Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
Skip to content

hpaluch/pic32mx-pwm

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

16 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

PIC32MX LED PWM

Introductory project for Microstick II with PIC32MX.

It glows on-board LED using PWM.

Here is example output from Digilent AD2 scope/analyzer:

PWM Example output

There are these important signals:

  • Yellow Channel 1: Toggles on Timer1 interrupt where we change OC1RS, in our example it changed from 15 to 20 (see decoded number on Bus)
  • Blue CHannel 2 is actual PWM output to on-board LED. You can see various measures on the measurements tab - PWM frequency is 1 kHz, and current duty (coresponding to OC1RS=15) is 27%
  • TMR2_PER (on digital output only) is whole PWM period (normally set to 61) and it defines 1 kHz PWM frequency (PR1 timer register)

And here is Workspace file for Digilent WaveForms software:

Used peripherals:

  • on-chip oscillator FRC 8MHz scaled down to 4MHz using FRCDIV by 2, so SYSCLK and PBCLK is 4 Mhz - for easy probing using affordable scope - in my case Digilent AD2
  • TMR1 - 100ms interval, flashing external LED on PIN3 (for debug) and increasing PWM duty (causing glowing effect on LED)
  • TMR2 - 1ms interval, PR2=61, clock source for PWM (OCMP1)
  • OC1 - PWM output to on-board LED D6 on Pin 2 (primary function RA0)
  • we also send current value of OC1RS register to "data bus" using RB7,8,9,14,14,15 GPIO pins

WARNING!

I intended to use PMP (Parallel Master Port) interface for this data-bus, but was really annoyed by Chip errata, where PMP completely overtakes all Address Pins, even when they are configured in GPIO mode in PTEN register(!)

Other problem is that PMD0 and PMD1 (Data 0 and Data 1 bits) signals conflict with Programmer/Debugger pins PGED1 and PGEC1. Programming switch does not help, because alternative PGDx,PGCx pins are on unsupported location (these alternate pins are used for USB signals by this chip and are not available for Programmer/Debugger)

Used Pins on Microstick II:

Pin Signal Note
1 /MCLR Reset + Programmer/Debugger
2 OC1/RA0 PWM - red LED
3 RA1 TMR1 - Debug LED
4 PGED1 Data for Programmer/Debugger
5 PGEC1 Clock for Programmer/Debugger
9 RA2 TMR2/PWM period Toggle
10 CLKO/RA3 PBCLK Output
16 RB7 DATA0 (output of OC1RS)
17 RB8 DATA1
18 RB9 DATA2
24 RB13 DATA3
25 RB14 DATA4
26 RB15 DATA5

Pins 16 to 26 are useful to show real OC1RS value on scope/analyzer.

Controlling PWM is easy. Its period is equal to source timer period (in our example TMR2 with period 62 => PR2=61 (-1). PWM duty is controlled with OC2RS register. Here are few examples:

  • OC2RS=0 - Duty 0% (LED off)
  • OC2RS=31 - Duty 51%
  • OC2RS=61 - (=PR2) - Duty nearly 100% - 1clock tick
  • OC2RS=62 - (=PR2+1) - Duty 100%

More information is available on PIC32 Output Compare data sheet - PDF.

WARNING!

Older datasheets for PIC24 clearly stated that PRx registers contain actually PERIOD+1, because they start counting from 0, but Interrupt/Reset occurs after reaching value in PRx (period compare registers match triggers on next clock edge). However, for some weird reason this important information is missing from PIC32 datasheets(!)

Compare this datasheet where is no useful information on period and PRx register: DS61105F (they use careful wording "TMR count register matches the PR period" without clear explanation)

And look here: DS70205D on page 13 where the text is clear:

This code generates an interrupt after counting 1000 rising edges in the T1CK pin.

...
PR1 = 999;  // Load the period value

This explains why MCC Harmony 5 now sets PERIOD-1 to PRx registers while old MHC Harmony 3 just used PERIOD (which is wrong)

There is now After build hook generating assembler listing .lst file under:

firmware\pic32mx-pwm.X\dist\pic32mx_pwm\BUILD_TYPE\pic32mx-pwm.X.BUILD_TYPE.lst

Where BUILD_TYPE is debug or production.

Verification

Here is system clock 4MHz measured on PIN10 CLKO with Digilent AD2 using BNC probe with 10x attenuation:

CLKO 4 MHz

And here is Workspace file for Digilent WaveForms software:

Please note that even 4 Mhz clock is quite high and its testing this 100 Mhz scope to its limits. It is main reason why I did not use 48 MHz system clock (which is now default in MCC Tool), because it is practically impossible to measure such clock without suitable Counter/Frequency meter.

Requirements

Hardware requirements

Software requirements

  • XC32 compiler - tested version v4.30
  • MPLAB X IDE - tested version v6.15
    • code regenerated with MCC Harmony (was MHC Harmony 3)
    • notable changes: both Timers have now set PRx less -1 (!) - see above notes
    • had to manually create (from other project) file firmware/src/config/pic32mx_pwm/pic32mx_pwm.mhc/settings.yml otherwise it was attempting to again and again import and convert old MHC Harmony 3 project...

Notes

PIC32MX is 32-bit MIPS32 M4K based CPU Core with peripherals from Microchip. Please see

Releases

No releases published

Packages

No packages published