Power supply build quality pictorial. part 2

[goodpsusearch]

Re: Power Win PW-330ATXE

IIRC, only Sirtec had problems with oscillations when you deviated too far from the original cap values.

[momaka]

Re: Power supply build quality pictorial - Lite-On PS-6161-2H1 (188 Watt ATX PSU)

No fancy introduction this time. Just a plain-looking Lite-On ATX power supply, model PS-6161-2H1. It came stock with my dumpster-picked HP Pavilion 8756c desktop PC. Though only rated for 188 Watts, the build quality is excellent. I'll let the pictures speak for themselves.

First the case: gray steel with good thickness. It is also a little shorter than most standard ATX PSUs.
https://www.badcaps.net/forum/attach...1&d=1460862225
https://www.badcaps.net/forum/attach...1&d=1460862225

Moving onto the label…
You can easily tell this is an old ATX v1.1 PSU. While the 5V rail is rated for 25 Amps, the 12V rail is rated for merely 4 Amps. Yes, four Amps is supposedly all you get. So obviously there is no 4-pin 12V CPU connector.


Next, an overall view with the cover off:
https://www.badcaps.net/forum/attach...1&d=1460862225

On the reverse, we see the solder side. Apart from that black wire, it looks clean. No SMD components anywhere here. Also, the solder joints are nice and shiny.
https://www.badcaps.net/forum/attach...1&d=1460862225

And the solder side of the logic daughterboard – more clean soldering here, though this one does have lots of SMD components.
https://www.badcaps.net/forum/attach...1&d=1460862225

Unpacking and removing everything from the case reveals the arrangement below.


Although the PSU is very compact, it has everything on there. Quality really is top-notch (I’m sure many of you PSU freaks have already noticed the “T” vents on the secondary, but more on that in a bit ). Even the fan is premium – a Panaflo FBM-08A12M (80 mm diameter, 25 mm thickness). And here is the filter board for the input:
https://www.badcaps.net/forum/attach...1&d=1460862225

Now let's have a closer look at the primary side under the heatsinks (which, by the way, are definitely more than adequate for a 188 Watt PSU). Topology is single transistor forward with a 2SK2654 transistor in a TO-3P case. Main transformer appears slightly smaller than “33” size, but larger than “28”. Personally, I think the primary could handle up to 200-250 Watts continuous.
https://www.badcaps.net/forum/attach...1&d=1460862225

Finally, the best part for last – the secondary side. Look at those tall and beautiful Panasonic caps!

In case anyone is wondering, they are HFQ series, which are now obsolete (almost equivalent to FC, though). The only real let-down is that tan/brown conductive/corrosive glue. But then again, I doubt this PSU will need a recap anytime soon. Only the smaller caps on the logic board and the filter caps for the -12V rail are not Japanese. Still, this PSU is barely “broken in”, with only slightly more than 25000 power-on hours (according to the HDD that came with the PC). For a computer from 2000/2001, that’s hardly much use.

Anyways, the only thing I can't explain is what that TYN610L silicone-controlled rectifier is for. Might be for the fan controller, but I don't know for sure and I haven't traced it either. There's another TO-220AB part right next to it, and I couldn't see what that was either. Perhaps another TYN610L? If anyone has a clue, I'm curious to know.

Primary Side Summary:

- three 4.7 nF Y2-class certified caps
- three MKP brand, GMF series X2-class certified caps (1 uF, 0.33 uF, 0.1 uF)
- two common-mode chokes, two MOVs
- D3SBA60 bridge rectifier (4 Amps, 600 V)
- two Rubycon USP bulk caps (in series): 200 V, 470 uF, 22 x 35 mm (dia. x h)
- 2SK6540 MOSFET in STF topology
- “33” size core main transformer, “19” size core 5VSB transformer
- 18 AWG input wiring, rated for: 600V before fuse, 300V after fuse

ICs
- UC3843BN primary driver
- KA3501 supervisory
- TNY255P for 5VSB generation
- four “815” optocouplers
- 7912 linear regulator (-12V rail)
- TYN610L (unknown purpose)

Secondary Side Summary:
3.3 V rail:
- mag-amp regulated
- B2545 (25 Amp, 45V) schottky recrifier
- two Panasonic HFQ 6.3 V, 4700 uF, 12.5x30 mm with PI coil (4 mm core, 3 turns) between them.
- 91-Ohm load resistor (?)

