Converted ATX power supply load resistor question

[Kilroywashere]

Just a quick one I converted an ATX power supply into a lab bench power supply I followed the instructions from this link .....

My question is this guy has a load resistor of 5 ohms about 5 Watts I think it was...(is this even required to keep the power supply stable if I'm only using the + 5  and + 12 volt rails?) I was going to do the same thing as great Scott but I did not have that particular type of resistor. But what I did have was a three olm 17 watt resistor.

It seems to run okay , and the resistor seems to get a few more degrees warmer than the case of the ATX power supply.

Is this going to cause a problem will this not be stable power supply because I did not get the required olms?

I can provide pictures of the final product if everybody wants to see them....

[Paul Moir]

Those resistors will be fine on the 5v or 3.3v rail.  Wastes a bit of power but that's not important.

I didn't watch much of that video but it seemed like some of the soldering was pretty nasty.  Don't take that as how it should be done.

EDIT:  Oh, and welcome!

[Shock]

Use ohms law to calculate how much current and power is dissipated through the resistor. Then you will know how big a resistor to use (e.g. one with a higher wattage). Since all the power supplies are different it's hard to be precise but aim for at least 1A on the 5V and 500mA on the 12V.

I've not looked at any modern ATX power supplies recently but have noticed inbuilt load resistors on other SMPS. If that is the case they may have smaller minimum load requirements.

[Kilroywashere]

Use ohms law to calculate how much current and power is dissipated through the resistor. Then you will know how big a resistor to use (e.g. one with a higher wattage). Since all the power supplies are different it's hard to be precise but aim for at least 1A on the 5V and 500mA on the 12V.

Okay so here's what I understand so far: I have a 3 Ohm 17 watt resistor and the voltage is 12 volts 12/3=4 I'm going to assume that that's 4 amps. From what I understand that will not be a very good load resistor for the 12 volt rail


5/3=1.6 amps and that would be okay for the 5-volt rail...right?    (The resistor I have is really only good for the 5-volt rail?)

Question: this is considered a dummy load and in the video according to what rail draws the most amperage( in my case it's the five volt rail ) is where you place the dummy load to make the power supply stable during low current draws. Is this really necessary to have this dummy load?

[Shock]

Depends on the supply it's a pretty good idea to place one on both. You can always monitor the output voltage to ensure it's stable. The main goal is to ensure the load meets the supplies minimum load requirements which allows you to then add a very small load up to the supplies limit (I would normally recommend no higher than 80-90%). An old cheap no name power supply would be more of a concern than a quality high performance one.

[jaycee]

Typically ATX power supplies these days regulate based on feedback from the +12V rail. You dont typically need to draw a lot of current though - 100mA is usually sufficient. SO...

(V/I = R ) 12/0.1 = 120 ohms. (P = IIR) 0.01 * 120 = 1.2W. Using a 3W resistor would be best.

Do the same calculations with 100 ohm and you get 120mA of draw, and 1.44W dissipated in the resistor. A 100 ohm 3W wirewound resistor is easy enough to get.

[Shock]

Since there are modern ATX supplies that require higher loads for stability than 100mA and ones that list various rails requiring a load I think it's easier to cover all bases as a precaution.

[jaycee]

It's true that there are ATX supplies which supply higher currents and probably have more complex minimal loading requirements, but those are the sorts of supplies that are where they belong - in a high end PC. Typically if you're converting an ATX supply for hobby use, you're talking about a 250-300W cheapie salvaged from an old PC. The lower end ones are pretty much all TL431 off the 12v rail, so loading the 12v rail as I described is fine.

Of course, YMMV, and you should check the quality of the outputs... especially if for example, your salvaged supply has been around the block a bit and the output capacitors are well and truly past it. You dont want a 5V supply with 500-800mV of noise on it's output

[homebrew]

Hmm - That's a typical DI-Why? project ...

