Power supply build quality pictorial. part 2

[ben7]

Re: Power supply build quality pictorial. part 2

Quite dated design and the psu is not really able to output 400Watt. It is a 300W power supply.

Notice that it has a -5V rail. This thing is old!

The 5vsb controlled by viper22a is a great thing though.

Yeah it is old looking.

I figured that it might not be able to do 400W.
At least it has decent heatsinks!

I also noticed on the label that most of the power output is on the 5V rail.

Output caps are only a measly 1000uF each, but there are 6 of them.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

And topower psu design is so and so. I would not waste GOOD caps for it.

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

Seems like a P4 or even P5.
Don't bother with it, Topower is quite bad...

[ben7]

Re: Power supply build quality pictorial. part 2

Seems like a P4 or even P5.
Don't bother with it, Topower is quite bad...

What is all the hate about it for? Because of crappy caps, incorrect labels, ...?

[Stefan Payne]

Re: Power supply build quality pictorial. part 2

What is all the hate about it for? Because of crappy caps, incorrect labels, ...?

no something like overshoot on power on.

[Behemot]

Re: Power supply build quality pictorial. part 2

I think with two fans, it can do 400 W no problem. The 2SC2625 is 10A part and the P30D20C is 30 A/200 V. Just give it decent 680-820uF primary caps and some good Samxons RS 3300/16 and NCC KYA 3300/6,3 on secondary and it would be all good. I would not go for much lower ESR, those 1000uF caps will be quite high IMO and these 3300 will be just OK for the driving circuitry.

Would not use it for too modern PC though because of being +3,3/+5 V heavy, it won't handle 12 V crossload.

[pdavid]

Re: Power supply build quality pictorial. part 2

I've got these two AT psu that I have no idea who the OEM might be. Both have the same casing and label, but it looks nothing special, very generic. Heatsinks are kind of unusual, a bit cheap looking. The beefier unit has a "temperature controlled fan installed" sticker. The main swithcers are 2SC2625. It has two SMBR3045PT in parallel for 5V rail. Also there is an unused "ALT Volt" rail that has no components installed at all. I replaced the main and secondary capacitors in this one.
The main trafos might be giving some clue who might have made these, though I never seen these markings before.
The fan controller with the ne555 looks like someting of a home made DIY circuit. Sorry for the crappy pictures...

[fanbus]

Fake Tacens: The first improvement session.

Hello again.

In my PSU I did:

- Replaced Cheng 470μF / 200V primary caps with OST 680μF / 200V.
- Replaced 13007 transistors with 13009.
- Replaced 5V rectifier SB2045CT with 2x SBL2040CT
- Replaced 3.3V rectifier SB2045CT with SB3045ST
- Replaced 12V rectifier F16C20C with 2x SB2045CT (The old ones from the other channels)

I would like to have:
- 1x 3300μF + 1x 2200μF (before and after the pi coil, respectively) for 5v and 3.3V.
- 2x 2200μF for 12V.
- 1x 1500μF + 1x 1000μF for 5VSB.
- 1x 1000μF for -12V

I will also check the caps powering the IC's. There is one (the lone BH) right next to the 5VSB diode. Not the best place for a cheap cap, unless it wants to tan.

Now the power supply looks better except for the soldering. My crappy soldering iron is so powerful that it sometimes obliterates traces. Even the tip shortens every day!

[c_hegge]

Re: Fake Tacens: The first improvement session.

- Replaced 12V rectifier F16C20C with 2x SB2045CT (The old ones from the other channels)

That might work for a few minutes, but it's not recommended. The maximum Peak Repetitive Reverse Voltage (VRRM) on those is only 45V (as given away by the 45 in SB2045CT). It should be at least 60V for the 12V rail, as most designs generate large reverse voltage spikes on the 12V rail, which the rectifiers have to filter out. Most PSUs use parts with a 100V or 200V maximum VRRM rating.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

What c_hegge said... 12V should use rectifiers with VRRM 60V or higher. I learned it the hard way...

All the other changes you made seem ok to me.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Linkworld LEC-991 1999

Someone tried to fix that psu. Some of the caps have been replaced.

It's a mini ATX psu. It came without a cover, so I don't know its rated power. I guess 200Watt.

