Thermal pads - "clean" cautionary tale - UK Vintage Radio Repair and Restoration Discussion Forum

Cre8anet · 1st Jun 2022, 8:01 pm

Just narrowly averted MOSFET heatsink failure by chance spotting some poor m/f... Maybe due to a dirty assembly line? or lack of inspection/QA??
When at Retrotech I picked up two 300W 24V battery-mains inverters from a jolly friendly table-seller who helped me out and I've finished adjusting design of my lowering the cut-out, so that they still work nicely down to safe-discharge of 19V (from my reclaimed EV battery)...
The low-side quad of IRF540s are heatsinked-to-case by isolating thermal pad. I happened to notice two tiny dinks in the pad (pic) but only on one side, though and was just about to reassemble.. but my bike-tube puncture experiences reminded to "check the tyre", so I looked at the IRF540s as well...

See pics! Turned out to be 2 tiny solder spray balls still on one of the IRFs, having been squashed between device and the pad on factory assembly in China, just like on a BGA package...
They pinged-off the IRF540 again nice'n'clean with a knife... just in time to avoid a potentially fatal flaw (and most likely some big bangs!).

joebog1 · 1st Jun 2022, 10:29 pm

Taking a battery down to 19 volts will kill it pretty quickly. I always set cutoff @ 22volts.
Just my experience with solar.

JOe

cmjones01 · 2nd Jun 2022, 9:14 am

Quote:

Originally Posted by joebog1

Taking a battery down to 19 volts will kill it pretty quickly. I always set cutoff @ 22volts.
Just my experience with solar.

That depends on the battery chemistry, surely. A battery of 6 Lithium-ion cells would be at about 3.2v per cell, which is acceptable. With 8 LiFePO4 cells each one would be at about 2.4V which is too low for comfort.

Returning to the thermal pad/solder ball issue, well spotted! I've experienced quite a lot of these inverters with the rows of MOSFETs screwed to an extruded aluminium case with thermal pads self-destructing even at low loads. I wonder now whether this defect could have something to do with the failures.

Chris

McMurdo · 2nd Jun 2022, 12:34 pm

A service bulletin was issued a few years ago from a very well known US manufacturer of active loudspeakers after a number of warranty repair reports showed that aluminium swarf had been left on the diecast heatsinks after the milling process and had been caught under the sil pads, shorting the IGBTs to earth after a time. They had been assembled by a third party contractor in China.

1st Jun 2022, 8:01 pm	#1
Cre8anet Pentode Join Date: Nov 2005 Location: Banbury, Oxfordshire, UK. Posts: 204	Thermal pads - "clean" cautionary tale Just narrowly averted MOSFET heatsink failure by chance spotting some poor m/f... Maybe due to a dirty assembly line? or lack of inspection/QA?? When at Retrotech I picked up two 300W 24V battery-mains inverters from a jolly friendly table-seller who helped me out and I've finished adjusting design of my lowering the cut-out, so that they still work nicely down to safe-discharge of 19V (from my reclaimed EV battery)... The low-side quad of IRF540s are heatsinked-to-case by isolating thermal pad. I happened to notice two tiny dinks in the pad (pic) but only on one side, though and was just about to reassemble.. but my bike-tube puncture experiences reminded to "check the tyre", so I looked at the IRF540s as well... See pics! Turned out to be 2 tiny solder spray balls still on one of the IRFs, having been squashed between device and the pad on factory assembly in China, just like on a BGA package... They pinged-off the IRF540 again nice'n'clean with a knife... just in time to avoid a potentially fatal flaw (and most likely some big bangs!). Attached Thumbnails __________________ ~~~~~~May your Waves be Long~~~~~~

1st Jun 2022, 10:29 pm	#2
joebog1 Banned Join Date: Jun 2015 Location: Mareeba, North Queensland, Australia Posts: 2,704	Re: Thermal pads - "clean" cautionary tale Taking a battery down to 19 volts will kill it pretty quickly. I always set cutoff @ 22volts. Just my experience with solar. JOe

2nd Jun 2022, 12:34 pm	#4
McMurdo Dekatron Join Date: Jun 2009 Location: Staffordshire Moorlands, UK. Posts: 5,270	Re: Thermal pads - "clean" cautionary tale A service bulletin was issued a few years ago from a very well known US manufacturer of active loudspeakers after a number of warranty repair reports showed that aluminium swarf had been left on the diecast heatsinks after the milling process and had been caught under the sil pads, shorting the IGBTs to earth after a time. They had been assembled by a third party contractor in China. __________________ Kevin