PSU Rectifier diodes

[Super Nade]

Guys, Why do PSU makers still use 4x1 rectifier diodes instead of using a bridge rectifier chip? Isn't it more convenient and cheaper?

[kc8adu]

Re: PSU Rectifier diodes

cheaper to use 4 than a bridge.
they can further cheap out by using 2 1a diodes and 2 3a
in doubler mode the 3a are used.in full bridge mode for220 the current is half so the 1a ones are adequate.
saves a few cents per unit.

[gonzo0815]

Re: PSU Rectifier diodes

saves a few cents per unit.

That`s a fortune in this business only the better brands like FSP and Seasonic usually are using hte bridge rectifier.

[yanz]

Re: PSU Rectifier diodes

That`s a fortune in this business

Well...premature optimization is the root of all evil...

only the better brands like FSP and Seasonic usually are using hte bridge rectifier.

Almost all respected brand use a chip bridge rectifier. Even the old ones like my Newton, Acbel, Astec, Hipro, and Delta (all under 200W) PSU have it. But i never see one with heatsink on it. According to the datasheet, a heatsink is needed if it will be put on heavy duty.

I have a LC from 2002 that indeed using a chip bridge rectifier. (Simbadda ATK 300W, recapped )

But the more recent L&C are cheaping out all it can:

AIBO ATK 380W




For those damn cheap deer/crap psu i use this:

RS405 -> $0.5
GSIB660 -> $0.25
D3SBA60 -> $0.75

They are all rated at 4A or 6A 600V. I think it is overkill to use 4A rated fast/smooth bridge rectifier diodes for this ~400W PSU. 4A x 220v = 880W. But for 110-120v like in the USA, it will have to be 4A at minimal.

I can use the cheaper 400V rated ones, but is it safe?

What about using them in dual/paralel? I know it is probably not necessary at all, but i see the enermax 1KW PSU use it. I want to know if it makes any good effect, perhaps like doubling the current ability?

[gonzo0815]

Re: PSU Rectifier diodes

It spreats the heat may be, but i think it is better to use one wich can handle all the load and heat alone. I would not use the 400v versions, as there are spikes on the line, e.g. if some inductive things like transfomer, fluorescent tube balast etc. are switched on or a severe short occures this can kill even the 600v versions easily. Thus i personally would not go under 600v, i would prefer 800v. The difference here are only a few cent`s not worth the headaches.
For the A rating you have to consider the inrushcurrent for the totaly discharged caps. And don`t forget, that normal non PFC PSU`s are do not have linear current draws. Thus the diode should be choosen with significant headroom.

[yanz]

Re: PSU Rectifier diodes

hmm, that is interesting. perhaps if i upgrade those cheap 330uf/470uf primary caps to 680-820uf or even 1000uf, upgrading the bridge rect is a must. my comp is sitting near the big laser jet printer, probably there are some effect to the line when that monster startup.

i was seeing some decent psu of 200-300w with rs405 so i assume that 600V is ok, at least i don't have any clue what is the safest/minimal limit on the voltage rate.

what do you think of deer 400-500W PSU still using those four single diodes, i thought the more wattage it is the cleaner the input line needed for the psu? or is it not significant wether to use the chip form and the single(s)?

[Per Hansson]

Re: PSU Rectifier diodes

i was seeing some decent psu of 200-300w with rs405 so i assume that 600V is ok, at least i don't have any clue what is the safest/minimal limit on the voltage rate.

Normal power draw for a laser printer heating up is 20aH for 20msec, that is basically just a short... Once I was lucky enough to catch that on my UPS (how likley is that btw :P) Anyway AC voltage (input voltage) went from 230VAC to 170VAC during that second... And mind you this is a 6kW UPS that is only loaded ca 10-20% and has a really really really big input bulk cap (a half arsed guess would be diameter 10cm and height maybe 15-20cm...)

That is why you never connect a laser printer to a UPS

[arneson]

Re: PSU Rectifier diodes

Ahh Haa.
Thanx Per.
Now I know how to shut up the UPS's.
Unplug the printers, Brilliant!
Under one desk is the home of five UPS's and one was always complaining it coulden't recover. That's the one, had three printers in the backed up sockets.
The other complaining UPS has two large LCD's on it, big power bricks, they are not wanted also.

[999999999]

Re: PSU Rectifier diodes

Although these bridges being mentioned typically have 4-6A or so ratings, their peak ratings are much much higher, often over 100A. That is usually going to give them enough margin on an ATX PSU, because it keeps the caps always charged, that surge isn't a continual event.

The voltage on the other hand, could depend on your AC line quality. Power surges could cook them or the chopper transistors, so either should have the highest voltage rating you find reasonable. Most often 600V is reasonable, 400V I don't see why you'd want to do it except if it's a 110V line. No point in trying to save a few cents when you're only doing a handful of PSU instead of thousands.

[yanz]

Re: PSU Rectifier diodes

Ahh Haa.
Thanx Per.
Now I know how to shut up the UPS's.
Unplug the printers, Brilliant!
Under one desk is the home of five UPS's and one was always complaining it coulden't recover. That's the one, had three printers in the backed up sockets.
The other complaining UPS has two large LCD's on it, big power bricks, they are not wanted also.

i never connected a printer to ups. Per Hansson is right, i just read the UPS FAQ/guide a minutes ago, it mentions about a warning not to connect a printer to ups.

[yanz]

Re: PSU Rectifier diodes

Although these bridges being mentioned typically have 4-6A or so ratings, their peak ratings are much much higher, often over 100A. That is usually going to give them enough margin on an ATX PSU, because it keeps the caps always charged, that surge isn't a continual event.

The voltage on the other hand, could depend on your AC line quality. Power surges could cook them or the chopper transistors, so either should have the highest voltage rating you find reasonable. Most often 600V is reasonable, 400V I don't see why you'd want to do it except if it's a 110V line. No point in trying to save a few cents when you're only doing a handful of PSU instead of thousands.

i take your words. so 4-6A is okay but 600V is the minimal (for 220V lines).

[999999999]

Re: PSU Rectifier diodes

Minimum is relative to the surges present on the line. If you had a hypothetically perfect line, 400V would be enough. Seldom does this perfect, surge-free line really exist (even surge protector has "some" higher threshold than 220V). The higher the spec'd part overshoots the line voltage, the fewer the surges you would expect that exceed the part spec. Lower voltage surges are far more common than higher voltage.

However, having your bridge able to withstand 1000V, isn't going to save your transistors in the next stage, and going very far above common parts to try to provide especially good protection will get expensive without diminishing expectation that it would matter. You as the (re)engineer have to decide what, if any, additional risks exist and how much money it's worth to guard against any possible failure mode. Overengineering is a black art, and done less and less often in modern competitive markets for commodity goods like switching PSU.

In the end if surges are that prominent a whole-site surge protector may be the best answer.

[yanz]

Re: PSU Rectifier diodes

Overengineering is a black art, and done less and less often in modern competitive markets for commodity goods like switching PSU.

i like that sentence