99.995% Pure Silver RBX (S-RBX)

[SysCrusher]

one time at band camp, a bear came, and i had this direct-die block thing that actually did pretty well, and then everyone hated the idea and kept telling me that it'd catch fire and die, but now that since its been six month and its still good, i think its safe still, and ever since i started lapping all the athlon die's that i can find, i'm getting a steady 1 celcius drop on every single barton die with the blue/grey/pink film stuff removed, and then using the latest arctic silver does nothing, i think that maybe its a bit pointless to go on and on about how it is if everything was silver and all...


i mean seriously, there should be nothing mythic about these waterblocks and just cause its silver doesn't mean it *has* to be better

ROFL I get what you mean and I agree.

and besides, its the overclockers forums, so plz speak in mhz increments/increases rather then temps from ur bios, cause that's not gonna help


e.g. cathar's silver gives him a better overclock

Sorry but thats pointless too. I got a crappy Barton 2500 here that doesn't do jack no matter how cool you keep it or how much voltage you give it. The best it does is 2200, but it does that with air cooling as well. Then you can't compare to some one elses CPU because just about each and every chip is different. Not to forget each users different setup will overclock differently. My old pally 1700 does 2000 no matter what block I use. You might as well throw all the voltage at it and see which one can keep it cooler. That would be a better comparison. The higher over clock then the other is pure BS!

Joe C's work is about the only thing I will put my trust in.

 

[nikhsub1]

3.5C better from copper to silver indicates a major measurement flaw. Or something else...

I have to agree with this. Rokk what board/CPU and temp measuring are you using?

 

[thorilan]

has anyone actualy tested the rbx to make sure its pure silver? i know many people are purchasing 92%- 93% silver and saying its 100% . i know cathars silver is 99.997% pure because i personaly have it tested

and some thing that a lot of people dont realize when they only expect .5 of a c improvement. it all depends on the block design a a whole. so some designs can show up to 1.5c diference with tweeking .
as for the animosity in the thread. we all know who did the research and came up with the best block and weather the other block is a ripp off or not since cathar is cool with it we should just let old dogs lay
( personaly i dont like it but thats me)

 

[SysCrusher]

as for the animosity in the thread. we all know who did the research and came up with the best block and weather the other block is a ripp off or not since cathar is cool with it we should just let old dogs lay
( personaly i dont like it but thats me)

Why not pick on the Evo and countless others? Heck, why not pick on the guys who started capping finned air heatsinks and running water through them? Seems to be the order of the witch hunt around here.

Did you know Intel has had a cascade design in it's white papers on a micro design? Do you know who originally had a cupped design water block on a grander scale? I can say no one here originally thought of these ideas including myself. The BS witch hunt fan boy stuff ticks me off. Does nothing for the community as whole one bit, foster new ideas and even further other ideas to make them better.

Whats everyone going to do when Intel and AMD actually have some stiff competition? Whine and cry some more? LOL

 

[thorilan]

Why not pick on the Evo and countless others? Heck, why not pick on the guys who started capping finned air heatsinks and running water through them? Seems to be the order of the witch hunt around here.

chill out dude
im not picking on anyone this thread is named rbx so thats whats being discussed

 

[Cathar]

Looks like some people can't take competition to well.

Interesting.

I've stopped selling blocks and the S-RBX isn't for sale. What competition?

I'm only interested in the truth.

 

[Cathar]

I have to agree with this. Rokk what board/CPU and temp measuring are you using?

From a post by Rokk1972 (aka CPU Killer @ Pimprig)

Originally posted by CPU Killer @ Pimprig
Well I spent part of today testing the S-RBX to the RBX. I know that the S-RBX should perform about 10% better than the copper version, but actually seeing the numbers is always better to me than "it should perform better".

So I set up a test rig... Nothing flashy or fancy:

PIV 1.6 Ghz OC'ed to 1.9 Ghz with Vcore @ 1.85
512MB Corsair PC3200
Abit IT-7

All temps were abtained locally via Hardware Doc.

