Bench power supplies can sometimes be frustratingly expensive and also kind of limited. If you’re enterprising and creative, though, you can create your own bench supply with tons of features, and it doesn’t have to break the bank either. Do what [Maker Y] did—grab an ATX supply and get building!
ATX power supplies work as a great basis for a bench power supply. They have 12 volt, 3.3 volt, and 5 volt rails, and they can supply a ton of current for whatever you might need. [Maker Y] decided to break out these rails on banana plugs for ease of access, and fused them for safety, too. But the build doesn’t stop there. [Maker Y] also added a buck-boost converter to provide a variable voltage output from 1 to 30 volts for added flexibility. As a nice final touch, the rig also features a pair of USB A ports compatible with Quick Charge 3.0, for keeping smart devices charged while working in the lab.
[Caelestis Workshop] also designed a fully enclosed version if you prefer that style. Check it out on Instructables.
No matter which way you go, it’s a pretty simple build, with a bunch of off-the-shelf parts tossed together in a 3D printed housing. Ultimately, though, it’s got more functionality than a lot of cheap off-the-shelf bench supplies. You can build it just about anywhere on Earth where you can get cheap eBay parts via post.
If anyone is going to take inspiration or instruction from this, please think twice a few times. Computer power supplies are not designed for this (mis)use. I’ve directly seen several students and colleagues get seriously injured using such “hacks”. Designing a good and safe power supply is not trivial. Excellent, inexpensive, safety-rated benchtop supplies designed for laboratory and experimental uses are widely available and should be used. Yes, I too made something similar in college 20 years ago, but I simply can’t recommend that anyone else do the same. Your life and the lives of others is simply not worth the negligible cost savings.
“Seriously injured” usually means sent to the hospital and staying for days or weeks. Specifically how did your students and colleagues get seriously injured from the DC side of an ATX power supply? Please share so we can avoid their fate.
Indeed. Burns have been the most common injury I’ve seen. One notable burn required skin grafts over a hand, forearm, and face due to an internal short circuit outside of the fuse’s protection catching the supply, the bench, and the clothes of the person at said bench on fire. Those computer PSUs can put out a lot of amps. It went from cold to red-hot in seconds. Using a smaller PSU can help a little, but the power resistors needed to idle the supply can fail easily. I’ve seen a couple 20W resistors go up in flames (luckily no injuries there). You can’t take the low-voltage high-current lines for granted.
I’ve also seen an electrocution due to failed internal grounding. That particular incident required the use of the lab’s AED after some internal electrical tape came loose and contacted the case and someone grabbed it firmly with both hands. The lesson there was never rely on something like tape to insulate. Always use properly rated termination methods. Never leave any wires dangling. Always physically isolate high and low voltage sections. Always use a GFI socket. Never operate it while open.
I’ve seen cheap, falsely rated Amazon/Ebay/Alibaba DC-DC converters explode without warning and startle people working nearby on other tools that you don’t want to be startled while using. Also, all of those incidents destroyed the circuit under test too. Buying parts outside of a trusted distribution chain is like rolling loaded (not in your favor) dice. Doing the research and buying quality components from trustworthy suppliers is recommended.
Lastly, I’ve seen 3d-printed plastic components that are not chemically compatible with the wire or other components inside lead to other failures due to softening at room temp. I’ve seen outgassing from resin prints destroy line filter capacitors, discolor circuit boards, and turn wire to sticky good. If 3d printing must be used, check the material compatibility.
Literally all of those issues except the “3D printed” part would happen *inside the computer* too. It’s not a case of it “not being designed” for it, it’s someone mucking about without having experience at all.
If the grounding in a PC power supply fails, you’re in just as much danger as if you’re using it as an external supply. And nothing’s preventing an ATX supply from causing a fire (or catching on fire itself!) while supplying hundreds of watts *normally*.
This is like saying “don’t use an oven, you might get burned.”
Every one of the failures I mentioned happened outside of the contained computer PSU and within the bench supply that was hacked around it. Computer PSUs are fine for their intended use. Making a naive bench supply from one that could be subjected to unusual abuse may not be fine. A lot can go wrong with high currents. There are better options.
I thought it was fine enough 20 years ago to make one myself. I used it in college. I’m sympathetic to the “that’s cool” and “bench supplies are expensive” sentiments. Experience of witnessing other get injured from similar devices changed my mind.
I was warned. I didn’t listen. Others got hurt. I’m just passing on the warning and alternatives.
“Every one of the failures I mentioned happened outside of the contained computer PSU and within the bench supply that was hacked around it.”
