Cheap PSU: Piece of junk or semi-decent?

[TELVM]

Greetings from Spain.

Got this cheap (€16 , about $21) PSU, opened it expecting a hopeless gutless horror show, but after having a look I'm now in doubt, maybe it's not so terribly bad after all. I would like to know the experts' opinion (I'm just an amateur with pretty limited knowledge).

The thing is called 'TOOQ TQEP-480S', and pretends to be 480W no less.

--> Larger pic --> Larger pic

Transient filtering stage: --> Larger pic --> Larger pic


Bridge recitifier: --> Larger pic --> Larger pic


Primary switchers, left to right 2N60 / MJE130007 / MJE130007 : --> Larger pic


Transformers: --> Larger pic


Secondary rectifiers, left to right 2045CTP / 20150CTP / 2045CTP : --> Larger pic


Other pics: --> Larger pic --> Larger pic


What do you think?

[Pentium4]

Re: Cheap PSU: Piece of junk or semi-decent?

Not terrible but not good, it won't do 480W that's for sure but for that price that's to be expected. I would definitely recap it with higher capacitance caps. That thing could probably do 300W continuous, the heatsinks look decent. But I'd be curious to know the specs on the switching transistors and output rectifiers.

[mariushm]

Re: Cheap PSU: Piece of junk or semi-decent?

Looks good enough for a 250-300w psu.

The GBU406 is a 4A part, so theoretically you can do 220v x 4 = 800 watts.

The MJE13007 can do 350w on a good day, with good cooling.

The 2045CTP and 20150CTP are all 20A parts... so you're looking at fake values on label ... 3.3v x 20 = 60 watts 5v x 20 = 100 w , 12v x 20 = 240 watts

You can even see on the PCB on the 4th picture that the fuse is listed for two models, one for 230w max, one for 300w max.

[goodpsusearch]

Re: Cheap PSU: Piece of junk or semi-decent?

The input filter needs an extra coil.

It is a 250watt unit that might do 300W for some seconds.

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

I hate this idea of "more coils = better". A manufacturer wouldn't go half-arsed on it. There's no point. You either pass CE or you don't. They passed CE with one coil, it will be fine.

[goodpsusearch]

Re: Cheap PSU: Piece of junk or semi-decent?

I hate this idea of "more coils = better". A manufacturer wouldn't go half-arsed on it. There's no point. You either pass CE or you don't. They passed CE with one coil, it will be fine.

Obviously you are unaware of the minimum required components for input filtering circuit of smps.


http://www.hardwaresecrets.com/artic...Supplies/327/4

http://www.techpowerup.com/articles/...king/psu/160/3

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

Obviously you are unaware of the minimum required components for input filtering circuit of smps.


http://www.hardwaresecrets.com/artic...Supplies/327/4

I am perfectly aware of that but it is written from a theoretical perspective. You don't necessarily need two coils to achieve the required filtering.

My point is: filtering components cost money. To sell something in a store, you need to pass CE EMC requirements, that is, under normal conditions (which will include a normal load profile, such as a 300W computer) the electromagnetic emissions must be below a certain threshold. They will find the minimum number of components to achieve the thresholds required.

If they do it with one coil they get the CE badge and proudly display it on their product then they get to sell it.

If not, they go back and redesign it, perhaps with an additional coil, or a different coil.

If they are unscrupulous they will make it with no coils at all, because that costs the least amount of money. They of course will not pass CE, and they will not technically be allowed to sell it (of course, they still manage it, but you're not going to find one in an average retail store.)

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

Thanks for sharing your knowledge gentlemen.

The MJE13007 can do 350w on a good day, with good cooling.

Excuse the noob question, how do you deduct that wattage from the MJE13007 specs?

You can even see on the PCB on the 4th picture that the fuse is listed for two models, one for 230w max, one for 300w max.

And the label says 'FUSE RATING: F6.3A', so 300W tops (if).

The input filter needs an extra coil.

I was thinking the same. Maybe placing it in that empty lower left corner with the wire? I could cannibalize ferrite coils from some old PSUs.


There is any MOV in this PSU? Could this little blue thing in the left be a MOV? Or is it an Y-cap?:


Any guesstimate for the % efficiency of this PSU?

[goodpsusearch]

Re: Cheap PSU: Piece of junk or semi-decent?

It's a Y cap. No MOVs here, except if they are placed behind the big caps.

