Weird issue with HEC Orion PSU

[Pentium4]

Originally, this PSU only got 5VSB (5.15V) and would not power on. I replaced the caps on the 5VSB circuit and after testing discovered that the diode (SB340) next to one of the 5VSB caps read open both ways. I replaced it with a tested good SB340 and now it still won't power on and the 5VSB only reads 1.9V

It was only plugged in for 30 seconds and the primary heatsink already got very hot. I assume it uses a 2 transistor design and that the 2N60L MOSFET outputs the 5VSB? ( )

That little PWM chip by the input caps is an ST UC3843B ( https://cdn.badcaps-static.com/pdfs/...d83cf69e0f.pdf )
Anyone have any ideas? Appreciate any help

[domas]

Re: Weird issue with HEC Orion PSU

could me anything. had two supplies with 5vSB working, nothing else, killed them by myself, now they are repaired. If I was you i would just go through every diode on pcb and power transistors with DMM, something got to be shorted. My supplies had 5V on SB rails though. They were not shorted

In my case it was some diodes power transistors, rectifier diodes (remember, they are not only heatsinked type diodes there, for low outputs like negative voltages or ~10v for PWM IC). Doesn't take that long to go through diodes. I have also had some resistors fried , they are little bit more difficult to find as you have to know what rating to expect..

In your case I assume that something is shortening just after 5vSB transformer, either something on main SB rail, or something on separate rail used for powering PWM IC. What is the PWM, i cant see in the pictures

[Pentium4]

Re: Weird issue with HEC Orion PSU

Thanks for the reply, the IC chip on the secondary is a Weltrend WT751002

[domas]

Re: Weird issue with HEC Orion PSU

now much voltage do you see on 7th pin?

[370forlife]

Re: Weird issue with HEC Orion PSU

Its a single transistor forward. I would take out the primary heatsink and test the primary switcher and the transistor for the 5vsb.

[Pentium4]

Re: Weird issue with HEC Orion PSU

Thanks for the replies, I will report the results hopefully today.

[everell]

Re: Weird issue with HEC Orion PSU

Double check those two capacitors you changed, that you did not get one installed backwards polarity. Also double check for possible solder bridge. The 5vsb was working before you started changing parts, so you should concentrate in the area where you were changing parts.

[Wester547]

Re: Weird issue with HEC Orion PSU

That very PSU, I believe, was reviewed on Hardware Secrets. I don't think it's single forward like the review says because I believe that requires the switcher to be 800V minimum and the one in the review is 600V. Whatever topology it uses, clearly, going by the results of the load tests, it's very inefficient and results in very high ripple levels even with relatively thick heatsinks and two 80mm fans in a push-pull configuration. It seems to use a method of DC-DC conversion for the +3.3V rail very similar to what's seen in many Hipros (the one in the review seems to have a separate transformer tap, and a separate rectifier, for the +3.3V rail as well). I'm not sure if it's linear regulation because the ripple on the +3.3V rail is a little too high (usually linear regulation is inefficient but simple and results in very low ripple and noise, even in gutless PSUs with outright deathly ripple), though that may be a byproduct of the unusually high ripple of the rest of the configuration. Makes me wonder if that's a custom type of DC-DC conversion. Sorry, I don't mean to steal your thread, I'm just curious about the above is all. ^^; But it uses the two-transistor +5VSB circuit (might even have the critical capacitor???), and it only uses a TO-220F part for the +5VSB transistor so I'm not surprised something that inefficient is shorting +5VSB diodes, nevermind causing other problems. Such a configuration is very inefficient which is why it was so hot... but maybe something's wrong with the 2N60L FET? Just a thought, though I could be wrong.

[momaka]

Re: Weird issue with HEC Orion PSU

Originally, this PSU only got 5VSB (5.15V) and would not power on. I replaced the caps on the 5VSB circuit and after testing discovered that the diode (SB340) next to one of the 5VSB caps read open both ways. I replaced it with a tested good SB340 and now it still won't power on and the 5VSB only reads 1.9V

If you had the 5VSB working before, then that diode couldn't have been open. Put it back and make sure it is in the right way.

In these cases where there is 5VSB but the PS doesn't power on, first thing you do is note if the fan twitches or the main PS voltages even try to turn on when you short PS-ON to ground. If not, then check if there is PS-ON voltage on the green wire. If not, check if the WT751002 is getting VCC.

Next, there are 3 optocouplers between the transformers.
One is for the 5VSB. If you have (or had) 5VSB, then it should be okay so leave it alone.
The other two optocouplers are for the main PS. IIRC, one of them turns power ON/OFF to the UC3843 PWM controller (or is used to disable/enable it in some way). The other is for feedback from the WT751002 to the UC3843 PWM controller to control the output voltages.

