Wednesday, December 22nd 2010

EVGA Intros Power Boost Gadget to Improve PCI-E Power Stability

EVGA released what it claims to be a handy little accessory that can increase PCI-Express slot power, improving overclocking stability of graphics cards installed. Called the EVGA Power Boost, the gadget is a tiny PCB that fits into the power-line notch of any PCI-Express slot (x1 thru x16), and draws in an auxiliary 12V line directly from the power supply unit. The little PCB draws in power from a standard Molex power connector, adding it to the motherboard's 12V line. EVGA also gave out a detailed HD video showing users exactly how to install the device, because inserting it into the wrong part of the PCI-Express slot will fry the motherboard. EVGA is asking $20 for it.
Add your own comment

65 Comments on EVGA Intros Power Boost Gadget to Improve PCI-E Power Stability

#1
qamulek
[quote="pr0nInspector]nitpicking: electrons flow from "ground" to the +12v line, so using more black wires makes sense, since electrons are exhausted after doing all that work in the video card.[/quote][QUOTE=Poisonsnak, post: 2130307"]I like that, maybe a few of them pick up a beer on the way home from work so there isn't as much traffic on the yellow wire as the black one.[/quote]If you send a billion electrons in then a billion electrons must come out or you build up a charge which can eventually lead to trouble.

My guess is the +12v wires are being load balanced, but the ground wires are simply connected to each other. As an example lets say each wire was rated to carry a unit of 1(going unit-less). Three +12v that are load balanced can carry a total of 3 units in. The ground units are simply connected together giving a variation of currents with the possible distribution of (.7,.8,1.5). One ground wire is now pulling 50% more current then its rated for. Add two more wires with the assumption that the two wires added pull the same current as the .7 above, then normalize the total current to 3: (.477,.477,.477,.545,1.02). The 1.5 current wire is now 1.02 which is just 2% over the rated current.
Posted on Reply
#2
kenkickr
This takes me back to the OCZ DDR booster. Atleast this contraption from EVGA will work on any board with aleast a PCI-E x1 slot.
Posted on Reply
#3
hat
Enthusiast
Is this a troll? What's with the 4chan smileys in the vid?
Posted on Reply
#4
rickss69
I suspect this was provided by request's from overclockers much the same as this was... http://www.evga.com/products/moreInfo.asp?pn=M020-00-000205&family=Accessories - Hardware&sw=4

I know of nothing much that does not work as advertised in the Evga line-up. They have always been on top for a reason with their products as well as service. I find it gratifying that Evga does listen to overclockers and acts on their wishes from time to time. Maybe someone will document the advantages of this device and report the findings.
Posted on Reply
#5
KainXS
I don't even know man

good board and good psu and this is not needed from the looks, hell i can't think of any way someone would need this unless they were runnin like 4 cards and other stuff

and 20 dollars, . . . . . damn
Posted on Reply
#6
OneCool
20 bucks for that!! :wtf:


dont forget the 8 bucks to ship it :rolleyes:
Posted on Reply
#7
newtekie1
Semi-Retired Folder
OneCool said:
20 bucks for that!! :wtf:


dont forget the 8 bucks to ship it :rolleyes:
EVGA is generous enough to give free shipping on it.:D
Posted on Reply
#8
[I.R.A]_FBi
sneekypeet said:
why did this come to mind as I read this? :roll:
i has one :)
Posted on Reply
#9
HillBeast
This is the stupidest thing I have ever seen in my life. A TRUE overclocker owns soldering iron. Why does EVGA think they don't know how to identify the 12V rail and solder on a plug of their own, and if they did, wouldn't it be better to use more than just a single molex plug? What's it add? 50W? Wow. That's a huge amount.

Here's what I'd do: go to the electronics store, pick up a $10 soldering iron, $3 of lead/tin mix, and $7 of wire. There you go, and you can use that however you like.
Posted on Reply
#10
[H]@RD5TUFF
So wait were taking a 5 v power source and using to to add "voltage stability" to video cards that take a 12v source?

