Prudent Way PWI-PR350 Rebuild

[Pentium4]

I don't think I ever posted this thing. I got it a long time ago. I think May, 2014. It looks like a decent 350W, even before I modded it. And I did, but those pictures will come later. Here's how it looked stock

6A bridge
680uF input caps
TO-247 13009 switchers
IC for 5VSB
40A on 5V
30A on 3.3V
two 20A ultra fast on 12V
nice 12.5mm cap slot for 12V
Young Lin DFS122512L fan

[c_hegge]

Re: Prudent Way PWI-PR350 Rebuild

Not too bad at all. It should be capable of 350W in spec.

[momaka]

Re: Prudent Way PWI-PR350 Rebuild

Looks pretty good indeed. The heatsinks in combination with that 12 cm fan should definitely be able to hold up to at least 300W. I think it will make it to 350W no problem.

What's with the darkened area around the resistors on the secondary side near the 5VSB transformer and PWM controller, though?

[Pentium4]

Re: Prudent Way PWI-PR350 Rebuild

Not too bad at all. It should be capable of 350W in spec.

Yeah doesn't look bad for a budget 6 year old unit. It was made in March, 2009. I think so too. It looks similar to some of the units that you've reviewed.

Looks pretty good indeed. The heatsinks in combination with that 12 cm fan should definitely be able to hold up to at least 300W. I think it will make it to 350W no problem.

What's with the darkened area around the resistors on the secondary side near the 5VSB transformer and PWM controller, though?

Yeah it's a nice fan too! It was getting a little dry, so I oiled it. I'm not sure what's up with those resistors. When I recapped it, amazingly all of the caps tested good! Even the BH caps on the 5VSB filtering. Both those were 1000uF 16V CD288 series. They were failing though, one of them tested 1,168uF. Here's the crappy thing though. I originally started working on this unit when I first got it, so I started pulling caps to test them and started removing glue. Because of my short attention span, I found a "more interesting" project and started working on that one. This one sat to the side for a long time. I remembered all of the values except the ones for the base drive circuit for the switching transistors. I know they were those red nicon caps, but I can't for the life of me remember their values, or find them in all of the caps that I've pulled and kept. I guessed 4.7uF 50V and installed those. The PSU runs but it creates a quiet high pitched whine at any load, either on or just 5VSB. I'm thinking I might have used the improper value, and in turn changed the frequency at which the circuit operates. If I did use the incorrect value, is this the likely result? I knew it was either going to be 2.2uF, 4.7uF, or 10uF. The pitch doesn't seem to change as the load changes. I used one of the red nicon caps in a silly experiment project. That would have the correct value but I'd have to find the darn thing

Besides that, I replaced all of the other caps, added MOV's, upgraded the input caps from 680uF to 820uF, removed all the glue the best I could, wired the primary permanently to 120V, added a 6 pin PCIe connector, and changed out the minimum load resistors on the output rails. None of the new ones were lower than 100Ω. I'm still on the fence on whether or not I should change the 12V rectifiers. Two 20A ultra fasts isn't terrible, but I was thinking that if I changed them to two 30A schottky rectifiers, the 12V would run too high unless there was a heavy load on it.
Edit: I also wonder what those jumpered resistors are for on the secondary next to the 12V wires. I wonder if that's their take on OCP?

[goodpsusearch]

Re: Prudent Way PWI-PR350 Rebuild

From my experience those capacitors usually are 4.7uF, 2.2uF or 1uF and rarely 10uF.

Don't hesitate to replace them with 2.2uF and test it. I have seen the same psu (Jou Jye 400W) using 1uF caps in 2004 revision and changing them with 4.7uf in 2005 revision and newer.


edit: c_hegge is right, didn't notice that.

[c_hegge]

Re: Prudent Way PWI-PR350 Rebuild

If it squeals with only the 5vsb active, then it's got nothing to do with those caps. Those are used on the main switching transistors and are only used when the PSU is actually running.

