Do you need top & side Case fans?

[usually_quiet]

I have seen horrible dust accumulation in a PC owned by relatives, and by that I mean worse than what is shown in Redwudz's photo. The PC in question was near a cage housing a pet c*o*c*k*atiel, which generated an amazing amount fine debris for one small bird.

You're making me think and wonder if it was my cats creating that "debris" in my machine that time from my experience mentioned about a machine having a performance boost after cleaning and upgrading the cooling.

I've been habitually blasting my machines with compressed air since, so I'm now wondering if what would pile up in there otherwise would indeed be cat residual, and not just "ordinary dust".

Like some other members of the parrot family this species has powder down feathers. Powder down feathers disintegrate into a fine powder. Every time this bird preened, scratched, or shook himself, he threw off feather fragments, dead skin and powder down. ...but he was a nice pet, and worth the extra mess.

I never had cats and a PC at the same time. I don't think much cat hair would get in. Dander and cat litter dust could.

[glenpinn]

I'll have to climb on the "sceptical" bandwagon for that one. I don't know how every Intel CPU works, but I don't think the principle has changed since my old quad core was released. The CPUs have a maximum Tjunction temperature at which point they start to throttle themselves.


[Attachment 34627 - Click to enlarge]

Someone who understands something about how the Intel CPU's do work, and raises some info regarding some of the more recent technological features in modern day CPU's.

I know this CPU Temps topic is actually getting way off topic in here, given the thread relates to case fans, and i didn't want to drag this along any further because of that, but after reading the post by hello_hello today, i thought i would toss this up anyway, and just get it out of the way.

If my memory serves me, ever since around 2005/06 there has been technologies such as Hyper Threading, Turbo Boost, EIST, Speed Step etc available to most of the intel chips, and all these things relate to how a CPU behaves in a computer, and has a bearing on CPU heating as well.

As long as i can recall (going back to my E8400 and Q6600 days at least) these technologies, or some of them, have been available to some degree or another, but for the purpose of this exercise, i will use my current 3770 system as an example of how these technologies behave in my system, and the same applied to my previous i7 2600 setup as well.

MY CPU has a default maximum frequency of 3.4ghz, with a turbo boost frequency of 3.9ghz, and i have SpeedStep and Turbo Boost enabled, so my CPU runs at 50% of its default frequency of 1.7ghz while idle or not doing anything extreme, and as i add more stress to the CPU its frequency rises until it hits the maximum default frequency of 3.4ghz, and in certain situations (depending on the load and how hot the CPU gets) it will run anywhere up to the 3.9ghz in turbo boost mode, and it will stay at whatever frequency it reaches, and perform at 100%, unless it starts to crank up on the heat, and when it starts getting too hot (i was told it should be around the 80's mark) it will then start to throttle back on the turbo boost frequency back down to 3.4ghz, and continue to run at that speed without affecting any performance.

However, as i have always been led to believe (this has not happened to me yet) should my CPU get to this 100 degree TJunction threshold (that is what mine is set to by default) the CPU will then apparently drop back on the frequency slightly so it has some time to cool down, and as soon as the system detects a drop in temps, the CPU will gradually throttle back up to its normal frequency of 3.4ghz, and if it hits that TJ threshold again, it repeats the same cycle etc etc etc.

In this scenario, if i was encoding a video file, it would drop a few fps in output speed as a result, but pick back up as the frequency rises again, but i know for a fact on my past 2 systems, and on my Clevo editing laptop, my encodes never slowed down at all while the CPU's were running between 55 and 90 degrees, which would be about the hottest any of them would have reached under full load, and i have had up to 3 or 4 encoding jobs running just to stress the CPU and get the temps rising just to test this out.

There is also what is known as a TJunction Max threshold, roughly around 105 degrees, and if the CPU temp does happen to go over the 100 degrees TJ threshold and hit this 105 degrees TJunction Max temp, this is when the system will, or should just shut down to protect from causing any damage.

During these temp rises in my CPU, my cooler fan ramps up in speed to try and keep the CPU cool, fortunately mine seem to do a decent job, but in cases where the fan is not able to cope with cooling the CPU enough, this is when the CPU temps could get too high and possibly hit the TJunction threshold and throttle back, or in extreme cases it may hit the Max threshold and shut down.

