The Last Power Supply (Charger) You'll Ever Need

This box also fits a 12S 6A charger, so when it is charging you remove the series connector(short!) and you can charge the two 12S packs in parallel. in order to not disconnect them in series you would need a 24S charger like for an EUC.

Ideally it would have two+ different cccv outputs but with just one it is still possible to charge more than one board as once as long as they are all close enough in charge. There is some math so that could be determined.

mine just stopped working. not sure if leaving it on for a really long time lead to it or what.

1 Like

this

1 Like

Heat death maybe?

1 Like

I suppose. I would have hoped that it would not just fry itself and instead limit but that is a lot to ask. Should have been monitoring it since it was the first time it was in the case running. Welp lesson learned for you all.

2 Likes

The converter died?

1 Like

yes. I think somehow converting 90V down to 25V at 10A with low airflow was just a bad idea so bewarned.

Oh I finally looked close enough to actually spot the burned trace here


Looks fine topside.

Gate trace is blown so the switching FET died. Could replace the fet and it should love again. Likely does from thermal load or just a bad FET

1 Like

http://www.smc-diodes.com/propdf/MBRF20150CT%20N0146%20REV.D.pdf

https://www.arrow.com/en/products/mbr20150ct/littelfuse

Diode… Hmmm not totally sure at that point.

High current means it’s probably still part of the buck circuit. Makes sense. Did you have a large capacitance on the load by chance?

1 Like

Looks like it could be just simple bad design and overheating of the trace. It occurred just where the trace necks down, where the big copper plane can’t pull away heat anymore and before the heat sinking effect of the diode’s beefy leg.

If replacing the diode fixes the problem you can use a beefier wire than what that trace provided so there isn’t a heating choke point at high current levels anymore.

Check around for other signs of heating and beef up whatever you can there too. More solder can help a little but it’s 10x the resistance of copper so a lot of solder only helps a bit.

2 Likes

I like this explanation. It’s also only on the narrow part of the trace

1 Like

looks… higher quality build than the Juntek

2 Likes

Somehow when I was shopping for Server PSU I missed this site.
https://www.servermonkey.com/1-dell-12g-1100-watt-psu.html
https://www.servermonkey.com/1-dell-12g-750-watt-psu.html
https://www.servermonkey.com/dell-750-watt-psu.html
https://www.servermonkey.com/hp-750-watt-psu.html
https://www.servermonkey.com/hp-499250-201-460-watt-power-supply.html
https://www.servermonkey.com/hp-500-watt-psu.html

Anyone want to weigh in on which one to pick?

If I was going to go for 8 in series 96V total that leaves only 187W for each PSU to supply with the 110v 1500W circuit limit, so the 400W is more than enough for that. Although for near the same price I could get higher power ones which is good if I reused them in smaller units.

3 Likes

Keep in 400w psu will only do 33A even at lower pack voltage. So like if you wanted to charge a 10s , you’d be limited to 400w/12v = 33 A instead of 1500/36v which is 41A.

Might matter, might not. But if you’re charging under 30 Amps, are you even charging?

2 Likes
1 Like
1 Like
1 Like