[ben7]

Re: Prudent Way PWI-PR350 Rebuild

No UL number! Can't tell what OEM it might be. (if it isn't designed by "Prudent Way") The layout looks familiar. (Especially with the 4 caps in a square shape on the secondary, instead of being staggered to fit them in a smaller area)

Does it feed the fan 12V, or is it speed/temp controlled?

Because of my short attention span, I found a "more interesting" project and started working on that one.

Happens to me too

The PSU runs but it creates a quiet high pitched whine at any load, either on or just 5VSB. I'm thinking I might have used the improper value, and in turn changed the frequency at which the circuit operates. If I did use the incorrect value, is this the likely result? I knew it was either going to be 2.2uF, 4.7uF, or 10uF. The pitch doesn't seem to change as the load changes. I used one of the red nicon caps in a silly experiment project. That would have the correct value but I'd have to find the darn thing

Hm, it might be the 5VSB that is doing that. Perhaps you used a cap with too low of an ESR? It might have done that since it was made. Maybe they didn't get the compensation network(s) correct.

[momaka]

Re: Prudent Way PWI-PR350 Rebuild

If it squeals with only the 5vsb active, then it's got nothing to do with those caps. Those are used on the main switching transistors and are only used when the PSU is actually running.

Exactly.

The squeal is likely from the 5VSB circuit. Now this may or may not be something to worry about. Here's why:

5VSB circuits usually employ flyback topology (i.e. no output inductor), regardless if it's a chip or 2-transistors circuit. Actually, many TVs and monitors (both LCD, CRT, and LED) regularly use this topology, among many other devices. And from what I remember reading a while back, since there is no output inductor, the ripple current on the output of such PSUs is usually quite high. Therefore, it is actually possible to use very low ESR and high ripple current caps on these PSUs without problem - even polymers. Now, I can't tell you why (because I don't know either), but this often results in a high-pitched squeal. I wouldn't worry about it though. I have seen this both from factory-made PSUs that used low ESR caps (such as Chemicon KZE) and through my own experiments. Almost all of the LCD monitors that I recapped with Panasonic FR and FM (and in one case also tried Rubycon MCZ) would produce a high-pitched squeal afterwards. But output voltages were stable as a rock - much more than with the stock crappy caps - so I left them alone.

There are cases, however, where that high-pitched noise could indicate a problem. It depends how the 5VSB feedback is implemented. Normally, most PSUs just take the 5VSB, run it though a half-half resistive divider, and connect it to a 431 shunt. But I have also seen a half-bridge PSU where they took both the 5VSB and the secondary side auxiliary rail from 5VSB and used a combination of both for feedback. So that's why it is usually important to recap both the output of the 5VSB rail and the secondary auxliary rail.

Looking at the pictures you provided, I think you already did that (brown UCC KY/KZE cap on the 5VSB rail and a small Panasonic FC? cap next to it). So in that case, I think you should try to measure the voltage of the secondary side auxiliary rail. Should be 10-20V without load on the 5VSB. With load, it may go up to 25V - it varies quite a bit between different PSUs, though, so I can't give you concrete numbers. Just make sure that even with a heavy load on the 5VSB (like 1 to 1.5A) that the secondary side auxliary rail's voltage does not go over the voltage rating on that Panasonic FC cap you used. In some cases , I've seen improperly specced caps on that rail pop (*cough* Deer *cough* ).

The secondary side auxiliary rail on your PSU comes out of that smaller diode next to the bigger diode for the 5VSB. The 5VSB diode appears to be labeled "5D3" on the board. There is a smaller diode just behind it, using this picture as a reference:
https://www.badcaps.net/forum/attach...3&d=1432017436
From what I can tell, it is possible that Q4 and those overheated resistors are used for some kind of linear voltage regulator for the controller. So that may be why they run so hot.

Or it could also be that they overheated if the PSU had bad caps on the output of 5VSB and before you recapped it - especially if it was used like that for a bit.

So just check the voltage on the secondary side aux. rail, just to make sure it's not up to something funny. Again, its voltage should stay under that voltage rating of that Panasonic FC cap, even with moderate 5VSB load. And hopefully, not higher that 30V. Even 25V is a bit on the high side.

