computer SMPS overload

[barathbushan]

I had this old 24 pin atx smps , i shorted the power_on and power_ok pins , to get the smps working without the mother board , and it worked fine , i checked all the rated voltages of +12,+5,+3.3,-5,-12,-3.3 . To test the output of the SMPS i decided to connect a 120 watt bulb at the +12 and -12 terminal (making it 24v ) , the bulb worked normally and after 30 seconds the power supply shut itself down , Any solutions guys ??

 

[Resqueline]

Never woke up again, eh? You may have overloaded and shorted a diode or something on the secondary then. Does it tick when you try?
Check the rated current rating of the -12V, I figure it's only around 0.1-0.5A, not exactly the output to load in that way..
And when you say a 120W bulb, what are you referring to? 24, 115, or 230V?
Besides, PC PSU's regulate the 5V, and have a minimum load requirement there. Loading the 12V only is the opposite of what it needs - for stability.

 

[barathbushan]

Never woke up again, eh? You may have overloaded and shorted a diode or something on the secondary then. Does it tick when you try?
Check the rated current rating of the -12V, I figure it's only around 0.1-0.5A, not exactly the output to load in that way..
And when you say a 120W bulb, what are you referring to? 24, 115, or 230V?
Besides, PC PSU's regulate the 5V, and have a minimum load requirement there. Loading the 12V only is the opposite of what it needs - for stability.

Thanks for your reply Resqueline

1)yes it never woke up and i never got any readings again , just the 5 volts on the +5vsb pin

2) Yes your absolutely right , it ticks every time i connect the pwr_ok and pwr_on terminals

3)the rated current on -12v pin is 800ma , and after checking it now i realize my mistake

4) i used two of my car's spare headlight halogen lamps each rated at 12v 5 amps (~60w) , by putting them in series i got the 120w load

 

[(*steve*)]

If you're lucky, the rectifier diode(s) for the -12V have failed short circuit. (But check the diodes for +12V too).

Fingers crossed, that will be all.

 

[barathbushan]

If you're lucky, the rectifier diode(s) for the -12V have failed short circuit. (But check the diodes for +12V too).

Fingers crossed, that will be all.

Thanks for the tip steve but i don't know where to find the -12v rectifier diodes on the
board , however taking your word for it i checked all the diodes on the board (painful ) , and found out around 6 diodes not behaving properly.

one of the diode is not showing a forward voltage drop , and the other 5 diodes are showing a forward voltage drop of about 0.05 v , which is not reasonable

 

[(*steve*)]

Schottky diodes will often show a very low forward voltage drop especially at low current. Look for a DC path from the anode(s) to the -12V output.

Follow the traces back from the -12V, you will probably find a capacitor, maybe an inductor, and then a pair of diodes (almost certainly in the same 3 terminal package).

Here is the sort of thing you might expect.

 

[barathbushan]

Yes i;ll try it out but still one thing that puzzles me the click sound (like relay switching) , any ideas about that ??

 

[(*steve*)]

The SMPS detects the massive overload in the first cycle that is applied to the transformer. It then shuts down.

The sound you hear comes from the transformer (I believe) as it absorbs a very large overload.

If you think about it, a transformer and a speaker have a lot in common. You can quite easily hear large transformers humming, or even buzzing if they have loose laminations. This is an example of that, except it's just one pulse rather than 20,000 or more per second. (I wouldn't be surprised if some SMPS's are quite noisy -- it's just that the frequency is too high to hear.)

 

[barathbushan]

The SMPS detects the massive overload in the first cycle that is applied to the transformer. It then shuts down.

The sound you hear comes from the transformer (I believe) as it absorbs a very large overload.

If you think about it, a transformer and a speaker have a lot in common. You can quite easily hear large transformers humming, or even buzzing if they have loose laminations. This is an example of that, except it's just one pulse rather than 20,000 or more per second. (I wouldn't be surprised if some SMPS's are quite noisy -- it's just that the frequency is too high to hear.)

I never knew that transformer's make noises , thanks for your answer steve , also can you tell me what comes after the main bridge rectifier (230v) , In this diagram there is an IC (similiar to but not T0-220 package) which has a hole through it , so i must replace it , but cannot identify it , the input to the ic is around 325v DC from the bridge , but zero output , also i went through the schematic of the 200w SMPS , and still am not able to recognize it , does this look any similiar to you ??