It was only plugged in for 30 seconds and the primary heatsink already got very hot. I assume it uses a 2 transistor design and that the 2N60L MOSFET outputs the 5VSB?

Yes, it's a 2 transistor design.
Make sure to run it in series with a bulb while doing any testing. If anything goes wrong, it can save you a lot of parts from burning.

[Pentium4]

Re: Weird issue with HEC Orion PSU

Double check those two capacitors you changed, that you did not get one installed backwards polarity. Also double check for possible solder bridge. The 5vsb was working before you started changing parts, so you should concentrate in the area where you were changing parts.

They're good I just checked them, and I was wrong I tested the diode incorrectly, I'm going to put the original one back in there.

That very PSU, I believe, was reviewed on Hardware Secrets. I don't think it's single forward like the review says because I believe that requires the switcher to be 800V minimum and the one in the review is 600V. Whatever topology it uses, clearly, going by the results of the load tests, it's very inefficient and results in very high ripple levels even with relatively thick heatsinks and two 80mm fans in a push-pull configuration. It seems to use a method of DC-DC conversion for the +3.3V rail very similar to what's seen in many Hipros (the one in the review seems to have a separate transformer tap, and a separate rectifier, for the +3.3V rail as well). I'm not sure if it's linear regulation because the ripple on the +3.3V rail is a little too high (usually linear regulation is inefficient but simple and results in very low ripple and noise, even in gutless PSUs with outright deathly ripple), though that may be a byproduct of the unusually high ripple of the rest of the configuration. Makes me wonder if that's a custom type of DC-DC conversion. Sorry, I don't mean to steal your thread, I'm just curious about the above is all. ^^; But it uses the two-transistor +5VSB circuit (might even have the critical capacitor???), and it only uses a TO-220F part for the +5VSB transistor so I'm not surprised something that inefficient is shorting +5VSB diodes, nevermind causing other problems. Such a configuration is very inefficient which is why it was so hot... but maybe something's wrong with the 2N60L FET? Just a thought, though I could be wrong.

Do I need to upgrade the 5VSB FET then? Hmmm, does zero input filtering have anything to do with the high ripple? It doesn't seem like the ripple should be so high. I would assume this thing shouldn't get lower than 75% efficiency as well but that's just guessing

If you had the 5VSB working before, then that diode couldn't have been open. Put it back and make sure it is in the right way.

In these cases where there is 5VSB but the PS doesn't power on, first thing you do is note if the fan twitches or the main PS voltages even try to turn on when you short PS-ON to ground. If not, then check if there is PS-ON voltage on the green wire. If not, check if the WT751002 is getting VCC.

Next, there are 3 optocouplers between the transformers.
One is for the 5VSB. If you have (or had) 5VSB, then it should be okay so leave it alone.
The other two optocouplers are for the main PS. IIRC, one of them turns power ON/OFF to the UC3843 PWM controller (or is used to disable/enable it in some way). The other is for feedback from the WT751002 to the UC3843 PWM controller to control the output voltages.


Yes, it's a 2 transistor design.
Make sure to run it in series with a bulb while doing any testing. If anything goes wrong, it can save you a lot of parts from burning.

I will put the original diode back in. I think I found out the source of my problem....check it out, just doing my normal desoldering and the Drain leg of that MOSFET just fell off!!! It must have barely been connected because I was gentle with it... It's a P10NK60Z ( http://pdf1.alldatasheet.com/datashe.../P10NK60Z.html )

Would it be ideal to replace it with a more powerful ST W12NK90Z with a ring around the Source leg? ( http://pdf1.alldatasheet.com/datashe.../W12NK90Z.html )

[Wester547]

Re: Weird issue with HEC Orion PSU

Do I need to upgrade the 5VSB FET then? Hmmm, does zero input filtering have anything to do with the high ripple? It doesn't seem like the ripple should be so high. I would assume this thing shouldn't get lower than 75% efficiency as well but that's just guessing

Input filtering only affects transient filtering, or EMI/RFI, whatever's before the rectifying bridge - it has nothing to do with output filtering. The high ripple even with linear regulated DC-DC conversion could just be the fact that the whole thing runs too hot to touch (the topology is extremely inefficient) - the gutless PSUs in PSU reviews, over the net, with very +3.3V low ripple (though HORRIBLE ripple otherwise) because of +5V to +3.3V linear regulation, did have their fans running at full speed. (they lack a fan controller). The +5VSB FET might be the problem, but I think it's more the two-transistor +5VSB circuit as a whole, especially since it has the critical capacitor.

