Piggy-backed PC power supplies?

[CRE]

Hey guys, I've got a kind if odd one here. A friend asked me to help him take two computer PSUs and merge them into one... the goal: higher wattage. Input 270W per PSU output >500W. Now I know that simply tying the two rails together would do the trick, but I suspect it could also result in a severely shortened lifespan, so I was figuring that a diode coupled with a nice big cap as a buffer would help eliminate any interference should one push less than it's counterpart. This will be on all the rails (-12v, 3.3v, 5v and 12v). Will this arrangement work?

 

[Thomas1]

Well... What did mean by "higher wattage"? Do you want to have more current or more voltage?

More voltage can be done quite easily by just connecting them in series, but I assume that's not what you wanted...
More Current, on the other hand, is very difficult to achieve because you have to connect them parallel. So both psu try to regulate the same voltage, causing their output to get very instable. Maybe a diode in each rail of each output could help, but you also get much heat and less output voltage.

I don't think you get happy with this one...

 

[CRE]

Hi, thanks for the response....

Wattage, lol, well, that's what I get for posting while half asleep. Yes I actually meant higher amperage. What I want to do is connect them in parallel, but filter them both so they aren't subject to any feedback from one another.

Thanks

 

[audioguru2]

The power supplies are already filtered and get their feedback directly from their outputs. Therefore you must isolate them from each other, maybe with just a 0.1 ohm resistor in series with the output of each one. If each one is supplying 10A, the power dissipation of its 0.1 ohm isolating resistor will be 10W and will be more if the supplies don't have exactly the same output voltage. 10A through each 0.1 ohm isolating resistor will drop the load voltage 1.0V.
They might become unstable with no load or a light load.

 

[ante1]

If the two PSUs are identical I think it will work with just the resistance in the harness but a minimum load is required to maintain regulation.

 

[Alun]

How about connecting a Schottky rectifier in series with the outputs and connect them all togeather?

Some of these units have forward voltage drops of less than 0.35V.

 

[CRE]

Yeah, I was thinking something along the lines of that Alun, but is that sufficient? I need to aim for a minimal loss of voltage and the rectifiers seem like the best bet. Depending on how the PSUs are built, I may be able to up each of the rails to compensate for that small of a loss... but a 1v drop kills the project.

 

[Kevin Weddle]

I would never suggest paralleling a power supply because the loads become variable and that could lead to poorer regulation. Some might say it is neglible when the load current demands are a small variable as opposed to large change in current which is a large variable. The output of the regulator is supposed to go directly to the load as the voltage there is regulated. Whenever you append something to the outside, like a voltage follower, you won't get regulation. However, for your purpose, I would suggest you just parallel the supplies.

 

[Alun]

Yeah, I was thinking something along the lines of that Alun, but is that sufficient? I need to aim for a minimal loss of voltage and the rectifiers seem like the best bet. Depending on how the PSUs are built, I may be able to up each of the rails to compensate for that small of a loss... but a 1v drop kills the project.

if you know how the PSU's schematic then you could remove the voltage sense connection and reconnect it to the output side of the diodes to compensate for the voltage drop. If this is inpratical you might be able to tweak the voltage up by adjusting an internal potentiometer or changeing the value of a resistor.

To do any of these suggestions you will have to reverse engineer the circuit, have a look at the ICs used normally there's one main controler look at the part number. Search the web for the datasheet, post a question in the datasheet request section.

 

[audioguru2]

you could remove the voltage sense connection and reconnect it to the output side of the diodes to compensate for the voltage drop.

Hi Alun,
You will be causing a conflict with the two supplies:
#1 PS senses that the voltage is too high but it is coming from the other PS, so #1 PS lowers its output voltage more and more. Then the other PS supplies the entire load current. The supplies must have exactly the same output voltage for it to work.

 

[ante1]

Hi Audioguru,

So, you don’t think it will work as is then with a minimum load?

 

[audioguru2]

Hi Ante,
When the supplies have diodes or resistors in series with their outputs and the voltage sensing is connected to their combined output, I think the supply with the slightly highest voltage will handle most of the load. At full load its output voltage will drop enough for the lower voltage supply to begin conducting some current.

 

[ante1]

I have paralleled DC-DC converters a couple of times on electric vehicles with no problems. These are similar in the regulating circuit as far as I know. ;D

 

[audioguru2]

Hi Ante,
Your DC to DC converters didn't have voltage sensing right at their outputs for feedback did they?

 

[ante1]

Audioguru,

No, not on the outside if that’s what you mean. ;D  But they have a regulated output voltage (13.2Volts) with very good accuracy and the input Voltage can vary between 48 and 96 Volts (80 nom.). They put out 40 A at 13.2V and if more is needed just add another one in parallel, no problemo! 8)

 

[CRE]

Well, after talking to my friend who asked about doing this, he's taken some time to think about it and decided to do the sensible thing..... buy a case big enough to fit a full ATX PSU.... Still, one of these days I might be inclined to experiment on this with a couple cheapie's.... just to find out what works best. Thanks alot guys!