Power supply build quality pictorial. part 2

[c_hegge]

Re: Power supply build quality pictorial. part 2

Nice Macron!
I'm guessing the 12V rectifiers are not in parallel, though. Would have been nice if there really were two totally separate 12V rails. Then the thing could probably do 20A on each (provided the transformer and primary could handle that). Of course, you'd need two sets of filters for the 12V rails, and that takes a lot more space.

They are in parallel. The thing separating them is the two OCP channels.

[Pentium4]

Re: Power supply build quality pictorial. part 2

Wait, am I awake? Is that really a Sun Pro worth keeping!?
Indeed it does seem like the best Sun Pro I've ever seen too. The cap selection may be horrible, but the rest of the PSU looks clean and well-built. And look at that - the PCB is connected to the AC line with a proper connector! You also have a proper ceramic fuse. For $6, it's quite a worthy PSU.

Now I don't know if it will really do it's rated 450W rating. With those heat sinks, I'm more inclined to think around 350W continuous is where it will top out. Possibly peak at 400W for a little bit. I just don't quite trust Sun Pro designs. But at least the individual current ratings for each rail seem quite fairly rated.
Another thing I'm going to pick on is the trace spacing on the primary side of the PCB - it really is too small. Seems like many Sun Pro PSU are like this. Also, I only see 1 cap spot for 12V rail. I would have expected 2 (keep in mind that I am spoiled, though, by HiPro PSUs with multi-12.5 mm cap spots ).

Yeah, the soldering could use some work but I didn't expect good soldering from Sun Pro. The reason I say I think it could do 400W no problem, look at this review: http://www.clubedohardware.com.br/ar...0-560-W/1904/1

That one did over 425W on 120VAC with a 4A bridge, smaller transformer, and weaker rectifiers. I imagine that 35 transformer saturated, asking over 425W from it in a half bridge design is asking a lot, I think the bigger transformer would help a lot.

I will have to load it up and see how the fan acts, although that little thermistor is not glued to anything, and never was! Super glue doesn't go conductive, correct?

[momaka]

Re: Power supply build quality pictorial. part 2

Interesting. So it did make it past 400 W. Good job Sun Pro!

Yes, super glue is not conductive and will not become conductive as far as I know.

[ncovert]

Re: Power supply build quality pictorial. part 2

Lead Year FP-150N60. Rated for 150 watts.
Primaries are HEC, secondaries are G-Luxon. Heatsinks are rather thin, but this thing is only 150 watts! As you can tell, this is one of those shallow-case (now known as Micro PS3) units.

[c_hegge]

Re: Power supply build quality pictorial. part 2

And here's the Macron recapped. The cool part is, it even has PCB spacing for 12.5mm caps in some areas, so I was able to use Nichicon HE to replace the 3300uF 10V caps, which are cheaper than the 10mm Panny FKs

[Behemot]

Re: Power supply build quality pictorial. part 2

Arent's Samxon RS even cheaper and yet in D10?

[Pentium4]

Re: Power supply build quality pictorial. part 2

^ Yeah, but if there's space for the 12.5mm caps, might as well use them and get the better ripple current rating so you can save the 10mm caps for when they're really needed! Looks great c_hegge! What are your plans for that Macron?

[c_hegge]

Re: Power supply build quality pictorial. part 2

^ Yeah, but if there's space for the 12.5mm caps, might as well use them and get the better ripple current rating so you can save the 10mm caps for when they're really needed!

This, and Samxon aren't Japanese, so I only use them I there's no way I can fit anyting else in.

What are your plans for that Macron?

I originally purchased it to review, but that's not going to happen now that I've had to recap it. TBH, I don't know what I'll do with it now. I already have more PSUs than I know what to do with. I'm sure I'll find a use for it one day, though.

I put it on the load tester too, and it did the full 400W in spec. The ripple was very good, only about 30mV on the 12V rail, 15mV on the 5V and 10mV on the 3.3V. The efficiency was about 77%. My only real gripe was that the fan was running too slow for my liking, so I swapped it for a higher speed ADDA 2BB fan.

[ncovert]

Re: Power supply build quality pictorial. part 2

Here we have a 200W Compaq/LiteOn power supply, pulled from a Deskpro EN. As you can see, I have removed this unit from its case, so it's just the board. Rubycon primaries, a mix of Matsushita and Rubycon on the secondary.


Now, this unit doesn't have a standard color scheme. Orange is +12 instead of +3.3, and brown is +3.3 here. And it has a 24-pin main connector! Of course, it's a different pinout than today's 24-pin, but still cool that a larger ATX connector existed back in 1999.