Would it be ideal to replace it with a more powerful ST W12NK90Z with a ring around the Source leg? ( http://pdf1.alldatasheet.com/datashe.../W12NK90Z.html )

I recommend against it since W12NK90Z's on resistance is higher.

[momaka]

Re: Weird issue with HEC Orion PSU

I will put the original diode back in. I think I found out the source of my problem....check it out, just doing my normal desoldering and the Drain leg of that MOSFET just fell off!!! It must have barely been connected because I was gentle with it.

Interesting. I wonder what caused it. Possibly a screw up at the factory??

I would say just check the MOSFET for shorts. If it's fine, just solder a thick wire to the Drain terminal and put it back in the PSU. Should be fine as long as the wire you solder to the drain is thick enough.

[Pentium4]

Re: Weird issue with HEC Orion PSU

Oh okay that's what I figured about the input filtering...I'm still going to add it though if I get this thing working. I don't know much about MOSFET's, so I'm not sure what the on resistence affects. Another FET I have lying around is a Fairchild 11N90, although it looks like the on resistance on that one is even higher.

Interesting. I wonder what caused it. Possibly a screw up at the factory??

I would say just check the MOSFET for shorts. If it's fine, just solder a thick wire to the Drain terminal and put it back in the PSU. Should be fine as long as the wire you solder to the drain is thick enough.

I've never done that before. Just a plain old bare wire?

When testing the P10NK60Z I'm only getting a reading with the red probe on Source and the black probe on Drain and it reads a consistent 0.497

[momaka]

Re: Weird issue with HEC Orion PSU

^Yup. Anything that conducts and that you can solder will do . Just has to be thick enough to handle all the current. 18 gauge (or thicker) stranded or solid core will both do.

The ON resistance affects the efficiency. The higher the ON resistance, the higher the voltage drop through the MOSFET when it's conducting and thus the more heat it has to dissipate. Of course, when comparing the ON resistance, also check at what gate-to-source voltage (V(gs)) they are measuring it. For example, one MOSFET may say "0.5 Ohms at 5 V(gs)", while another may say "0.7 Ohms at 3 V(gs)". That doesn't mean that the second MOSFET is inferior, though, because in order to compare the ON resistance, you also need to match the Gate-Source voltage. With MOSFETs, the more you increase the Gate-Source voltage, the lower the ON resistance becomes (of course, provided you don't exceed the Gate-Source voltage limits - then you get a shorted MOSFET ). So with the above example, if you raise the gate voltage on that second MOSFET to 5 V(gs) (same as the first MOSFET), you may find that it's ON resistance becomes 0.45 Ohms. In that case, the second MOSFET would be better.

Another very important parameter, particularly for these single-transistor forward PSUs, is the reverse breakdown voltage or working voltage. Basically, you do not want to go lower for that voltage than the original MOSFET.

[Wester547]

Re: Weird issue with HEC Orion PSU

In pretty much every datasheet I see, the V(gs) is always 10V for every MOSFET. The drain current ("Id=") does vary a bit more but I believe that ON resistance doesn't change that much with the current draw but with the temperature and with the V(gs).

[Pentium4]

Re: Weird issue with HEC Orion PSU

Thank you guys for the awesome info! I love the detailed responses. I decided to try and use the one still on the heatsink, this should work with no issues?

[momaka]

Re: Weird issue with HEC Orion PSU

^ Yes, as long as the MOSFET is fine. Did you check it for short circuits first.

Also, I suggest when testing the PSU to have it hooked in series with a 60W incandescent light bulb. I think I posted diagrams in another of your threads but I'm not sure. Let me know if you need them.

[Pentium4]

Re: Weird issue with HEC Orion PSU

I checked it and I couldn't find any shorts. I just soldered it all back together and it powers on now!!! you guys are awesome, thank you for the help! Now I'm going to recap it and add input filtering. Oh and those possibly fake Teapo SW's I posted in the fake primary cap thread came out of this PSU, I replaced them with 680uF OST caps

[momaka]

Re: Weird issue with HEC Orion PSU

Interesting. I have the same PSU but it's the Thermaltake branded version with all of the input filtering. Mine has Teapo input caps as well. I wonder if they are fake like yours.

Glad you got it working though. You really need to thank yourself only. After all, you found the problem when you desoldered the primary heat sink .

So, good job! Post some pics when it's done.

[Pentium4]

Re: Weird issue with HEC Orion PSU

Really? So yeah they just wanted to save money by not adding it. If you get a chance you should post a picture of the input filter. I will definitely post some pictures when it's done! Should be slow at work tomorrow so I can get it done Among a few other projects!