5 V rail:
- MBR4060P (40A. 60V) schottky rectifier
- two Panasonic HFQ 6.3 V, 2200 uF, 10x25 mm caps with PI coil (4 mm core, 6 turns) between them
- 100-Ohm load resistor

12 V rail:
- UG10DCT (10A, 200V) ultrafast recovery rectifier
- one Rubycon YXG 16V, 1200 uF, 10x25 mm cap with a PI coil (6 mm core, 17 turns) before it.
- no load resistor

-12 V rail:
- 1.5 or 2 Amp diode (as rectifier)
- one G-Luxon SM 16/25 V, 150 uF, 6.3x11 mm cap before 7912 regulator
- one Rubycon (can't see series) 25 V, 47 uF, 5x11 mm cap after 7912 regulator

-5 V rail:
- 1.5 or 2 Amp diode (as rectifier)
- one Rubycon YXG 16 V, 1200 uF, 10x25 mm cap
- 91-Ohm load resistor

5 VSB rail:
- 3 Amp (?) diode (as rectifier)
- one Rubycon JXA 10 V, 1000 uF, 8x200 mm cap before PI coil
- one Panasonic FA 10 V, 1000 uF, 8x200 mm cap after PI coil

Output wiring is all 18 AWG (with exception of floppy and two unknown auxiliary cables specific to that HP PC) and consists of a main 20-pin motherboard connector, 5 standard drive connectors, and 1 floppy connector.

[Pentium4]

Re: Power supply build quality pictorial. part 2

Nice PSU! Being 16 years old, it's still in good shape because of the caps and nice fan Do you use it at all?

The thing is compact but they did good as far as orientation, putting both heatsinks close to the fan. The label states it can run at 50C ambient. Should be fine if you don't live in death valley with your computer outside

This is a 188W PSU....and it has more filtering capacitance than some 400W! They also used nice and beefy toroid/PI coils.

What's the wire on the bottom for? Soldering looks good. Not perfect, but good.
It won't need a recap but it could use some glue removal!

[Wester547]

Re: Power supply build quality pictorial. part 2

The conductive glue and tiny G-Luxons and Jamicons are sort of... well.... show stoppers. It's a nice unit either way. 25,000 hours is a fair amount of use if it was power cycled much. I like how +5VSB is generated with a TOP switch in such an old unit. Strange how they don't linear regulate -5V from -12V and use a 16V capacitor on the -5V rail.

Hmm, 8x200mm... perhaps you mean 8x20mm? A 200mm tall capacitor would be one EPIC capacitor.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

12 V rail:
- UG10DCT (10A, 200V) ultrafast recovery rectifier
- one Rubycon YXG 16V, 1200 uF, 10x25 mm cap with a PI coil (6 mm core, 17 turns) before it.

Replace the UG10DCT with a 20A rectifier.

I would try to replace the Rubycon YXG 16V, 1200 uF with 2200uF (should work) or even 3300uF capacitor (might give oscillations).

-Add an ATX 12V 4 pin connector

Excelent psu!!!!

[ChaosLegionnaire]

Re: Power Win PW-330ATXE

It's not uncommon for a PSU manufacturer to have one PSU design/platform, and then make several "models" off of that, usually just putting cheaper components in the lower-end models. That includes smaller caps too.

yupp. one example i know of is the superflower leadex platform. it comes in silver, gold and platinum models. the silver uses all capxons throughout the psu for example and is the cheapest. while the gold uses 105°C chemicon caps on the primary and secondary side but the modular panel board uses capxons to smooth the output. the platinum is the most expensive model and uses japcaps all over the primary, secondary side and modular panel board.

