TMS570LS1224: NHET+HTU or ETPWM for 1MHz based sine wave generation?

Hi Geert,

The appnote below shows how to generate sine waves using N2HET and HTU:

https://www.ti.com/lit/an/spna217/spna217.pdf?ts=1636409773614&ref_url=https%253A%252F%252Fwww.google.es%252F

Yes, we have already implemented this app note.
However we seem to be suffering from jitter.

This jitter is a problem for the digital amplifier.

Our base for our PWM frequency is 1MHz to generate a 20kHz sine, is this 1MHz too fast for the NHET+HTU combination?

(see second link in post 1)

Thanks,

Geert

Hi Geer,

1MHz should be fine. 

If ECMP is used to generate PWM, the PWM frequency is determined by LRP and CNT_Max value:

PWM period P = (CNT_Max + 1) * LRP 

We put our custom board aside and used the XL2-TMS57012 Evaluation board as a reference since the lines to the N2HET2[4] ouput we are using are much shorter on the XL2-TMS57012 board.

What we see is that there is a 0.5V "undershoot" on some edges (as shown in the middle of the screen).

This occurs less often at lower base frequencies (50kHz is ok, 200kHz is ok), but when raising the frequency to 440kHz we see that the overshoots and undershoots are more present. This causes problems when connecting our class D amplifier module.

What is the cause of these overshoots/undershoots?
- Pcb layout? (Would be strange if the very compact XL2-TMS57012 board already suffers from that?)
- Software of HET timer?

How can we get rid of this?

Thanks,

Geert

Hi Geert,

I think the overshoot and undershoot are caused by the impedance mismatch of the transmission line. 

You can reduce the overshoot or undershoot by making the line length short compared with the rise and fall times of the waveform. You can add a series termination to the line or add a LPF to reduce the overshoot and undershoot.