Revolution GPU and CPU STILL in development.
- Thread starter SoVos20
- Start date
- Status
A bigger fan can move a higher volume of air at a lower rpm than a smaller fan. Think of a big water pipe vs a small water pipe. A bigger water pipe can move a lot more water at a lower pressure. A smaller water pipe has to be vey high pressure to move the same amount of water. Another example is Voltage vs Current. Voltage is analogous to the air pressure and Current is the volume of air. To move the same amount of air, a small fan will be noisier because of higher rpms/turbulence. Now with respect to the Revolution you obviously cannot put in a huge fan not only because of the size but because you do not want a HUGE gaping hole taking up half the side of the console.
.
Yes, but jvd was talking about a bigger processor fan vs a chassis fan. A 120mm fan will have a big dead spot in the center so it might not be ideal for a processor.
randycat99
Veteran
Given the same drag load (that of the air path through the entire device), the same airflow implies the same pressure head (as Ty has presented). You can't just read-off the cfm rating on big fans and little fans and expect that is what they will deliver when power is applied. The drag load of the air passage (and any dissipation surfaces therein) will greatly impact how much "cfm" the fan actually delivers. Generally, going from a larger enclosure to a smaller enclosure, while dissipating the same amount of heat, the smaller enclosure observe higher air velocities (to maintain a target flow rate through a confined air passage), which will likely require an array of smaller fans working quite hard to deliver a considerable pressure head AND net flowrate. All of that together, suggests more noise than a larger enclosure that could use a larger fan and slower air velocities.
Large fans deliver higher flowrate, but typically at lower pressureheads, hence will not achieve optimal flow for high restriction, highly confined airways. This can be compensated somewhat by driving to higher rpms, but then you lose all the benefit of quiet, lowspeed operation.
Small fans deliver less flowrate, and arguably are not much better at forcing a pressurehead. So you really have not many options other than to run at higher rpms, with associated impact to noise.
Large fans deliver higher flowrate, but typically at lower pressureheads, hence will not achieve optimal flow for high restriction, highly confined airways. This can be compensated somewhat by driving to higher rpms, but then you lose all the benefit of quiet, lowspeed operation.
Small fans deliver less flowrate, and arguably are not much better at forcing a pressurehead. So you really have not many options other than to run at higher rpms, with associated impact to noise.
Well if you're stupid, and left the airflow passge to flipping a coin then yeah, that would be the limiting factor with regards to CFM.
This is where your assumptions get iffy. A smaller case doesn't equal to more "drag" if designed intelligently. Take a look at how the GCN's airflow is routed if you need proof.
The flow depends on how you route the air passage not whether or not it's a confined space. Again look at GCN if you need real proof.
For those who do not know what GCN's airflow looks like, it's basically a wind tunnel running straight from one side of the GCN to the opposite side. There is an air inlet on one side of this tunnel and an air outlet on other side. The fan is mounted directly on the outlet side pulling hot air out through this wind tunnel. Inside this wind tunnel is where the heat sink is mounted. The fins of the heatsink offers very little airflow resistance due to the fact it runs parallel with the direction of the the airflow through this tunnel.
The GCN isn't just a marvel of semiconductor design efficiency, it's an equal marvel at cooling design efficiency. 8)
randycat99
Veteran
That really is the key, isn't it- how much heatsink componentry is exposed in the "wind tunnel"? The more heat you have to get rid of, the more heatsinking you need to expose, and the more turbulence helps you eek out the most efficiency in the heatsink. The GC has a clean design, but then again, it isn't called upon to transfer the kind of heat output that we are anticipating in PS3/XB2/whatever. Once you add that condition, you are right back to square 1, where you literally are blowing a hurricane through a confined channel where extensive heatsink geometries must be exposed in order to exchange a considerable amount of heat ==> hence the conditions quite ripe for noise issues (i.e., no "free lunches" here whether or not you are using clever and elegant wind tunnel contraptions). To address that, there are few options other than opening up that channel and significantly increasing heatsink area, just so you can reduce air velocities and load demands on the fans... If you are stuck with a compact case, this will place a certain upper limit to "opening the channel", so it becomes a distinct balance of noise vs. cooling challenges.
To think that someone actually believes fancy GC windtunnel design will spontaneously "quiet cool" the sort of heat bohemoths coming down the line to us next generation...well, all I can say is, "good luck with that, slick!"
The irony here is that "windtunnels" happen to also make terrific tuned hemholtz chambers. So any noise generated within the conduit tends to conduct sound at the exit quite effectively. The "magic" of these designs really comes from doing things that prevent noise from being too obstrusive/distinct and making the tunnel as attenuative as possible (relative to a "tunnel", anyway). They are no means a "magic bullet" for quiet cooling, and it would be naive to suggest that they don't bring a number of additional issues to overcome for the benefits they provide. Once you add in heat transfer targets in the realm of 100's of watts, it becomes just as much an engineering challenge to overcome as anything else.