HP Pavilion 8756c desktop PC

is this link on cnet the right specs of the pc u got? if its right not bad... its a p3 coppermine 850mhz. it also states it is a micro tower pc, so that explains the shorter than normal psu casing. the only killer is the intel chipset which supports 512mb of ram max which is too little for today's web browsing as some of today's bloated flash sites can use over 1gb of ram.

this PSU is barely “broken in”, with only slightly more than 25000 power-on hours (according to the HDD that came with the PC).

25k hrs is like more than 2 years of 24/7 use, almost 3 years. i'd call that more like a moderately used pc.

-Add an ATX 12V 4 pin connector

another (easier) way is to just get a 2x 4-pin molex to 4-pin atx12v converter to convert 2 molex plugs into an atx12v connector. u can then use it on a p4 williamette or northwood pc (after changing the 12v rectifier, of course) but not p4 prescott or any p4 faster than 3ghz as the transformer may not be able to handle such high power draw.

that psu is also great for running a radeon 9700/9800 card due to the card being heavy on the 3.3v and 5v rail.

[goodpsusearch]

Re: Power Win PW-330ATXE

another (easier) way is to just get a 2x 4-pin molex to 4-pin atx12v converter to convert 2 molex plugs into an atx12v connector.

Those adapters are sometimes unreliable. Solder a connector from a dead psu to be sure

u can then use it on a p4 williamette or northwood pc (after changing the 12v rectifier, of course) but not p4 prescott or any p4 faster than 3ghz as the transformer may not be able to handle such high power draw.

I agree with that. It may also be good for AMD 64 / Sempron etc

[momaka]

Re: Power supply build quality pictorial. part 2

Nice PSU! Being 16 years old, it's still in good shape because of the caps and nice fan Do you use it at all?

Thanks.
Yes, I do. It's the PSU from this PC:
https://www.badcaps.net/forum/showpo...&postcount=541
(You can even see in that thread that I promised to post pictures when I retire it... but it wasn't until December 2015 that I finally replaced it with another PC, due to being at the bottom of a pile of desktops next to my desk. And now it looks like I won't be retiring it and putting it back in service as my backup PC, so this Lite-ON PSU will continue to serve me)

The label states it can run at 50C ambient. Should be fine if you don't live in death valley with your computer outside

That statement is standard on all Lite-ON PSUs from that era and possibly even modern stuff too. I can't say how valid that is, but there is a chance this thing will run in 50C Ta environment.

That said, I am really considering moving down to Florida or Georgia (or some other hot and humid southern state) if I can't find work here. I think that will be as close as I can get to "death valley".

What's the wire on the bottom for?

I think just making a ground trace thicker.

It won't need a recap but it could use some glue removal!

Yes, it could. I actually did remove a good chunk of the conductive glue from the primary side and some parts of the secondary side. Unfortunately, there is too much glue between the caps on the secondary, so I didn't remove that.

The conductive glue and tiny G-Luxons and Jamicons are sort of... well.... show stoppers.

Okay, now you are just being picky.

Hmm, 8x200mm... perhaps you mean 8x20mm?

Replace the UG10DCT with a 20A rectifier.

I would try to replace the Rubycon YXG 16V, 1200 uF with 2200uF (should work) or even 3300uF capacitor (might give oscillations).

-Add an ATX 12V 4 pin connector

Nah, this PSU is just NOT made for a 12V PC. Granted the Pentium 3 it is currently powering does not give the 5V rail a very heavy load (like 30 Watts max.), the 12V rail never goes above 11.85V. With the CPU not loaded (5 Watts in Stop Grant mode), the 12V rail sags even lower - down to 11.7x Volts, IIRC. And that's with only a single HDD and 80 mm fan in the PC (which I run on a much lower voltage too, so it doesn't consume more than 1 W).

