Are ATX power supplies based bench power supplies "good"?

[cybergibbons]

One thing to note about ATX supplies that hasn't been addressed is the accuracy and stability of the rails. For example, the PSU that I measured for the pic in original post has a voltage output of about 11.8V at OCV. This voltage will change depending on the load put on the supply. If its a small load the voltage might go up to about 12V but a heavier load will drop it down to about 10 or so. I rarely require large loads and 99% of my projects use voltage regulators so for me this is not an issue but it could be for some.

10V is definitely out of spec. I think it's something like +/- 5% or +/- 10%, depending on which rail. Cross loads (eg. max load on 5V, min load on 12V) may cause the 12V to go up, but a properly designed power supply should stay in spec as long as your load is.

Latest spec is +/- 5% for the + rails and +/- 10% for the negative.

[TopherTheME]

10V is definitely out of spec. I think it's something like +/- 5% or +/- 10%, depending on which rail. Cross loads (eg. max load on 5V, min load on 12V) may cause the 12V to go up, but a properly designed power supply should stay in spec as long as your load is.

Yes a "properly designed" power supply will, but a cheap chinese supply I pulled out of the recycle bin probably won't. I think the supply I have on my desk in front of me was made in the late 90's and couldn't have cost more than $30 back then too.

[edude03]

Sorry I haven't had a chance to weight in myself, (being OP) but it seems the discussion has really taken off.

So basically it boils down to: No, a PSU can work in some cases but it isn't the right tool for the job.

What I mostly need is a way to power anything quickly, for example a phone that I have no battery for, or a board I just built. If bench power supplies are only $100-$400 as one poster mentioned, then I'd rather pick one up than risk having problems with my home made PSU.

Thanks to everyone who commented.

[sonicj]

Sorry I haven't had a chance to weight in myself, (being OP) but it seems the discussion has really taken off.

So basically it boils down to: No, a PSU can work in some cases but it isn't the right tool for the job.

What I mostly need is a way to power anything quickly, for example a phone that I have no battery for, or a board I just built. If bench power supplies are only $100-$400 as one poster mentioned, then I'd rather pick one up than risk having problems with my home made PSU.

Thanks to everyone who commented.

if you have a old psu, convert it! add reasonable fuses on the outputs and you have a safe, stable, cheap source for 3.3, 5, 12, -12 & -5v with plenty available current.

"What I mostly need is a way to power anything quickly"
this is the ATX power supply's strong point; you can power anything quickly! with a bench supply, you can power some things quickly. if you are into designing and tuning your own circuits, get a bench supply. if you are into building circuits(kits), arduino, charging batteries, repairing broken electronics, and other general enthusiasts stuff, then the ATX should easily meet your demands.
-sj

[vtl]

I don't know why you are all crapping on PC PSU's, for the money they are pretty dam awesome.

They are they wrong tool for the job. It doesn't matter if they are cheap.

An 860W PSU (a small one by todays standard), can pump out > 60 A continuously at 12 V. If something goes wrong when you connect it, everything gives in. Everything, but your PSU. Your board, your wires, your meter. Wana see you cables glow, your parts catch fire and your board becoming a piece of coal? No problem, these PSUs can make it happen.

Thats a lot of sweeping generalizations there. 860w small? You're crazy 860w is a huge overkill PSU. Around 500w is typical. Hardly any PSUs will pump out 60a on the 12v rail, theyll typically seperate them into multiple 17-20a rails. Besides, who would buy a brand new ATX psu to power their hobby stuff? Usually you grab one from an old comp which will be typically 300-400w (typically 12-17A 12v max)

If I wanna see my cable glow I better go grab some LEDs because if I short anything out, the PSU will automatically shut off because of the short circuit protection. I've poked around and triggered the short circuit protection a few times by accident (computer case is earthed), completely harmless.

Who cares if its the wrong tool for the job?  If it means I don't have to shell out 500$ to power a crappy hobby circuit then so be it.

[silicon_cockroach]

I'm considering building one mostly for use for experimenting and testing. I doubt I need a high end bench power supply but I wonder if a ATX based power supply is good enough?

So question is should I build a variable ATX based power supply or should I shop around for a "real" bench power supply?

This is interesting.Correct me guys if i am wrong.Why not build one regulated power supply from scratch ? It is not hard to do,almost anyone who needs one,can actually build one with some help,most of the guys/girls with no help at all.Just find some old transformer,and suitable box,everything else is so cheep.As mentioned before,AT or ATX supply's are sometimes dangerous,unless you want to power some car amplifier or similar beast.
edude03  if it would help,i can send you pictures and schematic for one simple variable power supply which i built some time ago.It serves me well.If you want , just let me know

[CodeDog]

i built mine into an old ammo case from an army disposal store ... (cheap and rugged)





... but I also use a real bench supply as well

cheers,

[Psi]

PSUs as a bench power supply are a fire starter risk. Do you haven an idea what 15A, 20A or more can do on your bench? Expect red glowing lab patch cables and vaporizing hookup wire, while the PSU still and happily pumps out its juice.

hehe yeah, that's exactly right.

They will totally destroy anything that shorts them out without a 2nd thought

[Zero999]

Yes, you need overcurrent protection for thin cable.

It's probably a good idea to have a power distribution PCB going to a few sockets, each protected with a 10A polyfuse, with a 100nF capacitor for decoupling and an LED indicator so you know which has blown. There could also be a couple lower power sockets with a 1A polyfuse for breadboards which won't take much current to melt.

