Artic Power 950w - Not 100%

[quattro alex]

Re: Artic Power 950w - Not 100%

For diode open means shorted. Good diode is closed from one side (unless it's zener), allowing current to pass only in one direction And even in such condition there is voltage drop. Open diode has the P-N barrier broken so it behaves like conductor.

But not always it just opens - sometimes it burns so hard it interrupts the path internally. Usually LEDs die this way.

Replacing the diode at D5 was all it took in the end to get this up and running. Thank you to everyone's input in getting this up and running.

Next up I have a Corsair 520w that keeps cutting out on the psu tester, and I have a Goodmans Tv with a burnt connector to the backlight lamps.

[Behemot]

Re: Artic Power 950w - Not 100%

For diode open means shorted. Good diode is closed from one side (unless it's zener), allowing current to pass only in one direction And even in such condition there is voltage drop. Open diode has the P-N barrier broken so it behaves like conductor.

But not always it just opens - sometimes it burns so hard it interrupts the path internally. Usually LEDs die this way.

[quattro alex]

Re: Artic Power 950w - Not 100%

+1 on Behemots post.

Switchers = primary transistors on primary heatsink (HV or hot side )

I don't see the diode you are talking about as i am in rush right now, but desolder it and check it again for accurate reading. I would just replace it with similar value, from, lets say, another scrap SMPS. But please do check components around to make sure it doesn't destroy same components again after powering up.


//unrelated:
okay so i was right )
i use nokian hkpl R for winter driving ) does the job.

Thanks for the heads up on the "switchers" I'll take that heat sink out and test them on the DCA55.

Oh, I looked again and find the diode. So it has been already removed and tested out-of-circuit it seems? Than it's definitelly bad if it reads open.

Yes, I've got it out of circuit to be sure, but it reads shorted rather than open. Thanks for checking

[Behemot]

Re: Artic Power 950w - Not 100%

Oh, I looked again and find the diode. So it has been already removed and tested out-of-circuit it seems? Than it's definitelly bad if it reads open.

[domas]

Re: Artic Power 950w - Not 100%

I've just tested the diodes and found D5 (UF1007) to be shorted ( I get a reading in both directions). Should I replace with the same or go up a notch?

Also can you confirm what you mean by switchers please?

I managed to get around 330v at Q6 but not Q5 does this help at all?

Yes I am very much a fan of AWD and I am on some of the car forums. But if you want to make the most of that system you need the right rubber... Dunlop wintersport 3D's when it gets cold enough

+1 on Behemots post.

Switchers = primary transistors on primary heatsink (HV or hot side )

I don't see the diode you are talking about as i am in rush right now, but desolder it and check it again for accurate reading. I would just replace it with similar value, from, lets say, another scrap SMPS. But please do check components around to make sure it doesn't destroy same components again after powering up.


//unrelated:
okay so i was right )
i use nokian hkpl R for winter driving ) does the job.

[Behemot]

Re: Artic Power 950w - Not 100%

have a look if there is some small-value resistor nearby, sometimes there is and it makes the reading seem the diode is bad.

[quattro alex]

Re: Artic Power 950w - Not 100%

quattro, did you test passive components as I mentioned in answer #4? switchers?
They can be tested onboard, typical failures will result in opens or shorts.

I had identical issue with another p/s - the resistors were cooked to fuck. they changed their rating from 1ohm to something like 1Mohm There wasn't any clearly visible damage to them though. I am not familiar with octocouplers, never had problem with them, but I have swapped them with known good ones as a precaution As it didn't help, i've put the original one back


//unrelated: is there any possibility that by any chance your forum name refers to quattro as and audi all wheel drive system and I have seen you somewhere else on the interwebs? Most likely car related forums. It rings some bell for me.

I've just tested the diodes and found D5 (UF1007) to be shorted ( I get a reading in both directions). Should I replace with the same or go up a notch?

Also can you confirm what you mean by switchers please?

I managed to get around 330v at Q6 but not Q5 does this help at all?

Yes I am very much a fan of AWD and I am on some of the car forums. But if you want to make the most of that system you need the right rubber... Dunlop wintersport 3D's when it gets cold enough

[Escort Eagle]

Re: Artic Power 950w - Not 100%

Escort Eagle - I have checked the switching transistors on the peak DCA55, and they came out ok. Can you advise me on how many volts dc I should be seeing on pins 1&2 of the optocouplers, as when I tested it was only 0.02v dc.

Is it 0.02V DC in both Off_State and On_state ? At On_State it should go upto 1.5V . That means your controller IC is not working properly .

