ATX bench power supply

[cryptlabs]

Hello,

I'm trying to build a bench power supply from an ATX power supply. I have a question about the risk of getting shocked from the output of the PS, Is it dangerous to touch let's say the 12v output with all that high amp rating coming out ?

Regards

Sent from my SM-N915F using Tapatalk

[tomlut]

No. Watch this for an excellent explanation: https://youtu.be/8xONZcBJh5A

[tomlut]

Also: https://youtu.be/XDf2nhfxVzg

And https://youtu.be/hp97GjuULX8

[MarkF]

On dry skin, you won't start feeling a shock until the voltage gets up beyond 24 to 28 VDC.

If you just want a plug-and-play solution, check out this ATX Breakout Board for $14.00.  It has an on/off switch and 1.25A poly-fuses on each output for current protection.

[Tim T]

Here is another option that also has two USB connectors as well as support for either Dell or generic ATX, available as a kit or pre-assembled.
Easy to solder, just be aware that with (some?) Dell supplies you will not get out 3.3v.

http://www.mysticengineering.com/atx.adapter/index.html

[eilize]

the resistance of your body(300<>50000ohm, take 2kohm to understand why you don't risk anything) limit the current

[cryptlabs]

Thank you all for your help, but if i use the breakout boards with the fuse i will lose the current output.

Sent from my SM-N915F using Tapatalk

[MarkF]

And that's a GOOD thing.  You should be able to do most projects with 1A.  Especially if you're hooking up things to a microcontroller and testing individual pieces and their corresponding code.  I have done a lot with less than 0.5A from a USB port.  The power breakout boards provide a switch to turn the power supply on and off.

If you need more power later, you have the extra cables normally used for the disk drives providing the full power of the supply.

In most cases you will want the current protection instead of melting wires and burning holes in your breadboard.  Especially since this all seems like it's new to you???


Take the typical 0-30 VDC bench power supply as a guide.  They will only have a 3-5 A current output.  Not the 20 plus amps the ATX power supply will provide.

[homebrew]

Two things to keep in mind:

1) As far as I know all the ATX power-supplies have a low resistance path between GND and mains earth.
-> Pay attention when connecting external instruments like oscilloscopes or the like.

Back in the days when I was a teenager I once turned my grounded soldering iron into an arc welder that way 

2) Use at least a name band one. Really! I wouldn't want to hang my life on one layer of mylar tape ...

Seriously I don't think that an ATX-psu makes an acceptable power supply. It's like using an axe in the kitchen to chop onions. Yeah, maybe doable but not a good idea either ...

Just my 2 cents

[eilize]

"Seriously I don't think that an ATX-psu makes an acceptable power supply. "

why?

ripple? 2% is not so bad
noise? everybody use wifi actually,
so the question is : Is that noise is a problem or not? does it make your circuit unstable?

 a hobbyist don't play with high frequency signal, buying a true lab psu is like using a scud to kill a moskito


a atx psu can be a problem when you want test a circuit when you didn't seriously think about how it would work...
trowing some ic on a broadboard hoping it work
testing blackbox

the best tools you can buy is a pen and some sheets of paper(ton of paper) 

[homebrew]

"Seriously I don't think that an ATX-psu makes an acceptable power supply. "

why?

Well, as I (and others, also in a number of similar threads) said:
1) Mains earth referenced
2) Absolutely no current regulation (yes, sure it will shut down above i.e. 20A, so there is a slight chance it will detect a dead short before meltdown...)
3) Needs a dummy load for stable regulation
4) Only fixed voltages

5) The procedure of modding the supplies often is dangerous in itself. I've seen a lot of these modes, where they attach binding posts directly to the sheet-metal case of the supply. Here often the low voltage wires touch the (not-insulated) hi-side heatsinks. 
I.e. in this video () at least some decent heat-shrink tubing is applied to address this problem. However i really wouldn't recommend drilling the case while the PCB is inside it  Mhhh, metal shavings wandering around inside the PSU ... AWSOME ...

