Topic: KNC just posted pics of their case, PCB samples coming next week \o/ - page 2. (Read 8273 times)

somebody please put down this dog.

Again, rather than actually refuse my claims directly you just dismiss them nonchalantly. No problem, I don't have KNC preorders. Ignore away.

Just banter.

Dude, a 'dogie'; a motherless or neglected calf, doesn't play fetch, a 'doggie' does, or 'doggies do'.

Incidentally, 'doggie doo', is pretty much most of what you'll find Dogie espouts...

Dear Non Engineer,

You don't have a clue about any of the technical information in this thread. Again you're just dismissing the logic spelt out line by line with the sentiment "they must be getting it right".

tldr: Greed/desperation > logic.

Ah dammit. I knew this!! I was trying to coerce Eve out of bot mode...

http://www2.warwick.ac.uk/newsandevents/pressreleases/researchers_apply_computing

Actually yes. It's already been proven that the chicken came first

And saying a case is like a egg and the chips = chicken, doesn't make any sense. They're not interrelated at all.

The case can exist without the chips.

But Eve, did the chicken really come before the egg?

Nice to see you totally ignored how I proved TDP in that test was 175W 75% of the time.

It doesn't particularly work that way in heatsinks, that cooling power is higher at 90C. The limit of heatsinks is the immediate transfer of energy from the base plate to the cooling fins, where THEN the cooling is increased. Heatpipes max out pretty quickly, and the chip temps just rocket after the limit.

This choice of heatsinks along with their dodgy/failed thermal simulations which they never properly answered after 3-4 pages are suggesting they either are getting it very wrong @ 250W, or they really don't believe it will be >200W and aren't catering for >200W at all.

As I've also said before, the need for them to be getting things right is blinding people from seeing logic set out line by line.

Had to edit just to add this quote, its too precious:

If you believe that you're a sucker. People sell overpriced stuff all the time. You're also not taking noise level into account, more expensive fans may be optimized for lower sound levels, rather then more air flow.

The only way you would have any relevant experience is if you'd actually fried CPUs.  Simply buying expensive heatsinks and thinking you needed too doesn't mean you know what the minimum actually is.


I don't see 175W anywhere on that page. The chip was massively overclocked and running at 1.385V, way out of spec.  Who knows how much power it was drawing.

And any event, the higher the temperature difference the more energy can be removed. Cooling power at 90C higher then cooling power at 79C.


Then... it's probably the intake? That's how Avalons are setup as well, airflow from front to back.

Anyway, what exactly is your argument?  You think KnC doesn't know what it's doing because they're using fans that cost a few dollars less then other fans? They don't need to buy the best fan possible, they only need one that gets the specific job done. And likely, the chips will put out way less then 250W anyway.

You keep hammering these points over and over again, but even if you're right it doesn't prove that KnC's chips won't ship.  It makes no sense.

Darwin might have something to say about that. But why am I talking to a bot anyway? (HMMM, evolution in software, perhaps I'm contributing to the advancement of our AI overlords ...)

What's the point of showing us the case when you don't have the chips. Is like showing us the egg without the chicken. You cannot have an egg without the chicken laying/produce an egg. Show us the Chip Please.

P.S 250w for one chip is just the worst case scenario.

So I'm sure they'll be fine with the HSF they selected.

Yes, I am aware what fan it is.... If I said my car does 350mph, would you believe me? No, its a number on paper. When you look at the benchmarks or test them yourselves, they just don't perform up to the advertised specs.

Have a look at a table of proper fans. http://www.quietpc.com/120mmfans There is no reason why the dirt cheap would outperform the considered flagship 120mm parts, unless the specs are just overstated. Overstated specs -> almost no one will ever be able to say otherwise, and hey who would listen to them when you can buy these holy grail fans for so cheap?



Let me show you another way. This is a review/test of the heatsink by X-bit.
Methodology: http://www.xbitlabs.com/articles/coolers/display/arctic-freezer-i30_4.html#sect0

Look at the picture of the test, we can see that at the screenshot the TDP was 27W. If we extrapolate to find the peak, its ~6.5x higher = 175W. Note that wasn't a 100% load in the sense that it was only at 175W for roughly 75% of the time. http://www.xbitlabs.com/images/coolers/arctic-freezer-i30/zfullscr_big.png *not embedded because its massive*

If you then look at the cooling performance at this 75% 175W: http://www.xbitlabs.com/articles/coolers/display/arctic-freezer-i30_5.html#sect1 Room temp was 24C. At max fan speed the chip was still reaching 79C.

So lets recap:
Test was 175W for 75% of the time.
Ambient was 24C, which is fair.
Fan speed was max, which is fair for testing the max cooling capacity.

However the chip still reached 79C. We want 250W 100% of the time for KNC, and 320W 100% of time for the reported specs to be true. The rating specs are asking for more than double the actual tested heat output, so there is absolutely no way that heatsink is going to get away with 320W.



Oh, and I'm assuming you're aware that where those case fans are mounted is either the intake or the fan is mounted the wrong way.

^^ I believe if you look to the top of his head he has an antennae to the onboard GPS soo many have been asking for!

Except the PSU wires will be strung through the mouth, so it will look like it's throwing up a rainbow.

^ They should have one of these pillows with every order! This way we could all sleep better at night .

Dude I'm looking at this on an iPad, I can zoom right in.

Here you go, just used an app called 'instaeditor' and played with the levels;

So how come no light is getting into the case on the right? The lighting appears to be coming in from the right, so we should be able to see the aluminium on the bottom of the case?

It is, there's two cases there. The one on the right has both parts (bottom being scratched), the one on the left is just the base by the looks of things.