Servos don't always turn as expected

Alan Ogborne

I·worked out and tested the pulse commands to made my boe bot turn· 90 or 180·degrees left or right but when I use these commands in a maze running program that involves left and right turns the beo bot does not always turn as far as expected.·Any thoughts?

T-man

Does your boebot have I.R.·or other sensors working?

T-man···

allanlane5

The actual number of degrees your robot will turn for a given PULSOUT setting depends on how strong the battery is, and the material the robot is running on.

Thus you need some way to 'close the loop' to keep your robot on track. Some people use IR-LED's with IR-Decoders to measure the distance between walls. Some people use 'encoders' on each wheel, which tells how far each wheel has moved.

Alan Ogborne

Yes I do have an IR sensor that checks to see if the path is clear before the robot turns. Perhaps this is sucking some power form the servos- I'll check. I'll also consider the options suggested by allanlane5

Many thanks for your prompt and helpful replies

Alan

Mike Green

The IR sensor and emitters don't draw that much current (compared to the servos for example). The main problem is that continuous rotation servos like those used for the BOE-BOT wheels are inherently a little inaccurate. Ordinary servo motors have a position sensor, but they can only rotate 90 to 180 degrees. In the continuous motion servos, this sensor is disabled. As allanlane5 mentioned, you can add encoders to the wheels to track the actual amount of movement although they don't account for slippage on the underlying surface. In the end, in a maze or other situation where you have repeated movement, you need some way to recalibrate the "dead reckoning".

T-man

Consider also that your I.R. may be working just fine
detecting maze related obstacles and your servos are simply being instructed accordingly.

T-man