Photos and datasheets attached

[Behemot]

Re: Power supply build quality pictorial. part 2

What c_hegge said... 12V should use rectifiers with VRRM 60V or higher. I learned it the hard way...

All the other changes you made seem ok to me.

But as always, depends on design. I have been partially dismantling my huge pile of junk to reduce the space it takes and at least one of the units had 45V rectifiers even on +12 V. It worked for couple years for sure…

[fanbus]

Re: Power supply build quality pictorial. part 2

I agree. I wasn't sure if it was enough, but I didn't have any other matched pair.
I haven't got any matched pair able to handle the peak voltage you say and the only electronics shop in my island hasn't got any. Do you think a clamp will work?

[fanbus]

Re: Power supply build quality pictorial. part 2

OK, finally I found 2 MBR20100CT for the +12V.

I also replaced the capacitors in the outputs.
- 3.3V: 2x 1000μF 10V --> 2x 2700μF 6.3V
- 5V: 2x 1000μF 10V --> 2x 3300μF 10V
- 12V: 2X 1000μF 16V --> 2x 3300μF 16V

I still have to check the 5VSB caps. Do I need to replace the small caps around the IC's?

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

It depends! The 5vsb critical cap is called critical for a reason. If it goes out of spec the 5vsb could go as high as 8-12V and kill the motherboard!

The 1 or 2 caps that are found near the switching transistors/mosfet(s) are also quite critical for proper functioning of power supply and should always be replaced.

The least important tiny cap is the one that is part of the fan controller circuit. I leave that alone unless the brand is complete junk!

Some decoupling caps near the TL494 / KA7500 those you should replace if they are near heatsinks or load resistors and/or if they are Jenpo, JEE and similar JUNK caps. If they are Teapo/OST/Ltec and they are not near heat sources you can leave them alone.

[fanbus]

Re: Power supply build quality pictorial. part 2

It depends! The 5vsb critical cap is called critical for a reason. If it goes out of spec the 5vsb could go as high as 8-12V and kill the motherboard!

I said this PSU didn't have any critical cap in the primary part. I also had a PSU with a failing cap on the 5VSB filtering stage which managed to kill 5 motherboards.

The 1 or 2 caps that are found near the switching transistors/mosfet(s) are also quite critical for proper functioning of power supply and should always be replaced.

They are used to drive the switchings a bit harder, aren't they?

The least important tiny cap is the one that is part of the fan controller circuit. I leave that alone unless the brand is complete junk!

I think it may have one, but I'm not sure if the thermistor I found hidden in a coil is from the fan controller.

Some decoupling caps near the TL494 / KA7500 those you should replace if they are near heatsinks or load resistors and/or if they are Jenpo, JEE and similar JUNK caps. If they are Teapo/OST/Ltec and they are not near heat sources you can leave them alone.

ALL caps there are NICON except for a single BH next to the 5VSB diode. Luckily this power supply hasn't been used much, otherwise that cap would be a bit burnt and swollen.

[Compgeke]

Re: Power supply build quality pictorial. part 2

Wintec WP609F11 300W power supply from a 1U Rackable Systems server.

Starts out quite decent.




But...Teapo! Not the absolute worst you could find but not really the greatest either. At the same time they aren't bloated or blown so maybe they're fine.

[Behemot]

Re: Power supply build quality pictorial. part 2

Is it DC-DC or group design? And isn't there a Hermei too?

[ben7]

Re: Power supply build quality pictorial. part 2

Is it DC-DC or group design? And isn't there a Hermei too?

It looks to me like a group design.

Either way, it is quite nice actually, there is a proper fiberglass (?) PCB, not a cheap paper/phenolic type. Also, plenty of EMI filtering!

That hermei vent could be a teapo, I think I recall someone finding them before.

[momaka]

Re: Fake Tacens: The first improvement session.

It should be at least 60V for the 12V rail, as most designs generate large reverse voltage spikes on the 12V rail, which the rectifiers have to filter out. Most PSUs use parts with a 100V or 200V maximum VRRM rating.

With STF topology, you can use 45V rectifiers on the 12V rail in most cases (at least, on units with less than 350W of power output). But on H-brdige, you definitely don't want to go under 60V. In many cases, the reverse voltage will probably be at least as high as high as 40V.