Archived from groups: alt.comp.periphs.videocards.nvidia (
More info?)
Look what happens if you would raise the voltage; under load you would have
the perfect voltage but when not under load?????
Then the voltage would be too high, in mine opinion just buy a other
powersupply with more power, 350 is the minimum, even if the powersupply
will work , it will work very overstressed and can fail with possible damage
to your components.
"JC" <jocombs45@fuse.net> schreef in bericht
news
fkpt0pm97du17l7npnofdks4pirdkml7t@4ax.com...
> On Wed, 5 Jan 2005 17:19:24 -0500, "dvus" <dven1invalid@adelphia.net>
> wrote:
>
>>JC wrote:
>>> On Wed, 5 Jan 2005 07:45:00 -0500, "dvus" <dven1invalid@adelphia.net>
>>> wrote:
>>>> NoRemorse wrote:
>>>>> "RaceFace" <nospam@myplace.com> wrote in message
>>>>> news:mMGCd.8754$Ka6.63197@news1.mts.net...
>>>>>> "NoRemorse" <thebummer@hotmail.com> wrote in message
>>>>>> news:hJadnUMJOphGv0bcRVn-3w@comcast.com...
>>>>>>>
>>>>>>> "rms" <rsquires@flashREMOVE.net> wrote in message
>>>>>>> news:RWECd.4998$8i5.1805@newssvr31.news.prodigy.com...
>>>>>>>>> Ok I just installed a new BFG 6800 GT OC card in my computer.
>>>>>>>>> The card worked fine yesterday (tested with HL2 and WoW), but
>>>>>>>>> today I started to notice weird problems with the computer.
>>>>>>>>
>>>>>>>> Be sure the card is on its own power cable, not shared with
>>>>>>>> other devices.
>>>>>>>
>>>>>>> Hmmm. Definitely worth a try.
>>>>>>
>>>>>> Not worth a try - it's required. The 6800's pull a lot of juice,
>>>>>> and need a stable supply of power. Having other devices on the
>>>>>> same feed can decrease the current enough to cause the card to have
>>>>>> troubles.
>>>>>
>>>>> Only if the free power cable is on a separate DC rail. If the
>>>>> current cable and the free cable use a common DC rail, then it
>>>>> won't make a difference.
>>>>
>>>> I don't know, but my PSU, for instance, is rated at 52A on the 5VDC
>>>> buss. That's a lot of current to push through #16 conductors (or
>>>> whatever it is), even though they're extremely short. Have to dig
>>>> out the tables giving the resistance of copper conductor to see if
>>>> it really has any effect.
>>>
>>> Someone mentioned tweaking voltages as you create more demand but
>>> that's not quite true. Voltage fluctuation or tolerance problems
>>> usually indicate a poorly designed power supply or failing components.
>>> Current is the real problem with increased load.
>>
>>Well, sorta, although the two are pretty well intertwined. Most power
>>supplies will experience some voltage drop at the source if the output is
>>overloaded. I do agree though, that damage sustained by overloading a
>>supply
>>will be the heat caused by excessive current.
>
> Since computer power supplies are designed to have several voltages
> all of which are fixed and voltage drop is due to the resistance of a
> conductor or component when current flows. My point was that one
> shouldn't try and adjust voltages to compensate for load. A good
> power supply should have rock solid voltages. Those pots are for
> tweaking the voltages after assembly to compensate for varying
> electronic component tolerances in the power supply circuit. I'd
> prefer a power supply that incorporates 1% tolerance components rather
> than 5%. If voltages deviate it either indicates the power supply:
>
> a) is poorly designed
> b) has faulty components
> c) has reached it's maximum operating range
>
> You should buy a bigger and more stable one, not tweak the voltage
> rails.
>
> Of course I'm sure that some power supplies get past QA which are
> stable but slightly out of tolerance. In this case I say sure, tweak
> it. But anyone that thinks they need to tweak it as load increases is
> a fool. Buy a bigger one!
>
> More Power!!!!! ooohhh ooohhh ooohhh :-D
>
>>> As far as the advice on conductors being used for only one device well
>>> while it's true that conductors do have their current limits. A real
>>> easy way to see if that's a problem is by simply feeling them. A
>>> conductor that's over loaded will always get hot. When you draw too
>>> much current through a conductor you're creating a light bulb
>>> filament, in essence. It's also important to note that all conductors
>>> of a given voltage potential are connected to the same rail of the PS
>>> so all conductors are created equal.
>>
>>The only problem with this is that since the leads are inside the case
>>most
>>people don't have easy access to feeling the conductors. In addition,
>>short
>>intervals of high current draw might not heat the wires too much but any
>>resultant voltage drop could corrupt the output of another component. I
>>really don't see much chance of this in a typical PC though, HDDs are
>>usually already spun up when playing games and as far as I know they and
>>CD
>>players don't draw that much anyway. An HP burner I have is rated at
>>1200mA
>>at 5VDC and 800mA at 12VDC. That's a total of around 15 watts which
>>doesn't
>>seem like a lot to me.
>
> It was assumed, because of the nature of the thread, that I was
> talking to guys that are not afraid to open the case.
>
> NoRemorse's first sentence:
>
> "Ok I just installed a new BFG 6800 GT OC card in my computer."
>
> CPU runs cooler with the case opened... ;-)
> Of course I firmly believe that one should button up the case after
> testing is complete to ensure there's no FOD.
>
>>Now, if the 5 VDC buss on my PSU were loaded to its full potential of 52A
>>through a foot of #18 wire the resultant voltage drop would be about 1/3
>>of
>>a volt. That might be enough to cause problems on a voltage sensitive
>>circuit but I'd guess that the #18 wire would melt before the problem was
>>noted.
>>
>>> Overloading a PS's potential to create enough current will not
>>> necessarily cause the wires to heat up but it will cause the internal
>>> components of the PS to heat up.
>>
>>In general, I'd agree.
>
> It's true! Whether the wires heat up or not is dependant on the
> current flow to wire gage ratio (resistance), the only exception of
> course is conductance (heat). This of course has nothing to do with a
> PS's current potential.
>
>>> Unfortunately measuring current draw is complicated and time consuming
>>> with computer power supplies. You'd have to place the ammeter in
>>> series for each conductor of a given potential and then do some math
>>> or open the PS and find the common point for each potential so you
>>> could add the ammeter prior to the split so you'll get a real time
>>> current draw for a particular voltage potential. Of course you'd have
>>> to place each component at max operating draw in order to reach the
>>> continuous current requirement for a worst case scenario.
>>
>>Heh, lots easier to just get an oversized PSU.
>>
>>> Just my 2 cents.
>>>
>>> I'd probably buy a new power supply if I were you (original poster)
>>> That's one upgrade that you should go all out on because it will
>>> usually ensure plenty of future system upgrades.
>>
>>Absolutely.
>>
>>> Good luck!
>