RBX is for sale now.

Just bought my RBX. Very good job DD for getting a product out in a very short period of time!

I'll tell you how it performs when I get it, probably Wed.

Just thought I'd let you know.

can't wait to see some benchmarks. What blocks are you gonna compare it to?

WW and evo perhaps?

Soja said:

WW and evo perhaps?

Click to expand...

OK, I suppose....

That's a nice price. Might have to look into one of those come income tax return time...

I want one. I need something else to break. Cant wait to see how it does.

Edit: Plus it has a .025in thickness at the "heat voids".

There is a waterblock round up over at

http://www.liquidninjas.com/reviews.php?op=showcontent&id=50&title=Waterblock+Round-Up+2003

It is showing that the RBX is a pretty good block.

oi it is doing quite well.
cascade came out on top for all the tests except restricted flow with a high OC, hmmm. mind you the cascade does like lotsa flow.

heres the conclusion for lazy people (from the Liquid Ninjas review)

With that said, there are some things I've learned about the blocks.

The Cascade consistently performed in this review, and has to be one of the best waterblocks I've tested.

The MCW5002 definitely lacks in performance due to the fact that SwiftTech opted to use the socket-tab mounting system, although if your motherboard lacks the holes needed to mount the other waterblocks, this would be a fine choice.

The brand new RBX waterblock is also an outstanding performer, and because of it's price and availability in the states should really work well. I would urge anyone with a high-flow system to purchase this block with one of the more restrictive jets for the best performance.

The ThermoChill is certainly a strong contender for being such a new block to the market, and is certainly a good choice for someone in Europe having a hard time finding other blocks. I know ThermoChill is working on an impingement design for this block which should boost performance.

I'm not sure I can make any performance claim about the SlitEdge, except to say that I couldn't get good results out of my system. It may need much more heat input to really shine, like on an overclocked P4C.

The AquaGold block also has a strong showing for a new block design, and I would love to pit this up against the other blocks in a Peltier match, as I'm sure it's design would excel for TEC cooling. I might even pick this block over the MCW5002 for TEC cooling an Athlon due to it's use of the 4-post mounting holes which can allow for more torque and pressure.

Lastly the Maze4, which can be considered the budget block of this review. The design has certainly been refined over the old Maze3, and the performance is on par or slightly better than the results I got with it before. It's no longer one of the best performing blocks, but for a system where price is everything, I would certainly give this one a go.

Click to expand...

themodguy said:

There is a waterblock round up over at

http://www.liquidninjas.com/reviews.php?op=showcontent&id=50&title=Waterblock+Round-Up+2003

It is showing that the RBX is a pretty good block.

Click to expand...

There is one HUGE problem I see with the numbers in the review. WB's scale linearly in terms of heat input and flow rate. That is if water block 'x' has a C/W of .20 at 'y' flowrate, that C/W WILL NOT CHANGE when the heat load is increased or decreased. Such is NOT the case in the above review. In fact, block performance changes dramatically, some blocks actually change POSITIONS with a higher heat load. This is not possible. The differences in deltaT should be greater among the blocks as you up the heatload, not get closer together.

A quote from one of BillA's articles: "A good wb, or a poor one, is such at any and every applied heat load (for the same sized heat source)"

So what kind of conclusion should one come to after the "HUGE" problem you just pointed out Nikhsub1 in regards to the review results?

I don't think I would draw many conclusions, except where the temperature difference is very large.

As nikhsub1 said, blocks wouldn't change order (in terms of performance) from the varying load and flowrates.

In the test - no mention is made of how stable ambient temps are, and they even comment near the last graph that the effect of the cascade dropping to third could be experimental error (it had dT of 10.9 and the RBX #1 had a dT of 10.4). so if 0.5 degrees delta T was in the margin of error, most of the closely ranked blocks in that test are indistinguishable from one another (with that particular test setup). I would feel that the margin of error is most likely even greater than that, maybe plus or minus .7 degrees C, meaning any blocks within 1.5 degrees of each other couldn't be ranked differently in the test results.

--Of course that's just a guess and I have no idea how the testing was done, or what the ambients were. It did say however that up to 4 mountings were used - a plus, but I think Bill A used to do 10 mountings when he tested blocks, and probably would've liked to do more if time permitted.

While I have plenty to be happy about with such a review, it does still show its margins of error to be fairly large.

Let us take for example the RBX with the #1 nozzle.

At 2000MHz/1.65v the difference between low flow and high flow is 0.8C in favor of the high flow.

Looking at Bill Adams' perfectionist test results and one can predict that the difference would only get larger at a higher heat load.

Now look at RBX #1 at 2400MHz/1.85v. The difference between high flow and low flow is now just 0.1C, which is certainly quite a deal less than the 0.8C seen for the low heat load tests. Further, the #1 config leaps ahead of the #4 config at higher heat loads despite trailing by 1C or more over the #4 config in every other test, which again is counter-intuitive.

