PBP projects for R/C models

[Kenjones1935]

This PWM exercise is designed just for what I have done.

http://www.melabs.com/resources/samples/pbp/hardpwm.bas

After loading the .hex into the RC car (carrying my PICkit2), hooking up another battery and connecting CCP1 to the electronic speed control, The wheels spin!!

Ain't technology grande!

Ken

[mackrackit]

COOL!!!
Another piece down!

Did you get Walter's PWM reading code to work?
If so then maybe now you can combine the two.

If incoming PWM is X then wheels turn sort of thing?

[Kenjones1935]

As I said, the wheels go round, but...

The wheels are not going round and round in the pattern I expected. In particular they only go backwards. The code claims to sweep from 20% to 80% cuty cycle.

SOOO, can I use my USB connection and PICkit 2 Logic and Analyze Tool to see what is actually happening?

Ken

[ScaleRobotics]

I believe you are using the PWM example here: http://www.melabs.com/resources/samples/pbp/hardpwm.bas

This is a PWM that is not supposed to work with RC servo's. It's wave form is different. It is used to dim LED's etc. I am a little surprised that it can turn the wheels at all. I have a scope to help me view wave forms. I have tried for a few minutes to see what the pickit's logic tool can do, but I have not had any success. I am pretty sure it is just a problem with the operator (me).

[rmteo]

This is what the program header says:

' PicBasic Pro Program to demonstrate hardware PWM.
' Output is a 1Khz signal with duty cycle sweeping
' from 20% to 80% once per second

Since the frequency is 1KHz, its period is 1mS. The duty cycle is sweeping from 20% to 80% therefore the pulse width is varying from 0.2mS to 0.8mS. There are 2 problems here. For R/C signals, the period should be 20mS or 50Hz. Second, the pulse width should be 1.5mS for neutral, 1.0mS for extreme left (or full reverse) and 2.0mS for extreme right (or full forward). So your car is responding to a very short pulse (even less than full reverse). You have to modify the values in the program to get it to do what you expect.

[Kenjones1935]

Thanks gang.

I came to much the same conclusion myself. I have been trying to figure out what exactly this PWM code does so that I can modify it meet the RC specs. The comments in the program are fairly clear. My first attempts have not worked but I am optimistic.

Interesting how rusty my brain has become.

Ken

[Kenjones1935]

Here is the code that makes the wheels spin.

http://www.melabs.com/resources/samples/pbp/hardpwm.bas

There are no PBP commands that overtly make pulses. They are created by appropriately loading CCP1CON and CCPR1L. I found CCP1CON easily enough. It is listed in the INDEX of the 16F887 Data Sheet. CCPRxL and CCPRxH are not in INDEX. They are quietly mentioned on page 127. But why is there no HPWM command per page 86 in the PBP manual?


Although I see how this might be a low cpu overhead way to create PWM, I don't want to learn about this now. I want to use PBP and ASM commands to control my motor and servo. Maybe tomorrow all this will make sense.

Ken

[Kenjones1935]

The code in

http://www.melabs.com/resources/samples/pbp/hardpwm.bas

makes the wheels of my car go backwards in bursts. The bursts last, maybe 3/4 second then the wheels stop for a 1/4 second. I look at the code . I think that it is supposed to sweep through all pulse sizes in a second. I surmise the short ones are too short so the wheels don't move. As they get longer the come up near the low end of the electronic speed control spec. They start to move. Then the second is over.

I would like to pump PWM pulses out CCP1 that have a frequency of one every 20 msec (50 per second) and a pulse width that I can (by changing the code) vary from 1 msec to 2 msec. Am I correct in guessing that the frequency of the pulses is dictated by the PAUSE command. One pulse per loop through mainloop:

So far all my attempts have created no forward wheel motion at all! If one of you has the time please modify hardpwm.bas so that I understand it. Or, better yet, show me how to use PULSOUT. That does not work for me either.

Ken

[rmteo]

I gave you that info in post #44.

Since the frequency is 1KHz, its period is 1mS. The duty cycle is sweeping from 20% to 80% therefore the pulse width is varying from 0.2mS to 0.8mS. There are 2 problems here. For R/C signals, the period should be 20mS or 50Hz. Second, the pulse width should be 1.5mS for neutral, 1.0mS for extreme left (or full reverse) and 2.0mS for extreme right (or full forward). So your car is responding to a very short pulse (even less than full reverse). You have to modify the values in the program to get it to do what you expect.

[Kenjones1935]

Thank you all for your help. I posted those quotes because they back up what you have been saying and what I thought I coded.

Trouble is my PIC code only makes my car wheels go backwards and stop. I have not been able to create a PWM stream that drives the wheels forward. I know the ESC works because the wheels go fine under radio control.

Oscilloscope, here I come.

Ken

[ScaleRobotics]

Oscilloscope, here I come.

Great idea Ken. I don't know how much you have modified hardpwm.bas, but here it is in it's native form. First pic is about 20 percent, and second pic is about 80 percent (of a 1ms pulse):





You will notice that the min is about 0.2ms and max is 0.8ms.

Your servo or speed controller requires a pulse width between 1.0ms and 2.0ms. It would probably like to see these pulses every 20ms or so, not every 1 ms (as seen here).

And here is what you want it to look like (from my rc receiver):