Do you want examples of failures that happened *within* the contained computer?Just google “power supply caught on fire” or “shocked by computer power supply”.
You’re pretending that the problem is somehow moving the supply outside rather than “using electricity.”
Kind of crazy he names all of these examples and you still disagree with him. Using a psu in a computer involves plugging in fitted connectors into components that have been tested to comply with safety regulations. Using a psu in a homemade power supply requires exposing wires and varying complexities.
“Using a psu in a homemade power supply requires exposing wires and varying complexities.”
There are exposed wires in a computer normally!
I’m not disagreeing power supplies can kill or start fires. They do that inside or outside of a computer!
Like I said it’s like saying “don’t use an oven to warm something up, you might get burned, only use it for cooking.” You might get burned period!
These computer power supplies are basically death traps. I’ve seen flame throwers do less damage to unsuspecting students. On a side note a Chevy bolt caught on fire the other day. It literally took out most of Missouri. You can’t really tell because Missouri just kind of looks like that. But at least 2 billion people died I saw it myself.
Thank you 😊 Best reply of this entire thread.
I’m in Missouri and use a repurposed ATX power supply for my R2-D2 when it’s on standby. I tested the voltage on all of the wires and have an Arduino Nano controlling the fan and monitor the PG signal wire. The only thing I hadn’t considered before reading this article is the current, which I’m learning about while building my lightsaber because it’s hard to find the regulator I need to power 2 LED strips on a lipo battery.
YUP, drop a #6 screw across the output of a PSU….
“seen an electrocution” – they died? Electrocution means death by electric shock. Normally that’s pretty rare unless they touch the input side of the power supply. 12v tends to run over the surface of the body – 120v runs through it and can stop the heart.
Electric shock (non-lethal) isn’t anything to laugh about – the 12v side of the ATX power supply can deliver 60 amps, which is more than enough to start fires, melt metal conductors such as screwdrivers, and cause serious burns. And the power supply contains capacitors that can discharge and cause severe injuries. It definitely has the potential to be lethal on several fronts.
It’s not the volts it’s the amps that kill. Also DC is much worse then AC. When the AC dips you have a slight chance to get free. DC on the other hand you’re just going to keep tighting tighter and tighter.
Also there are extremely high voltages inside of a SMP. There is no current protection in cheap ATX power supplies. Also they are earth referenced so you can’t safely use test equipment that is earth ground referenced or you will short things out and dump a ton of current into the test equipment such as a scope.
Dunno why people want 20amps to run an Arduino to blink an LED anyway.
“It’s not the volts it’s the amps” is the siren song of someone who knows nothing about electrical engineering. Even at 30A the 12v isn’t providing enough electrical pressure to overcome skin resistance and deliver the Amps. Yes the amps are what kill but ignoring basic electrical principles is just silly. 12v can’t deliver even 1 A instantly and kill you.
A little internet searching unearths claims that 12 volts can kill under very unfavorable conditions: large contact area wetted with salt water, current flowing through the chest causing either paralysis of breathing muscles or ventricular fibrillation.
The problem is not that you will get electrocuted but that 30A (and many of these supplies are capable of 80+ – it could probably crank a car!) is enough to turn plenty of things into a pile of red hot molten slag that will burn you and set things on fire.
Ignoring that is equally bad as not understanding Ohm’s law.
“is enough to turn plenty of things into a pile of red hot molten slag that will burn you and set things on fire.”
This is a problem with *any* supply capable of that much current. If you’re suggesting “don’t use a high-current supply unless you know what you’re doing”…
well… yes…?
Other things that can turn plenty of things into piles of red hot molten slag: pretty much any lithium or car battery.
Electroboom on YouTube has a whole channel of caution for you. Man is lucky to be alive.
Nobody is designing a power supply here. The PSU is off-the-shelf, connected by a bunch of wires to a breakout box with a second-stage regulator. The only modifications I’ve noticed on the video are the addition of a dummy load (likely to keep the PSU from shutting down) and hardwiring the “enable” PIN.
How are they not designed for this? An ATX psu provides certain voltages with certain current capabilities, and as long as you stay within those parameters the supply doesn’t know the difference and does not care. Unless you are thinking of the old S-100 unregulated supplies, I don’t see the problem.
They are designed for powering computer components, not for random benchtop experiments. A lot can go wrong on the workbench that is not a consideration inside a computer. They can output a lot of amps with very little protection.
“put out a lot of amps with very little protection”
Do you understand how voltage and amps relate to each other? Even if these were 10,000 amps at 12 volts it’s not going to be able to deliver to those amps to your body quickly enough to actually do anything more than get hot.