1st up was the RBX. Ran Prime95 for 45 min and recoreded temps...Ambient (case) 28deg C - CPU Diode 33deg C - Diff. of 5.0deg C

2nd up was the S-RBX. Ran Prime95 for 45 min and recoreded temps...Ambient (case) 29deg C - CPU Diode 30.5deg C - Diff. of 1.5deg C

So we can see that the S-RBX performs right at the 10% mark better than the RBX (3.5deg C better overall)

So there's no monitoring of the water temperature at all, and case temperature, as we all know, does not reflect the water temperature in the slightest.

 

[sganimefan]

Yes we have been playing with some silver RBXs but as far as pricing that has yet to be determined. Very expensive, so as the saying goes if you want one and price is no object then the price does not matter. I will ask our engineer about the production cost of the silver RBXs and if it really is going to be a real block and not just a novelty.

Best Regards
Dan Stephens

Yep... That's the email I got from dangerden. If it's really gonna be very expensive then there're just not going to be able to compete with SS Cascade... Especially since the SS Cascade uses Silver of higher purity if I remember correctly.

 

[vonkaar]

There really isn't any chance of it 'competing' with the Cascade-SS anyway... The SS wasn't ever really on the market. Cathar made a few, posted pics and then received enough feedback from people to make a few more. Unlike Dangerden, Cathar didn't have to make any 'investment' in a risky 'high-cost' waterblock. For the most part, those that got the SS were able to afford it because Cathar is enough of a masochist to not charge us for labor... he even lost money on the 'parts'.

If Dangerden wanted to enter the market with a few of these, they would have to pay someone to make them... pay for the parts... justify the manufacturing based on a projected profit... it really would be a risky investment.

 

[Graystar]

There is about a 10% increase in thermal conductivity between the copper and silver. What you'd really expect is to have 10% greater heat energy being transferred to the water (and even that depends on the input water temperature remaining the same.)

What happens to the actual temperature of the CPU depends on the radiator's ability to deal with the heat. If the radiator is overloaded and simply can't do any more for you, then it may just leave most of that energy in the water and you won't see a benefit. If you have a really good radiator with lots of air flowing through, you'll get rid of the extra energy and the CPU will be cooler.

What I'm saying is that there are too many factors involved to simply assume any kind of 10% difference anywhere. It's just not that simple.

I've see 5C+ difference is a single system between two of the same exact heatsinks. The difference was that one had a flat base and the other had been "lapped" by someone. The flat base was cooler.

Frankly, it's just difficult to control all the factors.

 

[fafnir]

no, cause again, the heat dissapated is affectly ONLY by effective surface area used, where better heat conductivity of the metals will only help to get the heat there better, while assuming purity is constant;

10% is not 10%

the heat will not leave the metal faster just because its more conductive, but because of a temperature difference over surface area, which instead is what the heat is also dependant on to dissapate


e.g. 1cm^2 is 100watts, regardless of material, given the temp is constant, where the better material can only help the heat get there faster;



so do not simply expect 10%



and besides, 10% is pretty hard to characterize or judge in this case where it really went...etc

 

[RickCain]

One more thing to throw into the testing of the RBX is what nossle was used by Rokk1972? If the stock one wasn't used wouldn't this be misleading the customer? There is also no mention of what pump / rad was used during the testing.

I don't understand why SysCrusher is getting so mad about this thread. When someone "claims" inovation or better results, shouldn't it be questioned before being hailed as the king; or better yet true? Nothing bad about hard questions about a product. The problem is nothing is getting answered or the data given is next to useless.

 

[Graystar]

fafnir, can you explain that more clearly? It didn't make any sense.

Also, we have no idea what your initial "no" was referring to. Can you put in a quote?

Thanks!

 

[lclark2074]

have you got temps yet and o/c

 

[SysCrusher]

I don't understand why SysCrusher is getting so mad about this thread. When someone "claims" inovation or better results, shouldn't it be questioned before being hailed as the king; or better yet true? Nothing bad about hard questions about a product. The problem is nothing is getting answered or the data given is next to useless.