I know on the SR2, and on the classified and on the Rampage 3, they have this feature built into the boards. But given that video cards have their own power source, but is this really going to work or is it needed, when the board doesn't come with it. I just don't see the point.:wtf:


$immond$ said:
I call it B.S, just like those gaming nic cards.
The killer 2100 actually works . .
Posted on Reply
#11
Mussels
Moderprator
lol, the original image has one of 4chans smiley faces XD



I think the idea behind this is that if you have a cheap shitty PSU with poor power distribution over its 12V rails, you're using another rail to help boost the power to the PCI-E slot, reducing the chance of overloading one rail.


that said, i doubt a little thing like this could even handle enough power going through it to make a noticeable difference.
Posted on Reply
#12
Wile E
Power User
chuchnit said:
No and No!! The original EVGA classified was having 24-pin connectors melt because of excess power draw from the PCI-E lanes with quad 4890's. This only applied to extreme overclockers running 4x4890's over 1000mhz. From what Shamino says, ATI is the only ones that draw power out of spec from the PCI-E slots. The classified's PCI-E power traces are/were actually designed stronger than any other board at the time. Shamino released a hardmod to eliminate the melting by adding auxiliary 12v power to the 24-pin. In the end this is why the 4-way classified included the 4-pin molex.

http://www.xtremesystems.org/forums/showthread.php?t=226942



Yes only on the 4-way. They were the first to do this, but Asus and Gigabyte soon followed, but went as far as having two 4-pins.
Ok, so it was quadfire, but the point still stands, It was 4 OCed cards causing problems. I wasn't focusing on the specifics, just the overall situation. Could this help with that instead of modding the 24pin? Please note, this does not mean I support this product at this price.

And my old ECS KA3-MVP had a power connector for the PCIe lanes, evga wasn't the first, as far as I knew, unless of course you are talking only within the scope of this generation of hardware?
Posted on Reply
#13
chuchnit
Mussels said:
lol, the original image has one of 4chans smiley faces XD



I think the idea behind this is that if you have a cheap shitty PSU with poor power distribution over its 12V rails, you're using another rail to help boost the power to the PCI-E slot, reducing the chance of overloading one rail.


that said, i doubt a little thing like this could even handle enough power going through it to make a noticeable difference.
No it is that when really pushing the voltage and clocks under load some, mainly ATI/AMD, cards pull out of spec power from the PCI-E slot. By pulling that much power through the traces and 24-pin it generates enough heat to melt the connector. You could have the best PSU in the world and it would still happen. It's just too much power for the connector to deliver and stay cool. There is a way to hardmod the card and feed power to the PCI-E on the card with a molex, but hell if I can find that mod. I have to have a guide with nice pics.


Wile E said:
Ok, so it was quadfire, but the point still stands, It was 4 OCed cards causing problems. I wasn't focusing on the specifics, just the overall situation. Could this help with that instead of modding the 24pin?

And my old ECS KA3-MVP had a power connector for the PCIe lanes, evga wasn't the first, as far as I knew, unless of course you are talking only within the scope of this generation of hardware?
I have no clue if this thing works. In theory I guess it should and I would imagine EVGA had tested it. It's not like it needs to supply a ton of power. You only need enough to lower the load placed on the 24-pin.

Wile E
Well, wasn't it the Evga boards having Quad-SLI power delivery issues because their PCIe power sub-system was not up to snuff?

I bet this works for those setups, but that begs the question, why not just do it right from the beginning?
There is what you said earlier. You must remember that the original classified were designed for 3-way X-fire/SLI + physx card if using nvidia GPU's. It wasn't until the xtreme overclockers got a hold of the board did they try to run 4-way by modding thermalright HR-03's or using PCI-E extenders. Plus they were running four cards, under load, at 1050-1150mhz. Another thing to note was that this was the only board to even have this ability at the time. Not until the Classified 4-way and P6T7-WS boards came out did another X58 have the slots to do this. If another did it wasn't used because it couldn't perform as the Classified did on the CPU clocks.
Posted on Reply
#14
[H]@RD5TUFF
Mussels said:
that said, i doubt a little thing like this could even handle enough power going through it to make a noticeable difference.
This is why I was confused.
Posted on Reply
#15
Poisonsnak
qamulek said:
If you send a billion electrons in then a billion electrons must come out or you build up a charge which can eventually lead to trouble.

My guess is the +12v wires are being load balanced, but the ground wires are simply connected to each other....
Hmm interesting idea on the load balancing, from looking at the PCBs I always thought they were all just connected together but maybe they aren't.

I guess the joke about the electrons being exhausted after work and going for beer wasn't too funny.
Posted on Reply
#16
Wile E
Power User
chuchnit said:
There is what you said earlier. You must remember that the original classified were designed for 3-way X-fire/SLI + physx card if using nvidia GPU's. It wasn't until the xtreme overclockers got a hold of the board did they try to run 4-way by modding thermalright HR-03's or using PCI-E extenders. Plus they were running four cards, under load, at 1050-1150mhz. Another thing to note was that this was the only board to even have this ability at the time. Not until the Classified 4-way and P6T7-WS boards came out did another X58 have the slots to do this. If another did it wasn't used because it couldn't perform as the Classified did on the CPU clocks.
Right, but some do now.