If that weren’t the case then you would die every time you tried to jump start a vehicle.
Hmm, I’m not sure that amps can be delivered quickly, maybe that’s only possible during the shorter seconds?
Having those 10,000 amps delivered to an accidental short circuit that you’ve created a foot or two away from your face would be quite exciting though, I’m sure.
I think this is the point that the other correspondent is eluding to.
JSA sounds like a green EE at peak on Dunning Kruger effect.
“Computer power supplies are not designed for this (mis)use.”
They aren’t designed to… supply.. 3.3V/5V/12V?
“I’ve directly seen several students and colleagues get seriously injured using such “hacks”.”
How the hell wouldn’t they get seriously injured using it *in the computer*?
They are designed for powering computer components, not for random benchtop experiments. A lot can go wrong on the workbench that is not a consideration inside a computer. They can output a lot of amps with very little protection.
“They can output a lot of amps with very little protection.”
Exactly what kind of protection do you think those “computer components” offer??
Those “lots of amps with very little protection” are happily supplied live on pretty much every motherboard on nicely exposed metal pins.
Computer components are usually contained within a case. A bench experiment is on a bench, likely in front of you. You may even be touching it. Benches are often covered in stuff not found inside a computer case.
If a video card is fried, I’m sad that I have to buy a new one. If I’m fried, I’m injured or dead. I can’t buy a new hand or life.
I’m not saying something bad is guaranteed to happen, but if it does, more protection is better than less.
“Computer components are usually contained within a case. A bench experiment is on a bench, likely in front of you. You may even be touching it. ”
you and I have very very different computer uses
You can grab both posts of a 12 volt battery and you’ll be just fine
Computer components usually don’t go short, don’t present random capacitive/inductive loads or don’t backfeed current into the supply. All common occurrences while working on a bench and trying to debug something – and all things that the supply is not designed to handle safely.
An ATX supply that starts to oscillate could make things go BOOM very fast.
The only protection a supply like this has is usually a short circuit protection that makes it shut off on a gross overload. But if you only have a partial short making 20-30A flow, the supply will be completely happy to deliver, easily setting stuff on fire on your bench. Current limiting is a must, a glass fuse is way too slow to prevent destruction of your device under test.
Furthermore, an ATX supply is usually not designed for clean outputs – having more than 0.5V ripple on the output is pretty common, especially if one “upcycled” an old supply with poor quality capacitors that are half-dry already anyway. Rather bad if you are doing anything sensitive.
Basically what is in the article is not a “bench supply” but a supply laying on the bench. Two very different animals.
“Computer components usually don’t go short, don’t present random capacitive/inductive loads or don’t backfeed current into the supply. ”
you and I have very different computer experiences
That really does sound and smell like the stuff that comes outs of the wrong end of a cow. So I’d really love to know how you could harm yourself or others with this and yet not do the same with any other supply…
An ATX PSU is meant to deliver voltages exactly like this, and should contain all the safety overtemp, over current, low ripple etc you could desire. They are probably actually substantially safer than ‘inexpensive, ‘safety-rated’ benchtop supplies’ – as if the computer PSU goes wrong and fries 3K worth of computer it costs a fortune in warranty and public relations type crap, where the cheap benchtop supply can be churned out for pennies and doesn’t even need to meet its rating, probably uses the same quality electronics as the adjustable module fitted to this thing etc – as 99% of folks that buy one probably don’t use it often or near its limit or actually care about the quality of power delivered, and then when it fries their project they will assuming it was their own incompetence…
And the folks that do actually care about a decent lab supply bought this “inexpensive” lab supply thing to tear it down BigClive style and find out if its actually remotely worth it. Dismembering the junk supply while ol’ reliable from some big name that costs 10x or more as much watches on from its usual place on the workbench…
“and should contain all the safety overtemp, over current, low ripple etc you could desire. ”
Yeah, OK, you had me until that part. You absolutely can get well built supplies with temp/current protection but they definitely aren’t standard. But you’re totally right that the idea that bench supplies are somehow magically “safe” is just absolutely crazy.
Everything I build is of the “assume the user is a total moron” type and insanely overprotected (fuses suck, use active components for a fast trip plus a fuse as a last resort), so every time I look at the schematic from a bench supply I cringe like crazy.
I’ve not seen an ATC PSU without all of them for a very very long time. I’m sure if you are trying to source the cheapest garbage one you can find it might still happen, but with the usual off the shelf easily available ones of the last decade or so I’d expect it to have them all.