What makes me mad is the whole fanboy thing. Who did what, who did this first, who's the fake and the list goes on. It doesn't achieve anything. Claiming something as "king" is useless because there is always going to be something better. Considering I like both blocks in question, I have the RBX and it's a nice performer. I run a maze4 GPU, Z-Chip in the loop and it still performs great. Lets test these blocks in a pelt setup against a Swifty and see. Thats where the Swifty truely shines. Just to many variables and different setups to make claims based on such. To me, I'd rather see someone stick as much heat to a die and see which one keeps it cooler.

Nothing bad about hard questions about a product at all. I, just like you, is very interested in the data also. Only thing I wish is if people would use more common sense before any false claims are made because their both great blocks. I really like to see hard data between the same blocks in copper and silver as I think the 5C better is a bit high. Then again, I could be wrong.

 

[Cathar]

I really like to see hard data between the same blocks in copper and silver as I think the 5C better is a bit high. Then again, I could be wrong.

Let's explain it in physics shall we, since this is what this stuff all revolves around.

The metal used does not affect convection (transfer of heat into the water). It affects conduction only. A metal that is more conductive for heat will spread the heat faster, and the temperature delta between a hot spot and a cold spot will be lower than a metal of lower thermal conductance.

Let's assume that C110 copper has a thermal conductivity of 375W/mK, and that pure fine silver has 410W/mK, which is about right according to most metal property texts.

For a ~100mm^2 CPU die running at a real 100W (very highly over-volted/clocked and under load) and a ~1mm thick base-plate, the C/W of the copper to move the heat that far is around 0.0254 (using the heat-spread models that I work with), and for silver it is 0.0232. Now silver will also spread the heat a fraction faster/further exposing slightly more surface area to be cooled, and the ratio here works out to around an effective extra 5% surface area.

So basically we gain (0.0254-0.0232) * 100 = 0.22C on the pure switch from C110 copper to silver in terms of conductivity gains.

For the extra effective convective surface area we gain around about a predicted 0.3C on that basis given the sort of convective efficiency that I predict is going on in the top-end blocks today.

Overall, we're looking at a physical reality of about a 0.5C gain when switching from C110 grade copper to pure-fine silver.

The Cascade SS actually changed a number of design elements in the silver implementation to trade off some conduction cost for greater thermal spread. This allowed it to realise a slightly greater gain. Some modifications were also done to the jets (and resultant flow distribution) and this resulted in a gain that was slightly more than the predicted 0.5C if the design was left unchanged. Not as large as 1.0C though (at least not for just a 100W heat source).

I was able to calculate, predict, and then ultimately measure in the final form pretty much bang on what the mathematics/physics was telling me was what would happen. These are the tri-goals that I strive for in my development.

It is also why, all competition aside, I am a fervent opposer when it comes to claims that seem to me to be more centered around sensationalist marketing through faulty (and undisclosed in these forums) testing. Having stopped making blocks (indefinitely), I feel that I am now more free to open my mouth to things that I see going on.

 

[SysCrusher]

Thanks Cathar! That makes perfectly good sense to me. I didn't think that switching between the two metals would bring such a high gain. One thing I think you forgot though. Change in block design can make a difference to those resulting figures - Better or worse. Depends on how well the block takes advantage of turbulance, velocity and not to forget flow.

 

[Cathar]

One thing I think you forgot though. Change in block design can make a difference to those resulting figures - Better or worse. Depends on how well the block takes advantage of turbulance, velocity and not to forget flow.

The Cascade SS actually changed a number of design elements in the silver implementation to trade off some conduction cost for greater thermal spread. This allowed it to realise a slightly greater gain. Some modifications were also done to the jets (and resultant flow distribution) and this resulted in a gain that was slightly more than the predicted 0.5C if the design was left unchanged.

You can make small gains by fiddling with the edges of the same basic design.