I am just curious if it's the 5VSB design or bad caps on the 5VSB that caused those resistors to run hot like that.

[Behemot]

Re: Prudent Way PWI-PR350 Rebuild

Leadman, Sun Pro or something like that. They have been labeled as Eurocase here. Not worst but not wonder either.

[Pentium4]

Re: Prudent Way PWI-PR350 Rebuild

From my experience those capacitors usually are 4.7uF, 2.2uF or 1uF and rarely 10uF.

Don't hesitate to replace them with 2.2uF and test it. I have seen the same psu (Jou Jye 400W) using 1uF caps in 2004 revision and changing them with 4.7uf in 2005 revision and newer..

You rarely see 10uF there? Hmm, seems to be fairly more common in my experience. I think 1uF seems more rare.

If it squeals with only the 5vsb active, then it's got nothing to do with those caps. Those are used on the main switching transistors and are only used when the PSU is actually running.

Alright, I'll test it again just to make sure I was remembering correctly.

No UL number! Can't tell what OEM it might be. (if it isn't designed by "Prudent Way") The layout looks familiar. (Especially with the 4 caps in a square shape on the secondary, instead of being staggered to fit them in a smaller area)

Leadman, Sun Pro or something like that. They have been labeled as Eurocase here. Not worst but not wonder either.

It's actually made by XHY Power http://www.xhy-power.com.cn/

Does it feed the fan 12V, or is it speed/temp controlled?

It's speed/temp controlled, and it's very, very quiet. Such a relief

The squeal is likely from the 5VSB circuit. Now this may or may not be something to worry about. Here's why:

5VSB circuits usually employ flyback topology (i.e. no output inductor), regardless if it's a chip or 2-transistors circuit. Actually, many TVs and monitors (both LCD, CRT, and LED) regularly use this topology, among many other devices. And from what I remember reading a while back, since there is no output inductor, the ripple current on the output of such PSUs is usually quite high. Therefore, it is actually possible to use very low ESR and high ripple current caps on these PSUs without problem - even polymers. Now, I can't tell you why (because I don't know either), but this often results in a high-pitched squeal. I wouldn't worry about it though. I have seen this both from factory-made PSUs that used low ESR caps (such as Chemicon KZE) and through my own experiments. Almost all of the LCD monitors that I recapped with Panasonic FR and FM (and in one case also tried Rubycon MCZ) would produce a high-pitched squeal afterwards. But output voltages were stable as a rock - much more than with the stock crappy caps - so I left them alone.

Very interesting. High ripple circuits, is this why they need a CLC filter even for such low loads? That would explain it, and why certain Leadmans have stupid amounts of ripple. Yeah, I've heard factory PSU's do that too. I'll have to measure the voltage again and see what it was reading.

There are cases, however, where that high-pitched noise could indicate a problem. It depends how the 5VSB feedback is implemented. Normally, most PSUs just take the 5VSB, run it though a half-half resistive divider, and connect it to a 431 shunt. But I have also seen a half-bridge PSU where they took both the 5VSB and the secondary side auxiliary rail from 5VSB and used a combination of both for feedback. So that's why it is usually important to recap both the output of the 5VSB rail and the secondary auxliary rail.

Yeah, and those darkened resistors might indicate a problem, or that there was one at one point with the possible bad/failing caps. Maybe the caps actually went out of spec when the load was increased on that rail. Pretty hard to trust BH caps It does look like a unique 5VSB design. Also, I found it weird how far away the 5VSB filtering caps were away from each other. The most annoying thing about this power supply was the glue! See how I had to remove the main IC to remove some, along with all of those little components.

