Q-Tec 550W Dual Fan Gold

[JEWilson]

Ok - enuff said - I have been advised these are crap

Th PSU was known good and I was using to test a number of dead mobos

I was using this to test, what appears to be a Trigem TG Crush 18D M
This mobo is based on a nForce2 chipset.

First, when the PSU was simply plugged into wall outlet - mobo s/bridge
was heating up - the PSU did not have any fans spinning.

As the mobo failed to start, I jumpered PS-ON to ground at the ATX
(2x10) connector, when power applied at the wall outlet, mobo
came to life, spinned up the PSU fans as well as the CPU fan.

I used a freezer spray to keep the temp on s/bridge down to see
if I could get the system to POST.

Nah no POST - shooting flame out of the lower PSU inlet fan though,
pop and the PSU shut down.

Pulled the cover of the PSU and find;
1. The fuse (10A 250V quick blow - I live in the UK) melted
Subsequently replaced
2. Q4 in the inverter has its face blown off!
The TO92 package details C5344Y.
This I determined once I fished that part of the device out of the PSU.

The part C5344 can be found at
https://www.auk.co.kr/kor/01_product...sp?pno=2SC5344

I suspect a s/c on the mobo created the problem, probably a cap or caps
or worse. I need an ESR meter to test them, probably at the 12V input.

I believe this part is used as a driver in the inverter. There are no other
areas in the PSU which detail thermal damage or immediate damage to
devices. Anything else I should check?

In fact, I have three of these PSU - same spec. and was intending to
use these as replacments in three Advent PCs - better, in part, than the
Bestec 250W crappers that caused the dead mobos in the first instance.

Also, it was my intention to recap the 10V 2200uF cans these PSU have
in the O/P filter. In there already, are Unite (never heard of these) and Koshin.
The primary has 2 Samxon LP 680uF 200V 85 deg C

Any help, insight and advice greatly appreciated

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

To bring you up to date on this item

The PSU is dead and when I apply mains power, remains dead, no +5VSB

Further to diagnostics, I have established what I believe to
be the +5VSB switching transistor has gone to meet its maker!
This may be consistent with the original fault, that is driving
the PSU into a s/c on the 3.3V rail of the mobo in question
so far as I forced PS-ON by jumper pins 14 and 15.

I took the primary heatsink out of cct to test the switching tsistors
These appear to be;
Fuji Electric 2SC3886
see http://www.datasheetcatalog.com/data.../2SC3866.shtml
Sanken 2SC5071
see http://www.datasheetcatalog.com/data.../2SC5071.shtml

I got all open with my digital mmeter on the 2SC3886
BE red lead to blk lead - infinte
BE blk lead to red lead - infinte
BC red lead to blk lead - infinte
BC blk lead to red lead - infinte
CE red lead to blk lead - infinte
CE blk lead to red lead - infinte

As well as the small signal tsistor referred to in the previous post I found
a 100R resistor o/c which connects to an opto coupler pin 3, a Sharp PC817
see http://www.datasheetcatalog.com/data.../1/PC817.shtml

Q - any suggestions as to what else I should be looking for in terms of
failures?

[linuxguru]

Re: Q-Tec 550W Dual Fan Gold

If they're authentic, those Fuji and Sanken are fairly good transistors. However, lots of fakes abound, and of course, bad caps.

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

Thanks for the info

Of course, looking for replacements might be onerous

For instance, the C5344Y has an odd pin out 1, 2 and 3 are E, C and B.
I used a BC33740 to replace it, these are 1, 2 and 3 as C, B and E

So I had to put it in the other way around (turned 180 deg) and
cross the B and C to correctly fit it - sigh...

The Sankens are ok - these test good

The nearest I appear to be able to get for the Fuji are 2SC3150 and 2SC2979
see http://www.datasheetcatalog.com/data.../2SC3150.shtml
and
http://www.datasheetcatalog.com/data.../2SC2979.shtml
respectively

My supplier charges almost 200% per cent more for the 2SC2979 - a 40W device.
The 2SC3150 is a 50W device. Both are Toshibas.
For all intents and purposes, they appear (at least to me) to have the same characteristics.