[eilize]

1) Mains earth referenced

that's why you have plastic on bananas connector

2) Absolutely no current regulation (yes, sure it will shut down above i.e. 20A, so there is a slight chance it will detect a dead short before meltdown...)

it has a protection against short-circuit
you can put a resistor to limit current

3) Needs a dummy load for stable regulation

10-30ohm  on the 12v rail  (5v rail for the very old psu)

4) Only fixed voltages

not a problem to put a buck converter
not a problem to make one with mofset and pic

there, you'll see the difference between who invest a lot in electronic-electrotechnic and other who just follow a tutorial finded on a random website

5) The procedure of modding the supplies often is dangerous in itself. I've seen a lot of these modes, where they attach binding posts directly to the sheet-metal case of the supply. Here often the low voltage wires touch the (not-insulated) hi-side heatsinks.
I.e. in this video () at least some decent heat-shrink tubing is applied to address this problem. However i really wouldn't recommend drilling the case while the PCB is inside it  Mhhh, metal shavings wandering around inside the PSU ... AWSOME ...

i agree, it's crappy..

but you can buy a old pc, trow  the motherboard, hd, and so on , keeping the tower and the psu.
you'll have a lot of place to put this dummy resistance, a lcd  etc.

[Dave]

1) Mains earth referenced

that's why you have plastic on bananas connector

What does that have to do with anything? You are not going to have a floating power supply. That's it.

2) Absolutely no current regulation (yes, sure it will shut down above i.e. 20A, so there is a slight chance it will detect a dead short before meltdown...)

it has a protection against short-circuit
you can put a resistor to limit current

Sure, it detects a dead short that pulls 30A. Good luck passing that through a breadboard.
Putting a limiting resistor in series is a pretty crappy solution, because that means your power supply no longer does its job properly, which is regulating the voltage regardless of load.

there, you'll see the difference between who invest a lot in electronic-electrotechnic and other who just follow a tutorial finded on a random website

Modding an ATX power supply is exactly that: Following an awful tutorial that some idiot thought of. Do yourself a favor and invest in a proper power supply, it will cost less in the long run.

but you can buy a old pc, trow  the motherboard, hd, and so on , keeping the tower and the psu.
you'll have a lot of place to put this dummy resistance, a lcd  etc.

Nice, compact solution you got there, mate.

[Christopher]

Prepare for your breadboard to smoke. If you must use one, put a series re-settable fuse in and beware even that might go pop

The only real way for your breadboards to not smoke is to get some kind of current limiting power supply.. What about a few AA batteries in series and a linear regulator?

[FlyingHacker]

I smoked a lot of stuff before getting current limited supplies. Look around of eBay for an old Lambda or something. I have two of their triple output models (LPT-7202-FM) I fixed up, and they work great. They vary wildly in price, but I got mine for ~30 + shipping. Acopian makes some decent bench supplies with current limiting as well, again old, used stuff on eBay. Sure, these are all garbage compared to modern digital HP or Keithly, etc. power supplies, but they fit the hobbyist budget if you look around on eBay, and make lowball offers, etc. I made a few lowball offers and eventually people accepted them.

You can also get a cheap Chinese supply for around $40 or so on Amazon or eBay. They are not fantastic, but they usually do the trick, and they have snazzy LED displays vs. the older analog meters. I actually use the Lamdas far more (set one up to do dual tracking supplies for op amps, etc. one positive, one negative). You can never have enough power supplies...

[homebrew]

I smoked a lot of stuff before getting current limited supplies.

Yeah, sure you did ... 

Sorry, I just had to!

[eilize]

Sure, it detects a dead short that pulls 30A. Good luck passing that through a breadboard.
Putting a limiting resistor in series is a pretty crappy solution, because that means your power supply no longer does its job properly, which is regulating the voltage regardless of load.

Quote

    there, you'll see the difference between who invest a lot in electronic-electrotechnic and other who just follow a tutorial finded on a random website

Modding an ATX power supply is exactly that: Following an awful tutorial that some idiot thought of. Do yourself a favor and invest in a proper power supply, it will cost less in the long run.


Quote

    but you can buy a old pc, trow  the motherboard, hd, and so on , keeping the tower and the psu.
    you'll have a lot of place to put this dummy resistance, a lcd  etc.