Of coarse there are lots of other reasons why a CPU will cook and hit those 2 thresholds, dust in the heatsink, a faulty CPU fan, a badly seated heatsink, poorly applied or inferior thermal past on the CPU, and the list goes on.

Apparently most, if not all Intel chips have been designed to run at temps higher than the recommended TJ Max thresholds, and in some cases that i know of, a user has actually bump their TJ settings up to allow for this, and added some pretty extreme cooling to their systems (mainly overclockers) but i certainly wouldn't do it myself, but i do recall in one computer a few years ago i did bump those temps down a bit, but i think it was at a time when i was attempting to overclock the CPU and didn't have the best of coolers at the time, so it was only done as a precaution.

Anyway, that is my take on this, how i have always understood all this to work, and that a CPU just doesn't randomly slow down or throttle back on performance just because it starts getting too hot, and like i said, in my systems i can encode videos with my temps at anything up to 90 degrees and not have my frequency drop and cause the encode times drop because of it.

Obviously i am lucky that my Dell 2330 All in One editing pc and my i7 laptop (before i sold it) both run at default frequencies or even turbo mode because they just never suffer with poor cooling.

If anyone here wishes to dispute what i said, or believes that anything i said is incorrect, then please feel free to comment, i am only telling it as i believe it to be from my own experience, and from my understanding of what other more experienced people than me have told me.

Cheers

[PuzZLeR]

If anyone here wishes to dispute what i said, or believes that anything i said is incorrect, then please feel free to comment, i am only telling it as i believe it to be from my own experience, and from my understanding of what other more experienced people than me have told me.

I'm not disputing your post - not this one at least - and you're basing it on your experiences as I based it on mine.

Perononally, I don't bother with how much the numbers do what and when but I do know there is slowdown on certain occasions related to heat. I don't even care to research it like you probably did, and it would start to go off topic as well regardless anyway.

But I'm not disputing your post because you finally do have an understanding that this happens - the processor will slow down due to heat, whatever level that is, and that could indeed include dust if there's enough of it stifling the cooling system - which would have made our earlier exchanges unnecessary.

And yes, heat can even shut down the system too.

[glenpinn]

I didn't have to research anything, i have known all about this stuff for years, sometimes get a bit rusty with numbers, but as i stated, i didnt want to raise this until hello_hello bought the TJ Max threshold up in his post.

What i said is based on my own experience yes, but it is also based on the facts about how Intel chips were designed to work using certain technologies such as those that i mentioned, and if you have a properly working Intel CPU, then it should be working in the manner that i described it, or there abouts at least.

The other reason for raising this topic again was because i kept reading your earlier posts where you was simply going on and on about dust all the time, and stating that a hot CPU slows down performance and dust can slow down encoding, and i interpreted most of what you was saying (by the way you said it) as meaning exactly that.

It was only after hello_hello posted what he did that your whole tone seemed to change somewhat in your replies after that, and i then thought to myself why could you not have mentioned those facts in more detail, and possibly mentioned those TJunction thresholds as well, given that you say you know how it all works, and if you had of, then it might have been a bit easier to follow and believe.

Finally, back on topic, one thing i know for sure, and that is if you build your computer properly, with the correct case cooling and airflow thru it, and place your computer case in the right place, there is no way anyone would suffer with all this dust accumulation, and certainly nowhere near the levels that have been described in this thread, and if anyone does have to blow dust from a case more than a couple of times a year, then they need to learn how to build it properly, including how to ventilate it correctly, and as for all this stuff about cats, dogs and birds, why would anyone seriously keep a computer anywhere near animals, that is just insane, and definately asking for trouble.

Cheers

Oh, btw, the system should only shut down when the CPU hits the TJunction Max threshold, as explained in my post, if it shuts down before that, or if the CPU throttles back on its frequency (slows down as you call it) before hitting TJunction threshold, then you most likely have a sick CPU, or some other problem with your system somewhere.

[hello_hello]

I'll confess I'd never heard of the Intel CPUs having a temperature at which they shut down. I thought they just kept throttling, so I did a little research. If the information I found is correct they do, but not in the way I expected.