[ben7]

Re: Power supply build quality pictorial. part 2

Here we have a 200W Compaq/LiteOn power supply, pulled from a Deskpro EN. As you can see, I have removed this unit from its case, so it's just the board. Rubycon primaries, a mix of Matsushita and Rubycon on the secondary.


Now, this unit doesn't have a standard color scheme. Orange is +12 instead of +3.3, and brown is +3.3 here. And it has a 24-pin main connector! Of course, it's a different pinout than today's 24-pin, but still cool that a larger ATX connector existed back in 1999.

I've seen that orange 12V and brown 3.3V scheme on other compaq PSUs

[Th3_uN1Qu3]

Re: Power supply build quality pictorial. part 2

High Power HPC-370-P12S by Sirtec

Ye olde half bridge design... Done pretty good this time. Passive PFC, full input filtering, 560uF primary capacitors. 2SC2625s as main switchers, 02N60P for standby. Properly sized output inductors. Full pi filters on all outputs. I did not bother getting to the diodes, but all three look beefy. Temperature controlled fan. All capacitors in this unit were made by Teapo, all were okay.

The only thing wrong with this PSU was that both MOVs were blown, shorting the primary capacitors. This has likely seen several power surges. Once the MOVs were removed, the power supply works fine. Now i get why they don't install them anymore, they have a finite life time.

Output cables:

• One ATX 24-pin (not 20+4, in one piece).
• One ATX12v.
• One chain with 3 Molex connectors and one floppy connector
• One chain with 2 more Molex connectors
• One chain with 2 SATA connectors.

All wiring 18AWG, except 3.3v rail which uses 16AWG. Overall, very nice for a low-end power supply. While i was in there i removed the wires for the 115/230v switch, just so that nobody does anything stupid (we have 230v in Europe). I also flipped it to 115v for shock value.

[japlytic]

Re: Power supply build quality pictorial. part 2

Sony APS-306 (from a PlayStation 3)

1000uF 10V Chemi-Con KZH before +5.5V (900mA) inductor with 680uF 10V Rubycon ZLH after it.
100uF 16V Sanyo SEPC x2 before +12V (13A) inductor with 680uF 16V Lelon RZH after it.
82uF 450V Lelon RGL x2 (in parallel) primary units.
47uF 35V Rubycon YXA unit on primary side control voltage for +12V output regulator.
47uF 35V Rubycon ZLH unit on primary side control voltage for +5.5V standby regulator.

13N60 (ST) PFC MOSFET with BYV25X-600 (NXP) PFC output rectifier.
K6A60D (Toshiba) x2 in a half-bridge configuration for +12V transformer.
TS10B06G primary side bridge rectifier.
MIP836 (Panasonic) +5.5V standby controller.

[japlytic]

Re: Power supply build quality pictorial. part 2

I forgot to mention about the +12V rectifier in the Sony APS-306, which is an ST STPS3045CT.

[momaka]

Re: Power supply build quality pictorial. part 2

I forgot to mention about the +12V rectifier in the Sony APS-306, which is an ST STPS3045CT.

I'm guessing this is a PS3-slim PSU. I have some of the older (fat) PS3 PSUs, and those pack some serious components inside. After all, they are rated for 32A on the 12V rail, so no wonder. I'll post pictures of those someday when I get a chance.

[Wester547]

Re: Power supply build quality pictorial. part 2

Dell branded Newton Power NPS-200PB-73M rev: 03, date code week 33 of 2000 (August), made in Thailand. It came out of a Dell XPS B1000. Yes, it has the notorious and dreaded proprietary ATX connector that Dell was a stickler for until 2001, as well as a proprietary auxiliary connector in addition to that. These were very common in Dell Pentium III machines from 1999-2000, with some having build dates as early as late 1998 and as late as early 2001. The underside shot, while definitely not great, shows that Delta/Newton's soldering quality is nothing to sneeze at. The unit is rated for 200W maximum power output with +5V being rated for 22A, +3.3V being rated for 18A, +12V being rated for a whopping 6A, -12V being rated for 0.8A, -5V being rated for 0.5A (0.8A combined), +5V and +3.3V being rated for 135W combined, and +5VSB being rated for 1.2A.