A schottky rectifier might help to bring that up to 12V. But I imagine that as soon as the 12V rail is loaded by any 12V-based PC, it will dip way too low. The main transformer in old and truly 5V-heavy PSUs like this just has a few less turns, so the 12V rail is always low to keep the 5V rail high.

is this link on cnet the right specs of the pc u got?

Yes, that's it.

the only killer is the intel chipset which supports 512mb of ram max

Actually, it supports 512 MB of RAM only with some tweaks. Namely, I have to set the PC APIC mode in CMOS instead of ACPI. And that is annoying, because I loose stand-by function and have to turn OFF the PC manually too (remember that "It is now safe to turn off your computer" message?)

If I don't do that, I get a BSOD that tells me specifically to turn APIC mode on. I forgot what the exact message is. So I'm actually limited now to 384 MB.

The real let down is the onboard video. And while it is light years ahead of the ATI Rage CDD (Crude Display Device ) chips, those ATI Rage video cards are light years behind everyone, so I can't say the Intel onboard video is good. Even calling it decent is too generous. But it can still do XGA (1024x768) at 85 HZ and 24-bit (yes, 24... not 32-bit), and that's good enough for me.

[ChaosLegionnaire]

Re: Power supply build quality pictorial. part 2

have to turn OFF the PC manually too (remember that "It is now safe to turn off your computer" message?)

oi, yea. long time no see that message i can tell u lol! i had that msg on my first rig when i forked up the bios settings cuz i was a big time nub lol! and i had to reinstall the os... again! after just installing it in order to get rid of the msg.

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

I think I'm the first with this one:

[Compgeke]

Re: Power supply build quality pictorial. part 2

Here's a bit of an unusual one, "PS08/RAD" out of a RAD RICI-T1 box. Little 5V 1.6A and 3.3V 2.4A unit, when running AC. Also can run off 40-120VDC.

Seems pretty well built, plenty of tracking slots and relatively decent components. A mix of UCC KMG, Rubycon MBZ and "Suncon" caps plus ICs that actually have markings. Was also able to find the TUV certification pretty easily.

[Compgeke]

Re: Power supply build quality pictorial. part 2

Next up is a Codegen "350W" model 250X1. Fan was seized and fuse is blown up, typical of Codegen.

I didn't grab a pic of the label, but claims:
3.3V 14A, 5V 30A, 12V 12A, 5VSD 2A, -5V 0.5A, -12V 0.8A. Ignoring the negative rails it does add up to 350W, but I'd not trust it when 250X1 brings up a bunch of 250Ws on google.

[whansard]

Re: Power supply build quality pictorial. part 2

why the mismatched diode sizes in this pic? are the 2 smaller ones considered just for blocking reverse voltages? does this actually do what it is supposed to? the 2 bigger ones are 3amp each.

[momaka]

Re: Power supply build quality pictorial. part 2

I think I'm the first with this one:

Oh, that looks like a ton of fun to troubleshoot, should something go bad.

Here's a bit of an unusual one, "PS08/RAD" out of a RAD RICI-T1 box.

Looks pretty good. Suncon owns Sanyo, so those caps you have are actually Sanyo. You can even see they still have the same "K" vent.

I'm not surprised to see Rubycon MBZ - after all, this PSU uses flyback topology, so ultra-low ESR motherboard-grade caps like Rubycon MBZ do well.

Next up is a Codegen "350W" model 250X1. Fan was seized and fuse is blown up, typical of Codegen.
...but I'd not trust it when 250X1 brings up a bunch of 250Ws on google.

That actually looks decent enough for a 250-300 Watts continuous if it had bigger/better output caps. Heatsinks are nice and tall with good silicone attached to it. Even the X2- and Y2- class caps are safety-approved. Definitely one of the better Codegen PSUs out there. I'd actually fix that up for general use.