[saturation]

Nice idea, I presume you ground the PSU to the metal casing?

i built mine into an old ammo case from an army disposal store ... (cheap and rugged)


... but I also use a real bench supply as well

cheers,

[PeterG]

There is an upside to using an ATX SMPS, they are noisy so you can make sure your project can withstand a little noise on its power rail, not a bad thing in my book.

My 2 cents worth.

Regards

[Wartex]

they are all chinese, some just happen to carry a "name" and cost a lot more

You are talking out of your ass. This is my computer PSU http://www.hardwarecanucks.com/forum/hardware-canucks-reviews/5438-pc-power-cooling-silencer-750w-quad-power-supply-review-8.html

16 mV ripple under high load, 30 mV at 60A load. Show me a bench PSU for under $200 that can do 60A at 12V.

It's designed in USA with USA parts, only soldered and painted in China, that doesn't make it chinese. It's a premium PSU will beat some cheap bench PSU in terms of output clarity.

[alm]

16 mV ripple under high load, 30 mV at 60A load. Show me a bench PSU for under $200 that can do 60A at 12V.

Yes, if you need lots of power at 12V (grounded) cheaply, it's hard to beat a ATX PSU. Most applications don't need bench supplies that can deliver 60A, however, especially in these days of low power design (how long is your iPhone battery going to last if the phone would draw 60A?). And 60A is a lot of current to kill your circuit if you make a mistake, it will easily vaporize bond wires in most ICs.

It's designed in USA with USA parts, only soldered and painted in China, that doesn't make it chinese. It's a premium PSU will beat some cheap bench PSU in terms of output clarity.

Do they still make components in the USA? Sure, some expensive/exotic parts, but what about standard passives or jellybean analog ICs? The kind found in ATX PSUs?

You have to screw up pretty bad to get a horrible output from a linear regulator. Even using a 78xx without any caps should perform OK.

[ndictu]

I don't like the term "Chinese". It's a huge country, with diverse fabrication houses. Sometimes the expensive, branded thing and the cheap knockoff are from the same production line.

The way it works is some company hires chinese production house for say 8 hours a day to make 100pcs of stuff. This is all made according to their plans, tested and everything. But they probably make a few pieces that won't meet the QC but still could be sold under price. Or, from what I've heard, the entire clothing industry works like this: they make a deal for 8 hours, make that and sell it,but really work for 12 hours, make the exact same stuff but without a brand logo and sell that too, with more profit margin for them.

Made in China is stupid, it can be the best, most reliable product you own and it can be a teddy bear stuffed with asbestos.

[saturation]

You are right since for the most part many good goods we buy are also probably made in China, but its changing as we speak.  A problem has been that China has led the world in violating intellectual property, because those factories making products identical to those ordered by the contractor but sell it outside of the contractors supply chain or lack of an official label is still intellectual theft. 

A process known as the 'ghost shift'.

http://money.cnn.com/magazines/fortune/fortune_archive/2006/05/01/8375455/index.htm

China is the #1 source of manufacturing of counterfeit goods in the US, for 10+ years in a row. 

http://www.cbp.gov/xp/cgov/trade/priority_trade/ipr/pubs/seizure/

There is an entire chapter in Wikipedia regarding IP theft in China.

http://en.wikipedia.org/wiki/Intellectual_property_violation_in_the_People%27s_Republic_of_China

By comparison, when Japan and Taiwan were emerging in the electronics field, their enormous output in the 1970s through 1990s were not blemished by counterfeit products from the same factories, and moreso, not to the level of individual components, unfinished products or parts, compared to China. 

Hong Kong, when they were a protectorate of the UK prior to 2000, was a common source of counterfeit goods, but it was not to the scale that, now under China, they are #2 in the world,as a source for counterfeit.

I don't like the term "Chinese". It's a huge country, with diverse fabrication houses. Sometimes the expensive, branded thing and the cheap knockoff are from the same production line.

The way it works is some company hires chinese production house for say 8 hours a day to make 100pcs of stuff. This is all made according to their plans, tested and everything. But they probably make a few pieces that won't meet the QC but still could be sold under price. Or, from what I've heard, the entire clothing industry works like this: they make a deal for 8 hours, make that and sell it,but really work for 12 hours, make the exact same stuff but without a brand logo and sell that too, with more profit margin for them.

Made in China is stupid, it can be the best, most reliable product you own and it can be a teddy bear stuffed with asbestos.

[westfw]

No, a PSU can work in some cases but it isn't the right tool for the job.

That's about it.  A "real" bench has variable voltage, variable current limits, and meters.  A low-end supply will do up to 30V at 3A and cost less than $200, which is enough for most practical purposes, and a very useful piece of equipment to have (useful for charging most kinds of batteries, too.)
A computer power supply has its uses as well, but they're different, and it's NOT a "bench power supply."

[AlsInd]

in my professional pc build & repair days i was always checking my builds with my scope. even the best pc psus are too noisy for precision electronic lab use but the majority of pc psus out there are cheap crap and really noisy. it is cheaper to buy a used lab psu from ebay than buy a high quality pc psu which still will not give you as clean a power. then you have to also add some voltage and current variable circuitry to it... it just does not make sense. you can get good enough lab psus used for under $50 on ebay... i just got another one, as if i needed it!

cheers :-)
al...