Nice, compact solution you got there, mate.

i don't see difference between this kind of idiot "Modding an ATX power supply is exactly that: Following an awful tutorial that some idiot thought of"
and the one doing idiot when hoping he didn't do something wrong on his breadboard when he put current on it


and..who care of compact  outside iphone fanboys ^^
you need space on your bench...not just enough space to put your breadboard on 



dave make a video where he fire an op amp, forgeting he didn't put protection to limit current..
where did you gain time and money in this case where your circuit must has a real use, outside a breadboard

[homebrew]

i don't see difference between this kind of idiot "Modding an ATX power supply is exactly that: Following an awful tutorial that some idiot thought of"
and the one doing idiot when hoping he didn't do something wrong on his breadboard when he put current on it

dave make a video where he fire an op amp, forgeting he didn't put protection to limit current..
where did you gain time and money in this case where your circuit must has a real use, outside a breadboard

Look! Even Dave is a human being! Who did thought. And he shares that with us to show that even professionals make stupid mistakes sometimes ... we all do!
But still better having an opamp wasted than an opamp + the breadboard + whatever.

According to your logic you could also argue to use a car battery. Then you have some 600A of fault current. So you might stand the change of being blinded by the arc, getting burns, burn down the house, the darn thing may explode due to hydrogen generation and shoot sulfuric acid everywhere etc. etc. etc.

But hey, if you know what you are doing ... 

In my eyes this distinguishes real professionals from experienced makers that became over-confident. Real professionals keep all their safety gear in place although they exactly know what they are doing...

One can find this behavior quite often. People don't use ear protection when using the grinder, don't put their safety glasses on when working on the drillpress etc. etc.

Regardless of any other arguments: YOU PUT YOUR LIFE ON THAT SUPPLY. Literally! How often are you fiddling around with the circuit powered? If something goes wrong (i.e. due to metal shavings inside the ATX-PSU) you will have a bad time.

In general the other arguments are more a question of practicability. As you said, often it is challenging enough to get the circuit behaving correctly. You don't want to have extra problems because of your gear. And I would have my strong doubts that these supplies are designed for the type of lab use at all. I.e. different types of load (capacitative, resistive, inductive), transients, reverse current etc. etc.

Like I said in one of the earlier posts: You can chop an onion with an axe but it still might not be the preferred tool for the job ...

[Dave]

i don't see difference between this kind of idiot "Modding an ATX power supply is exactly that: Following an awful tutorial that some idiot thought of"
and the one doing idiot when hoping he didn't do something wrong on his breadboard when he put current on it

If you are doing anything more than 3 components, screw ups aren't that uncommon. Silly mistakes which can end up destroying expensive components. If you claim that these never happen to you, you must be either lying or you just haven't spent enough time behind the bench yet.

and..who care of compact  outside iphone fanboys ^^
you need space on your bench...not just enough space to put your breadboard on 

Check out this thread. Pretty much everyone appreciates compact lab equipment.

where did you gain time and money in this case where your circuit must has a real use, outside a breadboard

I didn't quite understand this question. Please elaborate.

[eilize]

i agree  homebrew!!!

but..

If something goes wrong (i.e. due to metal shavings inside the ATX-PSU) you will have a bad time.

i drill the tower, not the atx psu

Check out this thread. Pretty much everyone appreciates compact lab equipment.

a tower has a surface close to the one a lab psu has
it's just higher

According to your logic you could also argue to use a car battery. Then you have some 600A of fault current. So you might stand the change of being blinded by the arc, getting burns, burn down the house, the darn thing may explode due to hydrogen generation and shoot sulfuric acid everywhere etc. etc. etc.
you don't care of that 600A of fault current if you do the thing correctly.

dave did his circuit, thinking there was a limiting current part outside.


once you fire something because of that kind of error, you try to don't do it a second time
thinking a lot before doing anything
don't play with blackbox like with lego

" If you claim that these never happen to you, you must be either lying or you just haven't spent enough time behind the bench yet."
i fire some, so i prefer thinking about before acting
how the circuit would work
how much is the impedance input
you don't know?=> blackbox=> lab psu
you don't have? don't touch or limit the current+make some protection before use it
you don't know how to? don't touch it

let the collegue fire it

I didn't quite understand this question. Please elaborate.

