is it possible to revive dead psu

[prongs12]

hello folks...

okay straight to the question... is possible to revive dead PSU and if possible increasing wattage..

because from my experience in my bussiness I had seen lot of PSU same product line but different wattage had same PCB design...

I had all components that I need. I even success in recapping some blowing caps and some of them works just fine...

thx

[klkmalik]

Re: is it possible to revive dead psu

Yes, u can increase little wattage. change from output diode, switching transistor and main caps. also most of psu not real wattage. some psu showing 450 watt but real watt 125,Brand psu only real wattage. so u can test through the amp meter.before u must connet 5v,12v bulb load in psu.

[eccerr0r]

Re: is it possible to revive dead psu

Sometimes different wattage versions use different transformers, and if this is the case you may not have an "upgrade" easily available...

[prongs12]

Re: is it possible to revive dead psu

Sometimes different wattage versions use different transformers, and if this is the case you may not have an "upgrade" easily available...

yes in some case that was the problem. and to add my headache, is design of second diode brigde.

i mean high speed diodes diode tha has three legde.. the code that show on their is too various..

[kaboom]

Re: is it possible to revive dead psu

yes in some case that was the problem. and to add my headache, is design of second diode brigde.

i mean high speed diodes diode tha has three legde.. the code that show on their is too various..

The center lead is the common cathode, each of the two outer leads is an anode.

[PeteS in CA]

Re: is it possible to revive dead psu

Several components in some way limit the amount of power you can get out of a particular P/S. The inrush limiting thermistor has a maximum current rating, and if you exceed that the part will overheat and de-solder itself, with results depending on its mounting angle (if upside down, it will fall out of the PCB - BTDTGTTS). PCB traces have a limit to their current carrying capacity - exceed that and the trace becomes a fuse. The input rectifier, switch device and output rectifiers all have maximum current ratings. The cooling for power devices - heatsinks and fan(s) - can affect maximum power. The main power transformer may be the most important factor - if the core saturates the switch MOSFET quickly becomes a ZorchFET. The core and wire gauge of the output inductor impose limits, though the problem may be more subtle and not immediate (if the core is too it deteriorates quickly, but may take several months for output ripple to cause performance problems). If it gets excessively hot inside the P/S - due to any combination of the factors mentioned - some of the electrolytic capacitors may swell and vent. If the capacity of the input electrolytics is too low, ride-through (the ability to stay in regulation through a power drop-out) will be affected, and the input ripple may be more than the regulators can compensate for, resulting in excessive ripple on the output voltages.

Most of those things allow room for improvement, but the hardest limits to improve will be the main transformer and the cooling (heatsinks plus fans). The chances of being able to find and have room for a larger transformer and more substantial heatsinks are not good. And while fans with better airflow may be available in the same package, they aren't inexpensive.

All that said, a 10%-20% boost in output power may be feasible, but beyond that it would be more of a learning experience and somewhat expensive.

[tmiha71]

Re: is it possible to revive dead psu

All that said, a 10%-20% boost in output power may be feasible, but beyond that it would be more of a learning experience and somewhat expensive.

You forgot to add explosive too.

Probably language barrier, I got impression, that it's about increasing efficiency, lowering wasted power, and expecting some percentage (probably in range of one digit) of output power...

I always wondered if they (manufacturers), did measured capacitors for input stage (serial connection) for capacitance and ESR ???

[Wester547]

Re: is it possible to revive dead psu

Several components in some way limit the amount of power you can get out of a particular P/S. The inrush limiting thermistor has a maximum current rating, and if you exceed that the part will overheat and de-solder itself, with results depending on its mounting angle (if upside down, it will fall out of the PCB - BTDTGTTS). PCB traces have a limit to their current carrying capacity - exceed that and the trace becomes a fuse. The input rectifier, switch device and output rectifiers all have maximum current ratings. The cooling for power devices - heatsinks and fan(s) - can affect maximum power. The main power transformer may be the most important factor - if the core saturates the switch MOSFET quickly becomes a ZorchFET.

Well, I have seen Delta manage to achieve an 80%+ efficient 500W PSU with what looks to be a size 35 transformer and an 80mm fan (not that an 80mm fan is bad if used properly). In saying that, though, the common OEMs all have their own custom transformers for a very large part so their actual transformer ability may not be indicated by the "size" level most PSU transformers are correlated with in terms of wattage ability. The only real way to find out is to test them.

