EK-TM4C1294XL: Synchronize PWM

Might your parameter, "PWM_GEN_MODE_NO_SYNC" for (both) PWM_GEN 3 & 2 (w/in PWMGenConfigure()) - effectively defeat/disable your (later) call to PWMSync() ? I've not used this vendor's PWM modules in quite some time - but believe that to be your issue...

Found past code - the parameter w/in PWMGenConfigure() which worked for us: "PWM_GEN_MODE_SYNC."     You got (most) of the hard code parts right - curious how you "missed" that "sync turn-off..."

Should this not (quickly) resolve I'll search/find our old code which clearly (did) sync 3 PWM Generators...   (further - years ago I posted that exact code on this forum...)

I had tried PWM_GEN_MODE_SYNC before, but that seemed to completely break the PWM signal.

"Completely break" inadequately describes your issue when employing "PWM_GEN_MODE_SYNC."

Our firm has many K units in the field - which properly "Sync" multiple PWM Generators - and employ that parameter.

Did you "humor me" by simply adding the suggested parameter change - and then re-running your code - and (properly) monitoring the outputs?   (that's completely unclear in your abbreviated writing)

You may benefit from imposing a delay prior to your call of, "PWMSyncUpdate()."

Eureka!

Remove:
GPIOPinConfigure(GPIO_PK4_M0PWM6);
GPIOPinConfigure(GPIO_PG1_M0PWM5);

You require (only) GPIOPinType()... Again - I've not visited this code for years - our "RUNNING/SYNCED" PWM avoids PinConfig calls.

Further - you probably want, "SysCtlPeripheralEnable(SYSCTL_PERIPH_PWM0);" to occur prior to your entire PWM Code Block.

We "PWMSyncTimebase()" immediately following PWMGenEnable() - only then call "PWMOutputState()."

Odd, even with these changes I've still got the same problem with the pins staying high.

I definitely need the GPIOPinConfigure, otherwise I don't get any output at all.

SysCtlPeripheralEnable(SYSCTL_PERIPH_PWM0);
GPIOPinConfigure(GPIO_PK4_M0PWM6);
GPIOPinConfigure(GPIO_PG1_M0PWM5);
GPIOPinTypePWM(GPIO_PORTK_BASE, GPIO_PIN_4);
GPIOPinTypePWM(GPIO_PORTG_BASE, GPIO_PIN_1);

PWMGenConfigure(PWM0_BASE, PWM_GEN_3, PWM_GEN_MODE_DOWN | PWM_GEN_MODE_SYNC);
PWMGenConfigure(PWM0_BASE, PWM_GEN_2, PWM_GEN_MODE_DOWN | PWM_GEN_MODE_SYNC);

PWMGenPeriodSet(PWM0_BASE, PWM_GEN_3, (BOARD_CLOCK / TIM5_FREQ) - 1);
PWMPulseWidthSet(PWM0_BASE, PWM_OUT_6, TIM5B_10US);
PWMGenPeriodSet(PWM0_BASE, PWM_GEN_2, (BOARD_CLOCK / TIM5_FREQ) - 1);
PWMPulseWidthSet(PWM0_BASE, PWM_OUT_5, TIM5B_10US);

PWMGenEnable(PWM0_BASE, PWM_GEN_3);
PWMGenEnable(PWM0_BASE, PWM_GEN_2);

PWMSyncTimeBase(PWM0_BASE, PWM_OUT_5_BIT | PWM_OUT_6_BIT);
PWMSyncUpdate(PWM0_BASE, PWM_OUT_5_BIT | PWM_OUT_6_BIT);

PWMOutputState(PWM0_BASE, PWM_OUT_6_BIT | PWM_OUT_5_BIT, true);

I'm bothered - given you much work - we've nothing (yet) to show for that.

Forgive me - but is there the chance that (other) code is "undoing" the new code block you present here? Or - might some external connection be impacting those PWM pins?

In addition - might you (for clarity & duplication by my team) replace the "unique" labels w/in "PeriodSet & WidthSet()" w/hard coded, numeric values. I've just noted that your MCU is 1294 - we've only used LX4F & TM4C123. (use higher speed MCUs from others - normally).

And - only those pins (officially designated as PWM) may be so used. Are you sure that "K-4 & G-1" are (legitimate) PWM (candidate) pins. You cannot "force" non PWM pins into PWM mode. (we chase our tails if I don't ask such...)

[edit]...I (now) note that newer MCUs (such as yours) DO require "GPIOPinConfigure()" - mea culpa.    (that call does NOT appear anywhere in our past (successful) Synced PWM code!   {we ran find w/out success}

I appreciate your cooperation/efforts - thank you. I have a local TM4C123 LPad - but NO 129 based parts.

Would you be so good as to (temporarily) "comment out" SyncTimeBase() & Sync Update - WHILE using parameter, "PWM_GEN_MODE_NO_SYNC" - and monitor & report. Does PWM Output then restore?  (yet it will remain UNSync'ed.)

Forgive me - clearly I'm NOT a multi-tasker.    The goal was to see if we could restore PWM Output upon both channels when employing, "PWM_GEN_MODE_SYNC."   (you've thus far reported the LOSS of PWM when that parameter is employed...)   I apologize.    (waiting for my flight @ an airport)

Are you triggering upon a single, edge event?    (if so the "Sync" may not have yet established)    Also - iirc - PWMSyncUpdate() should be called w/some regularity - as the "sync" may erode w/time.   (recall this finding was observed several years past...)

Perhaps worthwhile to try a different PWM pin set.    (from 2 different PWM Generators - to see if you've encountered a unique "pin limitation.")    

Even though it has been years past - I distinctly recall the (2nd & 3rd) PWM Channels, "Jumping into alignment" when we called, "SyncTimeBase & SyncUpdate()."

Thank you - I've now "twin" GREEN Rewards - one (kindly) from you - the second a "Shamrock Shake" upon landing.

You are NOT alone in stumbling on such detail. As I noted - the User Manual fails to adequately & properly describe - and NO example delves into this level of PWM Control.

You were extremely coachable - sorry for the "monkey motion" - I discovered that our 1st PWM efforts were made w/vendor's "LM3S" MCUs - which did NOT support GPIOPinConfigure()!    Later we moved to LX4F - and sure enough - PinConfigure() (as you thought) was right there!

Devil (always) delights in such detail.    Marketing's "push" upon poor tech writer(s) led to manual's inadequacy in this area.   (I'd bet)