I'm just trying to figure out what this thing's market is, and if it actually works.
Posted on Reply
#17
chuchnit
Wile E said:
Right, but some do now.

I'm just trying to figure out what this thing's market is, and if it actually works.
Honestly I don't know, hehe. Most extreme benchers, where there could be a use for this, will be using boards with aux power already since the top benching boards have them. I just don't see a regular overclocker needing this. I mean I guess it could work in theory by taking some of the load off of the 24-pin and splitting it with this gadget. I still doubt it would make any difference for the average Joe. Hell I'm scratching my head too. :laugh:
Posted on Reply
#18
pr0n Inspector
qamulek said:
If you send a billion electrons in then a billion electrons must come out or you build up a charge which can eventually lead to trouble.

My guess is the +12v wires are being load balanced, but the ground wires are simply connected to each other. As an example lets say each wire was rated to carry a unit of 1(going unit-less). Three +12v that are load balanced can carry a total of 3 units in. The ground units are simply connected together giving a variation of currents with the possible distribution of (.7,.8,1.5). One ground wire is now pulling 50% more current then its rated for. Add two more wires with the assumption that the two wires added pull the same current as the .7 above, then normalize the total current to 3: (.477,.477,.477,.545,1.02). The 1.5 current wire is now 1.02 which is just 2% over the rated current.
wires of the the same connector all start from the same point at the PSU side. not sure about the video card side though.
Posted on Reply
#19
pantherx12
Wait a second, don't most cards people overclock draw their power froma power cable or two and purposely stay within the pci-e line spec for the rest.

I.E won't draw more than 75w from the powerlines anyway?
Posted on Reply
#20
Wile E
Power User
pantherx12 said:
Wait a second, don't most cards people overclock draw their power froma power cable or two and purposely stay within the pci-e line spec for the rest.

I.E won't draw more than 75w from the powerlines anyway?
Some cards don't follow the spec when OCed.
Posted on Reply
#21
pantherx12
Wile E said:
Some cards don't follow the spec when OCed.
Aye but even then, they tend to overdraw from the power cables first right?

I mean even a pci-e 6 pin powercable can technically output the full 150w of the 8 pin if it wanted too ( the 8 pin just has two extra earths after all)

Although I suppose 480s would probably need to draw from everywhere when overclocked.
Posted on Reply
#22
Wile E
Power User
pantherx12 said:
Aye but even then, they tend to overdraw from the power cables first right?

I mean even a pci-e 6 pin powercable can technically output the full 150w of the 8 pin if it wanted too ( the 8 pin just has two extra earths after all)

Although I suppose 480s would probably need to draw from everywhere when overclocked.
Apparently, some ATI cards drew too much from the slots, which was causing problems with the 24pin connector overheating.

I don't know. Doesn't seem to me this has a very big niche to fill.
Posted on Reply
#23
voklskier4452
It was brought out because of cries from the masses that EVGA had to do something about nvidia and ATI not following spec on their cards.

As chuchnit said this was designed to supplement power from the 24pin. Some people had their 24pin connector melt because nvidia and ati cards were drawing to much power through the pcie slots.
Posted on Reply
#24
Wile E
Power User
voklskier4452 said:
It was brought out because of cries from the masses that EVGA had to do something about nvidia and ATI not following spec on their cards.

As chuchnit said this was designed to supplement power from the 24pin. Some people had their 24pin connector melt because nvidia and ati cards were drawing to much power through the pcie slots.
Ok, so that begs the question: Why didn't they give the board more power connectors for the PCIe slots to begin with? I mean, they had to know 4 heavily OCed cards might overdraw on the slots.

This is one of those things where I see the point, but at the same time, feel we probably shouldn't need to resort to this.

And the price is retarded.
Posted on Reply
#25
voklskier4452
They didn't give the board more connectors because they followed power spec. Not that EVGA is totally free of blame in this but IMO they shouldn't be held accountable for nvidia and ati making cards that draw to much power and fry boards.

As for the price I'm sure once you factor in the labor to make these, costs for making them, free shipping, and all the other costs associated with making these adapters they are close to the break even point.
Posted on Reply
Add your own comment