The whole Two Transistor Thing

[everell]

Re: The whole Two Transistor Thing

I attempted repairing the 5vsb - replaced the shorted diode and both bulging capacitors with Nichicon PW caps. I measured the transformer and both transistors. All looked good. So powered it up ... a second or two pause ... then a loud bang with a nice fireball. Both transistors shorted!

This is why I hate troubleshooting those 2 transistor 5vsb circuits. EVERYTHING has to be just right or else. No telling what other problems were there, and now additional components bad. So I removed the heat sinks and 5vsb transformer to take some photos ..... now TWO burned areas.

IF I decide to repair this power supply - I will "Bestec" it - that is, gut the 5vsb circuitry and install a pwm chip for 5vsb circuit. Instead of blowing up, they just shut down when there is a problem.

[POM_MJ]

Re: The whole Two Transistor Thing

Wow, too much damaged to PCB and nearby components. :-O
Once, I was repair Enlight VP-300 in same case, I found LT817C leaked also but after replaced the two transistors twice.

[everell]

Re: The whole Two Transistor Thing

The pwm chip on this Fortron FPS200-60ATV is marked 3528. This looks like a proprietary part number for Fortron. Does anyone know the generic part number for this chip? I couldn't find 3528 data sheet by googling on the web.

[370forlife]

Re: The whole Two Transistor Thing

I suspect there is no OCP on that, considering those open holes on the pcb labeled "OCP control board".

[everell]

Re: The whole Two Transistor Thing

The OCP control pc board is for the main pwm switching chip - not for the 5vsb circuit which failed. Therefore, when a two transistor 5vsb circuit fails.....it is spectacular.

If you want a reliable power supply - don't use one with two transistor 5vsb design.

[socketa]

Re: The whole Two Transistor Thing

Was reading through the threads on this fourm, and came across this:
post #16 shows a seasonic with a couple of 8-pin chips.
Although i can't completely read the markings the one on the secondary side, they appear to be the same as the ones on my antec true 380s

The one on the primary side (in between the heatsinks) is
UC3844B
LCS310

The one on the secondary side is
33C5R1M
TPS3510P

Noting that this seasonic is a 2-transistor design and a candidate for a dm311 mod (which is well beyond my current ability),
Could one safely assume that this antec is also a 2-transistor design?

Had a decade's use out of this, and all the voltages are still in spec - except for the powergood grey wire, which is 0.05V - shouldn't it be around 5V?
No bulging caps - there is there is brown/cream glue over, and in between, the 6 fuhyiyya caps under the heatsink)

The last 2 pics are close-ups of either side of the secondary heatsink near the 33C5R1M chip

[Behemot]

Re: The whole Two Transistor Thing

I think I can see a primary transistor there, the one with heatshrink. But try to clean the dust and glue first and take some more photos around the +5 V SB supply.

[socketa]

Re: The whole Two Transistor Thing

On the primary heatsink there is 3 mosfets
2 are 2SK2648's, each on opposite sides of the heatsink - they would have to be the primary switching transistors?
And the other one (4N60B) is smaller (on the 3rd photo in previous post)

After removing the glue, there is seen another transistor (i think you call these little ones BJT's)

So there is 3 transistors, and that small mosfet, in that area -
the one in the heatshrink,
the BJT that was under the glue,
the BJT next to the transformer and the optocoupler (you can see the optocoupler in the first photo in previous post),
and the mosfet on the heatsink

[Behemot]

Re: The whole Two Transistor Thing

The 4N60B is the first transistor of the circuit, you can tell right anway from the small current (4 A) it can handle, it cannot be used for anything else.

[socketa]

Re: The whole Two Transistor Thing

The 4N60B, is the one screwed to the primary heatsink.
Looked up the datasheet, and it says that it is a high voltage mosfet - not a transistor

It's Drain-Source voltage is rated at 880V
Continuous Drain Current is 9A

In a 2-transistor ISO psu, that i've partly traced out, the first driving transistor is a low voltage, TO-92 type - i.e., a small one.
The other one, that is being driven by the first one, is a high voltage transistor.

[Behemot]

Re: The whole Two Transistor Thing

MOSFET = Metal-on-semiconductor Field Effect Transistor

[everell]

Re: The whole Two Transistor Thing

The 4N60B MOSFET is a part of the 5vsb circuit. The small transistor under the glue is probably the other transistor in the 5vsb circuit.

The MOS in MOSFET stands for Metal Oxide Silicon (in this country!) So the 4N60B is a MOSFET Metal Oxide Silicon Field Effect Transistor.

I am including a drawing I made originally for a SmartPower SL400, and corrected to a SmartPower 380. Most of the components in the 5vsb circuit are the same. Hope this helps.

[Behemot]

Re: The whole Two Transistor Thing

Both acronyms are commonly used and basically mean the same though mine is more general - silicon is the semiconductor (but it can also be e. g. germanium). Oxide is insulator layer between the metal and semiconductor (which must be there by principle) but it is being replaced by more advanced materials, for example in CPUs it is for a long time as silicon dioxide still conducts quite a lot compared to some other materials.

[socketa]

Re: The whole Two Transistor Thing

OK got it.

Have thought for a long time that a transistor is only a device whereby a small current through the base-emitter, allows current to flow through collector-emitter,
and recently learned that mosfet current through source and drain was activated by a capacitance on the gate terminal (rather than a current)
mosfet and amplifiers operate quite differently, but now i know that they both fit under the catagory of transistors

Anyway, now i will call the the TO-92's amplifiers - transistors that are amplifiers

Thankyou for the diagram
i've being studying it, and reading up on transistors and zenners

What is the function of the two diodes at the top?
It doesn't appear that they would have any use,
since, together, they block current in both directions.

