SCS120PW-24 smps repair and components guessing

[SergiiB]

Hi All.

I'm attempting to fix high-end power supply but puzzled due to inexperience in this field. I would like to obtain some help and directions.

What I did so far:
I reverse engineered the control circuit of power supply as googling it failed:

I desoldered blown transistors (Q7 and Q5)
I checked all other transistors, they all were bad except Q9
I checked in-circuit all capacitors if they are not shorted (non so far)
I checked all diodes in-circuit
I desoldered zener D6 and created simple resistor + zener circuit and it did not hold anything close to 15V (it did not hold anything at all).

Now I have depopulated control circuit that made it easy to measure resistor values and find one that became open circuit - R13.

The problem for me here and the reason I'm writing this are blown beyond recognition transistors (or maybe they are not transistors)? To understand what they are I tried to understand their functions.


For Q7:
I googled zener + transistor circuit and found that Q7 is emitter repeater that repeats zener voltage minus 0.7V, so I guess Q7-D6-R18 form a stabilized power supply for all the control circuitry. Thus, even without any available designation on Q7 I can safely assume that it is NPN transistor and going to use 1P (MMBT2222A) as in many other places on the board.
Was I correct here?

For Q5:
I googled transistor-transistor circuits and it came up with different variations but somehow similar latches. They all use back-to-back complimentary transistors, like Q8-Q9 (1P-2F).
Function of Q8-Q9 latch became clear immediately -- upon signal from secondary side via optocoupler to pull down FB line of 1377B which according to 1377B datasheet disables switching of power MOSFET (probably some kind of overload protection).
Function of Q5-Q6 is not clear to me:
-- Can Q5 be not a transistor but some kind of voltage regulator?
-- If it is indeed a transistor, why didn't they use 1P like everywhere else on the board? Because I can clearly see second symbol is "2", not "P"?
-- If is is indeed a transistor, could it be a latch with a delay: it will be triggered when C12 is charged by pulses from Drv pin of 1377B. Resistor R7 is large, so it will take some time to charge C12. Once C12 is charged, it will open Q5, which will open Q6 and power will be supplied to Power Factor Correction IC thus enabling power correction.
Will I be right to say that Q5-Q6 is a latch that will delay power supply to power correction IC?

Thanx, Sergiy

[coromonadalix]

why ic3 has 2 solder blob ??

[SergiiB]

I started to desolder it but then I realized I hadn't done some photos of the board yet.

It is out now.

[xavier60]

Q5 and Q6 would be for delaying VCC supply to the PFC controller until the flyback  IC has started so that power comes from the transformer via Q7 rather than loading down the flyback ICs internal start up bias supply.
Q5 could be a MOSFET just for the higher turn on threshold.

[SergiiB]

Hi Xavier60

Thanx for explanation about the reason to delay power to PFC.

Now, as they used components mostly from ON Semiconductor, I should make sense that they used MOSFET from them too. So I went through all of their transistors.
I selected SOT-23 package, N-channel and tried to find matching marking with 2 symbols and last "2".

I found one "direct" match: MGSF1N02
Marking diagram is exactly the same as for 1P/2F, with that dot and date code (that can be rotated).

I redraw diagram with new Q5:


Could you please explain what means:

...just for the higher turn on threshold

?

[xavier60]

MOSFETs usually need a higher Gate voltage than  bipolar transistors need on their Base to be turned on.

[SergiiB]

OK, I see, bipolar transistor 1P has V_be = 0.6V, mosfet 2N has V_gs = 1.7V

Now onto replacement. I can easily buy N2 (although they won't sell me just a few). On the other hand on some scrap board I have 2N7002.

It has higher V_ds (60V vs 20V), it has higher V_gs (2.1V vs 1.7V). But does it matter in this particular application? A slightly extended delay?

[xavier60]

OK, I see, bipolar transistor 1P has V_be = 0.6V, mosfet 2N has V_gs = 1.7V

Now onto replacement. I can easily buy N2 (although they won't sell me just a few). On the other hand on some scrap board I have 2N7002.

It has higher V_ds (60V vs 20V), it has higher V_gs (2.1V vs 1.7V). But does it matter in this particular application? A slightly extended delay?