To think that someone actually believes fancy GC windtunnel design will spontaneously "quiet cool" the sort of heat bohemoths coming down the line to us next generation...well, all I can say is, "good luck with that, slick!"
The irony here is that "windtunnels" happen to also make terrific tuned hemholtz chambers. So any noise generated within the conduit tends to conduct sound at the exit quite effectively. The "magic" of these designs really comes from doing things that prevent noise from being too obstrusive/distinct and making the tunnel as attenuative as possible (relative to a "tunnel", anyway). They are no means a "magic bullet" for quiet cooling, and it would be naive to suggest that they don't bring a number of additional issues to overcome for the benefits they provide. Once you add in heat transfer targets in the realm of 100's of watts, it becomes just as much an engineering challenge to overcome as anything else.
MechanizedDeath
Regular
How does a smaller case have better airflow than a bigger case? AFAIK, there's a point of diminishing returns. The keys to cooling is CFM (cubic feet per minute) of air and surface area. If you've got a smaller case, you can get away with using the same fans as in a larger case. But this is assuming the heat conductor also has the same surface area. If you've got a smaller case, you've probably got less surface area to absorbe all that heat, which means you probably want to increase the airflow to compensate for the faster spread of heat. Crap, I think randy is posting the same thing as me.
I don't see Rev getting away with similar power to the PS3 and 360 without either going 65nm or just plain a weaker processor. If it's portable, it has to do this anyway, or it'll be a battery hog. I really don't buy it. I'd like to be proven wrong, but there's a reason laptops are always weaker than their desktop counterparts, and it's got nothing to do with the Wintel monopoly. PEACE.
I don't see Rev getting away with similar power to the PS3 and 360 without either going 65nm or just plain a weaker processor. If it's portable, it has to do this anyway, or it'll be a battery hog. I really don't buy it. I'd like to be proven wrong, but there's a reason laptops are always weaker than their desktop counterparts, and it's got nothing to do with the Wintel monopoly. PEACE.
Including the kind of physics and AI engines we have now in the developers kits is hardly going to maximize the power of a 115 GFLOP CPU. Unless your just talking about the same crappy physics and AI we have now times 200 (as in loads of bots). In which case its very easy to scale that down just by dropping the number of bots by a small percentage, nothing significant.
I think you need to take a time out and calm down. I just meant the 5th generation PowerPC architecture, like the XBox 360 CPU. I definitely should have been more clear but that's no excuse for your attitude.
I'm not saying developers will find it hard to use the full power of 360's CPU because its difficult to program due to multiple cores/threads. I'm just talking about being able to use all that power. If your used to £200 a week wages and suddenly you win the lottery and get £6 million would you know what to do with all that money? Not a perfect analogy but it'll have to do. How many developers will be able to produce a physics and/or AI engine on 360's CPU that would not be able to run on a 85GFLOP CPU without significant cut backs? Even a new licensed multi platform physics or AI engine would hardly max out any one system CPU.
Your just speculation as well, you have no way of knowing if I am wrong or right. But I think its reasonable to think that its going to be extremely tough for developers to produce the kind of engines which will max out a 115GFLOP CPU, especially looking at how lax developers have been at using current CPU's for AI and physics. IMO only the very best will even come close. Most developers will still focus on shiny graphics anyway.
What I'm witnessing is someone who isn't reading my post properly and is acting like a total ass. Sorry but I don't know why I'm being ridiculed for giving an opinion on next generation systems? Does this mean so much to you that it has to be taken personally?
1T-Sram was more expensive then standard DDR ram and both systems were releasing at the same time. But 1T-SRam-Q is more competitively priced now and 360 is using GDDR memory and coming out a lot earlier. Plus I did mention possibly using 448MB. Not that Revolutions design budget has to be portioned out the same as GC's anyway. For all we know Nintendo may cut back in one or two area's (no auxiliary ram for instance) and put that money into including more main ram this time (I hope so).
Anyway personally I don't think everything in a next gen system has to be the same or better as XBox 360 to produce results good enough that 99% of people won't tell the difference. Look at GC and XBox, GC is quite a bit lower spec but its graphics are very competetive and IMO Revolution could be a lot closer since it won't be released till a year after 360. In the end if Nintendo's system has a lower clocked GPU and its CPU is maybe 20% slower I don't think it will be important. The key IMO is equal ram and equal GPU feature set and close enough CPU and GPU speed. That's what I was getting at.
with water or this liqued metal cooler you can pipe the liquid to the rear of the case and have the radiator / heatsink by the air vents or the rear fan. That way the heat is released very far back in the case at basicly the exit and is pulled out with fans . WOuld be much more efficent than having a fan + heatsink in diffrent areas of the case .
Look at the vga silencer for graphics cards . It draws the air in from the case around it and pushes it out the back that way the hot air isn't just dumped back into the case .
