Auxiliary Power Supply question

[Mustang]

I have an old cheap ps that that I converted into an aux unit to power periferals (hd's cdrom and case fans) on a server board. The conversion went ok but the 12v rails seem too low. It averages 11.29v. The 5v rails are fine at 5.34v. Is there an easy way to increase the voltage on the 12v rail? Maybe a pot or a different rated resistor?

[Mustang]

Re: Auxiliary Power Supply question

There's a 370 ohm resistor connecting the 12v rail to ground. Would changing this affect the voltage? If so, can a pot be run on this connection to adjust voltage or do I need to look elseware?

[AK0R]

Re: Auxiliary Power Supply question

Most power supplies regulate based on a combination of the 5V and 12V outputs. Therefore, if you have no load on the 5V side, it stays above 5V, and the power supply regulates the 12V side down to try to get the overall regulation to a point where it's happy. The fact that the 12V side is loaded only makes the problem worse. Try loading the 5V line, and I bet you'll see the 12V line come up. Also, what kind and rating of PSU, what caps does it have, and what are their condition?

[Mustang]

Re: Auxiliary Power Supply question

Thanks for the reply AKOR. You're right it was checked under a no load condition. I'll check it with a load and get voltages. All caps 470uf and up were replaced with panasonic FM's. The 12v side is only a 1000uf 16v. I do have 1500 and 3300 16v caps but not sure if they would be overkill or possibly cause problems elseware. 5v side is 1000uf 10v. PS was overated at 450w but 200w seems appropriate.

[Mustang]

Re: Auxiliary Power Supply question

With 2 hd's and a cdrom the voltage increased to 11.55v. Then I added 3 fans to simulate the actual load this this unit will be under. Voltages are now 11.53v and 5.24v. This within tolerance for the drives? Perhaps run the fans at 5v and just add a few more. They'd spin slower and quieter and load up the 5v rail.

[Chris1992]

Re: Auxiliary Power Supply question

11.53 and 5.24 are well within spec (10% tolerance).

[Newbie2]

Re: Auxiliary Power Supply question

Yep, those voltages are within spec. Now what manufacturer made that PSU Mustang?

[AK0R]

Re: Auxiliary Power Supply question

Thanks for the reply AKOR. You're right it was checked under a no load condition. I'll check it with a load and get voltages. All caps 470uf and up were replaced with panasonic FM's. The 12v side is only a 1000uf 16v. I do have 1500 and 3300 16v caps but not sure if they would be overkill or possibly cause problems elseware. 5v side is 1000uf 10v. PS was overated at 450w but 200w seems appropriate.

Is there only a single cap and no inductor on each output? If so, you should be able to put in the highest capacitance cap that will fit in each location without any trouble. The problems come if there is a series inductor. Then you have to worry about the output inductor and filter capacitor(s) forming an L-C network that is resonant at or near the switching frequency.

[Mustang]

Re: Auxiliary Power Supply question

It's a pretty good size toroid inductor. 4 sets of legs with 2 sets +5v, 1 set +12v and 1 the -12v.

[PeteS in CA]

Re: Auxiliary Power Supply question

No +3.3V O/P? That's approaching ancient. Re using larger caps, both the LC and the ESR of the capacitor are important in the stability of the regulator's control loop. And in the kind of P/S that "doesn't have" an O/P inductor, in that topology, "flyback", the transformer primary functions as the "O/P" inductor.

[Mustang]

Re: Auxiliary Power Supply question

Here's the modified unit. Since it's only using the 12v and 5v lines, I guess there's no worries about the lack of an inductor on the 3.3v side. The ATX, P4 connectors and all 3.3v lines were removed as not needed. An input filter was added, primary side caps were changed from 470 to 680uf as well as the higher value secondary side caps. Finally, a ball bearing fan to cool the unit. I think it'll work well enough as intended.

[zepper]

Re: Auxiliary Power Supply question

What you need to do is load the 5V side. The 12V side will have significant load with a number of drives, but the electronics don't take much in the way of 5V - try to get the total load on the 5V side up to 5 Amps or so - better to get it up to 20% of rated if possible.

[tiresias]

Re: Auxiliary Power Supply question

What you need to do is load the 5V side. The 12V side will have significant load with a number of drives, but the electronics don't take much in the way of 5V - try to get the total load on the 5V side up to 5 Amps or so - better to get it up to 20% of rated if possible.

Yes, from what I can read it's a case of crossloading sending the 12V line out of spec. I've seen much worse actually with certain of the very cheap, but current, Fortron ATX300 units, in which a highly overclocked K7 machine (with +5v as CPU VRM input) could result in the 12V rail soaring as high as a potentially destructive 13V.

There isn't that much though nowdays in terms of modern hardware that draw much current from the 5v - apart from motherboards and PCI cards of course, but he's powering those with his "main PSU" already.

[gonzo0815]

Re: Auxiliary Power Supply question

I've seen much worse actually with certain of the very cheap, but current, Fortron ATX300 units, in which a highly overclocked K7 machine (with +5v as CPU VRM input) could result in the 12V rail soaring as high as a potentially destructive 13V.

I have seen exactly the same issue with an very cheap FSP350THN-P OEM unit. IMHO a very decent unit otherwise. This psu comes with an 24pin ATX main conector and an 130w 5v & 3.3v rail combined power.
I used it testwise in an socket A board, drawing the most power from 5v rail for a couple of days. Iwas glad, that my old Maxtor 6Y120L0 had not been blown up at that high voltage (they are prone for that....).
In generall, i never mind those problems, as i may be used that unit out of spec. But i have expected, that the overvoltage protection should kick in a little earlier. I locked at the datasheet, and AFAIk they stated a probably daedly high voltage of about 14v as the point, where the protectin will shut off.

[999999999]

Re: Auxiliary Power Supply question

You can add a load to the 5V side, OR you can hack the feedback circuit.
The feedback circuit's reading of the 5V rail is causing the 12V to be too low, if you cause it to think 5V is too low, it raises 12V without the extra real load or heat from more loading. Key is to not keep any equipment hooked up while testing so nothing is subjected to trial voltages.

Generally you'd put resistance inline or after a resistor and before the controller, a pull-down resistor to ground. Not having your schematic or unit for measurements I can't suggest a value resistor to use.

Also, 13V is not "potentially destructive". Parts using 12V are far more tolerant to an extra volt than something running off 3.3 or 5V would be, after all it's only 8% higher than the ideal value, though it'll make any fans which were somewhat loud, noisier still.