So I am working on making my own cc/cv 48V 40A lab bench power supply. I am designing it for a maximum of 40A at 45V, as there are some inevitable losses. There is already a thread for it. https://forum.allaboutcircuits.com/...-supply-from-atx-power-supplies.146806/page-4. I plan to use 4 atx power supplies in series with the - to earth connection removed (I know about the HV caps and know to take proper precautions). The microcontroller and switch parts will be controlled through the 5V rail of one of them. I am looking for some cheap atx power supplies that can be modified easily and can reliably supply 40A at 12V while delivering a few amps to the 5V rail. I found some $32 ones on amazon that are 550W and have a 12V 40A rail. I almost got them, but then I found a cheaper one on ebay that can supply more power.

https://www.ebay.com/itm/EVGA-80-PL...5?hash=item467ae781cb:g:drwAAOSwl8Ja2ecg#rwid

The problem: I cannot read how many amps the 12V rail can supply (look at the 6th image). In fact, I cannot read the current ratings for any of them. It is just too blurry. So is there somewhere else I could find it, or can anyone decipher it? Also, with cheap atx power supplies like these, will I encounter problems when some rails are unloaded but others have a heavy load? I have heard about many such annoyances. I am ok with having a dummy load on the used rail if the real load is low power, but I do not want to have to balance loads or any other annoying stuff. So thoughts on this cheap one here?

 

Or actually does anyone know a computer program that can read the current ratings accurately?

 

Looks like +12 @ 49 amps

 

Looks like +12 @ 49 amps

I would agree - it looks like 12V @ 49A

 

It's definately 12V@49A.

Are you after a variable power supply or fixed at 48V?

 

I contacted the seller. Hopefully they will get back to me. I think 49A is about right. That extra current will be nice. Looking at it it says "5 something something watts" and 588 would make sense. I am going for variable cc/cv. I am not sure what exactly the best approach is. You can see the thread I mentioned for more info. So does anyone think that I would need anything more than a dummy load (power resistor controlled by fet) for low actual loads? Would I have difficulty removing the gnd to - connection? Thanks.