What is your testing PSU and what's its story?

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

I was fooling around blowing more caps, and when i reached the 10v 2200uF slightly bulged Forever that used to be in this PSU, it didn't blow. Because it had nothing left to blow - it was COMPLETELY DRY!!! I will change the remaining Forever caps in this thing as soon as i get back home, as they probably aren't any better than this one...

I'll probably change the primary caps to 680uF too.

[PCBONEZ]

Re: What is your testing PSU and what's its story?

Measure the resistance of that length of the old gauge wire.
Measure the resistance of that length of the new gauge wire.
Take the difference, use ohms law, and do the math.
Then tell me if that's going to result in that much of a voltage drop.

ATX12V connectors burning black due to heat from from ill fitting pins is not a terribly uncommon on some mobos and some PSU's.
Changing the cable end doesn't change the pins on the mobo which might be where the problem is.
- I'm just saying look at it...

You ringing may be coming from the VRM MOSFETS.
Try a different mobo as a load and see if it changes.

Your mystery wave form [with PSU off] might be related to the +5vsb and the flyback some how.
Might even be going through +5vsb, through the mobo, and back to the PSU.
- It it there with PSU off -and- mobo unplugged?

.

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

PSU off i mean PSU turned off by its power button, or unplugged from the wall. The ringing is there no matter what load i put on it, motherboard or not. And since this is a test PSU it'll probably have no time to burn the connectors... but i'd still like heavier gauge wire on it.

I can't measure the resistance of the wire because i don't have a meter that goes low enough. But this 20AWG cable does heat up a bit in use, and that tells me lower resistance cable would be an improvement.

[PCBONEZ]

Re: What is your testing PSU and what's its story?

PSU off i mean PSU turned off by its power button, or unplugged from the wall. The ringing is there no matter what load i put on it, motherboard or not.

If it's there with the PSU unplugged from the wall then it's a problem with your scope.
Are you using an isolation transformer?

And since this is a test PSU it'll probably have no time to burn the connectors...

But if the connectors are crappy you will have a voltage drop there even if it's not on long enough to burn the contacts.

I can't measure the resistance of the wire because i don't have a meter that goes low enough. But this 20AWG cable does heat up a bit in use, and that tells me lower resistance cable would be an improvement.

If the cable complies with the AWG spec:
20 AWG is 0.01015 Ohms/foot.
18 AWG is 0.006385 Ohms/foot.
Difference is 0.003765 Ohms/foot.

According to AWG spec in chassis wiring 20ga is good for 11 amps per wire and 18ga is good for 16 amps per wire.

Assuming 20ga on a ATX12v connector with 2x +12v and 2x ground.
2xwires x11 amps @12v = 264 watts.

Taking 264 watts and assuming a crappy 60% efficient VRM means the CPU would have to be over about 158 watts TDP to over load two pieces of AWG20 wire.

Now, if that 20ga wire is not actually AWG compliant that's another story.
It may be an 'illegal' copper alloy or have too few strands of wire in the bundle with reduces current capacity.

.

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

If it's there with the PSU unplugged from the wall then it's a problem with your scope.
Are you using an isolation transformer?

I was talking about the other signal i see, which is also a decaying oscillation, and which could have something to do with the behaviour of the power supply. It only appears on the scope when the probe is plugged in - so if the probe cable acts as an antenna, then it can probably influence the power supply as well. At least that's my theory. Will post a few pics to make things clearer.

The spikes i see from the power supply are there only when the power supply is on, and they appear on 5v and 12v rails. Didn't check 3.3, will do... I don't think they are from the 5vsb circuit.

Now, if that 20ga wire is not actually AWG compliant that's another story.
It may be an 'illegal' copper alloy or have too few strands of wire in the bundle with reduces current capacity.

See, you're getting somewhere. I've seen speaker cable with resistance of 8 ohms per 10 meters... Out of boredom, i decided to power a random mains appliance with that cable (a neon lamp AFAIR). As soon as i plugged it in, the wire that went in the socket vaporized faster than a fuse because of the inrush current. There's no way of knowing what kind of wire they put in this power supply, so if i can replace it with something thicker, i will.

[370forlife]

Re: What is your testing PSU and what's its story?

Did you put a 10uf electrolytic and .1uf ceramic in parallel with the scope probes? Intel's spec for testing power supplies calls for them when using a scope to check ripple to eliminate those spikes.

[PCBONEZ]

Re: What is your testing PSU and what's its story?

He's right.
This is the doc Intel says to go by for a ver 2.2 PSU.

- Page 23.

I've also seen a 10uF and a .47uF ceramic recommended.
Not sure if that's from another spec or just someone else's opinion.

.

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

Did you put a 10uf electrolytic and .1uf ceramic in parallel with the scope probes? Intel's spec for testing power supplies calls for them when using a scope to check ripple to eliminate those spikes.

Will do, thanks.

[momaka]

Re: What is your testing PSU and what's its story?

Yes. The power supply that i built runs at 100kHz. I never believed the EI-33 core can do more than 200-250W (although i've read about people achieving that), till i did it myself. Btw, it's not laminated iron, it's ferrite. Perhaps you were thinking powdered iron, which is what the output chokes are made of.

As far as I know, ferrite powder and iron powder cores can do more than 300W because they can be used at a higher switching frequency.

Anyway, here the pic. There's one large cap hiding behind the secondary heatsink that didn't make it in the pic but oh well. The brown caps in the middle are the Nichicons. They actually heat up quite a bit, i think it may need more of them.

