Pegasus ATX400P

[goodpsusearch]

The psu is powering a Pentium 4 3.2GHZ Hyper Threading (Prescott) for 3 years.

I was impressed by the big heatsinks, the large ERL-35 transformer, the 2x680uF 200V input caps and the full transient filter (2 X caps, 2 Y caps, 2 coils, 2 MOV).

The controller is UC3843B and the power mosfet is STP13NK60ZFP, capable of 13A continuous, 52A pulsed.

The biggest surprise to me was the secondary though:
STPS4045CW for 3.3V
MBR6045WT for 5V
STPS30100 for 12V

5vsb caps:
2x1000uf Teapo bulging

3.3V caps:
1x1000uF Teapo
1x2200uF Ltec

5V caps:
1x1000uF Teapo
1x2200uF Ltec

12V caps:
1x2200uF Ltec


Mods:

1)Added an NTC Thermistor at the AC input
2)Replaced 4 3A diodes with a 4A bridge rectifier (assuming 230V AC, it is ok for 450W)
3)Replaced the bulging Teapos with 2x1000uF 16V Nichicon HD
4)Added a 3300uF 16V Panasonic FK at 12V (there was an empty spot on pcb for a second 12V cap)
5)Added a plastic insulation sheet between the case and the back of the pcb.

What do you think? Who made this power supply?

[370forlife]

Re: Pegasus ATX400P

HEC/Crapucase

[POM_MJ]

Re: Pegasus ATX400P

HEC/Crapucase

agreed, not too bad after recaps.

change C31 too.

[momaka]

Re: Pegasus ATX400P

Yup, its a HEC/Compucase. Looks like the Thermaltake TR2 430W (XP550-NP). Pretty decent overall.
5vsb circuit is 2-transistor, though - get C31 replaced as POM_MJ mentioned (this is likely the 5vsb critical cap).

Also, note that the value of R47 - the load resistor on the 5vsb rail - is 21 ohms (red, brown, black, if I'm reading it correctly). This value is too low, which is not good because it will generate a lot of heat - probably the main reason why the Teapo caps on the 5vsb were bulged in the first place. Definitely get it replaced! A 100 Ohm, 1-2W resistor should be perfect for this.

*Edit*
Looks like the -12v rail has 100 Ohm resistor - again, that's way too low of a value for that rail (at 12v it's dissipating 1.44W of heat!). Replace it with 470 Ohms or more.
See if the PSU has other low-ohm loading resistors on the output rails, and if it does, get them replaced. Generally, you don't want a loading resistor to dissipate more than 0.5W of power, otherwise you're cooking the caps.

[c_hegge]

Re: Pegasus ATX400P

2)Replaced 4 3A diodes with a 4A bridge rectifier (assuming 230V AC, it is ok for 450W)

Actually, on 230V it would be good for 920W power draw and 690W output (assuming 75% efficiency)

[goodpsusearch]

Re: Pegasus ATX400P

Actually, on 230V it would be good for 920W power draw and 690W output (assuming 75% efficiency)

Do you think it is that efficient? I assumed 60% just to be sure and 220V AC (it drops to that value often).

[goodpsusearch]

Re: Pegasus ATX400P

5vsb circuit is 2-transistor, though - get C31 replaced as POM_MJ mentioned (this is likely the 5vsb critical cap).

Also, note that the value of R47 - the load resistor on the 5vsb rail - is 21 ohms (red, brown, black, if I'm reading it correctly). This value is too low, which is not good because it will generate a lot of heat - probably the main reason why the Teapo caps on the 5vsb were bulged in the first place. Definitely get it replaced! A 100 Ohm, 1-2W resistor should be perfect for this.

*Edit*
Looks like the -12v rail has 100 Ohm resistor - again, that's way too low of a value for that rail (at 12v it's dissipating 1.44W of heat!). Replace it with 470 Ohms or more.
See if the PSU has other low-ohm loading resistors on the output rails, and if it does, get them replaced. Generally, you don't want a loading resistor to dissipate more than 0.5W of power, otherwise you're cooking the caps.

About the resistors and the 5vsb caps, I did a fan mod today so that the fan spins faster and the unit doesn't get very hot during operation.

I advised my friend to switch it off completely when he is not using the computer. That's what I do with my own PCs too.

I also try to bend the resistors so they are not near caps and the caps so they are not near heat sources in general.


But I agree that the best would be a full recap and some resistor replacements.

[c_hegge]

Re: Pegasus ATX400P

Do you think it is that efficient? I assumed 60% just to be sure and 220V AC (it drops to that value often).

If it really was only 60% efficient, then it would do 528 watts output, but that would be a record for the least efficient PSU in the world. Even this 12 USD piece of junk which failed all of the hardwaresecrets tests managed over 70%

[POM_MJ]

Re: Pegasus ATX400P

Yes, This HEC unit should better, it's used low RDS(on) (0.55ohm) in single-transistor forward topology, I guess its efficiency is around 75-80%.