Gigabyte GP-PGM shorting on main FETs - what now?

[LukeSavenije]

notice: this does NOT affect GP-APGM/Aorus PGM units. This is the same OEM, but different components and main design has been used here.

 

GP-PGM

 

GP-APGM

 

Summary

Gigabyte's recently announced unit has an issue, where the OPP is set too high, causing the main FETs to explode because of it being unable to handle the load. There is no public statement from Gigabyte and I didn't reach out to them about this, but have reached out to Aristeidis Mpitziopoulos, better known as Aris about what exactly happened. All info provided is from his video and review below.

 

Quotes

Quote

Jilin Sino-Microelectronics provides the main FETs, and as you can see, they didn't survive my tests. They died in a glorified way, making one of the loudest bangs I have ever heard. The main FETs are arranged in a half-bridge topology, and an LLC resonant converter is also used to boost efficiency. The resonant controller is a Champion CM6901X.

My thoughts

I think Gigabyte's response to this is unacceptable and very worrying. They've been offered the sample back, they've had a request to send a second unit to let the reviewer test it out and declined both of these. This is a serious issue where there isn't even a statement given. All that's been done is a couple unit put into testing by R&D where they didn't find anything. I honestly want Gigabyte to call back all the units, change the FETs to a higher quality level and supply the new designs to the affected consumers. This is a serious issue that has very likely been caused by cheaping out on main components, as GP-APGM didn't show this.

 

Further information

The OCP is set at around 130%

The main fets are 2x Jilin Sino-Microelectronics JCS18N50FH (500 V, 11 A @ 100 °C, Rds (on): 0.27 ohm)

The failure happened after the unit succesfully shut down on OPP/OCP, where it was turned on again after 5 minutes.

The issue is likely caused by cheap main FETs, a too high set OPP/OCP, a design error or a combination of these.

It does not affect the similarly named GP-APGM or any other Gigabyte units from what is currently known

 

Sources

https://www.techpowerup.com/review/gigabyte-gp-p750gm-750-w/single-page.html#introduction

[williamcll]

Well, I choose my PSUs from the PSU list on linustechtips.com anyway.

[LukeSavenije]

1 minute ago, williamcll said:

Well, I choose my PSUs from the PSU list on linustechtips.com anyway.

And as you can guess I'm taking this thing straight down

[Sunscrape]

I'm really worried about this PSU, I also commented on his vid, I have a 3080 + 10700k and it's been working fine for nearly 2 months. I didn't buy it cause I was looking to cheap out, I just have good experience with Gigabyte and this was a new unit. However, now that there's this mess Idk what to do. Also, my GPU is a Gigabyte with the connectors issue that I personally didn't experience (yeah that's a thing, the pins can become bent). I fear that if I disconnect the cables from the GPU the pins will become bent... Big paradox and I honestly don't know what to do.

 

Sorry for the long post just had to get it off my chest.

[Trik'Stari]

On 11/29/2020 at 8:33 AM, williamcll said:

Well, I choose my PSUs from the PSU list on linustechtips.com anyway.

Stuff like this is why I look not just at the product but also at the company and their customer service record.

 

EVGA hasn't done me wrong yet. And they have a reputation for good customer service as far as I know.

[StDragon]

The SOP of any company is to sandbag the issue first. Either way, the truth shakes out in the end.