ATX psu turned into an adjustable voltage bench supply

Here’s a fancy way to convert an ATX powers supply into a bench supply. [TG] didn’t just cut off the motherboard connector and add banana plugs, but improved the functionality. Right off the bat you’ll notice that he’s added a control panel. There is an Ammeter and Ohmmeter to let you know what the unit is putting out. He added an MIC29152WT adjustable voltage regulator so that he’s not limited to the fixed voltages of the psu. As a final touch he added an external voltage probe which can be used with the flick of a switch. It’s no replacement for a proper bench supply, especially since it doesn’t have adjustable current limiting, but it’s a nice improvement upon previous psu hacks.

Comments

  1. Fallen says:

    I’ve been using ATX power supplies to make adjustable power supplies for a while now. But I have never been able to make one look that good! Kudos! That and I only had an LCD readout for voltage, not amperage. All in all, this is a very well done hack. Once he gets the side grills on, I’ll have nothing to complain about.

  2. Fallen says:

    Excuse the double post. When I’ve made adjustable PSUs from ATX supplies, I do it by putting a pot in the feedback path of the SMPS. Not by putting a LDO voltage regulator between the output rail and the binding posts. This also lets me get more than 12V out. But make sure the other rails have upgraded components if you’re running it at more than 12V.

  3. Miroslav says:

    @Fallen, could you post more details on your variance with tweeking the feedback loop of the switcher.

    I suppose that works to some extent, even though very limited (define limited :) ) as there is a very tight relationship between the input voltage to the switcher, output voltage, current and the serial inductor.

    I suppose that you have PSU with ferite traffo and optocoupler in the feedback loop, but I suppose similar restrictions exist there too

    Regards!

  4. jeditalian says:

    verynice!(in Borat voice)

  5. andres says:

    Ohmmeter?

  6. mjrippe says:

    Ohmmeter ≠ Voltmeter

  7. BiOzZ says:

    adjustable PSU power is nothing new but there is an easier way to do it … simply remove all the caps and inductors from the low voltage side than find the feedback voltage divider and by changing the value of the divider you change the voltage … you can get up to 47 volts out of it (well my 500W can do that) … you need to put a higher voltage filter cap and higher watt inductor to keep the noise low

  8. Doug says:

    Great hack but all who are going to attempt this, please be aware that you can only source 1.5A off the variable output. This is a limitation of the regulator chosen

  9. strider_mt2k says:

    This should be a kit.

    Maybe some instructions on how to tweak the supply and an ATX connector, make it bolt-on to existing holes.

    Plug in and win! (you can use that)

  10. ???? says:

    @ Fallen

    I’ve also tried making my own and it’s usually amateur. This one looks great.

  11. Robert_A says:

    I’d be interested in a kit. something like a fancier (metered) version of this:

    http://www.sparkfun.com/products/9774

  12. Robert_A says:
  13. George Johnson says:

    Ditto the “Ohmmeter” comment, Ohmmeter?

  14. strider_mt2k says:

    Ohmmeter, why hast thou forsaken us?

  15. Ken says:

    By cascading the output of the ATX PSU into an adjustable linear regulator, he is pretty much eliminating superimposed voltage ripple from the SMPS section. I would be damned if there was 1mV voltage ripple left.

    Just messing with the feedback circuitry is a good way to throw the power supply out of regulation.

    Not to say that that technique doesn’t have it’s place, but if someone is after something akin to a lab supply, the linear cascade route is going to be better.

  16. Fallen says:

    Agreed Ken. For my application I didn’t really need a rock steady voltage with minimal current. I was using the PSU to power a car audio amp that I was repairing. Hence I wanted 14.4V.

  17. Ekaj says:

    If you really want more current out of it, just add some power transistors… I’ve got a stack of TO-3 packaged ones on one of my home built units, I’ve never tested the max current on it, but it regularly delivers ~6-7A with no problems. It’s powered by a 48V switcher from a dell tape autoloader. Works like a charm, and the autoloader supply is about the size of your average CD drive, just a little bit longer than one. Has 2 internal fans, it’s been working for more than a year without trouble!

  18. glagnar says:
  19. strider_mt2k says:

    Power supplies and power supply hacks are ALWAYS awesome.

    I forgot to mention that this project is indeed beautifully carried out.

    Well done!

  20. smoker_dave says:

    I love the flying voltage probe lead. Awesome!

  21. iPodder says:

    People have been doing this for ages, but this is really amazing; the case is very professional looking and doesn’t take up a ton of space like most other psu power supplies.

  22. Peter says:

    I modded a PC supply to create a high current 13.8V supply for a mobile radio. If you’re careful, you can get almost the rated power out of them.

    The hard part is finding a schematic, because it’s a lot easier than spending your time tracing the PCB etches and drawing your own. I discovered, though, that they are all basically the same design, and once you identify the regulator chip, you can find sample applications schematics on the chip manufacturer’s website.

