# Making my own lab power supply from a pc powers supply (atx)...

• posted

I made my own lab pc power supply (atx) from a spare atx power supply according to one of the many Web Pages on the web and my 4 10 ohm 10 watt

10% tolerance parallel connected resistors as a legitimate load to allow the power supply to function gets too hot (after about 1 to 2 minutes of being on the resistors are too hot to touch). I thought the cumulative wattage of this resistor "pack" would be sufficient to handle the power, especially after doing the calculations on the ohm's law calculator web page (one of them at the very least) and the parallel resistor calculator web. Is the 10 % tolerance unbalancing the resistor load (each resistor being different just enough to improperly handle the current/power?)? Does anyone have any resistor recommendations that I could get at radio shack or some other easy resistor arrangement? I don't want to take apart the main unit of the power supply because of the electrocution/shock danger. Thanks - I thought for sure these parallel resistors would be able to handle the power dissipation, later!
• posted

That's one of the errors people make. They think PC power supplies will make great general purpose power supplies, but the reality is that for many uses they provide way too much power. You waste the power in the dummy load, because you don't have need for enough current to load the supply properly. If you did need that much current, then you'd not need the dummy load.

Better to grab a transformer (and high value electrolytics) out of a piece of consumer electronic equipment, and build up a power supply that won't need a dummy load. You can take that PC power supply, strip it (the bridge rectifier can be used in the new supply) and use the case for your new linear power supply. The hole for the AC input jack is already there, as is the power switch.

Resistors in parallel do not increase their wattage handling. And since they are in parallel, they end up being a lower resistance, so when you put that across the load, they have to pass a lot of current, and hence wattage.

Michael

• posted

watt

the

being

of

of

10

any

easy

power

dissipation,

What about ohm's law and parallel resistors, p=p1+p2+p3+p4.... etc.? My main intention is not to argue but to learn, isn't the power dissipated over each resistor equally, thus 4 resistors = 4 times less the amount of watts/heat dissipated over each resistor in parallel? Interesting, what about in series, would that be better? Thanks for the comment.

And

Hmmm, what if I up the resistance and use the same power of resistors in parallel? Later!

• posted

\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\

Daniel, you're letting yourself get confused by the fact that the resistors are getting very hot. Power is measured in watts, and watts are experienced in the form of heat. The fact that your resistors are in parallel doesn't affect the power that each one is dissipating. In a parallel configuration, each resistor has 5VDC across it. Since each resistor is 10 ohms, then each one is dissipating E^2 / R watts, or 25 / 10 = 2.5 watts. Since there are

4 resistors, they are dissipating a total of 10 watts. 10 watts is a fair amount of heat, and if the resistors aren't mounted to a large metal surface (a heat sink) or cooled by a fan, they will get hot to the touch.

The effect of connecting resistors in parallel or series has no effect on ohm's law. Each resistor will behave the same, whether it is in series with other resistors, or in parallel. But, if you connect your four 10-ohm resistors in series, then your total resistance is now 40 ohms. E^2 / R =

25 / 40 or 0.625 watts TOTAL!!! Each resistor is now dissipating only 0.156 watts. Why?? Because you have divided the voltage across each resistor by 4. 5 / 4 = 1.25 volts. E^2 / R = 1.56 / 10 , or 0.156 watts in each resistor.

Looking at the current load that a series connection will provide, I = E / R = 5 / 40, or 0.125 amps. That's not enough to provide a minimum load for your power supply.

