Power supply build quality pictorial. part 2

[stj]

Re: Power supply build quality pictorial. part 2

nice foto's - why the hell arent all designs that clean?!

[japlytic]

Re: Power supply build quality pictorial. part 2

Boschert PS-9165 (from a Unisys S40DISK unit):

2x 220uF 200V Panasonic primary units.
2x 2200uF 16V Chemi-Con SXC before +5V inductor and 1000uF 25V Chemi-Con SXC after it.
680uF 16V Chemi-Con RX before +12V inductor and 470uF 25V Chemi-Con SXC after it.
220uF 25V Jamicon WG on -5V and -12V regulator inputs and 100uF 25V Jamicon WG at each negative regulator output.
Other electrolytic capacitors not mentioned are Nichicon and Chemi-Con.

MBR1645 +5V rectifer.
BYW29-150 +12V rectifier
1N5822 rectifier each before -5V and -12V regulators.
KBL04 input bridge rectifier
BUW12 primary switcher.

[ben7]

Re: Power supply build quality pictorial. part 2

Boschert PS-9165 (from a Unisys S40DISK unit):

2x 220uF 200V Panasonic primary units.
2x 2200uF 16V Chemi-Con SXC before +5V inductor and 1000uF 25V Chemi-Con SXC after it.
680uF 16V Chemi-Con RX before +12V inductor and 470uF 25V Chemi-Con SXC after it.
220uF 25V Jamicon WG on -5V and -12V regulator inputs and 100uF 25V Jamicon WG at each negative regulator output.
Other electrolytic capacitors not mentioned are Nichicon and Chemi-Con.

MBR1645 +5V rectifer.
BYW29-150 +12V rectifier
1N5822 rectifier each before -5V and -12V regulators.
KBL04 input bridge rectifier
BUW12 primary switcher.

Wow, that is a beauty!

[stj]

Re: Power supply build quality pictorial. part 2

i used to see dozens of psu's built like that from ASCOT.

[Wester547]

Re: Power supply build quality pictorial. part 2

Sort of off-topic to the thread but I have wondered for the longest time if the circled piece of metal in the attached image is considered a "fin" along with the rest of them. It is part of the, what I guess could best be called, "striped metal" design that's very common on the primary and secondary heatsinks of OEM Bestec, Hipro, and Newton/Delta power supplies. This isn't my image, it's from a Hipro larrymoencurly posted a few years back. I think it's too small to be considered a fin but I'm curious to know what benefit it contributes to the overall design of the heatsink, besides surface area and what have you.

[Behemot]

Re: Power supply build quality pictorial. part 2

It's too damn small even on 1280x1024 display…do you mean those jags on top?

[stj]

Re: Power supply build quality pictorial. part 2

the wavy surface increases the surface area to disipate the heat.

[Wester547]

Re: Power supply build quality pictorial. part 2

It's too damn small even on 1280x1024 display…do you mean those jags on top?

Yes, I was talking about those jags. I was trying to accentuate it by just using that excerpt from the whole image. I wanted to know if those qualify for "fins" or if they're anywhere near as convenient.

[Behemot]

Re: Power supply build quality pictorial. part 2

Well they can double the surface area without using any extra manufacturing processing and material…why not, I personally take it as „surface area for free“

[SIDMX]

Re: Power supply build quality pictorial. part 2

Not exactly double the area but going by the small pic I counted 24 two-sided fins with roughly 1mm per side (just guessing) that adds 4.8cm more surface that of course needs to by multiplied by the "depth" (from the pic perspective) of the heatsink and then take into acount the airflow that the fan should be pushing through them, so yes even small fins like that are useful.

[Pentium4]

Zumax ZU550W (X3)

Topower seems to be dead, but I guess Zumax still makes some decent PSU's (Although still overrated) A lot of "Topower" PSU's with the Topower UL number actually have Leadman units in them...Leadman at their worst too.

Anyways...