It shouldn't matter at all.

[xavier60]

With this sort of blow up, it is important to find all damaged components in particular the resistors near the MOSFET. Shunt resistor, Gate resistor and diode and CS resistor. R113.
If a damaged Gate resistor isn't replaced, the power supply is likely to fail again after some time due to increased MOSFET switching losses.

[SergiiB]

Yes, R13 is open circuit, that I have already found out before.
But now there is more:

R15 (through hole) shunt resistor on source of Q2 has a burn mark on the side that were facing the heat sink, so it was not visible at all:


R15 is completely open circuit.
And I'm struggling to figure out its designated value by its bands: brown-gray-blue-brown-black = 1-8-6-x10 = 1860 Ohm. Isn't it too big for shunt resistor? If I read them in reverse, gray band becomes multiplier of x1000000000 which does not make any sense for shunt resistor.

There is another exactly the same resistor in PFC, I desoldered it too to measure. It reads 0.1 Ohm though I guess my cheap multimeter cannot tell its precise value, but 0.1 is more likely than almost 2k.

So how to read those strips?

P.S. As for gate resistor and diode -- I'm busy desoldering shield to get there.


EDIT:
If gray is not gray but silver (which might be) and I read bands in reverse: black-brown-blue-SILVER-brown = 0-1-6-x0.01 = 0.16 Ohm
This makes sense.

[Johnny10]

I just repaired a flyback SMPS similar to yours except having all discrete parts.

Blown resistor allowed excess current to flow through VIn for Control IC.

Took out Q7, Q2, R24 and Driver transistor for Q2.
I don't see a small signal transistor driving the Q2 MOSFET
Believe this is quite common design for this type of power Supply you can read about them Flyback transformer SMPS designs.
http://micro.rohm.com/en/techweb/knowledge/acdc/acdc_pwm/acdc_pwm01/1261/

[wraper]

Certainly replace NCP1377B before powering anything. Otherwise power MOSFET is almost guaranteed to blow up again.

[SergiiB]

Believe this is quite common design for this type of power Supply you can read about them Flyback transformer SMPS designs.
http://micro.rohm.com/en/techweb/knowledge/acdc/acdc_pwm/acdc_pwm01/1261/

That information might be helpful as I cannot find locally 0E16 Ohm resistor in 1W form, but 0E15 or 0E18 are available. So would be interesting to see how these affect output.

Certainly replace NCP1377B before powering anything. Otherwise power MOSFET is almost guaranteed to blow up again.

Yes, this one has a big crack, it is already out.

These are all components that survived (except Q8 was already replaced, and R44 desoldered just to measure Q9):

[Chris56000]

Hi!

The difference between 0R15, 0R16 & 0R18 is that the maximum peak current that can be taken from the psu will DECREASE as you increase the ohmic value – you can safely use a 0R15 to replace them!

(I suspect the OEM would have had the 0R16's specially made for the job, which is why you can't easily get replacements!)

Chris Willams

PS!

Amazon have some 2W 0R16 wirewound in packs of 5 for £2.75 + 60p postage in uk – you might be able to order them from your local site!

[SergiiB]

The difference between 0R15, 0R16 & 0R18 is that the maximum peak current that can be taken from the psu will DECREASE as you increase the ohmic value – you can safely use a 0R15 to replace them!
(I suspect the OEM would have had the 0R16's specially made for the job, which is why you can't easily get replacements!)
Amazon have some 2W 0R16 wirewound in packs of 5 for £2.75 + 60p postage in uk – you might be able to order them from your local site!

I measured dimensions: 12x45mm. It matches to 3W ones, like this (5%) or this (1%) , although 1% is not in stock.

But I guess if I am changing from 160 to 150, 5% is actually good as it could swing the real value close to original 160

Another question is temperature coefficient: the original one has 100ppm/K, this one 80ppm/K. Does it matter?

[SergiiB]

Hi, just want to thank those who helped me, especially xavier60 and Chris56000.
Got that last resistor today, put everything together and it works!!!
As a load I tried 1.5k and 10E resistors.

Here is a picture:


Here is what was replaced: