Powering a video card with a Car Capacitor?

[Twigsoffury]

Can it be done?

It seems a large Car stereo capacitor could deliver ample..well amperage to the video card when it needs it...instantly and for long periods of time. freeing up the needs on the power supply for other duties.

http://i46.tinypic.com/16m15w6.jpg
http://i49.tinypic.com/2e4hr8x.jpg

I've got one exactly like that top one and have been pondering the question myself for awhile. they are both 12v, the cap itself has a hilarious potential. more then enough for a video cards needs. I mean the one i have is rated for a 2500W RMS amp. no way a card could pull that, and in theory wouldn't it just draw its power from the PSU unit if the capacitor is inline with the video cards power connection on the back.

should i put down the crack pipe, or am i onto something?

[Trace]

Lol!
Please video tape your experiment!

All jokes aside, I see no reason as to why it wouldn't work, but I'm not giving up my card to science.

[Twigsoffury]

Lol!
Please video tape your experiment!

All jokes aside, I see no reason as to why it wouldn't work, but I'm not giving up my card to science.

exactly, But i bet i can come up off a old 6600GT or something for s3ien3e. something with a basic four pin molex connection is what im thinking i should start on, then move over to the PCI-E pin out once i know what the f@#$ is going on.

I need more testing tools though. i think my current multi meter isn't accurate enough fo' mah needs.

hahha i'll try and tape my experiments, see how it goes in the future.

[x88x]

Hmm, an intriguing proposal. My electronic maths are a bit rusty, but I believe what will happen will amount to just about nothing. If you're only using your GPU for brief periods of time, then yes, it might work like you're wanting, but if you use it for any extended period of time, then it's not gonna really do you any good. Think about it this way; the PSU is like a pipe, filling up a water tower (the capacitor). Then imagine you have something draining the water tower at (presumably) a faster rate than the pipe is filling it. It works fine as long as the tower is only drained for brief periods of time, but once you pass a certain point, the tower is gonna be empty and you're just gonna be pouring straight from the pipe into the drain.

Like I said, my maths are a little rusty, but it should be possible to figure out what the discharge and recharge times would be given:
a) The power output of the PSU rail
b) The capacitance of the capacitor (in farads)
c) The power draw of the GPU

[Twigsoffury]

Hmm, an intriguing proposal. My electronic maths are a bit rusty, but I believe what will happen will amount to just about nothing. If you're only using your GPU for brief periods of time, then yes, it might work like you're wanting, but if you use it for any extended period of time, then it's not gonna really do you any good. Think about it this way; the PSU is like a pipe, filling up a water tower (the capacitor). Then imagine you have something draining the water tower at (presumably) a faster rate than the pipe is filling it. It works fine as long as the tower is only drained for brief periods of time, but once you pass a certain point, the tower is gonna be empty and you're just gonna be pouring straight from the pipe into the drain.

Like I said, my maths are a little rusty, but it should be possible to figure out what the discharge and recharge times would be given:
a) The power output of the PSU rail
b) The capacitance of the capacitor (in farads)
c) The power draw of the GPU

a = The one i have is 42A on the +12v
b = 2000 farads.
c.= that would be the variable as i don't know.

its a Evga 8800GTS SSC edition. (112SPU)

[slaveofconvention]

I'd be concerned at what the cap will draw while charging - last thing you want is for the cap to actually draw more than your PSU can cope with. I know next to nothing about this stuff, but I'd assume it'd draw "spare" power from the alternator/battery in a car - they're both finite resources and if they get maxxed out, it's no big deal - the battery drains or the alternator just keeps putting out whatever it can - your PSU however will be trying to provide whatever it's asked for and may well overexert itself in the attempt as it has a much higher and variable source of power, ie the mains.

[x88x]

b = 2000 farads.

O_o you sure that's not 2000 millifarads? A 200A 2000F capacitor would be about as big as a small car...

[Trace]

All you would need is a resistor then.

[mtekk]

I don't know why you would want to do this. In the best case it will deliver slightly "cleaner" power to the graphics card. In the worst case it will kill your PSU, graphics card, or yourself.

You won't be saving energy (actually you'll be more likely to waste it). Also, you won't be able to power the card off of that capacitor for very long (if the cap is not hooked up to a charging source). This is because capacitors are not very energy dense, and are suited for delivering large amounts of current instantaneously, not for driving something for a long time (which is what you'd want for powering a GPU). The capacitor is used in the car sound system to provide a buffer for when the sound changes from a lower volume to a much higher volume (and vice versa as speakers are inductive).

Also, a 2000 Farad capacitor would be possible, though it is an order of magnitude higher than the "Supercaps" you'd normally see. But, seeing this is for a car amp, it is probably just 2 Farads.

[x88x]

Thanks mtekk, I was hoping you would stumble across this ;) Slightly off topic, but ooc I went looking to see if 2kF capacitor exist currently, and stumbled across this little tidbit on wikipedia that shows the relationship rather nicely:
http://upload.wikimedia.org/wikipedia/commons/6/6b/Supercapacitors_chart.svg

[mtekk]

Yeah, if I was working this semester I could check PartMiner for large supercaps, and could tell you the largest commercially available supercaps (same thing as ultracaps, and it looks like they go up to 3KF). Ideally, on that chart you linked to we want to be as close to that upper right hand corner as possible. That way we'd have super fast charging batteries that can function like fuel tanks in current automobiles (the only way we're ever going to get truly worth while electric vehicles).

[Oneslowz28]

All I see this doing is smoothing out the ripple in the 12volt line and stressing the PSU much more than it would be normally. It will not supply gobbs of juice to your PSU for more than a split second and I doubt any modern day PSU that has its power matched to the gear is running would even grunt at a 100% load of 2 gtx295 cards and a hot power hungry i7 proc. Your PSU has a very large filtering cap for this reason.

The best case scenario is that your already exceptionally smooth 12v rail will be ever so slightly flatter and your PSU dies 1.5-2 times faster than normal. Worst case is the magic smoke gets let out.