Power supply build quality pictorial. part 2

[Pentium4]

Re: Power supply build quality pictorial. part 2

Don't know whether this one is worth recapping or not. CWT....I really don't like how crammed the secondary is...It looks like all that open space on the primary is space to put Passive PFC. It doesn't look like a 500W PSU but it could probably do 400W. Think it's worth a recap? the Primaries are 820uF 200V Jun Fu and all of the rest of the caps in it are Fulltec

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

It is similar to Thermaltake TR2-500:
https://www.badcaps.net/forum/showthread.php?t=19981

The model name of the Thermaltake is PSF450S-30.

It's okay. It wont do 500W but it can do 475W without a problem:
http://www.xbitlabs.com/articles/cas...oundup_23.html
page 23 & 24

[Pentium4]

Re: Power supply build quality pictorial. part 2

It is similar to Thermaltake TR2-500:
https://www.badcaps.net/forum/showthread.php?t=19981

The model name of the Thermaltake is PSF450S-30.

It's okay. It wont do 500W but it can do 475W without a problem:
http://www.xbitlabs.com/articles/cas...oundup_23.html
page 23 & 24

Awesome, thanks for the link! I think it's worth recapping then since I only have to pay for the caps to go into it.

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

I decided not to recap mine for 2 reasons. First, it would be a pain in the ass to get the caps under that heatsink and the second reason is that every cap in that thing is Teapo, not bad at all. I am waiting for new primary capacitors and I just had to replace the 5vsb critical cap, because it was an 85°C........

[Pentium4]

Re: Power supply build quality pictorial. part 2

I decided not to recap mine for 2 reasons. First, it would be a pain in the ass to get the caps under that heatsink and the second reason is that every cap in that thing is Teapo, not bad at all. I am waiting for new primary capacitors and I just had to replace the 5vsb critical cap, because it was an 85°C........

Yeah I see what you mean. Should I leave the Fulltec then and wait for them to die? I'm guessing Teapo are better though than Fulltec

[tom66]

Re: Power supply build quality pictorial. part 2

Interestingly, I've been studying the "ringing choke converter" 5Vsb circuit.

It's surprisingly efficient, typically 85%, and wastes only about 50mW in standby.

It is possible to design a 5Vsb self-oscillating circuit which doesn't go out of control when that cap fails. In fact, most of the variants don't seem to go out of control. Only a few do. Seems manufacturers didn't bother to simulate it. Basically, it will only go crazy when that cap is bad AND the load is near zero, because the cap provides power to shut off the oscillation. Under high load the supply is still available, it's under low load that with a bad cap the voltage drops and it can't stop the oscillation any more so the voltage just climbs and climbs... But if the circuit is designed properly, that cap going bad will prevent the circuit oscillating in the first place, or it will slow it down to the point that it will just pulse the output at a few Hz.

To be honest, it really only saves a couple of cents compared to a proper controller IC, so it's not worth it. But RCCs are fun circuits to play with.

Heck, I've seen LCD TVs with self-resonant circuits for the backlight! That's good few 150W for a 32 incher. Crazy!

[goodpsusearch]

Re: Power supply build quality pictorial. part 2

Yeah I see what you mean. Should I leave the Fulltec then and wait for them to die? I'm guessing Teapo are better though than Fulltec

I have no clue about Fulltec. But I did a quick search in the forum and there are incidents of them bulging

[c_hegge]

Re: Power supply build quality pictorial. part 2

I think Fulltec = AxiaX/JunFu/Fuhjyyu. I replace them on sight.

I have the 400W version of this PSU in my brother's PC (the one I reviewed at http://hardwareinsights.com/wp/chann...00s-30-review/). I have since re-capped it and the ripple halved.

[momaka]

Re: Power supply build quality pictorial. part 2

Interestingly, I've been studying the "ringing choke converter" 5Vsb circuit.

It's surprisingly efficient, typically 85%, and wastes only about 50mW in standby.

Keyword here is "typically".
Care to guess how much the one in this thread wastes?
https://www.badcaps.net/forum/showth...758#post272758
Me thinks 5-10W.

Heck, I've seen LCD TVs with self-resonant circuits for the backlight! That's good few 150W for a 32 incher. Crazy!