I don't know if you can add a coil when there is no empty slot on pcb. On the contrary, adding extra X cap or Y caps (when needed) is much easier.

[everell]

Re: Cheap PSU: Piece of junk or semi-decent?

I hate this idea of "more coils = better". A manufacturer wouldn't go half-arsed on it. There's no point. You either pass CE or you don't. They passed CE with one coil, it will be fine.

What do you mean "passed CE"? CE certification in Europe may mean something, but CE on a Chinese power supply means "Chinese Export".................

[mariushm]

Re: Cheap PSU: Piece of junk or semi-decent?

Excuse the noob question, how do you deduct that wattage from the MJE13007 specs?


And the label says 'FUSE RATING: F6.3A', so 300W tops (if).

Any guesstimate for the % efficiency of this PSU?

The 13007 transistors are very old and common (seen datasheets showing them as early as February 1999), found in lots of power supplies with similar designs to yours:

http://www.digikey.com/product-searc...376376?k=13007

In lots of reviews I saw, they tend to pop when going and staying over 320-350 watts for longer periods of time.

It varies on how good the heatsinks, the main transformer and other factors...

The fuse will always be a few Amps above the maximum load the psu can handle, to factor the high pulses of current when the psu is actually powered and the main capacitors charge. There's a burst of current for a second or so when you plug it in. During normal operation the current decreases.

As for efficiency, I'd say about 65-68% below 50-80 watts, about 70-75% up to about 200 watts and then going back down to about 70% above that.

[TELVM]

Re: Cheap PSU: Piece of junk or semi-decent?

... they tend to pop when going and staying over 320-350 watts for longer periods of time.

So no deduction but learned-the-hard-way empirical evidence, thanks.

As for efficiency, I'd say about 65-68% below 50-80 watts, about 70-75% up to about 200 watts and then going back down to about 70% above that.

Ouch, that low? Even this piece of junk / with discrete diodes / and non-Schottky +12V rectifier / manages 75% @ 90W.

This PSU is for a retro-PC I'm restoring which will never draw 100W, so I don't care at all if the PSU can't handle more than 250W, but I do care a bit on efficiency at low power.

... CE on a Chinese power supply means "Chinese Export"

[mariushm]

Re: Cheap PSU: Piece of junk or semi-decent?

I may be a bit pessimistic, but basically on idle, your computer will probably use less than 50-60 watts, if you say it's old. So that 65-70% could be very valid.

Well, if you look on page 7, you have the benchmarks there. But if you check each one, you can see that the accent is put on the 12v rail, gradually upping that load while raising the others much slower.

This is a valid procedure for modern computers, but might not always be the case.

With real world computers, it depends on what computer you have, how old it is, if the motherboard uses that 4 pin cpu power connector and other things.

An old system, let's say Pentium 3, or some of the first Pentium IV systems, would be heavy on 5v, deriving almost anything else from it... the 12v would only be used by hard drive motors.

These power supplies, like the one in review and this one, were not designed for 12v as main voltage like current computers are made, they had much less power on 12v but were tweaked for more modern systems by putting better transistors and diodes on 12v - you're sort of using "overclocked" 12v capabilities.

So when you pull what may seem a normal load on 12v, like let's say 10-15 A and the label says 22 but in reality the psu can barely do 18-20A, on a power supply that was designed with a transformer, coils and so on meant for even less amperage on 12v, then you're at the peak of efficiency for the 12v section. When you average the three, the high efficiency of the 12v counters the low efficiency it would have on 3.3v and 5v and so the overall psu's efficiency raises to 70-75%.

You can see on test 8 of that power supply how it starts to fail at 18A load on 12v and the efficiency already started to decrease - the peak was at 12 A on 12v and a bit of load on 3.3v and 5v

That power supply would probably get to about 80% efficiency with just 5v loaded at 20A but it was never tested.

[c_hegge]

Re: Cheap PSU: Piece of junk or semi-decent?

To sell something in a store, you need to pass CE EMC requirements...

No, you don't. Otherwise, we wouldn't see any PSUs with no input filtering at all.

[mariushm]

Re: Cheap PSU: Piece of junk or semi-decent?

c_hegge is in Australia so probably doesn't apply to him, but it applies to tom in UK.

Legally, to sell something in European Union, they need to have the CE mark, which theoretically is given only after testing the device.

However, it's not the job of the customs officers to check the CE documentation or test that the products are according to the way the CE mark was given when they come into the country.

