Help with PWM6 and PWM7 on 16F18857 ???

[picster]

Hello all,

I'm trying to get PWM6 and PWM7 working on the 16F18857, and unfortunately these two are not accessible with the HPWM command. I've got them set to 50% duty cycle and I believe set to operate on the same timer (TIMER2) as the CCP HPWM peripherals, yet I get NADA when I look at the outputs. All the HPWM outputs (CCP 1-5) work as anticipated.

It's not critical that the first timing period is accurate, so sloppy entry without an interrupt enabling at the end of a timing cycle period is fine.

Any suggestions on what I've missed here?

Thanks in advance for any brilliant ideas...




'Using PIC16F18857

DEFINE OSC 8 '8MHz
OSCFRQ=$03

TRISA=0 'all outputs
ANSELA=0 '1=Analog, 0=Digital
TRISB=0 'all outputs
ANSELB=0 'all Digital
TRISC=0 'all outputs
ANSELC=0 'all Digital

'PPS MODULE (Peripheral Pin Select)
RA2PPS=$0E 'PWM6OUT OUTPUT on RA2
RA3PPS=$0F 'PWM7OUT OUTPUT on RA3

' *** HPWM format as follows:
'HPWM Channel, Dutycycle, Frequency

HPWM 1, 0, 1000 '51/255 or 0% duty cycle pin 13
HPWM 2, 102, 1000 '110/255 or 40% duty cycle pin 12
HPWM 3, 153, 1000 '153/255 or 60% duty cycle pin 26
HPWM 4, 204, 1000 '204/255 or 80% duty cycle pin 21
HPWM 5, 127, 1000 '240/255 or 50% duty cycle pin 6

' Now, we need to select the same timer for PWM6 and 7 that is used for CCP1 (TMR2)
CCPTMRS1=%00010101

PWM6CON=%10010000 'enable and active high
PWM7CON=%10010000 'enable and active high

'set at 50% for both to test
PWM6DCH=128:PWM6DCL=0 'PORTA.2 (pin 4)
PWM7DCH=128:PWM7DCL=0 'PORTA.3 (pin 5)

keepgoing:
pause 500
goto keepgoing

end

[richard]

using the code configurator [mcc] for a 16f18875
yields the following settings , not unsimilar to yours. it works perfectly.
maybe your pps is locked , your code is incomplete so I can't tell.



1.5khz @32mhz fosc
PWM6CON = $80;
// DC 42;
PWM6DCH = $2A;
// DC 3;
PWM6DCL = $C0;
// Select timer
CCPTMRS1.2 = 1
CCPTMRS1.3 = 0
// T2CS FOSC/4;
T2CLKCON = 1;
// T2PSYNC Not Synchronized; T2MODE Software control; T2CKPOL Rising Edge; T2CKSYNC Not Synchronized;
T2HLT = 0;
// T2RSEL T2CKIPPS pin;
T2RST = 0;
// PR2 85;
T2PR = $55;
// TMR2 0;
T2TMR = 0;
// T2CKPS 1:64; T2OUTPS 1:1; TMR2ON on;
T2CON = $E0;
RA2PPS = $0E; //RA2->PWM6:PWM6OUT;
TRISA.2 = 0
ANSELA = 0;

[picster]

Actually that's ALL the Code... so I need to check the configuration word settings (I do this when prompted the programmer interface pops up) to ensure it's not locked, is that correct?

I'm ignoring all the Timer2 config stuff because I'm going with the assumption that it's already set up to the correct period via the HPWM commands.

[mpgmike]

I personally have not had very good luck assuming anything. I tend to spell everything out as a matter of routine and habit. You might have to manipulate your PWM6/7 outputs manually with the PWM6/7DCH/L SFRs. Not that big of deal once you delve in and learn how. The data sheets are pretty concise about how it needs to be done. I would list the TMR2 SFRs just to be sure. Also, the default timer for ALL CCP and PWM functions is TMR2. There are a few things pertaining to PWM6/7 in the CCP section of the Data Sheet.

[picster]

Might be best to abandon using the CCP HPWM at the same time until I get PWM6/7 working, that way I can config the timer registers independently and know exactly what they are. I can break off and use TMR4 later if I find that CCP PWM interferes.

[richard]

There is no technical reason that prevents simultaneous pwm use of all the CCP and PWM modules with the same timer .
I just made this example with mplabx/xc8 , it works exactly as expected.
I cannot try it with pbp as my version does not support that chip

[picster]

Good to know, thanks for the input!

I've probably missed something simple, but still can't see it in my code snippet.

[picster]

Ok, thanks all for verifying approach. I found the issue to be that the resolution of the duty cycle in bits was 9, not 10, (it seems it varies with the PR2 value predetermined by the HPWM command), so with the MSB of the PWMxDCH set at 1 (PWM6DCH=128), the result was "always on" at the output, whereas I was anticipating a 50% duty cycle waveform.

[picster]

Ok, thanks all for verifying approach. I found the issue to be that the resolution of the duty cycle in bits was 9, not 10, (it seems it varies with the PR2 value predetermined by the HPWM command), so with the MSB of the PWMxDCH set at 1 (PWM6DCH=128), the result was "always on" at the output, whereas I was anticipating a 50% duty cycle waveform.

Note that this was at 1KHz HPWM on CCP1 through 5. At 2KHz, you regain full 10 bit resolution again, according to my trials (using a clock speed of 8MHz).