Bench Power Supply vs ATX
Vložit
- čas přidán 21. 08. 2024
- Although there are tons of tutorials, I've always heard ATX power supplies don't make good bench power supplies. I wanted to do a little testing to figure out why.
This video was not sponsored, but the bench power supply was furnished by CircuitSpecialists.com for another video. You can check the Circuit Specialists CSI305 out at www.circuitspe...
The other linear bench power supply in my videos is this one and it has served me well over the last few years.
amzn.to/3mbQJGE
If you're interested in that ATX Breakout board, it's available at Amazon and Banggood
amzn.to/3GhhrW2
www.banggood.c...
Hey mate :) That PSU is either broken or out of spec because its an ancient design and cross-loaded. A-tier modern supplies are consistently under 20mV p-p.
It would be interesting to see the same test done on a newer ATX power supply to see how that compares to the older one.
I will absolutely do that.
Yup. I'll be checking that.
czcams.com/video/crLcfTKFBfg/video.html ee4e also probes one and gets similar results. Also, who wants to butcher their "A-Tier" Psu for a bench power supply 😂
@@AnotherMaker so, update?
You usually use bench power supply if device itself don't have linear regulator or DC/DC converter. Also it seems that ATX psu violates specs. 12V should have max 120mV ripple. Maybe that psu is older models that need also some load for 5V line. New ATX create 12V and then use DC/DC to do 5V and 3.3V
Yeah. I don't know if the PSU is going out or if it has always been out of spec. My guess is it's dying. I will be running more tests.
I watched a ton of your videos yesterday, and first and foremost, I loved them. Lots of good and interesting content. I wasn't able to find the follow up video to this if you ever made on testing a newer, better PSU. I am about to embark on making a bench top from a Seasonic (eww) 750w PSU (Seasonic SSR-750FX 80 plus gold) and I notice that in the very bulky 24 pin ATX motherboard cable, there are capacitors that clean up the ripple. Just wanted to come back and share that with ya! Keep up the good work!
Put a load on that ATX PSU and than check the ripples, normally in PC there is constant current draw for motherboard, harddrive gpu and than the regulation is much better. Other thing is that you use ATX and bench PSU in different ways, bench is designed to have low ripples so you can directly power your microchips. In PCs 12V is always used to drive motors (which doesn't need to be well filtered) or stepped down in peripherals to 5V, 3V or even 1,5V so proper filtering is done on that voltages. You Know that 12V is not powering directly any chips. 5V and 3V rails in ATX should have lower ripples.
Fantastic points. I will test all those things. I want to revisit this.
@@AnotherMaker I believe its still a thing that the +5V rail needs a dummy load across it to smooth out the rest of the output. On older PSUs it was a brown wire or a smaller gauge of the same color as the +12V rail.
Hey BART B, can you or maybe someone else explain this "put a load on that ATX PSU" ?
What does that mean and how is it done ?
@@kruskotv1311 by pustyni load I mean connect harddrives, cdroms, motherboard, gpu or other perpherals that draw current from PSU. If you want to perform some lab experiments you can use resistors, but keep in mind you have to calculate its value from ohms law and count the disipated Energy.
Sorry but it is unfair to test such a cheap ATX PSU without comparing your expensive Bench PSU with a 80 Plus Certified ATX PSU (what we normally use for gaming and as an alternative for a bench PSU) they are also affordable and very high quality. I tested mine (80 Plus Platinum certified) that I bought on Ebay for just $40 and it is very clean. Perhaps not exactly as an expensive Bench Power Supply but very similar for the price. You mentioned Thermaltake... the brand is irrelevant when talking about the quality of a PSU. At the end the certification is what is important because they need to have way better parts and components to pass each level of the certification. Thermaltake for example has cheap and low quality PSUs too, unless they are 80 plus certified and each level of 80 plus is night and day difference between Bronze and Titanium certifications.
I don't disagree with your point that a better power supply would probably be better. I will say that most of the videos I had seen up to the point of making this one... People weren't going out and buying specific power supplies for the bench power supply project. In my experience, it has always been grabbing some random silver power supply from a thrift store or old computer. That said, if you are going to go out and buy a specific power supply, you are probably better suited with just getting a real bench power supply. They start in the $50 range with no modifications needed.
I really do appreciate your post and would like to try this with a better power supply
@@AnotherMaker Thanks, very useful information. I always wanted a real bench PSU but they were very expensive in the past.
I have a bunch of different ATX-PSUs and most of them perform pretty well? Is your ripple under a certain load? Some old ATX need some rails loaded to work properly, typically at least .5 or better 1 Amps at least. Many of mine ATX have around 50...150 mV ripple - thats fine for many applications. Its not a linear regulated bench PSU of course.