This one looks okay. I hate the lack of an input filter though. Those are not proper Y caps. There are no spots for coils (or maybe there is, LF1?) and all my scrapped X caps are too long for the slots provided. So all I did was replace the 3 fake Y caps with real ones, soldered an X cap to the receptacle, and added a beefy heatshrinked MOV, they all came from a dead Bestec (It had so much tan glue that I deemed it not worth the fix) I know it's not a complete input filter but it's something. Those Jun input caps read 690uF and 700uF. It uses 13009 transistors. From first glance I thought it was double forward. I'm going to go out on a limb and say that I think this thing could do 400W even with 13009's because they have everything going for them. TO-247, good sized heatsink, no 2 transistor so it doesn't have to share more heat with the extra transistor, plenty of bulk capacitance, good sized transformer that shouldn't saturate under 400W, and two fans in push-pull configuration. What do you guys think?

The 5V and 3.3V are rectified by one MBR3045CT each, and the 12V uses two MBR30H100CT's in parallel. Not bad. It looks like this PSU uses the same base design as the CoodMax that c_hegge reviewed. That one looks significantly worse and did 260W (He's on 230V though)
Link: http://hardwareinsights.com/wp/the-2...ply-roundup/9/

[goodpsusearch]

Re: Zumax ZU550W (X3)

I'm going to go out on a limb and say that I think this thing could do 400W even with 13009's because they have everything going for them. TO-247, good sized heatsink, no 2 transistor so it doesn't have to share more heat with the extra transistor, plenty of bulk capacitance, good sized transformer that shouldn't saturate under 400W, and two fans in push-pull configuration. What do you guys think?

I agree. It all depends on the design now.

I can't believe that this thing did 330W+ using 470uF primary capacitors!
http://www.xbitlabs.com/articles/cas...400w-520w.html
This unit is sold for 25 euros here.

Meanwhile its US version did 420W using 560uF primary caps though:
http://www.hardwaresecrets.com/print...ply-Review/975

[Pentium4]

Re: Zumax ZU550W (X3)

I can't believe that this thing did 330W+ using 470uF primary capacitors!
http://www.xbitlabs.com/articles/cas...400w-520w.html
This unit is sold for 25 euros here.

Isn't the reason that low bulk capacitance limits the PSU's ability because when the bulk capacitors reach their limit, they can no longer properly smooth the pulsed DC current, putting more stress on the switching transistors?

So I found this strange on this PSU...I was redoing some of the "blobby" solder joints, one of them being that big blob you see in the center of the solder shot. I removed it and saw that the blob had a total of 4 joints under it, and they're clearly separated. The PSU was functioning before, so was it supposed to be like that by design or should they be soldered separately? Going from left to right the joints are part of 1. Jumper wire, 2. diode cathode, 3. diode cathode, 4. zener diode anode

[momaka]

Re: Zumax ZU550W (X3)

What do you guys think?

Nice PSU. Too bad the board silkscreen is such a mess. Everything else seems good. I like the dual 30A rectifiers on the 12V. That's a lot of beef for a half-bridge. Means it can do 30A on the 12V no problem, or 360W. The output filtering caps seem a bit small, though. With the PI coils, though, I doubt the filtering is bad.

It looks like this PSU uses the same base design as the CoodMax that c_hegge reviewed. That one looks significantly worse and did 260W (He's on 230V though)
Link: http://hardwareinsights.com/wp/the-2...ply-roundup/9/

260W for that gutless turd? With almost 80% efficiency before it died! - Now that is pretty amazing! Just look at the Sun Pro on the next page in that review .

So I found this strange on this PSU...I was redoing some of the "blobby" solder joints, one of them being that big blob you see in the center of the solder shot. I removed it and saw that the blob had a total of 4 joints under it, and they're clearly separated. The PSU was functioning before, so was it supposed to be like that by design or should they be soldered separately?

If it works and doesn't do anything strange, then I guess it was supposed to be this way. Looks like something to do with the secondary auxiliary rail on the 5VSB circuit. No exact idea, though.

