Sumo Bot Issue

simmel65

So I put together a sumo bot and can't get it straight. It veers to the right. We had it set for

PULSOUT 13, 850
PULSOUT 12, 650

It veered to the right.

Changed it to a few times then got to....

PULSOUT 13, 804
PULSOUT 12, 650

and it still veered right the same amount, then

PULSOUT 13, 803
PULSOUT 12, 650

made it veer left!

Changed the batteries and the servo and it has not resolved the issue.

Any suggestions?

Martin_H

Did you center the servos? Continuous rotation servos usually need to be centered to not move on a pulse of 750. After you do that they still tend to veer to one side, but it can be corrected with a slightly longer pulse to one side.

simmel65

Yes they are centered. That is what is strange...

erco

Something's hinky. Please show us your entire program, and a video if you can. Youtube is fine.

simmel65

Here is the code and a video that runs it. Had to pick it up at each pause so it wouldn't run into the wall, but you get the idea Now the breaking point is at 787/788

http://www.youtube.com/watch?v=cg13e3HmKVM

' {$STAMP BS2}
' {$PBASIC 2.5}
counter VAR Word

PAUSE 1000

FOR counter = 1 TO 2000
PULSOUT 13, 850
PULSOUT 12, 650
NEXT 'veers to the right

PAUSE 3000 'stop 3 sec's

FOR counter = 1 TO 2000
PULSOUT 13, 788
PULSOUT 12, 650
NEXT 'still right

PAUSE 3000 'stop 3 sec's

FOR counter = 1 TO 2000
PULSOUT 13, 787
PULSOUT 12, 650
NEXT 'veers left


END

erco

That is a strange jump. What changes happened between your first post (803/804 breakpoint) and now (787/788)?

Is anything rubbing on your drive wheels? I don't have a Sumobot, but I'm quite familiar with BoeBot. I'm assuming they're similar. Is there a caster ball on this, or do you have a bulldozer-style blade dragging on the ground? Have you tried swapping your servos L/R?

erco

At any rate, you have a non-linear servo response in at least one servo: strange, but not a deal breaker, as long as it functions consistently. Try leaving the servo on pin 13 alone (keep that pulsout constant) and tweak the other servo pulsout on pin 12 to get it to drive straight. When you get into accurate pulsout calibrations for matching servo speeds, you'll want to use a nice stable set of freshly-charged NiMH batteries. Alkaline batteries' voltage drops off linearly with use. They're a frustating variable you don't need in the mix. You might need to add a 5th NiMH cell since those are ~1.3V, either use the BoeBoost module or wire in an extra AA holder on your own. And add a jack or connector so you can connect a trickle charger easily to keep your batteries topped up.

simmel65

Great advice. My son is helping me and said the plow was not straight and he thinks that solved the problem. We also try fresh batteries and test later this week and get back to you. Really appreciate the help here!

W9GFO

First off, servos should receive a signal once every 20 milliseconds. You are missing the standard "PAUSE 20" in your code.

Try this;

' {$STAMP BS2}
' {$PBASIC 2.5}
counter VAR Word

PAUSE 1000

FOR counter = 1 TO 2000
PULSOUT 13, 850
PULSOUT 12, 650
PAUSE 17                      ' use "20" if you like but 17 will get it closer to a pulse every 20 milliseconds.
NEXT

Second, if it still goes one way or the other with a change in value of only one (787 vs 788) you could alternate between the two like this;

' {$STAMP BS2}
' {$PBASIC 2.5}
counter VAR Word

PAUSE 1000

FOR counter = 1 TO 1000

PULSOUT 13, 788
PULSOUT 12, 650
PAUSE17

PULSOUT 13, 787
PULSOUT 12, 650
PAUSE17

NEXT

Third,... there is no third.

erco

W9GFO wrote: »

First off, servos should receive a signal once every 20 milliseconds. You are missing the standard "PAUSE 20" in your code.

Good catch, Rich. I'm assuming that got edited out of the code listing since his video showed the bot moving...

And a great suggestion to try alternating 787/788. I think there may be another issue, but I'd be interested to hear the results if simmel65 can try it out and report back.