>Everything I build is of the “assume the user is a total moron” type and insanely overprotected (fuses suck, use active components for a fast trip plus a fuse as a last resort), so every time I look at the schematic from a bench supply I cringe like crazy.
Can’t really object to that POV, though for a piece of lab equipment that assumption really aught to be faulty, and if its not the minor dangers of a not entirely idiot proof lab power supply are way way down the list of things to worry about. Also by that argument this particular ATX based build even if you assume the ATX part is really garbage is already doing better than any cheap lab supply I’ve seen – its got fuses at all!
“though for a piece of lab equipment that assumption really aught to be faulty, ”
you haven’t been around many labs apparently
Also, to be clear, a ton of ATX power supplies don’t actually have true overcurrent protection: if you look up the supervisor chips for most of them, they just trip due to current causing the regulated line to drop (killing power good) – as in, they just monitor that the output rails are between a range. Of course the point at which that happens is *very* variable, so you can easily have other things horribly fail before it happens.
Indeed Pat, but its still ‘safe’ in that the PSU itself cuts out, maybe never to work again but without bursting in flames etc.
I’m not sayings an ATX supply is anything like a real Bench supply on protection potential – for one thing those tend to have the ability to set current and voltage limiting very arbitrarily (though that doesn’t make them idiot proof – set ’em wrong and the thing you are working on is just as fried). Just that the ATX supply inbuilt protections just keep the rails at the voltage they are supposed to be and prevent it from being dangerous itself. Which when in a lab setting and dealing with still low power DC that really should be good enough, as whatever it is you are working on that might be dangerous to touch etc would still be dangerous with a lab supply.
Yup. That’s what I’ve been saying too. And designing a “safe” current-limited DC output from an ATX supply is easy – way, way easier than designing one straight from scratch. The other trick that’s common is using a laptop or SFF supply: plenty of *commercial* components design to that 6-pin 12V connector that’s used by a bunch of SFF Dells.
All the complaints here are really “working with power supplies is dangerous” and it’s like… yes? and? You think a 30W bench supply can’t happily light something on fire? Pretty much the only criticism I would agree with is that “random ATX supply” doesn’t supply particularly clean power, and yeah, that’s true, but it’s also true for “random switching bench supply” too. You just need to test it.
I guess the other thing I could see is “don’t use an ATX supply as a bench supply for Baby’s First Power Supply,” which is a fair point. But if you’re looking for a 300-400W supply you’re past that point.
As mentioned in my reply to Gamma Raymond, the issues I have witnessed first-hand all trace back to improper integration of the computer PSU into a larger system operating outside of the expected environment of a computer using materials and methods not suited for the purpose. It’s not that computer PSUs are hazardous everywhere, but there are risks associated with using them outside the intended application of powering a computer. Good-quality, purpose-built supplies for bench work are designed to protect the operator and attached equipment in ways that a computer PSU and random DC-DC converter aren’t. Simple things like properly shielding jacks from inadvertent shorts when the PSU can source 20 amps can go a long way.
I have no problem recommending a robust, safe design of a bench supply based on a computer PSU that demonstrates that it meets some absolute minimum safety requirements. The one linked was not such a design.
“It’s not that computer PSUs are hazardous everywhere, ”
It *is* that computer PSUs are hazardous everywhere. If the grounding in one fails, it can kill someone – in a computer or not in a computer (probably more likely if it’s in a computer!). If they don’t have overcurrent/overtemp protection, they absolutely 100% can catch on fire and melt things – and again that’s probably *more* likely to happen in a computer!
I don’t know why you keep stressing this “outside the intended application” thing. It’s an appliance capable of sourcing hundreds of watts from an AC electrical outlet. It has *buckets* of ways of killing you, just like any other wall-plugged appliance.
There is nothing inherently safe about using *any* questionable electrical appliance. Period.
Whoever you’re talking to clearly doesn’t have any idea what’s inside of a PSU. They’re talking as if this thing has nothing in it except for electrical current coming from the wall to the box.
I mean, as I’ve said elsewhere very basic ATX supplies have weak regulation, little ripple control, and only pretty basic output protections. But… that’s true of high-current crap bench supplies, too. You can get super-nice supplies that have much more serious protection if you care. Or just… design them in yourself.
Saying “don’t use ATX supplies for bench supplies” is just silly. Explain what you need to look for, which ones to get, and what to test and possibly add afterwards.
All of which are good recommendations for using an ATX supply *in a computer* as well!