[Pentium4]

Re: Power supply build quality pictorial. part 2

Nice PSU. Too bad the board silkscreen is such a mess. Everything else seems good. I like the dual 30A rectifiers on the 12V. That's a lot of beef for a half-bridge. Means it can do 30A on the 12V no problem, or 360W. The output filtering caps seem a bit small, though. With the PI coils, though, I doubt the filtering is bad.

It is a mess isn't it :/ Yeah I like having rectifiers in parallel too, especially 30A Which is surprising on a PSU like this. The stock filtering caps are Asia'X TMX 2200uF 16V. I replaced them with 3300uF 16V Samxon RS from Behemot. Plus, if on the one c_hegge reviewed, that one did decent ripple suppression with two 1000uF 16V same series caps, and no coil. Indicating that it must be a good design.

260W for that gutless turd? With almost 80% efficiency before it died! - Now that is pretty amazing! Just look at the Sun Pro on the next page in that review

I was surprised too. And with mine having higher rated 12V rectifiers in parallel, the efficiency is probably the same or better! (Because I'm on 120V)

If it works and doesn't do anything strange, then I guess it was supposed to be this way. Looks like something to do with the secondary auxiliary rail on the 5VSB circuit. No exact idea, though.

I guess I'll just redo it as it was It's fully recapped now

[Pentium4]

Re: Power supply build quality pictorial. part 2

Well, I tried being experimental and it literally blew up in my face. I soldered them separately and when I flipped the switch the fuse instantly blew and there's a black char on the top side of the PCB under the MOV I installed. Gonna replace the fuse and MOV in hopes of it being okay, after bridging those joints the way they were before.

[momaka]

Re: Power supply build quality pictorial. part 2

The stock filtering caps are Asia'X TMX 2200uF 16V.

Not bad at all. They looked much smaller in the pictures.

Well, I tried being experimental and it literally blew up in my face. I soldered them separately and when I flipped the switch the fuse instantly blew and there's a black char on the top side of the PCB under the MOV I installed. Gonna replace the fuse and MOV in hopes of it being okay, after bridging those joints the way they were before.

That's sad to hear.

You may want to try this trick from now on when doing any "experiments"
https://www.badcaps.net/forum/showpo...4&postcount=70
Will save you a lot of explosions.

[c_hegge]

Re: Power supply build quality pictorial. part 2

Well, I tried being experimental and it literally blew up in my face. I soldered them separately and when I flipped the switch the fuse instantly blew and there's a black char on the top side of the PCB under the MOV I installed. Gonna replace the fuse and MOV in hopes of it being okay, after bridging those joints the way they were before.

I demand pictures of the carnage

[Behemot]

Re: Zumax ZU550W (X3)

Isn't the reason that low bulk capacitance limits the PSU's ability because when the bulk capacitors reach their limit, they can no longer properly smooth the pulsed DC current, putting more stress on the switching transistors?

Yeah that's part of it, OFC. But than also the capacitance is there for the PSU to withstand micro-blackouts in the power grid. According to ATX spec, it should handle at least 16 ms on full power. I wonder how many PSUs on the market don't…I should be able to measure this by the end of the summer holidays or so.

Anyway - looks like you've just called Mr. kaboom back to action

[Pentium4]

Re: Zumax ZU550W (X3)

Not bad at all. They looked much smaller in the pictures.


That's sad to hear.

You may want to try this trick from now on when doing any "experiments"
https://www.badcaps.net/forum/showpo...4&postcount=70
Will save you a lot of explosions.

I demand pictures of the carnage

As you can imagine, the MOV is bad it reads 9.2Ω both ways. Looks like it did its job though. Safe to assume the BR is okay? I'm replacing the glass fuse with a slow blow fuse same rating.
Edit: The bridge rectifier tested good

But than also the capacitance is there for the PSU to withstand micro-blackouts in the power grid. According to ATX spec, it should handle at least 16 ms on full power. I wonder how many PSUs on the market don't…I should be able to measure this by the end of the summer holidays or so.

Anyway - looks like you've just called Mr. kaboom back to action

Gotcha, the "hold up time"

kaboom is always in action