thing seem to work correctly on your board, your psu limit current, but the  environment is not the same once you leave the board...

what if your collegue  didn't put something to limit current..or you forget
or your colegue "short-circuit" some protection( i saw some short circuit fuse because it fire too often....)


you see where  i want to go?

you need to limit current, put it on the board, on the same circuit, not outside
my point of view


you need lab psu when your work on a circuit you don't know, can't know, a blackbox


anyway , you're hobbyist (?), you work on your own circuit,you don't play with blackbox
you play sometimes with scavaged part but you HAVE the datasheet



a lab psu give you a second chance to do something wrong

[homebrew]

i agree  homebrew!!!

Ok, then maybe the root problem is that English is neither mine or your mother tong.
Because otherwise this post has sooooo many BAD attitudes to it that it is beyond words :-(

i drill the tower, not the atx psu

That was only ONE extremely obvious example of dangerous mistakes that could happen.

you don't care of that 600A of fault current if you do the thing correctly.

Ok, that's up to you.
I do ...

i fire some, so i prefer thinking about before acting
how the circuit would work
how much is the impedance input
you don't know?=> blackbox=> lab psu
you don't have? don't touch or limit the current+make some protection before use it
you don't know how to? don't touch it

let the collegue fire it

Doesn't work because you cant assess the input impedance of a system with a static test. 

thing seem to work correctly on your board, your psu limit current, but the  environment is not the same once you leave the board...

what if your collegue  didn't put something to limit current..or you forget
or your colegue "short-circuit" some protection( i saw some short circuit fuse because it fire too often....)


you see where  i want to go?

you need to limit current, put it on the board, on the same circuit, not outside
my point of view


you need lab psu when your work on a circuit you don't know, can't know, a blackbox

No, I don't see where this is going at all ...
I wouldn't put any device into regular operation whose safety depends on the supplies current regulation capabilities.

anyway , you're hobbyist (?), you work on your own circuit,you don't play with blackbox
you play sometimes with scavaged part but you HAVE the datasheet

Yes, and I'm proud of my hobby and what I can do with it. The fantastic thing with hobbies is that you can take them as serious as you like. You have the luxury to do things the right way. Hobby doesn't mean that you have to ignore the last 100 years of research and professional experience.

[Daniel_Reyes]

The people saying don't use atx psu are only saying it because they care. Lessons learned the hard way. If you want to be stubborn, we can only hope you are ok in the long run enough to have a second chance to do it right.  Get a current limited supply

Sent from my SM-G900P using Tapatalk

[Landrew2390]

I'm going to have to side with the people saying don't use an ATX supply without some kind of over current protection.  I have ATX power supplies operating a few projects, but those projects were proven using current limited power supplies first.

A big ATX PSU can push a massive amount of current without tripping the OCP that is built into the power supply.  20 amps is a lot of power rushing through a prototype if something fails.  I just feel more comfortable knowing my bench power supply will prevent a minor failure from becoming a major situation.  Once the design is proven, I have no issues running it from an ATX power supply.  Until then, I'm going to err on the side of caution.  That's a personal decision, but I don't think it's one that should be made lightly.  If you're comfortable with the risks, by all means go ahead.  Myself, I know the risks, understand the risk and I won't run anything of a power supply without some kind of over current protection set to a low level.

As always, it's your money and your prototypes.  Do what you feel comfortable with, but make sure you understand the risks first.

[fivefish]

I'm trying to build a bench power supply from an ATX power supply. I have a question about the risk of getting shocked from the output of the PS, Is it dangerous to touch let's say the 12v output with all that high amp rating coming out ?

I say leave the ATX power supply for your computers.

Instead, grab some power transformers, bridge rectifiers, capacitors, LM317s, pots, resistors, and terminals and build your own bench PSU. You'd learn more in the process. It's not that hard... and safer too.

[cryptlabs]

Dont fight guys, we are all here to learn. Im not new to electronics i just wanted to try and mod old atx power supplies that i have and i also think that atx power supplies are not safe.  Thanks for all the help. I have a bench power supply for my work.

Regards

Sent from my SM-N915F using Tapatalk