[domas]

Re: Artic Power 950w - Not 100%

quattro, did you test passive components as I mentioned in answer #4? switchers?
They can be tested onboard, typical failures will result in opens or shorts.

I had identical issue with another p/s - the resistors were cooked to fuck. they changed their rating from 1ohm to something like 1Mohm There wasn't any clearly visible damage to them though. I am not familiar with octocouplers, never had problem with them, but I have swapped them with known good ones as a precaution As it didn't help, i've put the original one back


//unrelated: is there any possibility that by any chance your forum name refers to quattro as and audi all wheel drive system and I have seen you somewhere else on the interwebs? Most likely car related forums. It rings some bell for me.

[quattro alex]

Re: Artic Power 950w - Not 100%

Escort Eagle - I have checked the switching transistors on the peak DCA55, and they came out ok. Can you advise me on how many volts dc I should be seeing on pins 1&2 of the optocouplers, as when I tested it was only 0.02v dc.

[Escort Eagle]

Re: Artic Power 950w - Not 100%

I think you have switching problem . check those switching transistors and Opt-couplers shown in the image . If the switching transistors and optocouplers are OK, then give a voltage check . Measure the voltage of Optocoupler Pin 1 & 2 at normal state (I mean without connecting "Green + Black" condition) . Now connect "Green + Black" and measure the voltage of optocouplers pin 1 & 2 again . This time optocoupler's pin voltage will go higher than the voltage was at normal state . If it happens like that, then your controller section is working properly otherwise not .

If the secondary side is OK, then you have the problem in the Primary section and you have to search in the primary side .

I hope it will help you find out the problem side ...........

[Behemot]

Re: Artic Power 950w - Not 100%

It'll draw 95000000… by than

[kaboom]

Re: Artic Power 950w - Not 100%

belongs in the hall of shame.
950w my ass!

Now, now...

That rating is correct. It'll draw 950 when the PFC inductor shorts and fuse is bypassed...

[quattro alex]

Re: Artic Power 950w - Not 100%

This looks like HEC, or a HEC knock-off. May be the rectifiers like behemot suggested. When you try to power it on do the fans twitch for a second?

And define "meaty" graphics card

No twitching can be seen on the fans, and when I hooked up the multimeter's I am only seeing up to 100mv. I am going to take a closer look at those rectifiers.

The guy didn't mention the spec of the card, and now he's away for a bit, but I'll try and find out.

[Pentium4]

Re: Artic Power 950w - Not 100%

This looks like HEC, or a HEC knock-off. May be the rectifiers like behemot suggested. When you try to power it on do the fans twitch for a second?

And define "meaty" graphics card

[quattro alex]

Re: Artic Power 950w - Not 100%

Peak min/max on something like a Fluke 87V helps find the overload.

I have an, unsolved problem, ATX power supply and all rails (5 and 12V) grossly overshoot the specified voltages. "Normal" min/max was not fast enough to catch the problem. I had to use peak min/max and two multimeters to see this.

As a result, the supervisor IC shuts down the secondary voltage, but I have yet to find the culprit.

I've got a couple of multimeters, but they're nowhere near the same league as a fluke. However, I'll give this a go this evening.

This. I'd check rectifiers as the first thing - if somebody believed that sticker and pushed it too far, rectifiers would probably be the first to blow.

So I asked the guy how much he had connected and it was a lot.... 5 hard drives, x2 USB hubs, lots of case fans a decent processor, x2 DVD drives and a meaty graphics card. I believe he definitely maxed it out and was sold on the 950w rating,

[Behemot]

Re: Artic Power 950w - Not 100%

I hope you do not think you can get 950W out of this supply, you would be lucky if it could do half of that continuously.

This. I'd check rectifiers as the first thing - if somebody believed that sticker and pushed it too far, rectifiers would probably be the first to blow.

[senz_90]

Re: Artic Power 950w - Not 100%

I hope you do not think you can get 950W out of this supply, you would be lucky if it could do half of that continuously.

belongs in the hall of shame.
950w my ass!

LOL. this thread interest me because that 950W statement.

[retiredcaps]

Re: Artic Power 950w - Not 100%

When you attempt to turn on this supply, do you get any voltage out of the 3.3, 5 or 12 volt rails for even a fraction of a second?

Peak min/max on something like a Fluke 87V helps find the overload.

I have an, unsolved problem, ATX power supply and all rails (5 and 12V) grossly overshoot the specified voltages. "Normal" min/max was not fast enough to catch the problem. I had to use peak min/max and two multimeters to see this.

As a result, the supervisor IC shuts down the secondary voltage, but I have yet to find the culprit.

[kc8adu]

Re: Artic Power 950w - Not 100%

belongs in the hall of shame.
950w my ass!