New: Scythe Setsugen GPU cooler
Moderators: NeilBlanchard, Ralf Hutter, sthayashi, Lawrence Lee
New: Scythe Setsugen GPU cooler
Details here. GTX260 compatibility seems to be the starting point for Scythe, as no other GPUs are mentioned.
Looks a lot like the Accelero to me. Interestingly, from the photo it looks like the bundled Slipstream Slim 120mm is to be placed in between the cooler and the PCB board.
This might be an interesting idea for the Slipstream Slim and existing Accelero S1/S2 users...
Looks a lot like the Accelero to me. Interestingly, from the photo it looks like the bundled Slipstream Slim 120mm is to be placed in between the cooler and the PCB board.
This might be an interesting idea for the Slipstream Slim and existing Accelero S1/S2 users...
Might not be for heat reasons, since I'd say those are more powerful GPUs than the GTX 260. (Or it could just be Scythe lagging behind current GPY releases). Anyway, that's all the info I know, having just seen that news item!rpsgc wrote:I assume that therefore it's also compatible with the GTX275/GTX280/GTX285?
Check out all the mounting holes on the backplate. As you can see it looks just like the Musashi backplate. Also have a look at this link, scroll down and you can see it mounted on a 4850, right under the Ninja VGA pic.
Yes, they're saying it is made for the GTX 260, but it's only because of the model specific heatsinks that are included.
Yes, they're saying it is made for the GTX 260, but it's only because of the model specific heatsinks that are included.
-
- Friend of SPCR
- Posts: 279
- Joined: Sun May 06, 2007 10:37 am
Interesting. Does this mean that the musashi can also be mounted on GeForce GTX models? I'm guessing you'll probably have to find your own vrm/ram cooling solutions.
Also looks like the slipstream slim is mounted to blow through the fins and away from the card. I don't know how effective this is over a more traditional fan placement with the fan blowing down into the card. Seems like intake for the fan would be heavily impeded.
Also looks like the slipstream slim is mounted to blow through the fins and away from the card. I don't know how effective this is over a more traditional fan placement with the fan blowing down into the card. Seems like intake for the fan would be heavily impeded.
-
- Moderator
- Posts: 7681
- Joined: Mon Dec 09, 2002 7:11 pm
- Location: Maynard, MA, Eaarth
- Contact:
-
- Friend of SPCR
- Posts: 279
- Joined: Sun May 06, 2007 10:37 am
In my experience fans tend to provide more pressure when blowing rather than sucking in. With the fan blowing at the card, maybe more air can be forced into the small grooves and cracks between all the little chips on the PCB. I feel it's more important to ensure you actually have air flowing over components rather than overthinking whether you have cooler or hotter air flowing.
Not reversing the fan would mean that hot air coming from the RAM and VRAM module/s would be pulled through the fins and out at the bottom of the case, hot air rises.Chang wrote:Reversing the fan means that you will be pulling the hot air through the fins and directing it back towards the heat source. .
Reversing the fan means that you’d be pulling the cool intake air through the fins and over the RAM and VRAM module/s. Yes, since the fan is mounted “beneathâ€
My naive Physics knowledge tells me the following. When you create a pressure gradient (like at the back of a fan), air will find a way to get to it and nothing will impede it. On the other hand, when you blow air, you are giving it momentum which you hope will carry it far. If there are obstructions along the way then the air will turn back on itself and become turbulent. Turbulence has been rightly said to be the enemy of the engineer.PartEleven wrote:Also looks like the slipstream slim is mounted to blow through the fins and away from the card. I don't know how effective this is over a more traditional fan placement with the fan blowing down into the card. Seems like intake for the fan would be heavily impeded.
The Scythe engineers definitely know their Physics. By blowing air onto the graphics card you would be causing more turbulence than flow. By blowing it away from the card, Scythe guys are creating real flow, both in front of the fan and behind it. Very clever!
-
- Friend of SPCR
- Posts: 279
- Joined: Sun May 06, 2007 10:37 am
Not always. Turbulent flow is still as "real" as laminar flow. What it really means is that with turbulence, when you zoom down into the local level, you'll have fluid moving in random directions. However, the bulk fluid is still flowing in a particular direction. Turbulent flow is actually very important for mixing the fluid and is essential in certain heat exchange applications, such as watercooling.reddyuday wrote:Turbulence has been rightly said to be the enemy of the engineer.
The Scythe engineers definitely know their Physics. By blowing air onto the graphics card you would be causing more turbulence than flow. By blowing it away from the card, Scythe guys are creating real flow, both in front of the fan and behind it. Very clever!
