Cheapo Grid Tie Inverter Efficiency Mod

[PerpetualWalnut]

Okay, THIRD time trying to make a new post! Page just reloads with a blank body of text and with no error messages as to why. Do posts need to be approved? If so, then why not indication of that when I click post? Anyways... I'll try again with no pictures attached this time.

I bought one of these (1000W Grid Tie inverter Solar with limiter MPPT pure sine wave DC22-60V/45-90V) off ebay a while back. I found that it would run really hot and I was only getting about 80% - 85% efficiency out of it. It's advertised as upwards of 95% efficiency! Yah right! So I tore it down to see what cheap parts they put in it. I looked at the FETs, TRIACS, and Diodes and found them all to be decent except the diodes which were standard diodes rather than high speed Schottky???

The diodes they used: RHRP8120 https://www.digikey.com/en/products/detail/on-semiconductor/RHRP8120/965460

The diodes I replaced them with: SCS210KGC https://www.digikey.com/en/products/detail/rohm-semiconductor/SCS210KGC/3902856

The inverter runs MUCH cooler now and it looks like I'm getting an extra 75 watts of power! I have yet to do an accurate recalculation of efficiency, but based of my estimates and napkin math I'm getting closer to the advertised 90% - 95% efficiency.

[PerpetualWalnut]

Pictures...

[PerpetualWalnut]

Third times a charm!

[PerpetualWalnut]

Back of board.

[f4eru]

People from manufacturing or ignorant devs probably swapped some incorrect parts !

[nctnico]

Now check EMC compliance... it might be that the design didn't pass with Schottkey diodes...

[Monkeh]

Now check EMC compliance... it might be that the design didn't pass with Schottkey diodes...

It's bold to assume that passed any compliance tests at all.

[Wolfram]

That's quite a dramatic improvement. SiC schottky diodes are close to magic, but at a cost, which is likely why they weren't used originally. EMI is possibly lower as well, due to the lack of significant diode reverse recovery current spikes. These are commonly both a source of significant harmonic energy and switching device losses.

[PerpetualWalnut]

Now check EMC compliance... it might be that the design didn't pass with Schottkey diodes...

It's bold to assume that passed any compliance tests at all.

It passed QC didn't it? That's what the QC sticker is for right? RIGHT!? 

I didn't think about EMC with those diodes, I just figured that they saw that they could get away with cheap diodes in order to cut the cost by $50 - $80

[Circlotron]

Maybe the original sales sample that was sent to the importers had SiC schottkys and later they changed them and kept the cost difference. Probably wouldn't be the first time.