The officers often just open one container and pick a few random boxes and plug them in the socket and if they work, then they let them go through. They don't have the expertise or qualifications to do it.

So companies often build a power supply and then go at a laboratory in China that does the tests and if they don't pass they may just pay them out and get the CE mark anyway, then they do lots of versions with components taken out and replaced with jumper links to anyone willing to buy.

I remember seeing a french documentary about it with some french dudes going to China after all that fake toothpaste and milk with melamine and they bought some Christmas tree lights that were with the CE mark but obviously couldn't pass any tests (wires so thin you could break the cord in half with your hands)

It was also showing the French customs officers simply getting a microwave oven and plugging it in - it works so the shipment was released - they didn't have the authority and it's so many imports they couldn't hold the containers for weeks to make tests.

My Youtube search-fu is just not that great at this hour, can't find the documentary or "reportage" or whatever it's called.

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

No, you don't. Otherwise, we wouldn't see any PSUs with no input filtering at all.

Well, my point here is you'd only put filtering in if you -wanted- it to pass CE. If you didn't care about CE, you'd put -no- filtering in.

Obviously, the only way to measure this is with a spectrum analyser with IEC/CE limits set up (I'm not sure if the Rigol DSA815 supports this) and a defined antenna configuration.

[Behemot]

Re: Cheap PSU: Piece of junk or semi-decent?

The GBU406 is a 4A part, so theoretically you can do 220v x 4 = 800 watts.

The thing is, you have to not only count with wasted power, but also with reactive load, which goes all through the rectifier as well as active load. So with no PFC as it is here you have, say 50° PF, that gives you 400 W of usable power, and with 70 % efficiency you can get 280 W from it. So 300 W with some luck, but depending on power factor, input rectifier can go . And if not rectifier, than those 13007's. Or the output rectifiers.

And if you change all these components for higher-rated, than after 350 W, the higher you go, the closer you are looking towards transformer saturation

[Wester547]

Re: Cheap PSU: Piece of junk or semi-decent?

It's more likely the switching transistors would blow first. I think good cooling helps rectifiers lots, which is why you see some PSUs get away with using 4 diodes in place of a rectifying bridge. With decent cooling, a 4A rectifier is plenty since most people don't pull more than 200W (2A). As stated, as far as the PSU goes, it could use slightly better input filtering and better output rectifiers, and probably better capacitors too.

[tom66]

Re: Cheap PSU: Piece of junk or semi-decent?

The thing is, you have to not only count with wasted power, but also with reactive load, which goes all through the rectifier as well as active load. So with no PFC as it is here you have, say 50° PF, that gives you 400 W of usable power, and with 70 % efficiency you can get 280 W from it. So 300 W with some luck, but depending on power factor, input rectifier can go . And if not rectifier, than those 13007's. Or the output rectifiers.

And if you change all these components for higher-rated, than after 350 W, the higher you go, the closer you are looking towards transformer saturation

Another reason why manufacturers like using PFC for high power.

Many were using it before it was compulsory, it simply made things cheaper (even if the real power was higher.)

[Per Hansson]

Re: Cheap PSU: Piece of junk or semi-decent?

To sell something in a store, you need to pass CE EMC requirements, that is, under normal conditions (which will include a normal load profile, such as a 300W computer) the electromagnetic emissions must be below a certain threshold. They will find the minimum number of components to achieve the thresholds required.

You seem to think CE marking stands for something, it's infact just a load a bullshit, from the Wikipedia article on CE:

Self-certification
Depending on the level of risk of the product, the CE marking is affixed to a product by the manufacturer or authorized representative who decides whether the product meets all the CE marking requirements. If a product has minimal risk, it can be self-certified where manufacturers a Declaration of Conformity and affixes the CE marking to their own product.

So it basically means that this very honorable Chinese manufacturer insures that his product meets CE requirements!
No external testing is involved.
If it where do you think the crap seen for sale on Newegg would be possible?
Load any of those gutless wonders up to 50% of the rated label and they go "poof"
I'm sure that's part of CE conformity testing!
http://www.overclock.net/t/1310456/p...00w-psu-review
http://www.jonnyguru.com/forums/showthread.php?t=6714

Like checking if PFC transformers are made out of copper or cement?! (Note the CE approval marks!)
http://www.pcpop.com/doc/0/294/294565_all.shtml