Hi to all,

I’m working on a low-frequency inverter using a variation of unipolar SPWM. It uses a modified unipolar scheme similar to the EG8010: in the first half of the sinewave, one leg holds the negative-side MOSFETs ON while the other pulses with a complementary SPWM signal. In the second half, the roles swap — the opposite leg holds the negative rail while the first leg pulses. Only one leg switches at the carrier frequency per half-cycle, which helps distribute switching heat between both sides of the H-bridge. During AC-coupled charging, I also throttle the output frequency to control charging power. Frequency ranges from 60.00 Hz to 62.10 Hz, depending on system power levels. Carrier frequency change  (19.9kHz to 20.7Khz during AC coupling throttle) is used to change fundamental frequency. I'm using two output chokes, each made from three PJ ferrite cores stacked together, with 3 turns per winding, resulting in approximately 235 µH per choke. Deadtime is set to 2 µs.

The inverter outputs a clean sinewave under normal load, but I’m seeing a distinct notch in the waveform during AC-coupled operation (Enphase IQ7+ microinverters charging the battery through the inverter’s AC output).

Here’s what I’ve confirmed:

The notch consistently appears at ~¼ of the half-cycle (around the π/4 point).

Output RMS voltage is around 254 VAC during AC coupling. As solar power increases, I set the code to lower the SPWM modulation to maintain stable AC voltage. At ~5500 W of charging, modulation drops to around 53%.

The notching only occurs during AC-coupled charging and disappears as solar power decreases.

Immediately after the notch, I see a burst of PWM pulses — more switching activity than usual.



I suspect the output chokes may be saturating when current peaks from the backfed microinverters. My deduction comes from the visible ripple, increased pulse activity after the notch, and the rise in transformer temperature (154 °F) and MOSFET temperature (120 °F). Under normal load, temps are around 130 °F (transformer) and 105 °F (MOSFETs).

Has anyone seen similar notching behavior during AC coupling or strong backfeed? Could choke saturation be responsible for the extra switching after the notch? Any advice on how to confirm saturation?

Appreciate any insight or shared experience.









Thanks in advance for the help!
Edited 2025-06-17 15:13 by jony787

What's the grid profile you are using on the iQ7's?  If it's a default grid profile (for on-grid setup) then the voltage waveform distortion is most likely due to anti-islanding disturbance injections from the iQ7's.  Enphase anit-islanding disturbance injections are very aggressive especially with some older firmware versions.  To avoid problems you should set the grid profile to an off-grid one or one that disables anti-islanding all together.  Also, you should upgrade the firmware to the latest version if it's not already done.

Thanks for your reply analog8484

  analog8484 said  

What's the grid profile you are using on the iQ7's?

Im using a dedicated Off Grid profile " Off-Grid FW60"

 

  analog8484 said  

If it's a default grid profile (for on-grid setup) then the voltage waveform distortion is most likely due to anti-islanding disturbance injections from the iQ7's.  Enphase anit-islanding disturbance injections are very aggressive especially with some older firmware versions. To avoid problems you should set the grid profile to an off-grid one or one that disables anti-islanding all together.  Also, you should upgrade the firmware to the latest version if it's not already done.

My microinverters are on firmware 520-00082-r01-v04.40.01.

I will assume since this is a Off Grid profile, its not affected by the anti-islanding disturbance injections.