IF you pipe it out of the case then the air inside is cooler . That is the main thing about watercoolers . You put the radiator near an exit fan (or an intake but not as good for cooling ) that way the hot air is piped away from the hot components and is then exhausted out of the case . With a normal fan the air is just twirled around and you need two fans 2 remove it (power supply and rear intake fan ) so right htere its 3 fans . You can do away with it with water cooling .
In my old case i had 2 120mm fans . I had my northbridge , gpu and cpu all cooled with these two fans along with my case and it was only 5c over room temp . As i've said before i've heard my hardrives spin up over it .
So you can a very quiet x360 or ps3 . Its just that this will be much more costly than a normal heatsink
randycat99
Veteran
As has been cited ad nauseam, the advent of water/liquid cooling doesn't solve the problem- only relocates it. With the power outputs we are talking about, you will still need either a rather large external heatsink and considerable airflow, or various combinations in between. Once you are talking about a "smaller case", you automatically don't have the space to use the giant heatsink you would need. So it will be smaller with aggressive forced-convection. You can't cheat the piper, here.
randycat99 said:
I think its you who hasn't been paying attention .
If you can relocate the heat to some where like the rear of the case near the exit you will not need as much cooling because your not just pushing hte hot air back into the case your removing it from the case . This will be much better than just a normal heatsink fan combo . and it doesn't need to be anywhere near as big
Don't see whats so hard for you to get .
MechanizedDeath
Regular
You intend to have an external heatsink like an amp? You know how how these components get? We might not be putting hundreds of watts through this thing, but it's still gonna get damn hot. My 265W cap amp can get ridiculous at times, and it's got more than a square foot of heatsink as a casing. I don't want to think what a console with only a relatively small heatsink on the back would be like. Besides which, Nintendo is supposedly playing up the portability of this machine. I hope it's not for portable gaming, but you're not gonna want to wait for your Rev to cool down before you can pick it up or put it in your bag.jvd said:
It's generally better to keep the heat inside the case of a device you plan on dragging places with you. There's a reason laptops don't have external heatsinks. I can't put my 1.6GHz Pentium-M lappy on exposed flesh b/c the memory and wireless car bays get hot. Not warm, hot. I can tolerate the heat for a little bit, but I'm an exception. Most people would wince at the pain. I think you're being a bit naive about the cooling problems that form factor will pose. Again, either a smaller process (possible if it's a late 2006 release) or low-K/low-clock solutions...or some combination of the two. That and you'll still need temp-controlled fans. Just the facts of life. Until Nintendo created the philosopher's stone, they can't cheat the laws of equal exchange. Heat in and heat out.
PEACE.apparently you guys just don't read others posts .
Using the liquid metal heatsink that was linked early or watercooling you can move the radiator portion to the rear of the case near the vents . Not outside the case but at the rear by the exhaust vents with a fan pulling the air out of the case . Its much better than having 2 hot chips dumping heat into a very small case and needing more fans to cool it .
sorry but this has nothing to do with anything . THe rev will require a tv so its not a handheld you aren't going to drag it with u and use it while your dragging it around .
randycat99
Veteran
Let's not confuse "good airflow management" with power-in = power-out. Good airflow management definitely pays off in large enclosures and small enclosures alike, and should be an integral part to any thermal management design. What we have been discussing is the crucial implications of large finite power consumption dictates a large finite power dissipation which will be sizable or noisy (or somewhere in between), good airflow management or not. Dissipating 100's of watts doesn't happen by magic, whether it is 2 heatsinks inside a box, one external one on the back of the case, or one inside a cooling tunnel.
randycat99
Veteran
I guess it is obvious to you now.
MechanizedDeath
Regular
jvd said:
I read your posts just fine, though that's a nice catch-all excuse.
Moving the radiator around in the case doesn't change the need to have a certain amount of air flowing over it. So you still don't change your cooling solution, just relocate it. I don't see how it makes it better. CFMs should be the same. You'll just get to work with different ducting. Not to mention, the heatsink/conductor portion of the heat exchanger still dumps a certain amount of heat into the case anyway through radiation. So you still want some sort of wind tunnel effect. You know this for a fact? I assume the Rev was shown with a docking bay style device for a reason, and I have a hard time believing it's for portability along the line of the GC handle. I believe they're going with a small form factor for other reasons, and the scuttlebutt (which I admit is hardly conclusive) seems to suggest that Nintendo wants some sort of hybrid portable console device. One that can perhaps be played on the go from a backpack or so. You don't necessarily need a tv when Nintendo already makes a plethora of portable video screen-equipped devices. Perhaps that's what the big secret about the Rev controller is. Or maybe they're really making some VR goggles, and with that "trigger-equipped controller". A number of unknowns.
Otherwise, why would they need to shrink the system below what it is now? We've already seen how that form factor will be a problem with high-performance CPU/GPU solutions (heat dissipation). That form factor is important for more than mere aesthetics and (I hope) more than just taking the system to a friend's house. I'd like to think it's gonna be wireless for that reason.
Now...what uses a docking bay and a small form factor...hmmm. I wonder....hmmm. :? 2+2=5. PEACE.
- Status