Actually it doesn't make ANY sense at all.
For a beginner it is outright dangerous to perform the conversion. The conversion itself as well as the safety of the device afterwards. You literally depend your life on the thing - especially if used for 'lab' purposes.

1) Mechanical reasons: In a lab you poke around with bare wires, shoot around bits of copper clipped with your side cutters, splash solder etc. etc. That doesn't mix well with the ventilation slots of the ATX-case. It is simply not designed for the environment. If anything it is at best IP20 - finger safe and meant to be mounted inside some other case. 

2) Electrical reasons: Output is mains earth referenced (NO! DON'T EVEN THINK OF CHANGING THIS!) beware of ground current loops; beware of where your oscilloscope ground clip touches. It is quite noisy by design. Only fixed voltages with NO current limiting at all. Ok - it may switches off on a dead short ... after it obliterated ANY circuit AND the breadboard ...

So why would somebody try to do this... In any case doing this properly (safely) might be more expensive than just buying a quality lab PSU and stop worrying ...
There are some breakout boards around that offer at least some protection. However, the mechanical issues remain.

[jaycee]

This is true - I really would not recommend using an ATX PSU as a bench power supply, mainly because there is no sensible current limit. The kinds of current that an ATX PSU can deliver on the 5v and 12v line will roast parts and vaporise breadboard contacts in milliseconds if a mistake is made.

[xani]

I'd say do not use it *directly* as a bench power supply. At the very least fuse it and put some simple voltage regulator like LM317

So why would somebody try to do this... In any case doing this properly (safely) might be more expensive than just buying a quality lab PSU and stop worrying ...
There are some breakout boards around that offer at least some protection. However, the mechanical issues remain.

"mechanical issues" can be solved by some plywood and glue, the electric ones are worse.

I'd say it made sense some time ago, especially if you got one for free. But nowadays when you can get $20 pre-built 30/3A power supply module (that's actually not half bad as Dave tested) and then just connect any power source to it (even just junkbox transformer + diode bridge + few caps) it just isn't worth the effort and problems. Well unless you really need that 30-40 amps on +12 for something

[ChrisLX200]

Maybe use the ATX PSU with the 20$ module providing a front end? Got to be better than a rough transformer, caps and diodes feeding into it. I started out with an ATX supply but don't use it now, seems like a good way of bringing it back into service.

[homebrew]

I'd say it made sense some time ago, especially if you got one for free. But nowadays when you can get $20 pre-built 30/3A power supply module (that's actually not half bad as Dave tested) and then just connect any power source to it (even just junkbox transformer + diode bridge + few caps) it just isn't worth the effort and problems. Well unless you really need that 30-40 amps on +12 for something

Yes, true, as long as you manage to house the junkbox transformer safely ...

Maybe use the ATX PSU with the 20$ module providing a front end? Got to be better than a rough transformer, caps and diodes feeding into it. I started out with an ATX supply but don't use it now, seems like a good way of bringing it back into service.

But still, at least, the issue with mains earth referencing remains ...

[Kilroywashere]

Thank you guys... I know a little about electronics my job requires me to know a little ....  But I did have concerns about using this for a Lab Bench Power Supply .... After reading what you guys said I don't want to use this ....

I will just spend the money on a real LBPS, and be done with it ...

Seems like I'm hitting a dead donkey  with this one...


at the time it seemed like a good idea to make this, but now I am kind of mad that this project is out there for people to try out....

I looked on amazon for a good lab power supply today ... its just not worth making this project when it comes to your safety ....


Thanks again guys ... saved my life



Kilroywashere

[lisafig]

Depends on the supply it's a pretty good idea to place one on both.

Sorry for the basic question, but should we also add a load to the 3.3V?

If not, why?

It's strange because this rail will be the one that will draw a small amount of current (and we add a load to regulate this small current)?

[Shock]

Again it depends on the supply, some smps have a minimum load requirement. If you supply starts up fine and is output stable, then you have no problems.