No matter which way you look at it, the law can't be broken. Cheers!!!!!

```--
Dave M
MasonDG44 at comcast dot net  (Just subsitute the appropriate characters in ```
• posted

(one

have

other

will

supply

watts/heat

resistors

experienced

configuration,

each

are

surface

with

That is why I assumed that power behaves the same whether used in series or parallel. I guess power is separate and that now makes sense as power really isn't resistance (ohms). The funny thing is whether it is one of my books or information on the net the p=p1+p2+p3+p4 equation is in the same section as ohm's law, making me assume it is part of it.

But, if you connect your four 10-ohm

That E^2/r= equation lets me understand, thanks for explaining it to me!

I guess I need to give in away from perfection and just do what the web page says and zip tie one 10 ohm 10 watt resistor to the inside of the pc power supply and apply some heat sink compound before hand.... There are uses for these pc power supply/lab power supplies - every once in a while I come across a high current requirement, such as powering a particular cable modem, but I admit a lot of energy is still wasted but even then there still are uses (at the very least these pc power supplies does the job minus a somewhat higher electric bill).

Thanks guys - I need to regroup now. Later!

in

• posted

(one

have

other

will

supply

watts/heat

resistors

experienced

configuration,

each

are

surface

Wow, major revelation! Are you saying that these 10 watt resistors can handle 2.5 watts even if that means they get this physically hot? In other words - they are supposed to feel this hot and if I want them to feel cooler I need to attach them to a heat sink with heat sink compound? Should I just go ahead and use the power supply and ignore the heat? Should I install one

10 ohm resistor (10 watt, 10%) inside the power supply instead? I always thought a resistor would feel cooler than it's rating when the wattage needed is lower than the resistor's rating.

Sorry guys for all of the questions, I am relearning these basics that I originally learned in high school and when I was in high school I had only a brief introduction because we needed to cover so much material in such a short time (the story of my public school life - hence mathematics as well). Please bare with me, my goal is to get this project done with as low a heat output as possible (call me a perfectionist). Thanks.

with

in

• posted

"Too hot to touch" does seem a bit hot for a 10watt resistor at 2.5 watts.

I do not think you have ever told us the voltage across this resistor and it has been assumed to be 5 volts.

Could you please measure it so we could be sure of the rest of the calculations.

```--
John G

• posted

The voltage across the resistor pack (4 10 ohm, 10 watt, 10% resistors in parallel) is 5 volts (measured at 4.90 volts - close enough, plus - the voltage was known by me to be at 5 volts already, I just measured to be beyond a reasonable doubt). Any help would be appreciated!

• posted

Are you sure that a dummy load is still needed with ATX power supply?

I used many to power small electronic boards, and all I did was to short the Power OK pin with 3.3V pin on the ATX connector

Now, this is the right place to know if I did it right or if this could lead to strange problems. :)

semplificatamente, Cthulhu

```--

http://www.rlyeh.it/Rantolii/```
• posted

If your resistors are in parallel, then you can pretty much consider their power ratings independently. You have a 10 Ohm resistor, which will draw

5V/10 Ohm = 0.5 Amp from a 5V rail (I'm guessing that this is the one you are loading). That way, each resistor is dissipating 0.5A * 5V = 2.5W, and shouldn't get that hot if it is rated at 10W.

Are your resistors aluminium cased, a sort of gold colour with two screw holes in diagonal corners? If so, then the rating only applies when they are mounted on a heatsink. If they are instead white ceramic cased resistors, then in my experience when used at the rated wattage they will get very hot and smell a bit, and after a while cracks form. What I do is get a resistor with a much higher rating than required, because to me it seems that these ceramic cased resistors are given their power rating by someone in the marketing department of the resistor company, and this guy is a bit too enthusiastic.

If you put two resistors very close together, then the combined power handling capability will be less than if you space them apart a bit. This is because one resistor will not only heat itself, it will heat the other resistor too if they are close together.

If you want a cheap power resistor, what I use sometimes is a car headlight bulb, or car indicator bulb. If you change your headlight bulb usually there are two filaments are inside and one of them is blown, so you can't use the thing in your car any more, but the remaining filament is still useful for other purposes. Be careful that the bulb is not too close to flamable materials such as wood etc, since it could be a fire hazard.

Chris

• posted

Most PC power supplies need a minimum load on the +5V output, even ATX units. The actual minimum load requirement varies by brand/model. If the minimum load current isn't being drawn, the power supply won't regulate properly when current is below minimum required. The only way to know is to get the mfr's specs for the brand/model, or by trial & error.

Your procedure of shorting the PwrOK to 3.3V out is wrong. Pwr_OK is an output of the PSU, as is the +3.3V output. I don't know how you got it to run like that. You sure it wasn't the Pwr_ON input that you connected to Gnd to start the supply?

Look at the description of the ATX supply operation at

it will explain a lot to you. Remember that the schematic and description is for a specific brand/model PSU. Your may be similar to that one.. or it may be very different. But a few things are certain.. the connector is the same (except certain Dell models), and the logic that controls turn-on and turn-off is the same.

