Computer power supply max current question

Post Reply New Topic

I have a 1.7 amp pump, running between 4.5 and 12 volts. I wish to use my computers power supply to run this thing. I know the 12 volt lines will carry 1 amp, but 1.7? Anyone have any knowledge about this or experience?

not enough wattage output from that psu
12v at 1 amps = 12w
12v at 1.7 amps = 20.4w

I've never tried that, but at the very least the voltage would sag under that (over)load. So it could work, since hard drives and such are more voltage sensitive than a pump. The "1A" spec is likely means "able to maintain close enough to 12V for computer devices, as long as the current draw is less than 1A." And also the 5V line is usually draw from the same source as the 12V so if one drops, so does the other, so the 1A on the 12V line might even be in regards to how that affects the 5V line, which powers the electronics and is even more voltage sensitive.

As to how long the power supply would last that way, I have no idea. But I'd expect to it to fail eventually, from running that far out of spec.

interesting how you want to complicate things

Fuzzy, I burned TWO power supplies with my last pc from "excess" use
it wasn't fun
but do it, for science!

Ah, you guys are great. Funny too.

The pump is NOT a device for computers; it is a small bilge pump, intended for use with aquariums and whatnot. I've been reading lots on cooling a CPU with liquid and it seems that people do use these devices when they dont want to use the traditional fan on the CPU heat sink. I have previously installed a professional water kit in my brothers computer: this time I want to do it aspie style.

However, it is entirely possible that they are running it with a separate power brick rather than the PSU of the computer. This is an option.

and take pictures, right?

Giving it some thought in the dawn light, the decal that says 1.7A might refer to the maximum safe current. The positive and negative wires are only about 16ga in size. These are lighter than typical 12v wires coming off the PSU.

The thing that baffles me is the discrepancy between the packaging and the label on the pump itself. 200 G.P.H. (800 liters!) is quite a lot for that little pump. 1 G.P.M. is only 60 gallons per hour, not 200. Which is the reason I posted. Most of the actual computer pumps and forums seem to indicate that 1 GPM is typical.

The label on the device


The label on the package


For supplies I have made a reservoir from some 2 inch ABS pipe. Tubing will be some vinyl 5/8 hose. For the radiator I was thinking of going with a small heater core from a car. The kind that heats the cabin. Apparently those dissipate heat better than the computer supply ones. I am also thinking about an open cooling system: a 6 foot tower which breaks the heated water into droplets. This would negate the need for a fan, but it depends on my pump having enough head pressure. Very doubtful.

That leaves a need for a copper block for on the CPU itself. Buying is certainly an option, and probably the safest. However, I can mill my own, and know of a machinist that could do it as well.

Anyway. I could go post at a computer modder site, but its more fun here, and it breaks away from the typical 'your operating system sucks' pattern of we get here.

I could tell of the time when me and computerlove were out drinking and we met two twin Juanita's.. but that would be off topic and would get moved. Besides, what happens in Las Peñitas stays in Las Peñitas, right buddy?

Ooh DIY water cooling... good stuff. I made a set up like 3-4 years ago, but ripped it apart last year due to a computer upgrade. It was very quiet and lovely, and I miss it.

One thing about pumps -- the rated flow rate is probably with zero head pressure (I mean unless you already know that). In my system the heatsink/cooling-block was the point of highest resistance (and 5/8" tubing was a good size to feed it IME), and that determined the actual flow rate which was a lot less than the "free flow" rate.

That PSU situation seems pretty dicey tho. But that's the upside of experimenting with computers rather than, say, airplanes.

Can you tell me more about what you did?

To power the pump? At first I got a "wall wort" (plug-in wall transformer) from an electronics surplus place. But then that pump failed and I got another one that would just plug into the wall socket.

So what sort of pump? What did you do for a radiator? Did you purchase the CPU copper block? Did you make use of a reservoir, or was the total volume in the components sufficient for your use?

Oh ok, the whole thing. Off the to[p of my head:

* radiator was a car heater core, IIRC I bought a refurbed one at Kragen's pretty cheap

* 2 fans on the radiator -- large computer type, powered with another "wall wort" and a simple speed control to make them quiet

* the pump was a 110VAC fountain pump, specs I can't remember offhand. It was the submersible type.

* water block was hand-made, out of a copper pipe end-cap, brass plate, and some brass tubing (I know you're not supposed to mix metals, but that's all I had availible). Soldered it up, and then lapepd the contact surface as smooth as I could get with sandpaper (on a flat block). It actually ended up working pretty well, though if I were doing it again I'd buy a pre-made millled water block. Trying to construct a one from a copper block was just beyond my tools and abilities.

* the reservoir was a tupperware container (about a half gallon sized) with brass tubing soldered to a brass plate mounted on top, to allow the vinyl tubing to connect (4 ports: water in, water out, expansion hole, and thermometer in -- turkey/cooking thermometer).