You know it’s not magic right? there are people doing exactly what you saw in this video to create bench power supplies. In fact there are many bench power supplies that are basically rigged up just like this. And I’m pretty sure the person in this video knows more about electrical engineering than the people who assemble those Chinese made power supplies
I feel uniquely qualified to respond, since this was the project that got me electrocuted 17 years ago. It can happen with any project, not just ATX power supplies, but building a bench-top supply is certainly attractive to anyone starting out in electronics, as I was at the time.
That said, at the time I had used what I believed to be over-the-top safety precautions: build a little, cap and secure unused wires, plug in, test, unplug, discharge the caps through a small fan, continue building…
Everything went well during the build. I put the case together for the final time, plugged it in, and found that there was no power output at all… In a moment of frustration, I ripped the case apart to see what was wrong, forgetting to unplug it. I picked the assembly up to look, and just by holding the outside of the case I suddenly found electricity coursing through my body.
I don’t recall feeling any pain at all, but a very rapid “ticking” feeling through my entire upper body, heart, inside my head, and even eyeballs.
There was almost a sense of internal organs, especially eyes, feeling as if they were expanding towards the point of rupture. I felt stuck to the case, and even stuck to myself – not sure how to describe that…
I was completely unable to move: I couldn’t stand up, couldn’t let go of the case, couldn’t push away from anything, and couldn’t even move my eyes. I distinctly remember realizing that I was getting shocked, and could do absolutely nothing about it. I remember the thought passing through my head, “I am getting electrocuted. I am dying. This is how it ends.”
I was a college student living at home at the time, and this was one of the few times during this project that anyone was there with me. My mom was in the next room, and she said she heard me making a very quiet, very strange guttural noise… I believed that I was screaming at the top of my lungs… She came over and looked at me, and later said I was just sitting there looking completely normal, holding my project, but not moving and making this strange sound. She put her hand on my shoulder and asked me what was wrong, and later said she could feel the electricity flowing through me… She realized what was happening, grabbed the power cord and yanked it out of the wall… Total time elapsed was probably less than 20 or 30 seconds, but all sense was distorted, so I really don’t know. I do know the outlet was on a 20A breaker that didn’t trip…
I was able to immediately jump to my feet… I remember running to the bathroom and feeling the need to look in the mirror to see if my face, and especially eyes, were ok. They were, fortunately, but something made me look at my right hand, which was peeled open to the bones across the back of my middle and ring fingers, with a volcano-looking hole on the middle knuckle of my index finger…
Much time in the ER and a trip to the regional burn center… 3rd-degree burns on those 3 fingers, including blood vessels that were severed and burnt off, and many small 2nd degree burns on the tops and bottoms of both hands. It took many hours until the doctors could establish that there was sufficient blood flow to keep my fingers. Thank God, and thank Mom. I was incredibly fortunate in that skin grafts weren’t necessary either… some skin had to be removed, but scar tissue did eventually build up and fill in the missing areas, but it was about 4 months until I could leave my hand unbandaged.
Seventeen years later, scars are still present and I have some stiffness, feeling, and circulation issues, but that is by no means a complaint. I have my life, and I have my fingers. And eyes.
To this day I still don’t know exactly what I touched or where exactly the short was. I actually still have the project in a box of electronic junk, but I have never been able to bring myself to look at it to see what happened.
Thank you [Dave] for that sobering comment.
I had initially disagreed with you, but then decided to watch the video before commenting.
I think plugging a PSU into a board using a proper ATX socket just to break out the 12v and 5v supplies would be perfectly fine. But I would never consider opening one up and making modifications as shown in the video. But I’m not an engineer, and 12v is all I’m comfortable playing with.
If you want a useful bench supply then the computer power supply is probably best used just as a way to convert mains to an intermediate DC voltage inside a box before immediately feeding the bench supply’s DC input. For instance, you might get a 48V server supply and put it in a box wired up to power a RD6018 which is all you actually touch from then on.
There is so much misinformation in this thread and unlikely situations or incomplete descriptions of the issues that it’s hard to ignore. Firstly, the primary danger of a ATX power supply is the fact that they are designed for heavy current output. This means that if one of the rails, especially the low voltage ones like the 3.3 volt are shorted to ground, a lot of current can flow vaporizing, traces and possibly melting wires. Although the wires of the power supply shouldn’t easily melt because they were designed to handle that current. To begin with, including surge loads. The second big limitation of an ATX power supply is that you don’t have any constant current modes which most bench supplies provide. By the way, almost every power supply sold as a bench Supply on Amazon, for example, is actually a switcher as well and some of them are several hundred watt power supplies. Computer power supplies can be well over a kilowatt so they have a lot of potential to discharge energy there. However, that current is it’s in the low voltage rails, not the 12 volts.