Looking at the pictures you provided, I think you already did that (brown UCC KY/KZE cap on the 5VSB rail and a small Panasonic FC? cap next to it). So in that case, I think you should try to measure the voltage of the secondary side auxiliary rail. Should be 10-20V without load on the 5VSB. With load, it may go up to 25V - it varies quite a bit between different PSUs, though, so I can't give you concrete numbers. Just make sure that even with a heavy load on the 5VSB (like 1 to 1.5A) that the secondary side auxliary rail's voltage does not go over the voltage rating on that Panasonic FC cap you used. In some cases , I've seen improperly specced caps on that rail pop

Yeah, they're Chemi-Con KY 1000uF 10V before the coil, and 680uF 16V after the coil. The Panasonic FC cap is 47uF 50V. Thanks for sharing your knowledge with me. I will measure it when I find the thing. I think I know where I packed it

I am just curious if it's the 5VSB design or bad caps on the 5VSB that caused those resistors to run hot like that.

I'll let you know I'm willing to guess that it was from the poor shape of the BH caps. I'll also let it run with some load and get out my infrared thermometer to see how hot that section of the supply gets after a while.


Anyone have any idea why they used wire to fly that coil all the way to the feedback transformer?

[momaka]

Re: Prudent Way PWI-PR350 Rebuild

You rarely see 10uF there? Hmm, seems to be fairly more common in my experience. I think 1uF seems more rare.

Yeah, I find 10 uF to be more common as well. Maybe it's because we are used to seeing Macron PSUs . Although IIRC, L&C/Deer/Allied also uses 10 uF caps for those.

Very interesting. High ripple circuits, is this why they need a CLC filter even for such low loads?

Could be. I'm not entirely sure. In general, though, wherever you need to filter a high frequency circuit without much loses, a CLC filter is used. If loses are not a problem, you can do RC. The one thing you probably won't see on a flyback topology is a big inductor before the caps (i.e. LCLC). The primary side of the transformer servers as the output inductor, that's all I know.

Pretty hard to trust BH caps

Well let's Be Honest here - BH caps are crap.

Thanks for sharing your knowledge with me.

Most welcome!

Anyone have any idea why they used wire to fly that coil all the way to the feedback transformer?

Some kind of OPP?
My guess is they are *trying* to pick up stray flux from the core of the feedback traffo. But I somehow doubt that will work. If this really is for OPP, I will be very surprised if it works at all.
Which pins does it feed to on the PWM controller and what is the PWM controller?

*edit*
Never mind that last question. I see that the controller is an Az7500b (?)

[Pentium4]

Re: Prudent Way PWI-PR350 Rebuild

Yeah, I find 10 uF to be more common as well. Maybe it's because we are used to seeing Macron PSUs . Although IIRC, L&C/Deer/Allied also uses 10 uF caps for those.

Yep, that's probably why Same with Super Flower half bridge units.

Could be. I'm not entirely sure. In general, though, wherever you need to filter a high frequency circuit without much loses, a CLC filter is used. If loses are not a problem, you can do RC. The one thing you probably won't see on a flyback topology is a big inductor before the caps (i.e. LCLC).

It sure seems that way. Looking at c_hegge's review on the MPT-301 though, that thing handled 2.98A on the 5VSB, was running at 4.97V, and only had a max of 20.8mV when running at the max label ratings, and 27.4mV when being overloaded. Impressive! They use an abnormally good filter for the 5VSB though. Good sized PI coil, and a 2200uF cap before and after it.

Well let's Be Honest here - BH caps are crap.

I'm impressed though that all the filtering ones were okay on 3.3V, 5V, and 12V. One of the 2200uF 10V caps on the 3.3V was next to a 33Ω resistor, and the side of the cap had that "charred" look right next to where the resistor was. It tested 2236uF and 0.02Ω ESR

Some kind of OPP?
My guess is they are *trying* to pick up stray flux from the core of the feedback traffo. But I somehow doubt that will work. If this really is for OPP, I will be very surprised if it works at all.
Which pins does it feed to on the PWM controller and what is the PWM controller?

*edit*
Never mind that last question. I see that the controller is an Az7500b (?)