Perhaps someone blew the 12V rectifier when they connected it to a 12V-heavy PC, which in turn caused the PSU to blow?

why the mismatched diode sizes in this pic? are the 2 smaller ones considered just for blocking reverse voltages? does this actually do what it is supposed to? the 2 bigger ones are 3amp each.

It's a cost-cutting technique by low-end PSU manufacturers. Well, supposedly anyways. I don't see how it makes more sense to buy parts of different sizes than buy more parts of the same size and get discount on larger quantity purchased... or better yet a real bridge rectifier. Buy I digress.

Anyways, the two larger diodes are connected to Live, whereas the two smaller ones are connected to Neutral. In countries with 120 V AC, only the two bigger diodes are used to rectify the Live rail while the two smaller diodes are bypassed by the 115/230 switch and the Neutral connects to the common pad for the two caps on the primary side. This creates a voltage-doubler circuit, so the rectified 170 V DC across each cap produces about 340 V DC total. In countries with 230 V AC, the 115/230 switch does no bypass the two smaller diodes, all of the diodes are used for rectifying the line AC voltage.

Because of this, a PSU at 115 V AC uses about twice the current than at 230 V AC. But since the smaller diodes are not used when the PSU is set at 115 V AC, this doesn't impact the power handling of the PSU regardless if it is used in 115 V or 230 V countries.

Anyways, I've seen this only on PowMax/Leadman and Sun Pro PSUs. So I am curious what PSU you got there.

[Behemot]

Re: Power supply build quality pictorial. part 2

why the mismatched diode sizes in this pic? are the 2 smaller ones considered just for blocking reverse voltages? does this actually do what it is supposed to? the 2 bigger ones are 3amp each.

It's for voltage doubler where two diodes conduct twice the current while the other two conduct nothing. For ~230 V those smaller ones should be good enough.

[pdavid]

Re: Power supply build quality pictorial; AT Enermax E301P-V

I got two really nice AT Enermax units (plus several others that I need to unbox). On the label they are the same model, but totally different inside.

Enermax AT E301P-V no. 1 :

DOM: 1997

Half bridge topology-
Some EMI filtering, one MOV connected to AC Live / DC +
(Missing differential choke and ceramic y capacitors).
PBL 605 bridge rectifier.
2x 680uf 200V Panasonic primary capacitors.
2x 2SC3320 main switchers.
Size 35 main transformer with center tap.
KA7500 PWM for primary control.
LM339 dual comparator for secondary supervison and housekeeping.
Onboard thermostatic fan controller.
One D83-004 dual diode for 5V and one for 12V rectification.
CTC secondary capacitors.
Plenty of peripheral connectors, wires are 18AWG thick.

Enermax AT E301P-V no. 2:

DOM: 1996

Single transistor forward topology-
Decent EMI filtering, TVS diodes connected between each primary capacitors positive and negative trace.
BU605 bridge rectifier.
2x 680uf 200V Panasonic primary capacitors.
2SK2611 FET controlled by KA3842.
Size 35 main transformer.
LM339 dual comparator for secondary supervison and housekeeping.
Onboard thermostatic fan controller.
S30D40C dual diode for 5V and STPR1020 dual diode for 12V rectification.
KA7905 for -5V rail (No heatsink).
CTC and Rubycon secondary capacitors.
PCB has unpopulated area for possible 3.3V and standby rail.

If the other AT psus are any decent I'll post them too.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Very interesting power supplies. I would keep them.

[Pentium4]

Re: Power supply build quality pictorial. part 2

Wow! Those are some cool Enermax units. They look pretty beefed up for AT units. I wonder if the caps in them are still good or if they have dried up. Looks like they used to make their own PCBs? That's pretty cool. 15A switchers in a 300W unit, very overbuilt!

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

And remember that Enermax is a psu Maker, not reseller like Antec..

[Behemot]

Re: Power supply build quality pictorial. part 2

No.