But in my own observation, I have noticed that the biggest difference is made with how efficient the main switchers are and the overall efficiency of the PSU. Usually, I see 80%+ efficient PSUs achieve that stature by having main switchers and APFC transistors with an extremely low On resistance (and sometimes RG, RL, and RD values, too, along with low power dissipation), secondary rectifiers with very low voltage drops (or just synchronous rectification and if not that a buck DC-DC converter from +12V - such methods allow for less cooling and less filtering with cleaner outputs), a very powerful rectifying bridge (or two) mounted to heatsink(s), and maybe even 16 gauge wires or thicker. This does work (it means less wasted heat and a lower heat output) and also allows for the use of TO-220 parts instead of TO-3P or TO-247 parts (since the voltage drop is so low and the efficiency so high), allows internal components to run significantly cooler, and allows for thinner heatsinks and smaller fans, but I think a better way to achieve efficiency is to hugely overbuild the secondary side since I think that translates into an ultimately better unit.

Not to say the transformer size doesn't matter, though - as you said, if the core is saturated it can damage the main switchers which can in turn damage the input capacitors and/or voltage doubler and they also matter in determining wattage along with the secondary side (more current = more capacity all around), and the output filtering is indeed important in order to make sure everything stays in spec with the ripple and noise low and the voltage regulation good. Good cooling has lots to do with it as well, especially with regard to how fast the fan actually revs up in order to accommodate for the ever increasing heat output. Technically speaking, though, you could compensate for small input electrolytics with very powerful primary FETs (or a very overspec'd secondary, as mentioned before) but this still results in a power supply that has to work harder. It is better not to skimp on components.

[prongs12]

Re: is it possible to revive dead psu

The center lead is the common cathode, each of the two outer leads is an anode.

but how did I read current ratings? I mean is there any meaning on sb**** or mbr******, is there any standard for their code example sbr2100
(just my guess)
sbr = shotkky barrier rectifier
2100 is 100 amps?? or maybe 10 amps??
which one says diode current ratings.

as for the transformer is there any standard for their code too
I mean some trafo code is ERL 33 PSG blah blah blah
(my guess again)
ERL is manufacturer code
33 is current ratings or maybe wattage 330 watt maybe?? or wattage limit?

as far as I know if I can access those datasheet.. maybe is possible to increase PSU wattage by huge margin... or make the PSU become pure.... I have lot of component to be wasted in my search for knowlegde...

You forgot to add explosive too.

Probably language barrier, I got impression, that it's about increasing efficiency, lowering wasted power, and expecting some percentage (probably in range of one digit) of output power...

I always wondered if they (manufacturers), did measured capacitors for input stage (serial connection) for capacitance and ESR ???

lol blowing the object is everday breakfast for me. following failure after failure that I've made... but you had point.. that was exactly I try to do (for now) after I success maybe I'll step to increasing wattage

Several components in some way limit the amount of power you can get out of a particular P/S. The inrush limiting thermistor has a maximum current rating, and if you exceed that the part will overheat and de-solder itself, with results depending on its mounting angle (if upside down, it will fall out of the PCB - BTDTGTTS). PCB traces have a limit to their current carrying capacity - exceed that and the trace becomes a fuse. The input rectifier, switch device and output rectifiers all have maximum current ratings. The cooling for power devices - heatsinks and fan(s) - can affect maximum power. The main power transformer may be the most important factor - if the core saturates the switch MOSFET quickly becomes a ZorchFET. The core and wire gauge of the output inductor impose limits, though the problem may be more subtle and not immediate (if the core is too it deteriorates quickly, but may take several months for output ripple to cause performance problems). If it gets excessively hot inside the P/S - due to any combination of the factors mentioned - some of the electrolytic capacitors may swell and vent. If the capacity of the input electrolytics is too low, ride-through (the ability to stay in regulation through a power drop-out) will be affected, and the input ripple may be more than the regulators can compensate for, resulting in excessive ripple on the output voltages.

Most of those things allow room for improvement, but the hardest limits to improve will be the main transformer and the cooling (heatsinks plus fans). The chances of being able to find and have room for a larger transformer and more substantial heatsinks are not good. And while fans with better airflow may be available in the same package, they aren't inexpensive.

All that said, a 10%-20% boost in output power may be feasible, but beyond that it would be more of a learning experience and somewhat expensive.

heatsink and fan could easily modified... but for main transformer.. I didn't have anyway around this problem unless I replace it too.

Well, I have seen Delta manage to achieve an 80%+ efficient 500W PSU with what looks to be a size 35 transformer and an 80mm fan (not that an 80mm fan is bad if used properly). In saying that, though, the common OEMs all have their own custom transformers for a very large part so their actual transformer ability may not be indicated by the "size" level most PSU transformers are correlated with in terms of wattage ability. The only real way to find out is to test them.