Hi
Thank you for the video. Can you try the same test on another ATX power supply please? And one where you load the 5volt line when you checking the 12 volts output
The same goes for Amateur radio operators, there are a few that use ATX supplies but as soon as you really push output power then they just fall over. Also not the greatest to power LiPo Chargers for drone batteries. Been there done that
Absolutely. I want to revisit this with some more tests to understand these characteristics better.
Just stick a big capacitor across that ATX output, job done
You must remember that a PC PSU isn't designed to be bench top power supply anyway, that's just a hack, a really nice cheap one too, not bad for less than half a volt :) I noticed that at no stage did you compare prices, you can make a super one for a tenth of the price and you should clean the power first like it would be if it was plugged into a pc, those would normally be plugged into a UPS or an inverter, not directly into the wall 😇
It has to be a really crappy ATX PSU if the 12V rail has undervoltage on it even without a load. I'd throw away the cheap Chinese breakout box, too, and connect the PSU's own wires to bus bars, etc.
The 24pin MoBo-cable has very limited current carrying capabilities compared to grouping all the rail and ground wires in separate bundles and soldering them to hefty lugs. I'm building one at the moment using a basic 2007 460W HP unit and all the GND wires bunched together are thicker than my thumb. The Molex-boxes are for testing and repairs, not power delivery.
Is it true if you connect all 12 volt yellow wires into one lead will give you the total amps in atx power supply? Example the rating listed is 14 -20 amps on a 350 watts atx ps. I want to use it in electrolysis to remove rust and it needs higher amps on a 12 volt dc
Interesting. However it can be easily controlled using some bypass capacitors, almost all internal PC hardware must have some form of smoothing power intake Circuitry to overcome this.
You can absolutely use smoothing capacitors. Most motherboards have other things for further regulation and power clean up. Especially because the ATX standard is still old that there's a wide margin to what an ATX power supply can put out and still be in spec
@@AnotherMaker I agree with you. ATX and most power supplies/adapters including SMPS. Are not properly designed for optimal high quality clean power output. I guess there are many factors, one of those factors is to skimp on quality engineering of power electronics design, requires more effort and time but it’s not rocket science. It’s a subject I’m really interested on and wish to fully master on. AC and DC power electronics engineering.
Is it safe to put a scope on an at power supply? Are they isolated?
Also, I'm interested to test my high end at psu. It's a platinum rated supply, about a 450 dollar psu
ATX psu is isolated, yes.
i have a 5v power supply i have questioned for awhile. it burns out my tp4056 often yet a charger right next to it built the same using an atx converted supply never burns out a chip and is running at high voltage. is there a module that would help clean up my dc power supply? im willing to spend a little just to test the theory at this point. its become kind of a project lol
I've discussed it with other youtubers and we have decided that you should try a .1uF ceramic cap across the inputs and a 470uF electrolytic cap across the input together to filter multiple frequencies.
So your wires will come into the input and you will have 2 caps between + and -
@@AnotherMaker thank you! i will give that a try!
Does the desk multimeter go to standby all the time? I heard people complain about that
It does when there's no load. I need to check if that is adjustable.
@@AnotherMaker I've seen people complain about that, honestly that would ruin it's purpose as desktop multimeter
I confirmed in the manual that you can disable that feature. I use it on battery all the time though so I keep it on
classic, people reviewing products and making complaints about things specifically referenced in the manual. Always happens!
A bit late to the party, what you could do if you want to clean up the output of an ATX PSU, you can add a chock, (make sure it can handle the current though) and then another filter cap thereafter to that output, it does clean it up quite a bit depending on what you use, still does not clean up totally..... ATX is not a good bench top PSU for development or testing...
Do I see that correctly that your ripple is at 60Hz? czcams.com/video/fXk6HodtSmQ/video.html That cannot come from the ATX power supply because Switch Mode Power Supplies work at a much higher frequency. 60hz is picked up by the scope, typically by a bad or missing shielding on the test probes. i could be wrong,... but 60Hz and an SMPS? that doesn't add up...
/ChriV
Junkie power? My ATX is as solid as a rock and clean as heck...Ripple is as low as can be given its purpose (max 30mV) Logic, because CPU's tend not to like ripple...
Problem, most of the times is that ATX PS's used for this are almost always OLD. Most problems occur with older smoothing caps. ...BTW..Also, LIMIT your BW on the scope....Way cleaner results with these scopes.
You shouldn't compare DIY bench power supply from old atx to a well manufactured bench power supply🥴
Basically…….it aint worth the hassle making a bench PSU from an ATX.