See those large 2-3W resistors in that area, that what's causing the heat. I've seen many cheap PSUs with burn marks on the PCB and the caps due to those load resistors. They can dissipate as much as 1.5W of heat each. That's actually a lot.
Best solution is to replace them with less power hungry resistors. Normally I try to put 25 Ohms or more on 3.3v, 100 Ohms or more on 5v, and 250 Ohms or more on 12v rail.

Improvements look pretty good otherwise. Make sure to change the critical cap in the 5vsb circuit (usually 10, 22, or 47uF, on the primary side).

[kc8adu]

Re: What is your testing PSU and what's its story?

one of my regulars does not believe in repairing hardware.they only call when a virus is kicking their ass .
they bought 12 cooler master 430's.6 months later the boxes started having issues.showed them the bad caps and told them to pull all these and i can recap them.we have a bunch more and can just replace them.i got the 12 they were going to toss.8 are now in my test bank.they have every connector there is so no problem with testing any mobo and or video card.got 1 on each bench and 2 spares.they also tossed an antec 750 because the user of the system was afraid of the glow pattern on the wall from the led fan!now i am waiting for the new coolermasters to die.

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

As far as I know, ferrite powder and iron powder cores can do more than 300W because they can be used at a higher switching frequency.

All SMPS cores are ferrite or iron core... Ultimately, with the cores used in PC power supplies, it boils down to how much wire you can squeeze in there. They are always copper loss limited in a typical application.

See those large 2-3W resistors in that area, that what's causing the heat. I've seen many cheap PSUs with burn marks on the PCB and the caps due to those load resistors.

Isn't the minimum load mandatory? I was always taught so. The controller in the power supply that i built maintains regulation even with zero load, but that's because it runs in pulse skipping mode. I wouldn't trust a TL494 to be able to do that... 14v on the 12v rail is not my thing.

Improvements look pretty good otherwise. Make sure to change the critical cap in the 5vsb circuit (usually 10, 22, or 47uF, on the primary side).

Thanks. 5vsb is fine, but will check.

they also tossed an antec 750 because the user of the system was afraid of the glow pattern on the wall from the led fan!

Good one!

[momaka]

Re: What is your testing PSU and what's its story?

All SMPS cores are ferrite or iron core... Ultimately, with the cores used in PC power supplies, it boils down to how much wire you can squeeze in there. They are always copper loss limited in a typical application.

I think you're right about that. Just read an some random article I found on Google yesterday, and I remember that's what it said too. I'll see if I can find it later tonight. It had some interesting stuff in there.

Isn't the minimum load mandatory?

Technically it is, but I don't think it needs to be that high, at least not if you're running your PC with it.
As a matter of fact, my Macron PSU doesn't have a loading resistor on the 12v rail, just the 3.3v, 5v, -5v, and -12v and I changed all to lower values. Haven't noticed any voltage abnormalities, though - at least not when running a computer. PWM controller in the Macron is TL494. There's also a LM339 in there for UV/OV.
I also changed the resistors on a 250W Deer PSU. 3.3v rail had a 10 Ohm resistor, 5v had a 50 Ohm one, and 12v had 100 Ohms. Changed to 50 Ohms for the 3.3v, 470 Ohms for the 5v, and 500 Ohms for 12v. Runs my P3 computer happily with stable voltages on all rails.

[PCBONEZ]

Re: What is your testing PSU and what's its story?

Why would you change a load resistor to a SMALLER value?
Load resistors are in parallel with the real load.
A smaller one means more current through the load resistor and less available for the real load.

.

[Th3_uN1Qu3]

Re: What is your testing PSU and what's its story?

Why would you change a load resistor to a SMALLER value?
Load resistors are in parallel with the real load.

Nobody said that here...

[PCBONEZ]

Re: What is your testing PSU and what's its story?

.
momaka did but he went back and changed it with that new 2 hour rule.

This is why two hours to edit is too long.

Next time I'll quote it. I don't think edit will change a quote.

[PCBONEZ]

Re: What is your testing PSU and what's its story?

Actually that part is still there.

As a matter of fact, my Macron PSU doesn't have a loading resistor on the 12v rail, just the 3.3v, 5v, -5v, and -12v and I changed all to lower values. Haven't noticed any voltage abnormalities, though - at least not when running a computer.

""As a matter of fact, my Macron PSU doesn't have a loading resistor on the 12v rail, just the 3.3v, 5v, -5v, and -12v and I changed all to lower values.""

Easier to read now?

.

[Khron]

Re: What is your testing PSU and what's its story?

***Nevermind***

[momaka]

Re: What is your testing PSU and what's its story?

Why would you change a load resistor to a SMALLER value?
Load resistors are in parallel with the real load.
A smaller one means more current through the load resistor and less available for the real load.
.

Actually that part is still there.
""As a matter of fact, my Macron PSU doesn't have a loading resistor on the 12v rail, just the 3.3v, 5v, -5v, and -12v and I changed all to lower values.""
Easier to read now?
.

Ahh shit, another embarrassing moment for me . Meant to say *higher* values (as in higher resistance).
At least we got the right smiley for it.
Seriously, though, sorry for the confusion.

.momaka did but he went back and changed it with that new 2 hour rule.

No, not really. I use the "Preview Post" many times before I submit a post. The edit was to add this part:

I also changed the resistors on a 250W Deer PSU. 3.3v rail had a 10 Ohm resistor, 5v had a 50 Ohm one, and 12v had 100 Ohms. Changed to 50 Ohms for the 3.3v, 470 Ohms for the 5v, and 500 Ohms for 12v. Runs my P3 computer happily with stable voltages on all rails.

It wasn't more than 10 minutes after I submitted these changes.

This is why two hours to edit is too long.

I thick TC changed it to 30 minutes.

Next time I'll quote it. I don't think edit will change a quote.

Correct.