Since I was using a submersible pump, it was necessary, but I think a reservoir is good for a number of reasons:
1) makes getting the air out of the system on set up *much* easier
2) if you lose a little coolant from a leak or something, it takes a while to lose enough to cause problems
3) it buffers temperature changes -- if, say, the radiator fan failed, it still took a half hour for the computer to overheat
4) prevents any overall pressure build up due to expansion -- is an "open" system pressure wise

I had about 1/2 to 3/4 of a gallon (with anti-freeze/corrosion) total fluid volume.

* to get the water in and out of the computer case I made a connector out of brass plate and tubing and mounted it in one of the rear case slots.

* the water block was mounted by putting 4 threaded posts into some convenient holes at the corners of the CPU socket. A plexiglass plate pushed the block against the CPU. The plate was pressed down with springs put over the posts and held with wing nuts.



If I can dig the stuff out of the garage I might post some images, but I'm not sure I remember where all of it is anymore.

Yeah. All the dirty details make for an interesting conversation.



I'm tending strongly towards this, but I have an alternate idea as well.



it seems intelligent to power these devices separate from the PSU, though you then have to remember to turn them on and off.



I had looked at a few of these at home depot. They were about 25 dollars. I was hopeful to use my PSU instead however. With the pump I have I might run it for a few days on a spare PSU.



That is a very novel way to make one. I have a lathe, so I have a limited ability to work my own. In light of the simplicity and effectiveness of yours, I think I could make a passable waterblock. My brothers has all sorts of turbulence creating structures in it.. I couldnt do that advanced.



Mine is a length of 2 inch sewer pipe, an end cap for the bottom, a screw top fitting at the top, and two brass barbs for feeds. I may decide to not use that at all, as I will explain.



Yup. When I first set my brothers up it had a small 5.25 drive bay reservoir, which made filling and bleeding a nuisance. I replaced it with a larger version of my reservoir and it works much better.



That makes good sense, yes.



Right. One of the bad things about my brothers original reservoir was that it was internal to the case.. it just shed heat right back into the computer.



Right. Most of them are closed systems, leaving little room for expansion. You cant really compress the water/air in the reservoir, so you are more likely to pop a hose.



The fill line on my brothers tank is 3/4 of a liter, and about 1/4 more liters fills the tubes, pump and radiator. About 3.75 liters to a US gallon. Originally it ran fluid packaged with the kit, but when I changed the tank We diluted some anti-freeze from his work.



His case has dedicated holes for tubing, and the return line to the pump goes through the lowest PCI slot.



I'm still planning mine. I'll probably use the holes designed for the CPU fan.



I hope you do. I am very curious.

My alternate idea:

I saw a passive cooling tower that someone did. The fluid was pumped six feet up to the top, where a shower head divided it into a spray. The droplets fell down the inside of the tower, which was perforated with copper tubes. The aerosolized liquid shed heat, and the contact with the tubes would take out even more, as well as delaying the fall. The bottom few inches served as a tank, from which point the pump would draw.

This meant that a fan was not needed, which is my primary desire for water cooling. My computer isnt overly hot, but it is noisy. I'm not looking at overclocking.

Six feet is quite a job for a little pump, considering the resistance from the waterblock, though gravity helps with the radiator elements. So I am thinking of something along that scheme.

What I have in mind is more of a sloped cascade over a rippled or dimpled sheet of stainless steel.

It would be really cool to have that capability. But if IIRC my calculations left me with the conclusion that a few square centimters of a flat copper plate in contact with water was really good in the first place (the comp was only a 700MHz Celeron OC'd to 950, tho). I think it was assuming 1mm thick copper between the CPU and the water, about 1cm in diameter.



Very interesting.



That's an intreguing idea. Would it be enclosed, or like an indoor waterfall? And how do you distribute the water at the top so it's evenly spread across the plate? And would the idea be more to try to lose heat more through the plate or through evaporation/vapor loss?

...I was about to suggest an aluminum plate, but then envisioned the dandruffy, corroding aluminum... yeah stainless, defintely.

My main concern is that your 1mm of heat sink would fry your computer in short order if the pump failed. I'm going to go thicker.. maybe 10mm material on the sides and 3-5 thick.

Cascade:
I'll have to enclose it as I have cats.

On the top end will be a small pool area or something, with the feed coming in below the surface. The top surface will lap over the edge onto the spillway, and it will be divided and given turbulence by a sort of comb or fingers. from there it can cascade. The bumps will slow the water and give it more turbulence, and the bottom plate will soak the heat up. It should be considerably more surface area than the inside of a little fin style radiator. I don't much care if I can see it in action, so the top might be covered with steel too, depending on the delta of change between the inlet and outlet.

This is only the beginning of my idea. Steepness is exaggerated. And of course the bottom wont be boxed in like that.

[img][650:678]http://i12.photobucket.com/albums/a213/ThreeFingerPete/cascade.png[/img]


Positive benefit to opaque operation is that very little can grow. My brothers has clear hosing and it gets green and gunky after a few months.

Where do you get the 1A figure? Computer PSU usually has a much higher rating than 1A on the 12V line.

of course!



don't say "but she looked older than 18"...
btw You're the one with the restraining order... Señor Pedobear

1 amp was all that I was certain of. Thanks for confirming that my power supply should run that little pump fine.