 

[(*steve*)]

The hole looks like where the mounting screw is supposed to go.

Is there a spare screw floating around?

Find the part number for the SMPS controller on the board then google it. You'll probably find a datasheet with a simplified version of what you have.

Essentially the SMPR rectifies the mains, then switches it very rapidly (20-100kHz) through a transformer. The secondary windings produce voltages that are rectified and filtered. One or more rails is sampled and used to feed back (via an opto-coupler) the voltage to the SMPS controller which then adjusts things to keep the output voltage relatively constant with load and line fluctuations.

 

[barathbushan]

Sorry for the confusion , yes what you are talking about is a screw hole (i removed the screw to see the IC number) , but there is a smaller hole which is not visible in the picture , but anyway i am sure the IC is ruptured and not working since the output of the IC is 0 volts , so i want to replace it , the number i found on the IC was XW 13003 , i googled the part number but still could not find out about the IC , could this be a DC chopper ?? (forgive my ignorance as i am not well versed in power electronics)

 

[Resqueline]

Does the board say it's an IC?
I'd guess it's a transistor, MJE13003, and that it's the 5V standby chopper.
It does the job of Q12 in the diagram that steve linked to.
If you're trying to measure switching pulses with a multimeter; forget about it, you'd need an isolation transformer and an oscilloscope to do that.
Look, you have 5Vsb, and you have the ticking, both of which tells you that the primary circuits are working like they should.
Concentrate your efforts on finding that one shorted diode (or -12V regulator) on the secondary side.

 

[barathbushan]

Ok i'll try looking only at the load side , where around 6 diodes are to be replaced , since they have a voltage drop of 0.05v and one diode had broken down , to give a open circuit across it

EDIT : The pcb marking below the IC shows it's q105 , so is it a transistor ?? Also are all SMPS designed based on the schematic which steve has given , or are their slight variation's ??

 

[(*steve*)]

Qn (Q1, Q12, Q206 etc) is an almost universal designation for a transistor.

No, that was just an example I pulled from the net, specifically to show the secondary side. If you find the part number of the controller (as I mentioned) you will (possibly "may") be able to find the general design for your SMPS.

And also as I said, the diodes may exhibit very low forward voltage. Find one that measures (close to) zero -- both ways -- and you will have your culprit.

 

[Resqueline]

Looking around on the PCB I'm sure you'll see the connection between the marking & the part type. Some examples:
F = fuse
R = resistor
RA = resistor network
VR = variable resistor
C = capacitor
L = inductor
T = TR = transformer
Y = crystal
D = CR = diode
Z = ZD = zener diode
TH = thyristor
Q = transistor
V (German) = semiconductor (diode/thyristor/transistor)
U = IC = D (German) = integrated circuit
J = connector
JP = BR (German) = jumper wire
TP = test point
There is of course room for additions and variations but you get the point.

When diodes shorts they usually end up with less than 10 ohms resistance. If you ohm those 0.05V diodes you may find that it's actually a standard resistor value (100ohms).
What you're measuring then is most likely just the load resistor that sits on each output. It's not likely that 6 diodes have shorted, nor that one has gone open-circuit.
Feel free to post pictures of the secondary/output/load area.

 

[(*steve*)]

I've also seen some other synonyms:

X = crystal
LK = Link (same as jumper)
CON = Connector

Some other relatively common ones:

MIC = Microphone
SP = SPK = SPKR = Speaker
A = AE = Aerial
B = BATT = Battery
RLA = RL = Relay
SW = Switch
TH = Thermistor

And lots of really obvious ones for things that don't appear too often:

LED = LED
LDR = LDR (or R)
SCR = SCR (which is interesting because triacs are usually Q, as are diacs!)
REG = Regulator (or IC)

And there are some you won't see too often:

V = Valve (What did the Americans use?)
TR = Transistor (I haven't seen that for ages)

A rather disturbing trend I've noticed is to list the values of components (such as resistors and capacitors) on the schematic and the silk screen, but not identify them. It can leave you wondering which 1k resistor you're looking at Of course it can be useful if you don't have the schematic and you do have a toasted or exploded component.

 

[barathbushan]

Thanks for the pcb component label guys , i guess now i'll scan each component and find out which one is toasted , with luck i might find out the culprit