A ground coming off of the case inside the ATX power supply would not be a direct danger in and of itself. The only condition where that would be a problem, is if in addition to that, one of the hot legs of the AC touch the case, then the case would Be energized at 120 volts US or 240 volts? Most other places in The world. This would be an incredibly rare situation and probably not worth worrying about a whole lot. It could be tested very easily by testing the ground lug on the AC cable to the case of the power supply. The biggest problem with this loss of ground to the case would be radiation of RF switching noise to the surrounding environment.
One of the comments down below said it’s the current The kills, not the voltage and that half truth has lived almost forever in the misinformation universe. In truth, yes, it’s the amount of current that causes the heart to stop and potentially not restart again. However, in order for that current to flow, the voltage must be sufficient that the dielectric of the skin is broken down. This usually happens around 70 volts or more. Special circumstances where if you happen to have an open cut in each location where you’re touching The conductors then that voltage would be somewhat lower and possibly could suffer electrocution at a much lower voltage. This situation seems highly unlikely to occur in any normal situation. If you were concerned about this, then you should also be afraid of 9-volt alkaline batteries. They could also electrocute you and are capable of supplying multiple amps of current for short periods of time, if their output was connected directly to your bloodstream.
People should always be aware of the dangers of electricity, but over hyping it doesn’t serve to benefit anyone. If the external interface to the ATX power supply fused all of the lines, then overloading any particular leg would just cause the fuse to blow inside the external enclosure. The ATX Supply easily being able to supply a lot more current. If you are concerned about vaporizing traces, then you could fuse the external interface to an amp or two and be safe in most situations. If you happen to buy one of those variable power supplies off of Amazon, many of them are capable of 5 to 10 amps at 30 volts. One glaring missing feature of the external enclosure is the constant current control of the output. So you are more likely to be able to blow things up if you make mistakes. Blowing up components is part of the learning process of learning electronics and everyone does it more than once. Get a diode in backwards. Draw. Too much current through a transistor or FET and you can let the proverbial smoke out.
I was surprised that they didn’t use the negative rails from the switching Supply because they could be very handy in the external enclosure as well. They’re fairly low current but can be useful. The minus 12 volt rail comes to mind.
Using a buck boost converter in the external enclosure gives you voltages outside the fixed rails that the ATX Supply provides and is not a bad idea. It also limits the current that can be provided at those voltages and a lot of those converters have the ability for current control something you might want to add.
Electronics is always a potentially dangerous hobby or field to be in and one should always attempt to be as educated as you possibly can, but mistakes will always happen when you learn and there is a very small likelihood of electrocution on the output of an ATX Supply. The big advantage is that voltages are low even though currents are possibly high. The big danger is that with those high currents you can vaporize traces and components more easily and arcing can occur in certain situations.
After reading the full section of comments, I feel a little weary now at using the five volt rail on a PSU to drive some camera flashes w/AA battery eliminators. That doesn’t mean I fall either way, but still.
I’m not actually going to do it, but what safety measures should I be bearing in mind for this?
I would like to just say this, if anyone gets injured it’s most likely due to incompetence not because of faulty or even sub par quality and or missuse of hardware. Transformers are transformers, I would be comfortable using this variable smps in the configuration shown right here. Your always gonna have someone hating on your ideas and projects… The one you made 20 years ago must not be a fraction of the quality ths man has built. that why you want to bash him…. One more thing. Fine job! Really makes me want to get a 3d printer. Have a great day.
I’ve abused many a computer PSU to drive 12v audio equipment in a home setting.
While it’s ideal to offer some kind of buffer for high current draw devices (subwoofer amplifiers, for example), the process is really straightforward.
Jumper two pins on the main mobo atx connector, and you’ve got a 12v rail.
Everything else is niceties
I have done the same (except it was a commercial VHF transceiver and 13.8VDC). We took a 350W AT supply and bent it to our will. This took me an entire weekend of schematic tracing and PCB hacking and ended up with a very robust but redesigned 13.8V @ around 20A supply. Didn’t look pretty but had the original overvoltage and overcurrent protection, plus a few extra caps on the output. All in the original case. Can’t remember the details, but it might have been combining two secondary windings in parallel to get way more current at “12V” than the supply was designed to put out, and then tweaking the feedback voltage divider for 13.8V output. It ran for many years.
