EVGA W1 500W... Oops!

[Jon Gerow]

https://www.youtube.com/watch?v=60jV...ature=youtu.be

20g wire on the PCIe!!!! WTF?!?!!?

[dc-dc_conversion_police]

Wow. I’m just... Stunned.

Also this from comments: “Corsair PSUs are way better since there is a whole team supporting them, and Jon Gerow is their R&D manager, and this guy KNOWs about PSUs!”
Lol

[Jon Gerow]

Wow. I’m just... Stunned.

Also this from comments: “Corsair PSUs are way better since there is a whole team supporting them, and Jon Gerow is their R&D manager, and this guy KNOWs about PSUs!”
Lol

I'm flattered. ;-)

[Juular]

But it's EVGA !

[ashiekh]

OK, what am I missing

EVGA 500W Review - Another One (PSU) Bites the Dust - YouTube
"According to the specs, this bridge rectifier cannot exceed 3A at 100 degrees Celsius without a heatsink."

GBU1- Series.qxd (farnell.com)
GBU 1006
"Maximum Average Forward Rectified Currentat TC = 100°C" 10A


Would have been nice if Aris had actually tested to see if it really was the bridge rectifier that had failed.

[Jon Gerow]

OK, what am I missing

EVGA 500W Review - Another One (PSU) Bites the Dust - YouTube
"According to the specs, this bridge rectifier cannot exceed 3A at 100 degrees Celsius without a heatsink."

GBU1- Series.qxd (farnell.com)
GBU 1006
"Maximum Average Forward Rectified Currentat TC = 100°C" 10A

When mounted to a 100mm x 100mm x 1.6mm copper plate.

[ashiekh]

I was looking at

• Typical Thermal Resistance Per Leg (Note 1)(Note 2) 21 C/W

which makes the circuit board itself a plausible heat-sink; 4 legs yields 5.25 C/W
and so 53C for 10W and so 9A (voltage drop is about 1.1V).

Now the supply is 500W so only 6A, so the temperature rise might be a lot
less than 50C if the board is itself a good heatsink and that would keep the
junction temperature below 100C even with a high ambient.

So I am not at all convinced that Aris is right.

[Orion]

GBU1- Series.qxd (farnell.com)
GBU 1006
"Maximum Average Forward Rectified Currentat TC = 100°C" 10A


Would have been nice if Aris had actually tested to see if it really was the bridge rectifier that had failed.

Wrong datasheet. You're looking at a GBU1006 made by Multicomp.

The bridge in the EVGA is a GBU1006 made by HY. http://www.hygroup.com.tw/en-ww/prod...D=10251,11640-
http://www.hygroup.com.tw/upfiles/AD...U1010(GBU).pdf

"Maximum Average Forward Rectified Current @ TC=100°C (without heatsink) = 3.0A"

[ashiekh]

From the HY specs
54-GBU10005-GBU1010(GBU).pdf

• Typical Thermal Resistance to case (Note2) 2 C/W
• Typical Thermal Resistance to lead (Note2) 1.5 C/W

which still suggests that dumping heat down the leads is a good way to go, potentially better even than a case mounted heatsink.

[Jon Gerow]

From the HY specs
54-GBU10005-GBU1010(GBU).pdf

• Typical Thermal Resistance to case (Note2) 2 C/W
• Typical Thermal Resistance to lead (Note2) 1.5 C/W

which still suggests that dumping heat down the leads is a good way to go, potentially better even than a case mounted heatsink.

I can tell you, from experience, that it's not. I mean, it's not like the diode is more efficient the hotter it gets.