John said:
I cannot imagine any interference between the two.
I would unless it were explicitly warned against in my owner's
manuals. In my opinion, that's where you should seek guidance.
In the manual:
"If high power (75 W or greater) add-in cards are installed in
either or both the Secondary PCI Express x16 (electrical x8)
and the PCI Express x16 (electrical x4) bus add-in card
connectors, the Auxiliary PCI Express graphics power connector
must be used. Failure to do so may cause damage to the board
and the add-in cards."
First of all, to clarify the Intel statement, 75W happens to
be the max theoretical power available on a PCI Express slot.
That power number includes both 3.3V and 12V consumption.
Apparently, on video cards, little to no 3.3V is used, as
all the power is converted via switching converters, from
the 12V rail. The PCI Express slot has 5 pins rated at 1 amp
each, for a total of 12V @ 5A or 60W. Video card manufacturers
content themselves with drawing 4A from those pins, as a max.
The highest card I've read about draws 4.35A measured.
If you use only one video card (say a 6600GT), that would draw
4 amps from the video slot. Lesser cards (say a 6200 or something),
would draw less than 4 amps. For a single card installed in the
system, you have little to worry about. (12V * 4A = 48 watts).
If you plug two cards, in say a Crossfire configuration, then
you could be drawing 8 amps from the video card slot(s). That
would be considered a "high power" situation.
Now, the thing is, if you installed a 24 pin main power
connector, that has two 12V wires, and since each has a
capacity to carry 6 amps, the main connector is good for
12 amps. In theory, two video cards plus a few fan headers,
can be handled by a 24 pin power connector. The reason
this is possible, is the video card slot has a limit as
to how much current can be drawn through the slots pins.
And that is why video cards are "split in half". The 2x3
connector on the end is used, if the video card is going
to need more than about 4 amps.
So where does that leave you ? If you plugged in a 24 pin
main power connector, then in theory there is enough
power for two video cards. You plug in the 1x4 as an
aid to the distribution of the power, and by using the
extra connector, there is less resistance. On some
motherboards, the path is shorter, from the 1x4 Molex
on the motherboard, to the video slots.
One thing to be careful of, with all these power connectors,
is that they all draw power from the same rail. For example,
there have been power supplies, that drive the 1x4 Molex
connector with a different rail than the main 24 pin
connector. The 12V pins then come from different outputs.
It is not a good idea to connect independent rails
together (loop currents can flow). Some supplies
will actually shut off if they detect this. Other
supplies might not be so smart. This is generally
only a concern with quad rail supplies, as the one
three rail supply I know of (PCP&C), powers everything
off 12V3. So right now the quads are the one to watch.
And on a dual supply, 12V2 powers the processor, while
12V1 powers the 24 pin main connector and the Molex
connectors. With a dual supply (12V1/12V2) you are
safe to plug in the 1x4 motherboard connector in
all cases.
This is a rough picture of an SLI/Crossfire motherboard.
12V2 12V2 -----------+-----------> Vcore, for processor
2x2 12V2 -----------+
square GND
GND
24 pin 12V1 -----------+-----------> Video card slots,
main 12V1 -----------+ PCI Express x1 slots,
connector | fan headers
|
1x4 12V1 -----------+ (X16 Video slot powers
mobo half of the card)
plug
2x3 12V3 -----------+------------> Video card Aux conn,
PCI 12V3 -----------+ powers other half
Express 12V3 -----------+ of the card
The two halves of the video card are independent of one
another, for the sake of this wiring diagram. The 2x3
PCI Express connector could also be wired to 12V1 (on
a dual rail supply), instead of the 12V3 I jotted on the
diagram above. Since the 2x3 connector is electrically
independent at the video card, it really doesn't matter
where the PCI Express 2x3 comes from anyway. But notice
that the 1x4 Molex connector had better come from
the same source as the 24 pin main connector (and
definitely does on a dual rail supply).
The 2x2 versus 2x4 processor connector is tricky. Some
desktop boards use the same rail on all four 12V pins of
the 2x4. The BadAxe seems to split the connector in two.
Be real careful in selecting a supply, _if_ plugging in
a 2x4 connector, that the supply matches the Intel
philosophy (to use the 2x4, you need a triple rail or
quad rail supply). Plugging a 2x2 is infinitely safer,
and in the case of Core 2 Duo processors, is all you need
anyway. The safest solution is to use the more commonly
available 2x2 for the processor power. I show the
use of the 2x2 in the above diagram (mainly because
I don't know what the hell Intel does with dual rails
feeding a single processor slot).
The only evidence of how video card power is split
within the card, is this article. This article was
discussing the merits of using an independent power
supply, as a source of video card power.
http://www.xbitlabs.com/articles/other/display/fsp-vga-power_6.html
HTH,
Paul
According to reviews, the Radeon x1900 series cards can consume up