• Please review our updated Terms and Rules here

AT PSU repair

mR_Slug

Veteran Member
Joined
Nov 28, 2006
Messages
941
Location
UK
Hi, I have got a Seventeam ST-200W AT PSU that I'm trying to repair. Its the original Full AT style like in the 5170.

When i connect the PSU to a HDD, for a load, the fan spins, and then the whole thing shuts down. It doesn't try to restart.

attachment.php


I have in the past just put an ATX sized, AT PSU inside and hooked everything together but I would like to fix it if possible.

I'm reading http://www.repairfaq.org/sam/smpsfaq.htm, and have identified that the primary side seems to be working, I get 240V at the rectifier, ~480VDC across both the filter caps, ~240VDC between the center of the two filter caps, and either side of the caps, and ~240V at the input to the main transformer.

I'm using a DMM so I'm not entirely sure if the power transistors are switching to pulse through the transformer. In fact I think this is the problem.


From the source above I have narrowed it down to the control section. The PSU is based on the IR3MO2. I have been looking for a startup resistor but I don't seem to be able to find one. I'm looking for a large one with the third band as blue or green, in the Mohm range.

There are 2 ceramic power resistors. one is labeled 5W 1[ohm symbol]K and is connected between the AC input and one leg of the AC pin on the main rectifier. The other, 3W 51[ohm symbol]K, has one leg connected to one side of the main transformer T5, and via a ceramic capacitor to a tiny transformer T4. Its been a while since I read the values on a power resistor, is the placement of the ohm symbol supposed to indicate the multiplier?

The IR3MO2 is getting 7V at its VCC pin, which is the minimum voltage it is supposed to operate at.

The secondary from T4 goes back into the control circuitry, as does the secondary from T1. Each of T2 and T3 is connected to a power transistor, and the other side of T2 and T3 lead back to the control circuitry.

Photos: Sorry they are a bit grainy.


attachment.php


attachment.php


Does anyone have any ideas? Any help appreciated

Oh and BTW, yes, i know how to discharge caps:)
 
What are you using for a load?

On PSUs this old, it's not unusual for electrolytic capacitors to dry out--i.e., they don't short; they act as empty cans. One suspect is the cap labeled C4 (the blue one)--if I'm guessing right, it's usually a non-polar one and ripple current can run very high and heat it.
 
From what I'm reading the IR3M02 is designed to shut down if the 5V output supply drops below 4.3V but provides a trickle current of 40ma in the event of a short circuit on load.
I would think in an overcurrent condition you would be able to hear a ticking sound. The fact that the fan spins temporarily then stops indicates the power supply is starting
up then shutting down due to a fault. The power supply will not try to start again until the main caps have discharged so needs to be off then caps discharge to reset.
Weak caps can cause an undervoltage shutdown condition. Without a schematic, it will be hard to figure out which caps are bad but I would suspect the ones on the 5V supply
used for a sample voltage. I'm guessing the Vcc on pin 12 provides a 5V reference on 14 which is compared to the sample. A schematic and a scope would be needed to narrow
it down further.

Larry G
 
I would first replace all smoothing caps on the 5V rail, most likely the four large ones at the upper right of the first pic.
 
>I would first replace all smoothing caps on the 5V rail, most likely the four large ones at the upper right of the first pic.

I vote for those, too. Sometimes they pop out underneath with no evidence from above.

Larry G
 
A simple ohmmeter test between the common rail and each output should determine if there are any shorted filter caps.

The forum you pointed to obviously haven't had much experience with caps. You've got a 2.2 uF, 250V non-polar capacitor. with a 10% tolerance. If you've got a DMM with a capacitance scale, lift one of the legs from the board and verify the value.

Here's a similar 1.0 uF capacitor:
1mk_250v_cmc.jpg
 
I don't think any caps are shorted, just open. A power supply driving a direct short would immediately go into shutdown (tick mode) without giving enough power for the fan to start.
Open caps will not provide enough power to maintain operation other than a second or so but would torque the fan.

Larry G
 
Thank you all. Particularly Chuck(G). I have ordered a 2.2uF non-polar.

My multimeter doesn't have a capacitor test, and the short/open test doesn't seem to apply to non-polar caps like this, so I'm not 100% sure its failed. I had a look through my parts bin, to find a 2.2uF non-polar, but no luck so far.
 
Are you sure? Is that across the AC mains? Looking at schematics of typical AT type PSU show that to be 220nf which is .22uf 250V NP. It's just for RF suppression. If it shorts, fuse would blow instantly or it would be cracked / burned.
I don't think it would cause shutdown w/o fuse popping.

Retraction - looking further at it's location it looks more like a tuning cap by the transformer. I still advise being accurate on it's value. A wrong value could toast the output transistor. I would still advise trying the caps Paul mentioned 1st.

Larry G
 
Last edited:
The 2.2 uF value is, in my experience, very typical. 2200 pF would physically, be much smaller.

The image I linked to is verified as a 1.0 uF cap. That it's part of a tank circuit is the reason for the non-polar. You'd toast a polar cap in minutes.