To begin with, I looked up the specs of my old Q9450 CPU.
http://www.intel.com/content/www/us/en/intelligent-systems/mccreary/core2-qx9000-q9000...datasheet.html

The only spec listed is Tcase (maximum temperature at the processor integrated heat spreader) at 71.4 degrees. That seems to be the maximum, but it's a moving target related to power and I don't fully understand the power specifications yet (page 79 of the pdf for my CPU).
Edit Tcase max of 71.4 confirmed. http://ark.intel.com/products/33923/Intel-Core2-Quad-Processor-Q9450-12M-Cache-2_66-GHz-1333-MHz-FSB

Page 84 of the pdf:
An external signal, PROCHOT#(processor hot), is asserted when the processor core temperature has reached its maximum operating temperature. If the Thermal Monitor is enabled (note that the Thermal Monitor must be enabled for the processor to be operating within specification), the TCC will be active when PROCHOT# is asserted. The processor can be configured to generate an interrupt upon the assertion or de-assertion of PROCHOT#

PROCHOT# is a bi-directional signal. As an output, PROCHOT# (Processor Hot) will go active when the processor temperature monitoring sensor detects that one or both cores has reached its maximum safe operating temperature. This indicates that the processor Thermal Control Circuit (TCC) has been activated, if enabled. As an input, assertion of PROCHOT# by the system will activate the TCC, if enabled, for both cores. The TCC will remain active until the system de-asserts PROCHOT#.

So going with the assumption that TJunction max is the maximum core temperature (I still haven't confirmed it for sure), Tjunction max is the temperature where the CPU starts to throttle itself.

The temperature at which it shuts down though isn't directly related to Tjunction max.

Page 72:
In the event of a catastrophic cooling failure, the processor will automatically shut down when the silicon has reached a temperature approximately 20°C above the maximum Tc
(Page 75 says Tc = Tcase)

So my processor should shut down at 20 degrees above Tcase, which would be 91.4 degrees. The problem there is I don't know if there's any way to read the TCase temperature unless it's been implemented on the MB and you probably still can't count on that being accurate. SpeedFan displays a CPU temperature (as well as core temperatures) so for the moment I'll assume that's where it's getting the CPU temperature it displays. The fact that the Tcase max temperature seems variable still has me a little confused though.

I read a couple of forum posts that seemed to confirm the fact throttling starts when Tjunction max is reached and the CPU shuts down when Tcase max is exceeded by 20 degrees.
http://www.overclock.net/t/476469/the-truth-about-temperatures-and-voltages
http://www.overclock.net/t/275776/tjunction-max-thermtrip-prochot

Edit: So here Intel confirms Tjunction max is the maximum core temperature before throttling begins. They say Tcase max is the maximum remperature Tcase "should" reach. What does "should" mean?? They mention the CPU is capable of shutting itself down but don't give any details as to how that works. Why do they need to be so vague about this stuff?
http://www.intel.com/support/processors/sb/CS-033342.htm

[redwudz]

Not to confuse the CPU shutdown issue, but a year or two ago I had a server PC shut down unexpectedly. I did a restart, but after about five seconds it shut down again, before even reaching POST. Once more, same thing. Then I finally pulled the side covers and found one of the CPU cooler hold down ears had broken off. The cooler was only touching the CPU on one corner.

At least in this case, the motherboard apparently shut down the power supply and the PC. Probably from reading the CPU temperature.
I was happy. I replaced the hold down and the PC is still in use.

With an older CPU a few years before that, a similar incident just fried the CPU. Now apparently the CPUs and the motherboards both seem to contain some protection setup.

[PuzZLeR]

I didn't have to research anything, i have known all about this stuff for years, sometimes get a bit rusty with numbers, but as i stated, i didnt want to raise this until hello_hello bought the TJ Max threshold up in his post.

What i said is based on my own experience yes, but it is also based on the facts about how Intel chips were designed to work using certain technologies such as those that i mentioned, and if you have a properly working Intel CPU, then it should be working in the manner that i described it, or there abouts at least.

The other reason for raising this topic again was because i kept reading your earlier posts where you was simply going on and on about dust all the time, and stating that a hot CPU slows down performance and dust can slow down encoding, and i interpreted most of what you was saying (by the way you said it) as meaning exactly that.