Looking near the optocoupler,
What is the function of the zenner diode?
If the voltage from the transformer coil (at the bottom of the diagram) exceeds the zenner breakdown voltage, and the optocoupler is turned off, then there will always be some amount of current through the first transistor, and hence also the second one.
Is that correct?

When the optocoupler is turned on, then the combined path (of the zenner and optocoupler) resistance is decreased, and the amplifier 1 (C945P) receives a boost,
and hence, more voltage is supplied to 5vsb output

Assuming that the motherboard 5vsb circuit always requires regulated current, then the voltage from the transformer winding (at the bottom of the diagram) that hits the zenner, must always be some degree higher than the zenner's breakdown voltage

Am i getting the hang of this?

What is the reason for two transistors?
Couldn't the job be done with one?
i.e, just connect the optocoupler to the 2nd transistor

i also note that if the first amplifier was unsoldered and removed, the 2nd amplifier would still be powered from the 320v (from the 2 mains capacitors) through resistors R9 & R89
So is this a permanent base voltage that the other amplifier contributes to when the demand is required?

[everell]

Re: The whole Two Transistor Thing

As for the two back-to-back diodes at the top, they are the SNUBBER circuit which removes the negative spikes from the signal on the primary winding of the transformer. Sometimes the SNUBBER is made up of a resistor, capacitor, diode combination. Sometimes you will see the back-to-back diode configuration.

As for the rest of the components in the two transistor circuit, the best explanation I can give is ...........magic! All the components have to be good.....OR.....bang! That is why I hate to troubleshoot two transistor circuits. I prefer to gut the 5vsb circuitry and install a PWM chip circuit. For me an easy fix. For others too difficult. Some time ago I wrote a detailed thread on converting Antec SmartPower supplies from two transistor to PWM chip.

https://www.badcaps.net/forum/showth...ighlight=SL400

[i4004]

Re: The whole Two Transistor Thing

just about the only thing i figured out about 2tr +5vstb circuit is that no voltage on optocoupler (secondary side) means 2 transistors on primary side will blow...on first one i just forgot to put the tl431 back after checking it, and on 2nd i probably shorted 2pins of tl431 while measuring...i also removed the 2nd lytic on secondary that was slightly leaky...that brought the voltage slightly down, by some 200mV or so...dunno was that alone (cap removal) enough to blow primary side...lol!

checked most (better to say all) components on both sides and found zilch.
(it started by me hoping removing a damaged cca 100ohm resistor between +5vstb and ground would solve the issue...hehe..that was naive...then i looked colosely at optocopupler and tl431(feedback) circuit, then components on primary side....found nothing)

by now it seems to me this circuit can't be troubleshoot....

this one had 8V instead of 5V, and i was hoping to figure out what's driving it high (i had one psu at 8V and one good at 5V i thought i coudl just compare the two and see what's up)....but i never really came up with a good idea how to do this in a logical and reasonable way that would ease troubleshooting....

i wasn't doing it because i wanted to make money on this cheap garbage (i don't buy or use such crap), but i was just curious....why 1 transistor circuit goes haywire is easy, it has lytic on primary that craps out, but this?

i think i'll now try ring test and inductance of trafo (8v psu vs the good one)...just to be able to say i checked it all (i did ring test on trafo previously..passed it fine)

last 2 pix show slightly different circuit, where function of tl431 seems to be integrated into that sg6105d ic that's on the secondary side...

it now occurs to me that interesting thing would be to check ripple on the output (screwed vs ok functioning circuit), to see how that affects working of this rubbish...

[Behemot]

Re: The whole Two Transistor Thing

They really put the chloroprene glue on the input? Oh my

[rievax_60]

Re: The whole Two Transistor Thing

There is plenty to be found on the web when "flyback converter" is searched for. Things will then make more sense.

[i4004]

Re: The whole Two Transistor Thing

rievax, i think both everell and i are way passed the basics stage and working principles of smps are mainly absorbed....it's just that chinese make these things in a wrong way, removing all components they can in order to save some money on mfr.
(and that way produce electronic bastards that shouldn't even exist)
it's oscillator steered by feedback...or, it should be...heh...

voltage on optocoupler was lower on the psu that had 8v stb (2-3v vs. 4-5 on properly functioning one...ie one pin was 2v against ground, other 3v).

either way i think thing to look at would be to see how small fluctuations on secondary side affect primary (well, by now we know they can blow the primary, but it would be interesting to see just at what point..heh). this psu had about 200mVpp of noise on 5vstb output.
the tolerance of the output filter is probably very tight, that explains why chaning output caps messed it up so much.
(one needs to wonder how they make it work in the first place... do they try many different caps and coils? )

now i don't think (anymore) i had to short 2 pins of tl431 to blow primary transistors, as i remembered that even on the psu that had no tl431 it took some time to blow them, so i think here it was simillar; removing the slightly leaky cap threw it out of the balance and then primary tr. fail....it also took some time...

[everell]

Re: The whole Two Transistor Thing

Most two transistor 5vsb circuits I have worked on refused to be fixed. After blowing up several I tried to fix, I came to the conclusion that the modification to a pwm chip circuit is the way to go. Why waste time on a loser!

The last power supply I worked on overheated due to a stuck fan. The two transistor circuit refused to start, so was replaced with DM311 chip. Then it became a well regulated +17 volt circuit due to a bad optocoupler. After replacing the optocoupler, voltage dropped to 5.2 volts with no load but voltage dropped too rapidly with loading. Turns out the TL431 chip was also bad. So.......nearly everything in that 5vsb circuit was bad, except for the transformer.