We have some Ikea 12V track lights in our house. The power supply in one of them burned up. Turns out it was a 105W RCC-based switching PSU. Easy to fix, though.

[Wester547]

Re: Power supply build quality pictorial. part 2

I think Fulltec = AxiaX/JunFu/Fuhjyyu. I replace them on sight.

I hope that doesn't mean that Hermei/Asiacon are essentailly Fuhjyyu.... I know they're bad, but....

momaka, I think you linked the wrong thread. :P

[Behemot]

Re: Power supply build quality pictorial. part 2

I'd say Fulltec is utter crap, replace on sight. As for me, I replace on sight everything besides japanese caps and Samxon, but hey, it is not relevant?

[momaka]

Re: Power supply build quality pictorial. part 2

momaka, I think you linked the wrong thread. :P

Yes indeed. Bad copypasta . Thanks for pointing it out.
Let's try again one more time :
https://www.badcaps.net/forum/showthread.php?t=23019
This time I double checked it and it works .

[Wester547]

Re: Power supply build quality pictorial. part 2

^
Yes. I've pulled 20A from 20A rectifiers on PSUs with forward topology without incident for as much as 15 mins at a time (I did try for more, but the OCP wouldn't allow it).

Sorry if this is beating a dead horse, but out of curiosity, how much more did you ask for when it shut down, if you remember?

[c_hegge]

Re: Power supply build quality pictorial. part 2

30A. My load tester only goes in 10A increments on the 3.3V rail.

[Wester547]

Re: Power supply build quality pictorial. part 2

I have another question that regards a topic of discussion that has been encircling my head of late.

Does putting two rectifiers or diodes in parallel really double the capacity? Or is that just done to just reduce heat output? I ask because theoretically, the load would be less than even, I think, given the possibly inconveniencing variations in voltage drops. Won't those differences cause more current to flow through one rectifier rather than both, thus dispelling the notion that two 30A parts, for an example, in parallel, equal 60A?

[Behemot]

Re: Power supply build quality pictorial. part 2

World isn't ideal, of course, but more or less it is so that it doubles. Another thing is, in reality it seems than using two rectifiers reduces amount of wasted power more than by 50 %. That means that they may both handle more than twice the power of one, so when you take all these things together, yeah, you can usually put double the load on two rectifiers with no problem and with luck even more.

Generally, if you already put in two rectifiers, than most likely it does not matter for you to pay few extra cents for 5A higher rated rectifiers, and than you can sleep with easy mind :-)

[ben7]

Re: Power supply build quality pictorial. part 2

Its generally not a good idea to do that.

Yes, it would double the current capacity.

BUT! Since no two things are alike, one will have a smaller voltage drop than the other, causing the other to handle more current. When it handles more current, it's forward voltage drop decreases, causing it to handle more current! Once the diode gets out of it's max SOA temperature, it will self destruct from within a few minutes, to only a few millis. This domino effect is called 'thermal runaway'.

Using two diodes in parallel is just asking for failures!

-Ben

[Behemot]

Re: Power supply build quality pictorial. part 2

The difference is very small, really, both for the same types (same manufacturer too as well) and between temperatures. For example for SBR2045 the difference is only arround 0,05 ohm between 25 and 125 °C!

[tom66]

Re: Power supply build quality pictorial. part 2

You cannot characterise diodes with resistance.

I repeat, you cannot characterise them with resistance!

The only time it's a good idea to use diodes in parallel is if they are thermally coupled. If you can be certain they will be at nearly the same temperature, it is OK, though it is generally considered bad design practice due to it being very difficult to achieve a nearly equal temperature.

[Behemot]

Re: Power supply build quality pictorial. part 2

Oops, that vas forward drop - bad units;-) Just understand, that the tolerance is usually 0,05 V (typical min to typical max), this is the worst which could be between two different pieces. Also with temperature on 100 °C it is more or less arround 0,05 V. But temperature would be similar on both (shared heatsink).

Anway, I do not agree. Diode has different VA characteristic than resistor, but basically, in every point of the VA diagram, if the situation is static, you can work with that as with resistor of some parameters.