Re: Guiding a 14 year old nephew on building his own computer

On 2020-08-29 22:49, Paul wrote:
> s wrote:
>> On 8/29/2020 6:13 PM, bad sector wrote:
>>> On 2020-08-29 19:28, s wrote:
>>>> On 8/27/2020 7:13 AM, bad sector wrote:
>>>>> On 2020-08-27 00:37, s wrote:
>>>>>>
>>>>>> My nephew is 14 and living in another state who wants to build his 
>>>>>> own computer.
>>>>>>
>>>>>> I sent following YouTube videos for him to watch and learn how to 
>>>>>> do it.
>>>>>>
>>>>>>
>>>>>> Build a PC - Step-by-step
>>>>>> https://www.youtube.com/watch?v=IhX0fOUYd8Q
>>>>>>
>>>>>> Beginners Guide - Build a PC
>>>>>> https://www.youtube.com/watch?v=fVmcD8v3vR4
>>>>>>
>>>>>> How To Build a $500 Gaming PC
>>>>>> https://www.youtube.com/watch?v=dFyhn6seoow
>>>>>>
>>>>>> The idea is to assemble a basic PC using which he can do homework, 
>>>>>> learn new topics etc.
>>>>>>
>>>>>> Would the following parts be compatible to buy, assemble?
>>>>>>
>>>>>> https://www.amazon.com/Cooler-Master-Transparent-Adjustable-Ventilated/dp/B0785GRMPG/ 
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> I like that case, made me one ten years ago that draws filtered air 
>>>>> in via the raised bottom and 3 4" silent fans with all other fans 
>>>>> reversed if required. Has been working very well with cpu never 
>>>>> above 47c.  I have another idea now but before that I might just 
>>>>> get me one of those cases from amazon!
>>>>
>>>> Would the single fan in this 
>>>> https://www.amazon.com/Cooler-Master-Transparent-Adjustable-Ventilated/dp/B0785GRMPG/ 
>>>> be enough for cooling?
>>>
>>> Cooling always depends on ambient conditions and cpu and cards and 
>>> such. The fact that it's a best seller on amazon would I think speak 
>>> well for it under average or better conditions, power supplies, 
>>> CPU's, some video cards and drive racks have their own fans in 
>>> addition. I would have no problem ordering it. For a beginner who 
>>> isn't going to jump right into hacking and cannibalising hardware 
>>> it's plenty good enough :)
>>>
>>
>> Thank you! I appreciate the prompt clarification.
> 
> You use the system cooling equation for that.
> 
>     By now, you might be saying, "how can I figure out what fan to
>     use for the case ?". Well, there is an equation for that too.
>     Now, for this one, you need to know all the thermal loads inside
>     the computer case. We have the 125W processor, say a 100W
>     video card, two 12W hard drives, say a total of 250W. Our definition
>     of a well cooled case, is 7C, which is equal to 12.6 Fahrenheit 
> degrees.
>     Plugging in the values.
> 
>     CFM = 3.16 * Watts / Delta_T_degrees_F
> 
>     CFM = 3.16 * 250W / 12.6F = 62.7 cubic feet per minute.

Interesting, I never saw that before.

I have three relatively silent 4" fans sucking air into the box through 
the filtered bottom, then the power unit and disk racks boosting flow to 
the outside through them with their own fan and only the cpu fan doing 
little else than a bit of purely internal flow increase. My study hardly 
ever gets above 25c, 95% of the time 20c, and with that cpu or disks 
never went above 47c. I have seen a simplified version of the above 
formula where the wattage is 1/2 the power-supply capacity.

Agree about the noise, it's my next target, want my next box not just 
low-noise but TOTALLY silent :)



> It would take two rear fans of 35CFM each, to give 70CFM to cool
> the air inside the case to "only" a 7C rise above ambient.
> 
> If room temp is 20C, then case air is 27C under those assumptions.
> 
> You're trying to keep the hard drive temperature below 60C,
> if you need a metric. (The air can't be 60C! If the air was
> 40C, maybe the disk is 60C.) And you don't want the CPU getting
> so hot, that it throttles because you didn't have a good
> enough fan on the back. Using a utility like Speedfan (almico.com),
> you can monitor various temperatures and see how close to an
> ideal situation you're in.
> 
> For most people, a realistic value for system power is the
> hard part. If the fan were to be "temperature controlled",
> it could "rail" when the computer is gaming, and "loaf"
> when the machine is idle. I like to run mine at a fixed
> speed, so that if a fan is failing and slowing down, I
> can notice the change with my ear. But others like
> their temperature-controlled fans.
> 
> My newest machine is probably off the ideal value by
> around half. So my case cooling is only about 50% of
> what it should be. But the 110CFM fan that used to be
> in the computer case, was causing "hearing loss", so
> it had to go :-)
> 
> And I know what the power usage on the Test Machine is,
> because it has a Kill-O-Watt meter connected. It draws
> 180W when all cores are loaded. If the video card cuts in,
> it might hit 340W. If AVX is used, that's another increment.
> I try not to run it too long with synthetic tests on
> it (with the low capacity fan at least). I don't have
> a problem running it all day at 180W though (compressing
> an entire disk drive). I know the VCore regulator doesn't
> have a big enough heatsink on it, which is the limitation
> on my system (don't want to damage it). What's funny is,
> the computer I'm typing on, has a metric acre of copper
> to cool VCore, and it doesn't even get luke warm. The
> computer where they knew the power would be 4X higher, they
> put this shitty little cooler on it (probably just
> aluminum too, not a copper one).
> 
> When buying the motherboard, I neglected to note just how
> poor that cooler was. But I figured it out, the day I
> "built up" the computer on the kitchen table, without fitting
> it into the computer case. I run them first, without a computer
> case, then I "run a finger over the hot stuff", just
> to spot issues. And that's when I burned my finger on
> VCore :-/ I had to fit a small fan, just for Vcore.

hehehehe, the last time I did THAT was with a 486 and it literally 
burned my finger because it WAS actually cooking itself to death, just 
like an ironing plate!

> The clever guys at some of the review sites, they use a
> FLIR infrared camera to spot build issues. Typically
> the camera overlays a low-res infrared image on top
> of a slight-higher res visible image, and from that,
> you can spot issues with precision. For example, when
> someone spotted an eight-pin DIP running at 100C, that
> was dead easy with an IR camera. I would have missed
> that, using my finger, as I would never suspect such
> a small chip of running that hot. I'd hate for an
> engineer to have done that on purpose.
> 
>     Paul