Would there be a problem using the higher O/P part?
Of course, the amps, or at least capability to deliver a higher O/P on the
+5VSB rail would be possilbe but the device should be biased correctly
in any event as matters stand (with the PSU).

I did check the remainder of the discretes in the primary and they all
check out ok but this may change when powered up, of course.

Watcher' reckon on the Toshies?

[starfury1]

Re: Q-Tec 550W Dual Fan Gold

looking at these pdf's and thats if correct





I think maybe not but suppose it depends on the actual circuit
I suppose if it was heavy duty device a lower duty device
probably wouldn't cut it and I guess
if they used a big spec transistor a big spec transistor is probably needed.
(other words they wounldn't spend more then they had to I suppose)

2SC3886. N. Channel. 1400/600v, 8A, 50w

2SC3150 3A,800V,50W
(although peak is 10A)

Anyway I'll leave that to those that know

I did find this though JEWilson

Its a SGS thom cross ref Pdf might be of use to you

https://cdn.badcaps-static.com/pdfs/...508677f865.pdf

At cricklewood elec
(I know its not local to you but your region I suppose)

2SC3886A NPN F218 High Resolution 1500V 8A 50W

http://www.cricklewoodelectronics.co...at=178&page=16

I dont suppose thats cheap and then shipping

not much help to you mate but hope the cross reference is at some point

Cheers

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

I appear to have been able to source 2SC3866 from an eBay vendor.

Of course, as stories abound with regard to fake devices, how
can I establish whether (not) I have a legitimate device?

Also, on this repair, I have two 50V 10uF NKNC in the primary,
can size 10mm length, 5mm diameter.

They report ESR of 2.0 and 2.1 respectively? Are these bad.
I have a K-7204 and the worst case ESR values are dated
as on the silkscreen of the meter.

Can anybody offer any insight here as to the ESR?

[gonzo0815]

Re: Q-Tec 550W Dual Fan Gold

2.0 and 2.1 should be a good ESR value for a cap with this capacitance.
Even 2.9 ohms would be in the ok range.
At this position IMHO the capacitance is equally important as the ESR value.

The usual 0.0xx ESR values are probably not showing up in caps with a capacitance >470uF
If unsure i would suggest to checkout the data sheet of any of the good brand capacitors with similar characteristics and compare it with the same voltage and value part.

[starfury1]

Re: Q-Tec 550W Dual Fan Gold

10uF small caps will read higher, if its in spec or not is another story.
(is the meter reading accurate now? did you get new probes just wondering?)

Well not sure if you will open a can of *worms by doing it but I would just replace them if any doubt exists
(* sometimes strange things happen with circuits)

Of course, as stories abound with regard to fake devices, how
can I establish whether (not) I have a legitimate device?

Now that is a very good question I have no idea really so maybe linuxguru might shed some light since he mentioned it (and a good point too)
(or anybody else for that matter)

I guess possibly logo printing might give it away or perhaps something about the package.

I'd hang on to the old ones for comparison visually

Also if in doubt post a photo of them when you get them someone would probably be able to tell.

I suspect by "fakes" its probably a lower spec device relabeled as a higher spec Device
So maybe if the price is too good to be true it probably is
(this is not to say the seller maybe selling them in good faith)
(but know nothing really of what his selling)

Anyone add to this or correct it

Cheers

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

I will be getting a new set of probes in due course and
a set of clips to go with them

Thanks

[linuxguru]

Re: Q-Tec 550W Dual Fan Gold

Quick tips to identify a fake transistor:

1) Finish - look at the plastic moulding, tabs, etc. Authentic Japanese transistors have clean moulding with no plastic chads, etc. The metal edges/faces are straight/flat with no burrs, scratches, etc. The legend/text should look clean and similar to the published pictures. The legend should not be easily erasable with common industrial solvents like IPA, acetone, etc.