Way back in the day, I used to drive around with a home cinema sub (trash repaired) in my car with an inverter hooked up to it for powering the AC input. Kinda like the opposite of the shenanigans you were up to.
Isn’t the negative rail on these supplies connected to chassis and line ground?
That can lead to some interesting surprises.
Only in the exact same way that actually using it in a computer is…?
Any sane power supply’s outputs are referenced to earth ground from the electrical input, so it just depends on your electrical wiring. If that “surprises” you so would literally any use of, say, a USB port on your computer, since that’s the same reference. Motherboards don’t have isolated supplies or anything.
haha my 3d printer’s USB connection (which I use to transfer the G-code in real time) was unreliable (glitching unrecoverably whenever a fan turned on or off anywhere in the house) until i figured out the expediency of putting both the host computer and printer power supply on the same power strip.
so yes i would agree, the fact that USB doesn’t have an isolated ground *is* surprising, if you ever happen to learn it :)
it all depends on how ignorant you are, but i remember how surprised i was to learn that the ideal MIDI transceiver is opto isolated. that kind of ignorance set me up for the next surprise down the line
“until i figured out the expediency of putting both the host computer and printer power supply on the same power strip.”
Yeaahhh… you should probably have the wiring in your house checked if that was the solution.
100 feet of 12 gauge wire is 0.2 ohms. 10 Amps rms through 0.2 ohms is 2 volts rms, 2.8 volts peak. That’s enough to upset most signals, and even damage components.
That sort of current should not be flowing through ground, and there’s the possibility that ground is shorted to neutral somewhere it should not be, either in the equipment or the wall.
In principle, Y capacitors could be conducting current spikes from neutral to ground when a fan starts. In practice, I don’t know.
“That sort of current should not be flowing through ground, and there’s the possibility that ground is shorted to neutral somewhere it should not be, either in the equipment or the wall.”
…Hence my suggestion of having the wiring in the home checked? If you’ve got 10 amps flowing through earth, your home’s a death trap.
Sounds to me like you have a ground fault in your house that you need to fix before you all die in a fire
Pat, if you think a bench supply’s negative rail should be connected to line ground you have a fundamental misunderstanding of the purpose of such a thing, and the risks associated with it. It’s an accident waiting to happen.
The fact that USB’s negative rail is connected to the computer chassis is a bad compromise, forced by the role and cost target of that interface. Ethernet got it right.
I personally think having the negative rail grounded is safer.
If the internal isolation between the Mains AC and DC sides fail (which can happen in poorly made power supplies), the DC output can have Mains AC voltages superimposed on them. If the DC output is floating, devices can appear to function OK until a person them touches and becomes part of the Mains AC circuit.
If the DC output is grounded it should trip Earth Leakage / Residual Current devices immediately instead of waiting to be touched.
Marcus, how many devices with two-prong power cords do you use every day? How many have failed to provide the required insulation?
There’s a reason legitimate lab power supplies have both a negative output connection and a separate, distinct chassis ground terminal — it’s so the supply *explicitly* is allowed to float, but grounded if desired, to *either* rail.
Isolated supplies are absolutely useful! Critical in many situations!
I *very much* disagree with *every* lab supply’s output floating – there should be isolated and earthed supplies, and ideally it should be obvious which is which (like red/green vs red/black or something), and you use the appropriate one.
Either that or like, a friggin’ obvious switch or something. Maybe with a big indicator light or something.
Keysight/Agilent/HP supplies typically have that little dumb “bracket” which allows you to easily earth the negative, but the number of times I’ve seen that stupid thing just hanging there because it fell off when someone loosened it to insert a wire…
Just want to second everything you said. It’d improve safety, AND make “stacking” power supplies much more straightforward. There really should be some sort of standard for this.
and how! i’m still not clear on how the negative lead on my bench scope is connected to line ground?? but i’ve certainly been surprised by it. what’s amazing i guess is how long i went not giving it a second thought before i found out.
Yeah, the scope ground is another of those compromise decisions. This one in the name of safety. But I have melted a scope probe lead and started a fire inside a power supply because if it. I carry a cheater cord now for those times I need it (it’s lighter than an isolation transformer!)
Just use an ungrounded socket, computer PSUs are plenty safe without a PE connection. Depending on where you live appliance grounding hasn’t even been a requirement in dry rooms for that long
Bench PSU was giving me headaches: first one (Thandar?) just couldn’t deliver anywhere near the amps it was supposed to, maybe it had a fault I didn’t bother investigate. I sold it and bought another cheap PSU, but it kept switching between voltage circuits no matter what voltage I was trying to get, which was really annoying and not safe for the powered device. Eventually got a DPS5005 because it was cheap enough on eBay (private seller), and power it from a laptop 19.5V 180W charger. Not the best power supply, but this “hack” suffice for my meagre needs, and it’#s more compact/gives me less hassle than the bench PSUs I had.