There's a guy in Kazakhstan who has a huge capacitor collection, whence this image came from. Here are his 1.0 uF ones. The whole site, http://www.wasp.kz, is a veritable treasure trove of interesting stuff.
 
Last edited:
Yup some manufactures put the value on the cap as is, 2.2K = 2.2 uf +-10%, Or using the coding system 225K = 2.2uf +-10%
It can be a bit confusing sometimes.
 
Hi, well i replaced the non-polar cap with a new one, and replaced the 5 electrolytic caps on the output side, but its still not working. I checked the electrolytic caps with my multimeter on the high ohms range and they started out low and went up. Not exhaustive but the best i can do.

I now get a much bigger kick of the fan, but the 3.5" hdd still wont spin. I get about 7V on the 12V and 3V on the 5V for a second and then it dies.

retrogear, I will check around pins 12 and 14 with the datasheet in hand and see whats going on.

One thing i did notice is that the main filter caps that have 240VDC across each of them are rated as 200V 680uF? I have checked the voltages switch and it's set correctly. As far as I can tell, it doesn't affect this part, I will have to check more closely.

Any ideas?
 
Line operated switch-mode power supplies operate in two stages. The first involves changing the line AC to to DC, so there's typically a bridge rectifier and a smoothing capacitor(s). The DC is then converted to high-frequency AC (usually between 25 and 40 KHz) and fed into a transformer (high-frequency transformers don't need nearly as much iron as line-frequency ones do). The outputs are taken from windings on the transformer, rectified and filtered. Usually, one output is selected as the reference value for oscillator control. Because the stuff coming off of the transformer is high-frequency, you don't need nearly as large a smoothing capacitor than you would if you were dealing with line frequencies.

That's the typical SMPSU in a nutshell. So other than dodgy caps, what can go wrong?

One overlooked item are the various rectifier diodes, particularly the high-power ones for the +5 output. They do have a failure rate in which they'll tend to fail shorted. Check those out. Another choice spot is if the secondary (-5, -/+12) use linear regulators (.e. LM7805) on their outputs. If one of those fails, it's usually as a short. Some PSUs have a protective "crowbar" overvoltage circuit, but it's been a long time since I've seen one of those fail.
 
Wait a minute, the DC caps should only have 150-160 VDC across each for a TOTAL of 300-320. Don't leave powered on in the overvoltage condition !! The caps will explode. Something is wrong with your voltage switching.
You are in the voltage doubler mode for 110VAC but feeding it 220VAC ? The faq you referenced should have the jumper J1 removed for 220VAC operation. Your power supply is shutting down due to an overvoltage condition.
Looking at your photo, I'm guessing the back connector by the fuse is your AC input and the connector by the transformer is a voltage switch aka J1?

PS - give us a photo of the PCB side
PSS - line up two photos: component side then PCB with same orientation so I can switch back and forth
 
Last edited:
I will try to do the photo tomorrow. Yes the equivalent of j1 is removed, but i am still getting 480VDC? Its a familiar number to me to do with 3-phase, so I will have to look here a bit more closely, perhaps the rectifier is screwy. I am getting 5V out of the IR3MO2 as far as I can tell so this seems to be working now. I replaced some caps around there too. The datasheet i found for it is dreadful quality. I will inspect it further.
 
That's way too much. Usually, on a supply of this vintage, there's a voltage selector switch mounted on the power supply case that selects between 120 and 240v. If you're running it on the 120V setting, the rectifier circuit is placed into a voltage-doubler configuration and output should be no higher than about 340V (1.414*120*2). When you're in 240 volt line mode, the rectifiers and capacitors are typically arranged in a full-wave bridge, which should yield about the same no load 340V.

Here's a typical example. Note that switching between 120/240 VAC mains involves only a single jumper. See how the input circuit shown matches with your PSU.
 
PS - give us a photo of the PCB side
PSS - line up two photos: component side then PCB with same orientation so I can switch back and forth

Here are two photos, annoyingly i forgot my camera again so its done with a phone. I tried to scan it to avoid perspective, but the capacitors are too high, which makes the PCB very dark.

IMAG0337.jpgIMAG0339.jpg

I am going to trace the circuit out, as so far i cant see anything tying the center of the two caps to the Neutral in. The voltage selector switch is set as open. (im on 240V here).

If you need a better photo let me know and i will re do it. Thanks everyone.
 
Just a thought have you checked the bleeder resistors, They can go open, Not only can they serve as a safety device they can also establish a predetermined level, 480 VDC is definitely not right, Around 340 VDC or 170 VDC per Cap as Chuck and Retrogear have said is about right, Might vary a bit as it depends on what you got coming out of your wall socket.

Is it my eyes or do those main caps look bulged ?
 
Is the nylon plug you have plugged in the lower left with red/black wires coming from the AC mains? The connector it is plugged into is the 110/220 switch. Your AC should be plugged into the other connector where the fuse is.
That plug that says C IN with L and N goes to the fuse. Look at the sample Chuck provided. AC goes to points L and N !!!

Larry
 
Back
Top