SERVOS.C Problem

Guest · Posted: Fri Jan 06, 2006 9:02 am

Hi,
I am using the servo.c driver on my 2 wheeled robot. The chip is a PIC18F452 running at 10Mhz.

I can't get thed driver to work correctly with the motors. No matter what spinning direction I choose or what TIMER_1_DIV I set, the servos are always spinning in the same direction.

I am sure this is a timing problem. Has any of you come acros this problem?

Here is a little test program I wrote

Guest · Posted: Fri Jan 06, 2006 5:16 pm

Hi,

What do the stop servo & set servo routines look like?

gael · Joined: 08 Jan 2005 Posts: 7

I posted the initial question. Here is the servo.c file:

Guest · Posted: Sat Jan 07, 2006 7:40 am

Hi,
That looks ok.

What kind of motors are you driving? Are they dc motors (hence H-bridge) or RC servos (3 wires, power, 0v and control).

Do you have access to an oscilloscope? Check that you are getting some PWM signals on the pins D6 and D7.

Also, try using "LEFT" instead of "LEFT_SERVO", and "RIGHT" instead of "RIGHT_SERVO" as defined in the servo.c file. I don't know where you have defined the left and right_servo words.

Se what you get, I hope that helps.

Guest · Posted: Sat Jan 07, 2006 9:21 am

Guest · Posted: Sat Jan 07, 2006 9:59 am

Do you realise how RC servos 'work', when used like this? A RC servo, will not automatically 'stop', if fed a 1.5mSec pulse. It will only do so, if the feedback pot (which is disconnected from the mechanical drive), is set to exactly the centre position. The movement for the other pulse widths is also dependant on the position of this pot. I'd suspect that you have either completely removed the pot (in which case the servos will seek continuously in one direction, no matter what pulse is fed to them), or it is set to one end of it's travel (which will have a similar effect on most servos).
You should write a much simpler program, that just outputs a fixed 1.5mSec pulse, and adjust the pot till the servo stops with this. Only once they are 'set up' like this, then try the code you are using.

Best Wishes

Guest · Posted: Sat Jan 07, 2006 10:16 am

Hi,

I inclined to agree with the last reply.

You mention that the servos are "spinning" twice. I assume you are using the servos to drive wheels instead of positioning some link between 0-~180 degrees.

The servos will stop if used in positioning mode. If you apply 50% duty (perfect square wave with off-time and on-times being equal), then the servo will position itself somewhere in the middle of its positioning range.

In speed mode (no feedback potentiometer) it will never reach a target position and will continue to run in one direction. There are website showing how to modify and RC servo for speed mode.

I'm not sure if the mod will give you both speed and direction control.

Here is a site that may help:
http://www.seattlerobotics.org/guide/servohack.html

Google "modify servo" and there should be more links.

Hope it helps.

gael · Joined: 08 Jan 2005 Posts: 7

Ttelmah · Guest

What is your clock rate?.
If you look at (say) the statement to calculate the 'ticks' needed for the longest pulse, this will only work, if the clock is below:
65535/(0.0021*8). 3.9MHz
Otherwise the counter value will be incorrectly calculated, resulting in an overflow.
You can handle higher clocks, by increasing the divider used on the timer.
As a comment, you are ging to be very likely to find the motor does not stop on 1.5mSec reliably. This is down to the 'deadband' setting of the servo amp, and on some chips, it is tiny.

Best Wishes

gael · Joined: 08 Jan 2005 Posts: 7

Thanks a lot for your help. I finally found what the problem was. It was the potentiometer. I modified it so that I can ajust its value using a screwdriver from the outside. Now, the servos seem to be working fine.

The PIC is running at 10 Mhz. I think a timer div of 4 should do fine.