LRWW vs SlitEdge - a comparison

[MikeTimbers]

Little River White Water vs. BeCooling SlitEdge




A quick mini-review.


I need my LRWW for a hot-running XP2100B which needs 1.95v to get to 12x200 so using it for an XP1700B that only needs 1.65v for the same speed seemed kinda wasteful.

I saw BillA's review of the SlitEdge and thought for the price, it had to be worth trying.

The block came on Thursday so it went in today. First off, to drain the water and remove the LRWW. I removed the LRWW and drained the water straight from the block into a bowl so I could re-use the water. Then I removed the out tube from the LRWW's Y-adaptor and the in tube from the centre barb.

Since I had no mounting hardware for the SlitEdge I thought I could just use the LRWW's but unfortunately the LRWW's mounting bolts are too big to go through the holes in the SlitEdge.

So that meant removing the motherboard to remove the LRWW's bolts. I have a Swiftech MC462B on the shelf (literally) so I checked whether the Swiftech bolts would fit and they did, so I attached the Swiftech mounting posts to the motherboard and replaced in the case. Fortunately, this is in my Lian-Li PC60 with the removable tray so it didn't take too long.

Then I removed the chip and re-applied some AS2 (I've been using the same tube for two years and it's still going strong!). The old tubes were then applied to the new SlitEdge and the block carefully placed over the chip. Since the SlitEdge is quite a bit fatter than the MC462, I used only the spring and one washer when trying to get the bolts through the SlitEdge and into the posts.

The first one went in OK and still pushing the block down hard against the chip, I tried to get the diagonally opposite bolt in. My tubing was in the way, but with a bit of jiggling, I got the second bolt in and partially screwed in. Then the other two went in and then I tightened them up using the standard bit by bit method. Soon all four bolts had bottomed out in the mounting posts so I knew I had a good contact.

I started re-filling using my fill tube which is a recent addition from when I tried it with just the BeCooling Pump/Res the first time. Once I'd got most of the water back in (since the new system no longer included the two outlets of the LRWW and the Y adaptor, the new system has slightly lower water capacity), I turned on the pump and watched for leaks.

The water was very frothy as it whizzed round the system. This was collecting in the fill tube as foam which I got out using rolled up kitchen towel, and then topped up with some of the left-over water.

Once the foaming settled down a little (and I didn't wait too long ) I re-connected the power and the externals and tried the power switch. It booted first time and a quick trip to the BIOS' health screen showed it wasn't going to burn up as the CPU temp settled at a couple of degrees over system. Then I re-booted and the 8RDA+ kicked in at its real settings of 12x200, 6-2-2-2, with the Vcore at 1.65v.

No problems with the boot and the health screen looked OK despite the still incredibly foamy water so I let it boot into XP.

No problems so far, so I kicked off my Distributed Folding client which loads the proc at 100% and is an incredibly tough test for a computer. MBM5 was recording the temps for me and I was very pleasantly surprised. Now the foam has gone so how did the two blocks compare?

Cooling components:

BeCooling Pump/Res with maxi-Jet 1000
Black Ice Extreme
Evercool 120mm with 120mm gutted fan as a shroud sucking air at 7v.

With the LRWW, the CPU temp was between 6C and 7C over system
With the SlitEdge the CPU temp was ... exactly the same!!!


I'm not going to claim that the SlitEdge is the equal of Cathar's wonder-block even though it is using some of his carefully calculated principles but I have to say that its results speak volumes. For £37, inc. VAT, this is one hell of a block!

 

[JFettig]

Hey man, Thats a neat little test you did. Most motherboards do not have acurate enough sensors, or the sensors are reading off the rong places. Like yours could be reading the air behind the cpu, wich is exactly what the 8rda does. I know cathar did a little testing with this on his, nothing lined up.

BillA's test shows how they do perform compared to eachother a little more acurately. But like your tests show, some people wont notice a difference.


Jon

 

[MikeTimbers]

Maybe the cpu itself is a completely different temperature but as you say, I'm only going on the basis of the 8RDA's in-socket diode so who knows? But the ain thing for me is that my overclock is still good and the LRWW can now go in my other rig.