Just looking at how the heatsink is designed, I don't see how you'll get any less turbulent flow with the fan blowing away. The card blocks most of the fan. Unless you find a special card with holes in the PCB, the airflow will have to take a 90 degree turn before going through the fan and heatisnk. That already is going to produce turbulence. I was worried that having the fan in this position would create a dead zone near the fan hub and chips closer to the center of the card will not get any air. At least with the fan blasting air down onto the card, turbulence may help reduce this effect by causing the air to hit the card and fly off in random directions.
Agreed. But when you use a fan to blow air through the fins of a heat sink, you are hoping that the air will extract heat from the fins and moves away. If it bounces against the PCB and comes back, obstructing the new air that the fan is trying to blow, then the fan's effectiveness is reduced.PartEleven wrote:Turbulent flow is still as "real" as laminar flow. What it really means is that with turbulence, when you zoom down into the local level, you'll have fluid moving in random directions. However, the bulk fluid is still flowing in a particular direction. Turbulent flow is actually very important for mixing the fluid and is essential in certain heat exchange applications, such as watercooling.
My feeling is that the turbulence behind the fan doesn't matter. By pushing the air forward, the fan causes low pressure behind its fins, which then sucks in air from all directions through various openings and gaps. This is vacuum suction. So the fact that the PCB is "blocking" the air doesn't matter. The only way to block vacuum suction is to seal it air tight.Just looking at how the heatsink is designed, I don't see how you'll get any less turbulent flow with the fan blowing away. The card blocks most of the fan. Unless you find a special card with holes in the PCB, the airflow will have to take a 90 degree turn before going through the fan and heatisnk. That already is going to produce turbulence. I was worried that having the fan in this position would create a dead zone near the fan hub and chips closer to the center of the card will not get any air. At least with the fan blasting air down onto the card, turbulence may help reduce this effect by causing the air to hit the card and fly off in random directions.
You are right that there is likely to be a "dead zone" behind the fins where the chips won't get much air. (The GPU itself is fine because it has the cooler mounted on it.) There is the lateral flow of air along the surface of the PCB as the fan draws in air. But this air is trying to move away from the PCB rather than towards it. So we can't be sure how much heat it can extract from the chips. It will need to be tested.
But on the question of impeding the intake for the fan, I don't think there is any problem there.
-
- Friend of SPCR
- Posts: 2887
- Joined: Mon Feb 28, 2005 4:21 pm
- Location: New York City zzzz
- Contact:
From my experience, so far at least, when something is "blowing" it really is sucking.PartEleven wrote:In my experience fans tend to provide more pressure when blowing rather than sucking in. With the fan blowing at the card, maybe more air can be forced into the small grooves and cracks between all the little chips on the PCB. I feel it's more important to ensure you actually have air flowing over components rather than overthinking whether you have cooler or hotter air flowing.
So that should be fine.
-
- Friend of SPCR
- Posts: 381
- Joined: Fri Aug 01, 2003 12:13 pm
- Location: Rochester, MN
It seems Quietpc in the UK will have stock soon too! They also state its compatible with the following!
Nvidia Compatibility: Geforce Fx5900, 6600, 6800, 7300, 7600, 7800, 7900, 8500, 8600, 8800GTS (G92), 8800 GT (G92), 8800 (G80), 9600 GT, 9600 GSO, 9800 GT, 9800 GTX, 9800 GTX+, GTS250, GTX260, GTX275, GTX280, GTX285
ATI Compatibility: 9***, X***, X1300, X1600, X1650, X1800, X1900, X1950, HD2600, HD2900XT, HD3650, HD3850, HD3870, HD4770, HD4830, HD4850, HD4870, HD4890, HD5850, HD5870
Nvidia Compatibility: Geforce Fx5900, 6600, 6800, 7300, 7600, 7800, 7900, 8500, 8600, 8800GTS (G92), 8800 GT (G92), 8800 (G80), 9600 GT, 9600 GSO, 9800 GT, 9800 GTX, 9800 GTX+, GTS250, GTX260, GTX275, GTX280, GTX285
ATI Compatibility: 9***, X***, X1300, X1600, X1650, X1800, X1900, X1950, HD2600, HD2900XT, HD3650, HD3850, HD3870, HD4770, HD4830, HD4850, HD4870, HD4890, HD5850, HD5870
Here's a very good review of the cooler, tested on a 5850 and 5870:
Translation:http://translate.google.com/translate?j ... l=de&tl=en
Original:http://ht4u.net/reviews/2009/scythe_set ... on_hd5800/
Translation:http://translate.google.com/translate?j ... l=de&tl=en
Original:http://ht4u.net/reviews/2009/scythe_set ... on_hd5800/
-
- Posts: 332
- Joined: Wed Jan 17, 2007 8:10 pm
- Location: Toronto - Ontario - Canada
Source you based this opinion on? I have one these coolers coming to me on Monday plan to run the fan at 1200 rpm on HD 4770 the current reviews although not using SPCR testing mythology prove it be quiet enough compared to stock coolers running 36dba +.pashap wrote:the stock fan even a 800 rpm is not silent
Nice coolers with a couple of drawbacks
I have two XFX 4890 cards in SLI mode. Things were very noisy with the stock cooler and couldn't take it anymore. Had to find something to replace it.