It was only after hello_hello posted what he did that your whole tone seemed to change somewhat in your replies after that, and i then thought to myself why could you not have mentioned those facts in more detail, and possibly mentioned those TJunction thresholds as well, given that you say you know how it all works, and if you had of, then it might have been a bit easier to follow and believe.

Finally, back on topic, one thing i know for sure, and that is if you build your computer properly, with the correct case cooling and airflow thru it, and place your computer case in the right place, there is no way anyone would suffer with all this dust accumulation, and certainly nowhere near the levels that have been described in this thread, and if anyone does have to blow dust from a case more than a couple of times a year, then they need to learn how to build it properly, including how to ventilate it correctly, and as for all this stuff about cats, dogs and birds, why would anyone seriously keep a computer anywhere near animals, that is just insane, and definately asking for trouble.

Cheers

Oh, btw, the system should only shut down when the CPU hits the TJunction Max threshold, as explained in my post, if it shuts down before that, or if the CPU throttles back on its frequency (slows down as you call it) before hitting TJunction threshold, then you most likely have a sick CPU, or some other problem with your system somewhere.

If you had knowledge of this prior, you would have understood the complications dust contributes, and would have understood what I meant by it in the context of heat, and why every computer store worth its salt sells compressed air, and in the spirit of this thread regarding important attributes of casing.

And you don't have to use Hello_Hello as your coverup, or resort to foul language, and don't have to mention things that are untrue like my so-called previous posts going "on and on" about whatnot. (Stalker.) And I never changed my tune at all in this thread.

Dust can indeed stifle cooling. Enough of it will slow down certain processors, and removing a particularly large amount of dust will once more boost performance. And yes, more than this will indeed even shut down PCs, and restarting them will reveal some black screen with some message about the PC having to shut down due to overheating. Just put some sawdust on your vents and try this.

Didn't you say you didn't want to respond to any more of my comments a few posts ago? (^F respond - speaking of searching - another blow to your credibility.)

Let me do you the favour then. I'm done discussing this with you. I've made my point already several posts ago. Have a nice day. Get some rest too.

[glenpinn]

Puzzler, you seriously need to get over you infatuation with DUST, all you do is talk about DUST, and it is very clear to me that you have massive problems with dust in your computer, and that clearly demonstrates a complete lack of ability and knowledge in how to build a computer and how to ventilate it correctly, because if you did, you would not be having any issues with dust inside your computer.

Just get over yourself for once and learn something

[usually_quiet]

Puzzler, you seriously need to get over you infatuation with DUST, all you do is talk about DUST, and it is very clear to me that you have massive problems with dust in your computer, and that clearly demonstrates a complete lack of ability and knowledge in how to build a computer and how to ventilate it correctly, because if you did, you would not be having any issues with dust inside your computer.

Just get over yourself for once and learn something

...or maybe he doesn't build his own computers. Most people do not. The very dusty PC I was talking about in my post was manufactured by a large company. It had one fan in the rear, plus a few air intake vents in the front, and no dust filters. That is still the most common ventilation design for the cases used in manufactured PCs today.

The gaming-oriented cases which are often used by those who build their own PCs are the sort that have lots of additional ventilation to cool a system containing one or more massive power-sucking video cards. Of course they have the potential to let in a lot of dust too.

[Nelson37]

Any case will let in dust, it has nothing to do with design or ventilation. Air filters will cut down on this dramatically but cannot eliminate it. I use very light foam covered with fine panty hose, with effort to seal other openings. This definitely keeps out larger particles, the very fine dust that does get in is easier to remove.

For cleaning, clean paintbrush, vacuum, and compressed air. No one, or even two, is adequate for accumulated dust. Suck, blow, physically dislodge in alternation necessary. A pass or two with vacuum and brush before blowing leaves much mess outside the case. Once pulled a dust blanket of a mobo where you could identify some motherboard features on the dust blanket. Sort of like a negative photograph. Was in an embroidery shop.

CPU's have had thermal shutdown for a long, long time. Thermal throttling I'm not certain but some definitely malfunction or run slower when severely overheating, eliminating source of overheating results in faster operation.

If you have read that modern CPU gets hot enough to blister your finger 3 seconds after power-up, I have verified this.