ty_ger07 
You state that at 0%, the fan may run 950RPM.  No. 

So you connected a 4 wire cable between a PWM header (that you set to 0%) and your gtx580 fan, right?  I mean you must have, otherwise  you couldn't say that I'm "wrong".  See, the problem you have is that anyone can connect any PWM fan to their mobo PWM fan header, set the speed to 0%, and see for themselves who is right and who is wrong.  So while your response to my post is "roller, you're wrong because I say you are wrong, and for no other reason" I have actually done the test so anyone who is interested can see for themselves.
Don't feel like connecting a fan?  This link proves you're wrong.  Scroll down to just below 'specifications' to the graph "speed v.s. Duty Cycle" 
 http://www.tigerdirect.com/applications/SearchTools/item-details.asp?EdpNo=6365753&CatId=2216
 
The graph shows 0% = 390RPM. 
I could post 20 more examples of your wrongness, but I'll do just one more.
http://www.newegg.com/Product/Product.aspx?Item=N82E16835426010
RPM ranges from 900-2000 RPM with this PWM fan, not 0- 2000RPM.   Look up any PWM fan you choose, it will show a minumum fan RPM of something more than 0 RPM.  Yet according to you, they should all be 0 RPM because minimum=0%.  Are all these fan manufactures lying, or are you wrong?

PWM operates on duty cycle and therefore 0% duty cycle can not physically create anything other than 0 RPM fan speed since no power would go to the fan at 0% duty cycle.

Clearly you have never designed a transistor amplifier as your two mistakes indicate.  The first mistake I addressed above.  Second, the fan does not get its power from the squarewave, that's just a high impedence control signal.  The drive current is from the +12VDC line which is present all the time regardless of duty cycle.  As the duty cycle approaches 0%, the driving transistor  will be driven by only the bias resistors that supply quiescent current to the base, so there will be some DC drive current, the amount  depending on the gain of the transistor and the bias.

If you mess around with electronics, like me,

Like you???  LOL!

I have stalled fans before as this has been an interest of mine in the past.  I have noted that a fan can operate at very low duty cycle, but that once it stalls, the duty cycle % required to get it back in motion was much higher than the duty cycle required to keep it in motion when not stalled.  I can record this for you if you wish. 

No need, that's the only thing in your post that is semi accurate. 

Preventing a stall prevents unexpected operating conditions from existing and thus wild changes in fan speed and temperature of the item which the fan is cooling will not occur.

Fluctuations are  normal in an GPU, they aren't a problem.  Heat is a problem, so  all that matters is the actual temperature, not whether or not that temperature changes over time.

what would not be fine is the behavior of the fan and GPU core temperature fluctuation once the card entered back into 3D load. 

Temperature fluctuations are normal under 3D load they don't cause damage.
 

Would this stalled fan operation be acceptable to the consumer?

Yes, of course.  All that matters is that the GPU stays within it's safe operational temperature.  Whether or not that temperature is constant is of no concern, ICs will always change temperature (fluctuate), doesn't hurt them.
 
The issue isn't how a PWM controler works, it's whether or not a GPU that is running cool is in jeapordy.   If the 30% restriction was changed to 0%, the only change would be on the cool end of the temperature scale.  The max temperature would not be affected * at all *.  Therefore there is no logical reason for the 30% restriction.

ty_ger07 
 If there were not restrictions, the fan would stall.  You argue that no restrictions are necessary and then explain to me why restrictions are necessary.  Are you trying to confuse me?

When I said there's no logical reason to have any restrictions at all, I'm sure you understood that I was speaking strictly about the "minimum fan speed" value in the fan control section of the bios.  That and nothing else.  
Your position is that there is something inherently 'bad' about a fan turning 0 RPM irrespective of GPU temp.  That is, even if a GPU was -100C you must still have a fan with some non-zero fan speed.   Fan and fan speed is not merely an implementation detail, it's an end unto itself. 
 
My position is that the fans only purpose is to cool the GPU to a safe temperature.  So if the GPU is cool, no need to force excessive fan speed.
 This argument  is beside the point because the fan speed is controlled by the profile, not the minimum fan speed value.  Evga could  keep the profile as is with its stall proof curve, and at the same time lower the minimum fan speed from 30% to 0%.  That way, the default profile remains as is, and the people who want the curve to start at 0% can  create their own custom curve in Precision or Afterburner.  There's no reason to deny the end user their right of choice, especially considering there is no safety issue given the built in override in the bios.  The fan will go to 100% if the GPU temp reaches 96C and there is no software that can override this safety feature. 
Therefore, there is no logical reason to have any restriction at all TO THE MINIMUM FAN SPEED VALUE IN THE BIOS.
 I made a big deal out of the fact that 0% fan speed does not mean 0 RPM cause I got tired of the "yur gonna burn up yur card, man!" crazies.  While true it's beside the point.  The lowering of the minimum fan speed value would have no affect *at all* once the temperature reached about 60C, and then only if someone created his own custom profile.

work like a charm. I overclock to 850/1700/2100 and was and need to increase fan speed. thank's EVGA.

Lets say one uses the SC version on a normal one, what would happen?
And WTH is this?
Make sure no programs are running in the background, and double click update.exe.
 
I mean what? Should i quit everything in my taskbar? Steam, skype, etc?

ok but can i use the SC version in a normal one? i would get SC clocks then? and OMG evga forum is SLOWWWWW

I already Oced to 900/2100. But i guess you are right

have you use the file for the standard version or the superclock version douglatins ?

Will this "Fanspeed Unlocker" bios feature going to be included with the next stocks of the GTX 580's cards? would be nice to buy a new 580 having this included already on the bios other than flashing it on.
 
Thanks Jacob :)

Can someone post the link for me again? I can't download..:(

Ok,
 
I have flashed : works, except minimum value, I have 34% and not 30%, but I don't care. I prefer to be sure to get 100% :)
 
Could someone answer to my questions :
 
- Does this flash remover warranty ? I hope not...
- Why can't I unlock voltage controol with MSI Afterburner...? I use NVIDIA Inpsector finally....
- Which max GPU voltage do you recommend ?
 
Thanks.

Hi,
thanks. I use last version of MSI afterburner (v2.0.0)....

Sorry, it only works on EVGA cards.

GTX580 hardware voltage limit is 1.3v, 1.15 should be no problem.

Hello Everybody.
 
I´ve just buy my first EVGA vga!!! . It's a gtx580 (1580 - vanila model).
 
I will flash it with this updated bios in order to improve its fan curve
 
Can I flash it with the SC bios in order to pre-overclock it from bios???
 
Thanks in advance and kind regards!!!

It only changes the fan restrictions