The DPS5005 is similar to the adjustable one seen in the video, but I don’t think it’s a step up PSU. Clearly the video shows he can get 24V out, so I presume the one shown is a different type than mine. I’d say “I wish I had one of those”, but I’m an amateur rarely using my PSU, and when I do, I typically only ever need 3/5/9/12V, which I have no issue getting from the 19.5V laptop PSU obviously.
Epic win for the form factor design, use of common standards, and barrel jack support, although I only very rarely need a bench supply, and don’t think I could really justify the cost and desk space of building one.
I have one of those little modules in a 3DP case somewhere but don’t remember when I last used it. I mostly just use an RC hobby charger with an adjustable DC output feature.
What, no -12 plug? It’s one of the best parts of using an ATX supply! So useful when doing analog work.
I just came here for the comments. Haha
The best part about this build is also the scammiest :(
Its based on $7 “ZK-4KX Buck Boost Adjustable CC CV 0.5-30V”, but you wont find that information anywhere because its all about that affiliated alibaba $. Author went as far as ripping the silkscreen label off the DC-DC module. Lame!
Wow.. So much angst here, it’s almost as if we were on a german forum (where every answer to questions anywhere near mains voltage will be about how only qualified people can ever work with mains power and that you have a 93.8% chance getting killed if you even look at an appliance’s insides).
That said:
– it’s total BS to install the load resistor INSIDE the PSU (if it’s even needed!)
– why not get a Modular PSU?
– and a ATX breakout board instead of all the cable cutting (There are 10+ Versions of these on AliExpress that even have USB PD and all that jazz for around 10 bucks)
– don’t use shitty terminal blocks for high current! ever!
– use fuses matching your wire gauges
And the you realize the video is OVER A YEAR old… geez…
Those DCDC current/voltage controllers seem to be really slow.
So if you expect your circuit to require 50mA at 12V so you set the lab power supply at 12V 100mA and then hook up your circuit-with-an-error, the lab power supply will quickly drop the voltage so that no more than 100mA is flowing. Say you shorted the 12V->5V regulator. So now your 5V circuit is getting the 12V. Now the 5V chips will start drawing more than usual at say 7V and there will be about .7W of energy flowing into your circuit that might go places where it breaks things. Still .7W is not that much and if you disconnect it in a few seconds, often “no hardware will be harmed”.
But these will enforce 12V at several amps for several seconds before dropping to the lower voltage. THAT is the most harmful thing about this setup.
I am slowly working towards building something that would be much better in this situation, but it is not ready yet.
ATX bench supplies have been around for a while… it’s all in the “presentation”…
I made one of these without the CV/CC module, and soon after replaced the whole thing with an AC to 48V DC converter and a CV/CC module design I found on YT. I wouldn’t want to fool around with a high capacity PSU without the ability to limit current.
Computer power supplies aren’t the best thing to use as bench PSUs, and safety is just one of the reasons not to use them: they emit all sort of EM trash, and the output is dirty; also mains insulation is often lacking.
This is a fuse blower. It always “seems like a good idea at the time” to convert a PC supply into a bench supply. Yes, its GREAT that all the outputs are fused. But the reality is without good current limiting and over current protection, there is going to be a lot of blown fuses. Eventually you run out, and start replacing with whatever you have on hand. Somewhere down the line you forget what the fuse should be, and now you are putting 20A fuses in there and hooking it up to your breadboard. One wrong move and you have a flaming pile of plastic in your face. Just one such example of “ways to ruin your day / house / life / neighbor’s life” with a “bench supply”.
Side note: one of the best “fuses” I have “discovered” accidentally is a cheap dupont breadboard wire, with like.. 4 strands of copper mostly infused with the insulation plastic. One too many amps later (from a real bench supply that I had left the OCP setting way too high) and it was more like an electric toaster for the briefest of moments, while leaving a deep burned line across 3 fingers.
There are special use cases where you might want/need something like this. But you should have a bit of experience and a safe environment. I STRONGLY advise anyone new to electronics trying to kit out their first bench to avoid these and all other such DIY bench supply kits. Especially since now you can get a decent SMPS based bench supply for under $100, with OCP, OVP, adjustable current limit, CC/CV modes, digital displays and USB data streaming.