Pwm with pulsin.c base

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

Hello I need to read pulses in 4 pic entries modulate this signal and put it in other 4 outputs I managed to control the high pulse however not low, the pulses are of servo engine 1 to 2ms.The pulse down this very long over 60ms I need 20ms, the PWM works perfectly without pulsin.c code but it does not work but I need it to measure the pulse of time.

temtronic · Posted: Wed Feb 10, 2016 7:40 am

you should post the 'pulsin.c' code or where you got it from. It's not on my version of PCM( 4.xxx).

Jay

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

1- format your code with the code button
2- list PULSIN.C as without it your code is NONSENSICAL

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

Pulsin.c

Obtained here in the forum

http://www.ccsinfo.com/forum/viewtopic.php?t=42353&highlight=pulsin

temtronic · Posted: Wed Feb 10, 2016 8:07 am

Perfect ! Now we can help....

Knowing PCM created it is has to work.....

first start off with ONE servo,get it working..then 2,3,4
It's a LOT easier and 'proper technique' to start small, build on working code.

Jay

asmboy · Joined: 20 Nov 2007 Posts: 2128 Location: albany ny

by "servo engine" i believe you mean hobby style pulse width controlled servo right ?

PCM's program was created for impaired PICS that have no CCP capability
but the 877 DOES have the hardware.

have you read the CCP section of the
16f877 datasheet to see how it could help you?

what IS your project and what are you trying to control?
Your post is very hard to make sense of unless you explain what your project IS and what you are controlling..
add some DIMENSION please

temtronic · Posted: Wed Feb 10, 2016 8:46 am

I _think_ he's trying to read 4 input signals ( 1-2ms) and then control 4 RC servos with the data..

It looks like he's bodged code from several sources....
I understand English isn't everyone's 1st language but we really need a better description of the project.

He should start with one input-> PIC-->one output

Jay

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

ok I'll work the signal 1 radio control RC receiver, PPM signal and the pic increase or decrease the duration of the positive pulse, the only problem and that I'm not able to control the negative part of the pulse with the above code, not to use the CCP as are 4 different signals

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

This part controls the negative pulse time should be in 20ms but this around 60ms.
If you use this code in alone I have a PPM signal 1 and 2ms 50Hz.

Code:

{
for(x=0; x<t;x++){
//x=x+1;
if(x>pulse0)
{
output_low(pin_c0);
}
if(x<pulse0)
{
output_high(pin_c0);
}
//************************************
if(x>=t) // 20ms
{
x=0;//reset each 20ms
}
}
}

Ttelmah · Joined: 11 Mar 2010 Posts: 19499

Take a deep breath, and step back.

What you currently have is never properly going to work.
Bits of code from various locations assembled without any understanding of what they do, and more importantly of how they will interact.
Big problem is that every single direction through a loop like this will differ slightly in time. So if you tweak everything to get your 20mSec, then you just change one thing (one of the time values), the timing will then be wrong....
The delay while the values are read, will also change according to the incoming pulses. and result in even more variation. This is why you are seeing the timing error for your 60mSec.

You need to be switching to actually using a hardware clock, and effectively multi-tasking. Use the hardware CCP's to generate the output pulses. I posted single pulse CCP code to generate a servo pulse only a few weeks ago, and for what you want (with the pulses sequential), this can very easily be modified to handle four pulses rather than one.
Then realise that 'pulsein', was published as a 'bodge' to allow a pulse width to be measured on a chip that did not have a CCP. To do the measurement without affecting the output pulse, I'd be suggesting you need to look at using the interrupt on change, on the four input bits (so feeding them to B4 to B7), and then recording the value from a single timer to measure the widths. The main loop can then sit and when these values have updated, calculate the new times, and feed these into the code for the outputs.

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

The operation of the code was understood I use this PWM in several other codes, the part that is not mine and the pulsin.c .The intention was to replace a reading done by a adc, and a pot for the values returned by pulsin code and thus controlling the PWM. I posted here because after several attempts with other PWM until even with Timer use not got success the pulse of negative time ever and larger than expected with any implemented PWM

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

And how sometimes the solution is simple and this is our view but we could not see by some oversight, asked for help here. But I see the PWM is really incompatible with the control made through the code pulsin.

Ttelmah · Joined: 11 Mar 2010 Posts: 19499

No.

The point is that pulsin takes _time_, and this varies according to the pulse widths involved, and the phase relationship between the pulses being tested (given that you test multiple pulses in sequence).
Your PWM code won't maintain accurate gaps, if anything in the loop involved changes in time.
You need to generate the timing between the pulses here using a timer, so that it becomes independant of the other things happening.

Laus · Joined: 16 Jan 2016 Posts: 17 Location: Brazil

The problem is that the pulsin disable the timer sometimes and this causes more errors in the PWM.

Mike Walne · Joined: 19 Feb 2004 Posts: 1785 Location: Boston Spa UK

I'm just plain lost.

Maybe some pictures/diagrams/ASCIIart showing us what you are trying to achieve will help.
Then we can come up with alternative ways of doing the job in hand.

Mike