAs.

But I'll try out your other suggestions and see what happens....

Abhijit

Depending on the mechanics of your drive system, performance can be a very strong function of battery level. It's a well known problem -- well enough known that it's a line item in the robot design scoring rubric.

If you are currently on run-of-the-mill alkaline AA's then drop in a set of top of the line lithium ion disposables and you can watch a lot of missions just go wildly off the rails. My first year I co-coached a team and the other coach wanted to drop in shiny new Energizer E2's on competition day --- kids ran exactly one of the their missions on a practice table before ripping them out and putting in regular AA's.

If the problem with the drift doesn't boil down to batteries check the motors for bad gears. The rotation sensor runs close to the motor but there is a long train gears to from the the motor to to the connection point. Especially if the kids are rough with the robot while dragging it around it is possible to mangle a gear.

[JHagan]

Originally Posted by abhijit10425

b. one of the B or C motors is dying - but I think this would cause the robot
to swerve to one direction which is not the case.

If someone could confirm my understanding I'd appreciate it, but I think that if you're using a Move block, the NXT brick will try to synchronize the rotation of the two wheels, which will mean that if one motor has slowed down because it's dying, the NXT will also slow down the other motor to try to keep the robot going in a straight line.

[Dave Parker]

Originally Posted by JHagan

If someone could confirm my understanding I'd appreciate it, but I think that if you're using a Move block, the NXT brick will try to synchronize the rotation of the two wheels, which will mean that if one motor has slowed down because it's dying, the NXT will also slow down the other motor to try to keep the robot going in a straight line.

Yes, the Move block will slow the other wheel to try to compensate, but this will also introduce a zig-zag pattern, often with an initial steer that messes up the original heading.

In more detail, here is what happens with the Move block if, say, the right motor is dying and only half as strong as the left motor:

1. NXT gives 100% power to both motors to start with
2. Robot will right away start steering to the right because the right motor is too slow
3. Move block tries to correct by reducing power to the left motor
4. Robot now turns to the left to try to make up the initial extra degrees on the left side. This motion will be quite slow since it will have to reduce the left motor to something like 30% to get less then the 50%-ish right motor to achieve a left turn correction
5. Once the left correction finishes and both motor now have the same total degrees, power will be increased to both motors
6. Sequence repeats, and robot will travel in a slight zig-zag pattern.
7. Because of the asymmetry in the order of things, the net heading will not be quite straight as aimed.

Dave Parker
www.nxtprograms.com

[GWhizNonKid]

You can simulate the kind of error of which Dave speaks by deliberately adding friction to bear on the sidewall of one of the wheels. Not so much that the geometry is seriously impaired, but rubbing more than incidental contact.

Another side note: If you see the slight zig-zag appear in a part of the program you know to be a straight move block, you are free to suspect a bad motor, or mis-matched motors, but only after you check for extra wheel friction.

[eaglesc]

Originally Posted by GWhizNonKid

You can simulate the kind of error of which Dave speaks by deliberately adding friction to bear on the sidewall of one of the wheels. Not so much that the geometry is seriously impaired, but rubbing more than incidental contact.

Another side note: If you see the slight zig-zag appear in a part of the program you know to be a straight move block, you are free to suspect a bad motor, or mis-matched motors, but only after you check for extra wheel friction.

My crew observed the friction issue a couple of weeks ago when a cable was dragging against one of the wheels. They came to me in a panic -- "the 'bots driving crooked!" This was really odd, not because the robot was doing something weird, but because of the panic; they are some of the better debuggers I've had the pleasure of coaching and generally solve their own problems. But it had been a while since they had faced a hardware problem so the first thing they did was blame the software; good wake up call.

[dahlboy15]

We have tried to drive our robot this year and it keeps wanting to go to the right. we dont know if its because of the drive system, the jaguars, or the programming. we did have the two old motors on one side from last year and the the new ones on the other side, so i dont know if thats the problem.