But in my own observation, I have noticed that the biggest difference is made with how efficient the main switchers are and the overall efficiency of the PSU. Usually, I see 80%+ efficient PSUs achieve that stature by having main switchers and APFC transistors with an extremely low On resistance (and sometimes RG, RL, and RD values, too, along with low power dissipation), secondary rectifiers with very low voltage drops (or just synchronous rectification and if not that a buck DC-DC converter from +12V - such methods allow for less cooling and less filtering with cleaner outputs), a very powerful rectifying bridge (or two) mounted to heatsink(s), and maybe even 16 gauge wires or thicker. This does work (it means less wasted heat and a lower heat output) and also allows for the use of TO-220 parts instead of TO-3P or TO-247 parts (since the voltage drop is so low and the efficiency so high), allows internal components to run significantly cooler, and allows for thinner heatsinks and smaller fans, but I think a better way to achieve efficiency is to hugely overbuild the secondary side since I think that translates into an ultimately better unit.

Not to say the transformer size doesn't matter, though - as you said, if the core is saturated it can damage the main switchers which can in turn damage the input capacitors and/or voltage doubler and they also matter in determining wattage along with the secondary side (more current = more capacity all around), and the output filtering is indeed important in order to make sure everything stays in spec with the ripple and noise low and the voltage regulation good. Good cooling has lots to do with it as well, especially with regard to how fast the fan actually revs up in order to accommodate for the ever increasing heat output. Technically speaking, though, you could compensate for small input electrolytics with very powerful primary FETs (or a very overspec'd secondary, as mentioned before) but this still results in a power supply that has to work harder. It is better not to skimp on components.

so if we upgrade the transistor and the second side we got some significant power ups? I have done that and didn't notice any difference.. maybe because my lack understanding.

what i had doing for now is.. recapping the secondary side with higher voltage and capacitance.. recapping main caps. with higher capacitance. replacing little three legged diode usually that has mospec sign on their body with bigger ones. and if they had any room on other side of bridge, i put another one with same number like other side i.e double sbr****, or mbr**** . is that right thing to do?


Thanks for your advice folks

[mariushm]

Re: is it possible to revive dead psu

but how did I read current ratings? I mean is there any meaning on sb**** or mbr******, is there any standard for their code example sbr2100
(just my guess)
sbr = shotkky barrier rectifier
2100 is 100 amps?? or maybe 10 amps??
which one says diode current ratings.

as for the transformer is there any standard for their code too
I mean some trafo code is ERL 33 PSG blah blah blah
(my guess again)
ERL is manufacturer code
33 is current ratings or maybe wattage 330 watt maybe?? or wattage limit?

as far as I know if I can access those datasheet.. maybe is possible to increase PSU wattage by huge margin... or make the PSU become pure.... I have lot of component to be wasted in my search for knowlegde...

There's no standard, but usually the name contains the maximum voltage and the current.
There's a picture a few posts above that says SBL4040 ... that's probably maximum 40v, 40 amps.

ERL is a type of transformer core type and a soft of "standard" of transformers suitable for switching power supplies.

Depending on the number after the ERL, the coils have various inductances and the higher the number the more wattage the transformer can usually handle (but it also depends on how the circuits around work)

A rough approximation, a real ERL-33 can do up to about 200-220 watts, a ERL-35 can do about 300-350w, ERL39 may do about 450w ...

But like you've already read, it really depends on the switchers and other things on the PCB - if one uses more high end ICs and transformer is a quality ERL-35 you might even do 500w with it.

Keep in mind a no-name Chinese manufacturer can lie on the label, i've had a 450w power supply that had a sticker saying ERL-35 on top but I desoldered the transformer and on the bottom the real sticker said ERL-33.

I kinda knew it wasn't ERL-35 because it was smaller in volume, when you see a lot of them you can usually tell from the volume how good it is.

[prongs12]

Re: is it possible to revive dead psu

There's no standard, but usually the name contains the maximum voltage and the current.
There's a picture a few posts above that says SBL4040 ... that's probably maximum 40v, 40 amps.

ERL is a type of transformer core type and a soft of "standard" of transformers suitable for switching power supplies.

Depending on the number after the ERL, the coils have various inductances and the higher the number the more wattage the transformer can usually handle (but it also depends on how the circuits around work)

A rough approximation, a real ERL-33 can do up to about 200-220 watts, a ERL-35 can do about 300-350w, ERL39 may do about 450w ...

But like you've already read, it really depends on the switchers and other things on the PCB - if one uses more high end ICs and transformer is a quality ERL-35 you might even do 500w with it.

Keep in mind a no-name Chinese manufacturer can lie on the label, i've had a 450w power supply that had a sticker saying ERL-35 on top but I desoldered the transformer and on the bottom the real sticker said ERL-33.

I kinda knew it wasn't ERL-35 because it was smaller in volume, when you see a lot of them you can usually tell from the volume how good it is.

that was very handful... thanks for your repply... must search more info about switching transistor...