Yeah it's an AZ7500BP-E. None of the coils' leads are directly connected to the 7500, but one is connected through that resistor with the black marking underneath it (stupid how they put the component numbers UNDER the components) and it's connected to pin 16 on the IC which is 2IN+

[Pentium4]

Re: Prudent Way PWI-PR350 Rebuild

Okay so I was wrong about one thing. It's not a high pitched whine, it's a dull buzzing sort of sound. Almost similar to those old school fluorescent lights in old buildings. With power plugged in, 5VSB runs at 5.13V. When just turning the PSU on with a paper clip, the buzz gets slightly louder, and 5VSB stays at 5.13V. When the PSU is just plugged in, there is about 9V going through the auxiliary cap. When turning on via just the paper clip, there is about 23V going through the same cap. Both voltages start a little higher, then quickly lower to the other voltages.
Edit: I forgot to add that when pulling out the paper clip, the buzz became pulsating, and almost beep like until the main capacitors drained, then went to the more quiet low buzz.

[ben7]

Re: Prudent Way PWI-PR350 Rebuild

Okay so I was wrong about one thing. It's not a high pitched whine, it's a dull buzzing sort of sound. Almost similar to those old school fluorescent lights in old buildings. With power plugged in, 5VSB runs at 5.13V. When just turning the PSU on with a paper clip, the buzz gets slightly louder, and 5VSB stays at 5.13V. When the PSU is just plugged in, there is about 9V going through the auxiliary cap. When turning on via just the paper clip, there is about 23V going through the same cap. Both voltages start a little higher, then quickly lower to the other voltages.
Edit: I forgot to add that when pulling out the paper clip, the buzz became pulsating, and almost beep like until the main capacitors drained, then went to the more quiet low buzz.

You could try changing the compensation network for the feedback. That's really the only suggestion I can think of.

Well let's Be Honest here - BH caps are crap.

Let's Be Honest, they are Brutally Horrible, always Bulging Here and there.

-Ben

[momaka]

Re: Prudent Way PWI-PR350 Rebuild

With power plugged in, 5VSB runs at 5.13V. When just turning the PSU on with a paper clip, the buzz gets slightly louder, and 5VSB stays at 5.13V. When the PSU is just plugged in, there is about 9V going through the auxiliary cap. When turning on via just the paper clip, there is about 23V going through the same cap. Both voltages start a little higher, then quickly lower to the other voltages.
Edit: I forgot to add that when pulling out the paper clip, the buzz became pulsating, and almost beep like until the main capacitors drained, then went to the more quiet low buzz.

Well, the voltage appear normal. Perhaps the swing on the auxiliary rail is a tad bit larger than what I normally see. But it should be fine. As long as those toasy-looking resistors don't run super hot anymore, I think it should be okay.

You could try changing the compensation network for the feedback. That's really the only suggestion I can think of.

Probably not worth messing with it. Not unless you really want to experiment "in the name of science", if you know what I mean . And it would be best to have some kind of load tester and of course a dim bulb device to make sure everything goes smoothly if you do experiment.

[ben7]

Re: Prudent Way PWI-PR350 Rebuild

Probably not worth messing with it. Not unless you really want to experiment "in the name of science", if you know what I mean . And it would be best to have some kind of load tester and of course a dim bulb device to make sure everything goes smoothly if you do experiment.

I agree, though I don't think messing with the feedback compensation could make the power supply blow up, unless you change the feedback voltage itself (basically, changes the voltage the supply regulates the output at).
But if you have good hearing like me, then I don't appreciate high pitch whining. However I can tolerate buzzing, like in this case. :P

-Ben

[Pentium4]

Re: Prudent Way PWI-PR350 Rebuild

Thanks for the replies! It seems like there is nothing wrong with the 5VSB circuit. By the way, I found that project where I used the nicon caps. They were indeed 4.7uF 50V, so turns out I guessed correctly

You could try changing the compensation network for the feedback. That's really the only suggestion I can think of.


Let's Be Honest, they are Brutally Horrible, always Bulging Here and there.

-Ben

[lti]

Re: Prudent Way PWI-PR350 Rebuild

I just noticed the umlaut in the brand name (Prüdent Way). That makes it better, right?

The soldering doesn't look good.

BH caps seem to have inconsistent quality. They can sometimes handle some abuse and last a long time.