Indeed, look at the MCW5002 results. For the low heat tests there is no difference measured between high and low flow. If this was accurate then one would predict that there would be no difference at higher heat loads too, but we see a 0.9C difference.

Look at the thermochill at high heat load. It's actually doing better at low flow than with high flow, which is again in total contradiction to everything which is understood, and in contradiction to the results shown at low heat load for the same block.

Let's look at across blocks now. Compare the high flow / low heat for Thermochill/Aquagold/Maze4. They are all equal (within 0.1C) so one would naturally predict that as we raise the heat load that this similarity would remain. Now look at the high flow / high heat for the same blocks and we see 2.6C range that separates the blocks.

Basically the results themselves would seem to indicate a margin of error of at least ±1.5C for any particular data item given, and possibly greater.

Any two data items within 3C of each other are therefore indistinguishable.

One may think it odd that for someone who makes the block that "wins" the review would be critical of the review, but I truly believe that for there to be any conclusions to be drawn, that one must understand the nature of the data being presented.

Albigger said:

As nikhsub1 said, blocks wouldn't change order (in terms of performance) from the varying load and flowrates.

Click to expand...

I think that is true only for varying heat loads. Differences in flow rate applied to a particular block show as the differences between the colored lines in the graph above. For any particular block, these differences could be bigger or smaller, depending on how the block reacts to the flowrate change.

As a very satisfied owner of a SlitEdge, I'm mystified by the failure of the block in the tests. BillA's tests showed it to be very close to a WW, yet at LiquidNinjas it appears to be extremely poor. Maybe the individual block tested had manufacturing faults?

FIZZ3 said:

I think that is true only for varying heat loads. Differences in flow rate applied to a particular block show as the differences between the colored lines in the graph above. For any particular block, these differences could be bigger or smaller, depending on how the block reacts to the flowrate change.

Click to expand...

Thanks - it was late. momentary lapse in judgement when typing.

And one more to show the same C/W at varying heatloads:

sbud said:

So what kind of conclusion should one come to after the "HUGE" problem you just pointed out Nikhsub1 in regards to the review results?

Click to expand...

The conclusion is that there are lots of errors in this test. Let me explain more:

deltaT at 2000Mhz, 1.65V
Cascade = 4.7
RBX = 6.0 (using the best value)
Difference = 1.3C

Now look at this:
deltaT at 2400Mhz 1.85V (heatload probably 25-30W more?)
Cascade = 8.6
RBX = 9.5
Difference = .9C

The above CAN NOT be correct, it is an impossibility. The difference between the Cascade and the RBX at the higher heat load MUST be greater than the 1.3C difference of the low heatload test. Same goes for the differences in the low flow test where the Cascade is #1 at lowflow/low heat but then drops to 3rd when the heatload is increased at the SAME flow, again not possible.

succinctly put Cathar
would that people could understand it

be cool

no, nikhsub1, the c/w value CAN change

because of the inherent design faults with things that need water flowing and also secondary heatpath effects

<<<and there's one more variable you're not thinking about>>>

for example, i would bet you that a cascade would just *stop working* if flow was cut down to the extent that the cross-flow coming out and exiting the impingement cups actually kill the impingment action

or to think of something else, there are "flow optimized" block designs, even though cathar keeps saying there aren't, but there ARE, and here's the example:



channels cut at 0.05mm vs. channels cuts at 0.5mm

over an <top view> area of 3cm^2

and so therefore the 0.05mm version obviously has more surface area right?

add an eheim pump into the equation

AND

now the 0.5mm is gonna come out on top right? cause the 0.05mm is just "too restrictive" right?

that little cute eheim thing is proally gonna treat the 0.05mm channels like 4 meters of head or something, and therefore flow rates end up like this:

0.5mm = ~maybe 90gph WHILE the 0.05mm ends up like ~maybe 3gph, which obviously isn't too cool for the 0.05mm right?

BUT

when you throw a bilge pump into the equation or that bosch fuel 10 bar pump i keep thinking of buying that supposedly does ALMOST NO PRESURE DROP AT 5m HEAD

THEN

with a maintained 100gph though BOTH blocks, OBVIOUSLY THE ONE WITH THE MORE SURFACE AREA IMPLEMENTED THE SAME WAY IS BETTER RIGHT?!?!?!

so BEHOLD;

now the blocks take on different characteristics, and therefore with more pressure, the impossible 0.05mm micro-channel block tops all


i really don't have any idea about what exactly the flow dependancies or otherwise characteristics of Cathar's cascade is, because only cathar knows that


BUT,

again i do believe that Cathar makes some very fine work and some of the best blocks out there, and i respect him for that

and also i TRUST Cathar's work, and i would expect that should i ever need to find the best, Cathar's the way


but when you force the porsche to run as a tractor, AND FEED IT DIESEL, obviously its not going to be very happy right?



p.s. sorry about the CAPS or the tone of the post, but otherwise i'm just trying to get the point across that c/w values do change, especially if there are other variables in the same equation, cause the pump characteristics must also be considered

nothing against you nikhsub1