    The key for getting high current output is to use the 5V or 3.3V output circuit. Sometimes there are two of these, and you can combine the transformer secondary windings in series to boost the available output voltage.

    The other tip is that usually, only the main logic voltage (3.3 or 5) is regulated (a sample of the output voltage is fed back to the regulator chip). The other voltages are designed to track the logic voltage and aren’t regulated themselves. So, if you’re looking for precise voltage control, use the main outputs.

    These don’t make great adjustable supplies, as the output voltages are designed to be fixed, with varying loads. However, with a schematic and a will to tinker, anything is possible. One thing’s for sure: you’ll come away from the experience knowing a lot more about how switching supplies are designed!

    #include [warning about messing with line-powered stuff]

  23. Peter says:

    Well, it turns out that getting a schematic is as easy as Googling “ATX power supply schematic”:

    This may not be the *exact* schematic of the supply you have, but I guarantee it’s probably close.

    Just a quick look at this image will tell you a lot about how to mod the supply. First, the “TL494″ chip is used in almost all of them, and the TI website has loads of helpful information. Pin 1 of this chip is obviously being used as the regulator feedback point in this design, so varying the voltage on this pin, or the voltage divider attached to it will change the output voltage. Note that the +5 and +12 outputs are both fed back to the regulator. The bottom left corner has a standby 5V supply for the soft power switch, and there should be an overcurrent detect circuit in there somewhere…

    One piece of advice I got when doing my design, was to strip out all the filter and diode circuits connected to the PSU DC outputs, and replace it with only the parts you need for your desired output voltage. If you are lucky, they didn’t try to save money by varying the size of the wire used in the secondary windings and you have four high current windings available to tap off of as you need to.

    Again, notice from the filter capacitor voltage ratings that there’s upwards of 500VDC of rectified AC line voltage on the primary of that output transformer, and treat it with the respect it deserves.

    As these supplies can often be had for the price of asking for them, at your local “old PC” recycling facility, they’re a very cheap way to experiment with switching supplies. You’ll save reverse engineering time if you can manage to get several of the same design. If the worst happens, and your project goes up in a shower of sparks and smoke, just set it aside and pick up another supply.

  24. Ragnar says:

    @Peter: your warning about threating it with the respect it deserves is very welcome. I’d say comparing this to something everybody would understand is: if you touch a capacitor loaded at 500V this would be like a punch from Mike Tyson at is best right into your face.

  25. BiOzZ says:

    @Peter
    i have only seen one TL494 PWM circuit used in a PSU out of atleast 30 i have ripped apart i dont know where you got that the idea that its used in almost all of them

    only cheep ass ones or old ass ones are actually used with that as most of them have a different transformer for each voltage rail requiring separate regulation

    most common ones use a quad comparator to regulate the voltage (either built in to the PWM or used sepretly)
    by modifying the divider’s in voltage at 12V you can make it go above 12V (as high as 48v) and by modifying the reference voltage you can make it go down (i dont suggest doing both on the same circuit) all the way to 0v

  26. Peter says:

    @BiOzZ: well, given my usual method of acquisition, I’m pretty much guaranteed to get old-ass ones! Another schematic I found on the web used an SG6105.

    Whatever the PWM regulator chip used, the manufacturer’s website is the best place to start when looking for a schematic. Once you understand how the regulator chip is supposed to work, you can start to figure out how those principles are implemented on the circuit board in front of you.

    In any case, these power supplies are very hackable, and well worth the effort, both for the resulting supply and for the knowledge you’ll gain.

    Just watch out for the primary voltage. Being plugged into a GFI outlet while testing and keeping one hand behind your back when the supply is on, are both good ideas.

  27. Concerned EE says:

    I hope this article isn’t implying that Hackaday doesn’t know the difference between an ohmmeter and a voltmeter. :(

  28. That Guy says:

    No… it’s implying that Mike Szczys doesn’t know the difference between an ohmmeter and a voltmeter. Or that he made a simple mistake and that hackaday doesn’t proof-read their posts. :(

  29. tgtomm says:

    Thanks for all the great comments people.

    Thanks a lot to all the people who discussed tweeking the actual psu circuit, I didn’t even consider doing that. I’m only getting started and haven’t done anything like this before so didn’t want to go too overboard.

    I am considering adding a higher current regulator so that the variable output can give a bit more. I’ve ordered one for my parts bin so might get around to doing that if i need it.

    I’m already getting ideas on how to do things differently. I’d probably start by getting the psu to give a higher voltage output then regulate it down for my needs. Anyway, i’ll save that for another project.

    thanks everyone

    #cough# ohmeter #cough cough# voltmeter ;-)

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Follow

Get every new post delivered to your Inbox.

Join 96,434 other followers