1. I had initially setup the cooler as documented, but hated the idea of having to 'manually' set the fan RPM rate. I was too afraid that setting it to a low setting for quietness would result in lack of cooling when it was required for higher usage (Crysis). I have read other people complain that you couldn't connect the fan power directly to the VGA board.
2. The other problem was that I use to have both SLI connectors on my cards. Don't know if this helped things or not, but now you can only connect one due to the height and width of the cooler. It takes one of the SLI connectors strips. At first, I thought you might be able to slip it between the cooling fins, but theres a wire that goes the height of the cooler to hold the fan in place. I'm sure this could be removed and utilize a different mounting device for the fan.
3. Another problem is the cooler 'closeness' to the VGA card. The cooler was actually touching some of the capacitors when it was first installed. While cooling was better than the stock fan and provided quieter operation, I didn't like the fan being so close and touching the capacitor.
So after giving some thought, I went back and conducted the following operations:
1. I went to Radio shack to get some shrink tubing and cut off the power connectors of the stock fan and this fan. I also cut off the fan control assembly and shorted the wires going to the control knob. I took the fan out and rotated 180 degrees so the wires could come out the bottom instead of the top. I then spliced the stock connector into the fan and plugged it into the board. Since this fan power draw was smaller than stock (.5 amps), I knew it would be safe. The only downfall was that there was no way of connecting the yellow wire to send back RPM speeds via Speedfan. Fan is as quiet as can be now and speeds up when it needs to as it is now controlled by the VGA card.
2. I am only using one SLI connector now. I haven't seen any difference in performance, but if it is advisable to have both ribbons connected, then I'll have to remount one of the fans. If you are running more than two cards, then you will have to remount at least the middle fan.
3. I used the included copper plate that is utilized in cases where the VGA chip may be set below the surrounding perimeter edge. I figured that this will give a bit more standoff and allow a little more cooling air to sneak in behind the cooler. It worked out much better and even gives me better cooling numbers now.
The only other thing I didn't like was the included backplate (the part that goes on the backside of the card). There was a single rubber cushion that would apply pressure to the card. The problem is that there were a lot of chips in this area of contact on my board. Fortunately, I was able to use the stock plate that have four contact areas that were around these chips. Works just fine and appears to disperse the contact pressure to four areas rather than just one.
1. I had initially setup the cooler as documented, but hated the idea of having to 'manually' set the fan RPM rate. I was too afraid that setting it to a low setting for quietness would result in lack of cooling when it was required for higher usage (Crysis). I have read other people complain that you couldn't connect the fan power directly to the VGA board.
2. The other problem was that I use to have both SLI connectors on my cards. Don't know if this helped things or not, but now you can only connect one due to the height and width of the cooler. It takes one of the SLI connectors strips. At first, I thought you might be able to slip it between the cooling fins, but theres a wire that goes the height of the cooler to hold the fan in place. I'm sure this could be removed and utilize a different mounting device for the fan.
3. Another problem is the cooler 'closeness' to the VGA card. The cooler was actually touching some of the capacitors when it was first installed. While cooling was better than the stock fan and provided quieter operation, I didn't like the fan being so close and touching the capacitor.
So after giving some thought, I went back and conducted the following operations:
1. I went to Radio shack to get some shrink tubing and cut off the power connectors of the stock fan and this fan. I also cut off the fan control assembly and shorted the wires going to the control knob. I took the fan out and rotated 180 degrees so the wires could come out the bottom instead of the top. I then spliced the stock connector into the fan and plugged it into the board. Since this fan power draw was smaller than stock (.5 amps), I knew it would be safe. The only downfall was that there was no way of connecting the yellow wire to send back RPM speeds via Speedfan. Fan is as quiet as can be now and speeds up when it needs to as it is now controlled by the VGA card.
2. I am only using one SLI connector now. I haven't seen any difference in performance, but if it is advisable to have both ribbons connected, then I'll have to remount one of the fans. If you are running more than two cards, then you will have to remount at least the middle fan.
3. I used the included copper plate that is utilized in cases where the VGA chip may be set below the surrounding perimeter edge. I figured that this will give a bit more standoff and allow a little more cooling air to sneak in behind the cooler. It worked out much better and even gives me better cooling numbers now.
The only other thing I didn't like was the included backplate (the part that goes on the backside of the card). There was a single rubber cushion that would apply pressure to the card. The problem is that there were a lot of chips in this area of contact on my board. Fortunately, I was able to use the stock plate that have four contact areas that were around these chips. Works just fine and appears to disperse the contact pressure to four areas rather than just one.