TMS320F28388D: EPWM sync in tms320

Hello Stevin,

I'm not certain how much experience you have with the ePWM, so I'll start at a basic level and work up, alright? If this is too long, a succinct version is at the end of the post, after my signature. Additionally, key notes are bolded to easily be noticed.

If we have 2 different ePWM modules, at the most basic level we just have 2 different counters and quite a bit of logic that reacts to those counters in the form of the CC, AQ, DB, etc. submodules. There is really only 2 things that affect those counters:

1. Time-Base submodule configuration
2. Sync-In signals from outside the ePWM

(This is a bit of an oversimplification, but I believe it works well for explaining the usage of the synch scheme)

Now, of these 2 options, the TB submodule configuration is done at the beginning of the program and is designed behavior. The TRM section 26.4 goes over all available options for this initial configuration. What's important to us right now is the following excerpts:

• Figure 26-5, an overview of the Time-Base Submodule from a logic perspective
  • Look through this image to find where the Sync-In or 'EPWMxSYNCI' signal is located. Notice how it's only in one spot?
  • If you follow the wire this is connected to, after the OR gate, you come to a switch controlled by TBCTL[PHSEN], or Time-Base Control[Phase Enable].
    • Effectively, what this means is that the only way for the ePWM Sync-In signal to affect the CTR is if PHSEN=1
  • Next, look at the name of the port on the counter that this wire connects to- "Load." This means that the Sync-In signal controls a "loading" functionality for the TBCTR. Remember this, it's important!

• Figure 26-10, an example of Time-Base Submodule behavior in count-down mode
  • There are numerous examples of these waveforms, but this one is particularly useful for explaining the behavior of the EPWMxSYNCI signal
  • First, look at the EPWMxSYNCI signal- the sync-in signal is a pulse that affects the system, not a constant signal.

Now, we finally arrive at the succinct explanation on how does ePWM synchronization work? As you can see from the figures and explanation above,

When an ePWM receives a Sync-In pulse, it loads the value stored in the TBPHS (phase) register, replacing the current value of the TBCTR.

For example, if '150' is stored in TBPHS, and the counter is at 250, counting up to a period of 350, we would expect a Sync-In signal to set the counter value to 150, counting up to a period of 350. By default, TBPHS is 0.

Each ePWM is capable of a receiving a Sync-In signal from numerous places (listed in the TRM section 26.17.2.3 ). If you look at that list, you'll notice that many of the options are EPWMx.SYNCOUT. If you look back at the first figure I posted, you'll see that each ePWM is capable of generating a sync-out pulse in reaction to various events, such as reaching CTR=0. If TBPHS is 0 and SYNCOUT is generated at CTR=0, then the receiving ePWM counter should be identical to the sending ePWM counter!

I hope this explanation helps,

Jason Osborn

Summarized version:

• Each ePWM can send or receive a synchronization pulse. When an ePWM receives a synchronization pulse, it replaces the current CTR value in the TB submodule with the value stored in TBPHS. This perfectly synchronizes the two signals. If TBPHS =/= 0, the receiving ePWM is offset from the sending ePWM by the amount stored in TBPHS.