```--
Dave M
MasonDG44 at comcast dot net  (Just subsitute the appropriate characters in ```
• posted

So I was lucky that time. Good to know. :)

A red led would be too mimimalistic load?

Gaaah. You're right, don't know what was thinking when I wrote that!

I shorted PWR_ON to GND.

Thanks, but don't overstimate my analog electronic capabilities. :)

limitatamente, Cthulhu

```--

http://www.rlyeh.it/Rantolii/```
• posted

Yes. Your hand is not a good thermometer. If you can touch it the chances are the temperature is less than 50C (Partial thickness burns will occur within 30 seconds at 55°C). Check the rated spec for the part noting that it may specify temperature several different ways.

cooler

Exactly. You could use the case of the PSU.

• posted

Typical minimum loads on PC power supplies can range from 2 - 5 amps, depending on the specific design. A red led would be totally inadequate for a minimum load, since they typically draw only a few milliamps.. Power resistors are the best, and most economical. your 2 amp load is probably minimally acceptable, but if the power supply works with that load, it's probably OK. Just remember to account for that current when you're running the supply cllose to its maximum ratings.

```--
Dave M
MasonDG44 at comcast dot net  (Just subsitute the appropriate characters in ```
• posted

watt

calculator

Does

and

Mine are the white ceramic cased resistors (i.e. sandbars).

What I do is

Anyone have any ideas on resistor combinations (including some resistors in parallel and some in series, etc.)? The reason why I ask is for one thing some of the reactions I have been getting on this topic is that my expectations are not achievable? Tonight I am going to sit down with a piece of paper and see if I can come up with a good resistor combination. Later!

• posted

in

No it's reasonably easy...

Look up the spec for you power supply and see what the minimum load is. Lets say it's 5A on the 5V rail.

This means you need a resistor or combination of resistors that totals to 1 Ohm (R=V/I) and is cabable of burning 25W (W=I*V).

You could use one resistor rated for say 50 or 100W but it's easier to dissipate heat if it's spread out. I would consider using 4 resistors in parallel....

Each resistor would need to be around 4 Ohms (4//4//4//4=1) and rated for more than 6.25W (=25/4).

The nearest standard value to 4Ohms is 3R9 so use 4 off 3R9, 10W resistors in parallel. Mount them on the PSU case somewhere where the PSU fan will cool them.

• posted

DaveM wrote this in : [...]

[...]

I've tried quite a few ATX psues, from 250 to 1100W, and none has needed the load described to behave properly. So the best advise is to measure output first, to check if it really needs this load to regulate.

```--
MVH,
Vidar```
• posted

Mid post.

thing

Lets

1

Thanks for the very poignant response! However, I do fail to see how what I did differs from what you recommend (i.e. except for the amperage of my resistor (pack) being 2 and the resulting ohm's being 2.5 I don't see any difference as the power supply works with my resistor pack except for the heat showing that I have met that power supply's load requirement (fan works as well) at 2 amps) - will these 1 ohm resistors you recommend run cooler? Thanks very much guys - I plan on going the 1 ohm resistor pack route and see what happens sometime soon....

• posted

Um, that is only an example..

You need to find out the minimum load required to run the supply, then pick the proper resistors. That will let it run while disspating the least amout of heat in the resistors.

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----