2) The pin assignments match the published specs., and the polarity (NPN/PNP) matches the published specs. (don't laugh - some fake transistors turned out to be NPN while the original with the same number was a PNP).

3) For genuine devices, the DC beta is in the specified range - for high-voltage power devices, that's usually in the range of 10-100, often near the lower end of the range. If you get a suspiciously high DC beta that's higher than the range, sometimes >100 or more, look and test carefully before buying at large lot. Fake power transistors often have a die from a lower power device with a higher beta.

4) If you want to buy a large lot, sacrifice and dissect a few from the lot before buying the large lot, if you can. The die size should be relatively large for a power device. A tiny die from a small-signal device usually indicates a fake.

[starfury1]

Re: Q-Tec 550W Dual Fan Gold

Thanks linuxguru
that puts it in a lot clearer light then I did
With regard to fakes

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

Ok - got a new set of probes for the ESR meter

For this PSU got a 2SC3866, mounted on the Primary heatsink
and remounted to PSU pcb

Replaced the failed 100R carbon 5% with a 100R metal film 1%
as this was all I could find

Applied power to PSU

The ATX specification details
+12.0V : +/- 5 % readings between 11.40 to 12.6
+ 5.0V : +/- 5 % readings between 4.75 to 5.25
+ 3.3V : +/- 5 % readings between 3.135 to 3.465
- 5.0V : +/- 10 % readings between 4.5 to 5.5
-12.0V : +/- 10 % readings between 10.8 to 13.2
+ 5.0V SB : +/- 5 % readings between 4.75 to 5.25

Powered up
Tested with PSU Tester2 (the one with LCD readout) for go/no-go.
Found +5VSB voltage fluctuates and lose readout above upper regulation threshold of 5.25

Tester 2 readouts;
+12.0V1 : 11.9 V ATX 20 pin ok
+12.0V2 : 11.9 V ATX 4 pin ok
+ 5.0V : 5.2 V ok
+ 3.3V : 3.4 V ok
- 5.0V : Unknown
-12.0V : 11.7 V ok
+ 5.0V SB : 5.2 to 5.8/5.9 V - sometimes blanks and returns to 5.5V (problem(s))
PG : 320 mS

The PSU has come back to life a bit, fans spin up.

As I believe I have fixed, possibly, the Primary - would the opto isolator
have got zapped also? Would this indicate the problem with +5VSB?

Again, any advice, insight and help greatly appreciated

[JEWilson]

Re: Q-Tec 550W Dual Fan Gold

Was the opto isolator - so it seems.

The in situ part was as Sharp PC817. Replaced with a 817C equivalent

Problem appeared to be - Sharp on test (out of cct);
pins 4 emitter and 3 collector on the primary side
CE RedBlk mmeter probes reads Infinte Z
CE BlkRed mmeter probes reads 0.835
CE should always be infinte Z both ways

Go to load test this PSU at some point. Will know the next time won't I!

Must not force ATX PS-ON into a s/c load - Lesson duly learned.

[Brian C]

Re: Q-Tec 550W Dual Fan Gold

Thanks for the info

Of course, looking for replacements might be onerous

For instance, the C5344Y has an odd pin out 1, 2 and 3 are E, C and B.
I used a BC33740 to replace it, these are 1, 2 and 3 as C, B and E

So I had to put it in the other way around (turned 180 deg) and
cross the B and C to correctly fit it - sigh...

2SA, 2SB, 2SC, 2SD are transistor naming according to (JEITA). All Japanese transistor follow JEITA method.
2SA and 2SB are PNP, 2SC and 2SD are NPN. The pin out is ( regardless of PNP or NPN ):
for TO92 package: [1][2][3] --- [E][C][B]
for TO220, TO-3P (big size): [1][2][3] --- [B][C][E]
Center lead is always the collector pin.