 

[OC Noob]

Having a great block is cool, but the rad is where its at. If you've got a good system for cooling the water down any block will do just fine.

 

[Soja]

Having a great block is cool, but the rad is where its at. If you've got a good system for cooling the water down any block will do just fine.

I agree. Having a good rad/heatercore vs a cheap radiator (ie rad w/ fins wrapped around the coil instead of bonded ) makes a big difference compared to a standard block and a fancy block that drops your temps a couple C. /me learned from experience

 

[Cathar]

I've always found that one of the biggest chores when comparing waterblocks is trying to eliminate the mounting variation factor. Mounting variations can very easily result in ~3C differences (sometimes even much greater).

BillA's testing involved mounting each block at least 10 times to ensure that such variances are measured and controlled, something for which a "1 mount of each block quick test" can't really account for.

The point being that if you spend the money on a good block because you wanted the best possible cooling performance, then it really is worth your time to remount the block a few times until you've figured out when you've got the block mounted correctly.

Sadly mounting variations are pretty much an unavoidable fact of life, and it really is up to the user to ensure that they are getting the most out of their block, and it partly explains why some people may not see results that match up with more exacting testing procedures.

 

[Soja]

Cathar, do the swiftech 3 lug mounting system cause mounting variations like the 4 hole/spring mounting style? BTW does swiftech use this system because their block is within amd weight spec?

 

[Cathar]

Cathar, do the swiftech 3 lug mounting system cause mounting variations like the 4 hole/spring mounting style? BTW does swiftech use this system because their block is within amd weight spec?

The true causes of mounting variations are many and varied:

1) Uneven clamping pressure across the die (admittedly more of an issue with spring mounts, than lug mounts)
2) Uneven thermal paste application initially
3) Uneven application of the cooling device during installation resulting in squeezing out some of the thermal paste causing an uneven thermal paste condition
4) Total clamping pressure - the more force that's applied the lower the temperatures - a well documented fact. The 4-hole method allows for more clamping force to be applied safely, while the lug method risks breakage of the socket
5) Hosing torque - all blocks are held onto the CPU through varying degrees of spring based pressure - the hosing pretty much acts as a lever to try and prise the waterblock away from the CPU. If the hosing is allowed to "hang off" the waterblock rather then be supported properly and evenly, then this can be a huge impact on final clamping pressure evenness, and hence cooling performance
6) Less common - but CPU not mounted flat in the socket - sometimes an issue for motherboards with in-socket thermistors that bend up and touch the base of the CPU

There are probably more that I haven't thought of from the top of my head. With such a large number of potential variations it's hard to say if the 3-lug, or the 4-hole method is better because there are so many factors that can influence both outside of the clamping procedure itself.

 

[Soja]

Good points. Boy do I like the ease of 3 lugs though

 

[OC Noob]

The true causes of mounting variations are many and varied:

1) Uneven clamping pressure across the die (admittedly more of an issue with spring mounts, than lug mounts)
2) Uneven thermal paste application initially
3) Uneven application of the cooling device during installation resulting in squeezing out some of the thermal paste causing an uneven thermal paste condition
4) Total clamping pressure - the more force that's applied the lower the temperatures - a well documented fact. The 4-hole method allows for more clamping force to be applied safely, while the lug method risks breakage of the socket
5) Hosing torque - all blocks are held onto the CPU through varying degrees of spring based pressure - the hosing pretty much acts as a lever to try and prise the waterblock away from the CPU. If the hosing is allowed to "hang off" the waterblock rather then be supported properly and evenly, then this can be a huge impact on final clamping pressure evenness, and hence cooling performance
6) Less common - but CPU not mounted flat in the socket - sometimes an issue for motherboards with in-socket thermistors that bend up and touch the base of the CPU

There are probably more that I haven't thought of from the top of my head. With such a large number of potential variations it's hard to say if the 3-lug, or the 4-hole method is better because there are so many factors that can influence both outside of the clamping procedure itself.

I had no idea mounting was so important. I always screw the block down pretty hard and was worried I may be doing more harm then good.

Thanks for the tips.


Wish I had the dough for a WW or Cascade block. Those extra degrees will have to wait tho.