I think it could do 250W with no problem, assuming the load was +5V heavy and that it wouldn't draw upon the +12V rail much. Those heatsinks are very nice for 200W (their base may not be that thick but this is rated for 200W...), and I'd reckon those heatsinks would be befitting for 350W with no issues. Primaries are 470uF, 200V Rubycon USP and the bridge rectifier, while not shown in the images, is D3SB60 (4A/600V) without a heatsink. The main transformer appears to be some relatively small variation of a size 35. Secondary side rectifiers are MBR3045PT for the +5V and +3.3V rails and STPR1020CT for the +12V rail. -5V has a linear regulator, L7905CV, soldered to the secondary heatsink. The output capacitors are:

+5VSB: Two 1000uF 10V Rubycon YXFs with a small ferrite coil after them.
-5V: 10uF 50V Taicon VT before the regulator.
+3.3V: 1500uF 10V Rubycon YXB before the ferrite coil with a 2200uF 10V Rubycon YXB after it.
-12V (not linear regulated): 1000uF, 16V Chemi-con LXJ.
+5V: 1500uF 10V Rubycon YXB before and after the ferrite coil.
+12V: 2200uF 16V Chemi-con LXJ and a 330uF 16V Rubycon YXF in parallel.

I would guess that YXB is relatively equal to YXF and LXJ in terms of ripple and ESR ratings. The +5V/+12V toroid appears to use -52 material and the +3.3V magamp -26 material. There's another small toroidal coil that appears to be after the voltage doubler but before the main switcher (2SK2746, so this is single switch forward topology). I don't know what it's for but it appears to be made of -52 material as well and I see it in almost every Newton unit (in the early 2000s) built for Dell or otherwise. The rest of the toroids in the PSU (for input filtering), besides one large one (for input filtering) and the small one deriving the +3.3V transformer pins from the +5V pins, seem to use -26 material. As you can see, there's another daughterboard soldered to the secondary side, which has some TO-92 parts soldered to it and I assume is part of the feedback loop otherwise with its network of smallish resistors and what not. The fan is a Delta ASB0812HH that draws 0.30A @ 12V. It's rated up to 3250RPM pushing 37.43 CFM, rated for .181", at 37 DBA. It only appears to spin at about 1600-1700RPM when temperature controlled by way of the thermistor, though, but at that point it's fairly loud and moves a great deal of air (though the fan seems to be producing an unusual amount of vibration at the bottom of the chassis only... not quite as much as those Sunon sleeve bearing fans of old do, the ones that seem to "rattle", but close enough). The input filtering stage is complete with MOVs, X capacitors, Y capacitors, the NTC thermistor, etc...

This power supply actually starts with no load and when that happens, the fan spins quite a bit slower, a little higher than 1200RPM, I would say, at which point it becomes much quieter, the aforementioned vibration lessens, and of course the airflow is quite diminished. DNA1001D controls the primary side with DNA1002C handling the supervision duties of the secondary side. The tiny capacitors are a mingling of Taicon VT and Rubycon CEW (85*C). Unfortunately, this power supply does seem to use a two transistor +5VSB circuit since I could not locate anything but a small FET right by the flyback transformer (after which three optocouplers lay). I couldn't make out what it said on the package even with a bright flashlight but it doesn't even have a heatsink, and I don't like that (in fact, Newton/Delta did that with their 200-250W units for Dell all the the way up to 2005...). Even worse, the "critical" capacitor is a 100uF, 25V Rubycon CEW. To that end, this is not a "forward looking" power supply (no pun intended). Okay, I get that many were doing this back in 2000, but still. The wires are all 18 gauge save by the wires going to the floppy connector which are the standard 22 gauge.

Aside from that, though, it's a good unit for what it is. This was before Delta and Newton units started being chock full of LTEC and Taicon, and worse yet, later on, CapXon. Sorry for the long winded post.

[ncovert]

Re: Power supply build quality pictorial. part 2

Has FSP gone downhill lately?
This FSP VR400 isn't that impressive. Teapo primary caps, a mix of CrapXon, Teapo, and OST on the secondary. It's been powering my main PC for several months now, but I decided to replace it with an AcBel API4FS06.

[Agent24]

Re: Power supply build quality pictorial. part 2

Has FSP gone downhill lately?
This FSP VR400 isn't that impressive. Teapo primary caps, a mix of CrapXon, Teapo, and OST on the secondary. It's been powering my main PC for several months now, but I decided to replace it with an AcBel API4FS06

The impression I got was that FSP have good designs but use useless capacitors in their low-end designs. If you replaced the capacitors with decent ones you get a really good PSU.

[momaka]

Re: Power supply build quality pictorial. part 2

Has FSP gone downhill lately?
This FSP VR400 isn't that impressive.

Yeah, for 350W it doesn't look that good, indeed. Claims are it also does over 75% efficiency. But even so, at 300W DC power draw, this thing will have to be dissipating around 75W of heat. I can see those Ost, Teapo, and especially CapXon caps, easily cooking in there.

[ncovert]

Re: Power supply build quality pictorial. part 2

I think I'll give it a recap. Should be good for 250-300W.

[Behemot]

Re: Power supply build quality pictorial. part 2

FSP continuously goes down the hill for quite some time. Is that Fortron/Source or some OEM supply?