I remember when you got the first one. You were fairly proud of it. 

I think it was the subject of a "Happy Birthday, Jafo" post. Could be wrong, though.

edit: Wasn't his birthday...he was uninstalling W10.

I actually have Windows 10 in that box...but only as a VM...

This was "back in the day" when M.2 'drives' were available, but the Win 7 OS couldn't be installed on one - however there was "a way" to do it...so I did.

These boxes are inherently so cool and 'solid' the fans are almost never on...and if the are you don't ever hear them anyway.  Even the PSU fan totally shuts down after boot, never to be heard from again.

You can see the section/thickness of the material used in that photo.

The little plate at the bottom of the MoBo 'box' is the offset for venting as well as clearance for the GPU.  I had to reinvent that with spacers so it's now out about twice as far [12mm] to clear the RTX3070 which isn't small.

One thing I have been improving is all six physical drives in the 'rack' are now SSDs ... a wee bit lighter...and cooler as well.  It's not really an extravagance to replace the drives...those old platters have quite a few uptime hours on them.... by 'few' I mean 3546.8 days [not hours]... and the M.2 drive claims it wants to retire in 2.5 years time ... 

Hence the jackhammer.  

Didn't Starkers have one too? Or maybe he wanted one. Speaking of him, whatever happened to him? I haven't seen him about in quite some time?

No...his rig was different. His last post was in 2-3/2022...he'd been hacked. I also remember a computer disaster. Also, he'd gotten terminally po'd at bankers/lawyers.

https://forums.stardock.com/517834/page/1/#3883730

IM'd with Shaunna...she tells me he's been ill and he's still trying to get his computer up and running.

Looks like I asked that question already. Thanks for the reminder, Doc. 

Starkers' was a Thermaltake Level 10, much the same layout but not a 'heavy metal'  Limited Edition one.

The spreading out of the components makes for a pretty darn big case but it's an advantage for cooling.

When I built mine I opted for "bigger is better" when it came to PSUs so it's running an extremely unstressed 1200 watt Corsair.  Looking at potential 'bits' I have at hand for this new one... I have a Thermaltake 500 watt or perhaps I'll pinch the Antec 1000 from another machine.

It all depends on what condition the MoBo/CPU is in and in which direction I take it...

I'm also upgrading my 'spike' protection as apparently an existing board doesn't last forever and the last time I suffered from one it was a $2500 mistake on my part ...

I think I remember he had a problem putting it in the desk (?) where he wanted it...or a problem getting at it when something went wrong. I certainly could be wrong in this.

Quite probably.  I have mine directly on the desk by my left shoulder.... 2 reasons...it's away from [most] dust and it's pretty much too heavy to lift.

The empty case was about 50lb on its own....

To give an idea of the job we left those guys (the package designers) to solve, a typical CPU has a greater power density (in watts per cubic meter) than if you were standing on the surface of the sun, or if you were in the middle of the Hiroshima explosion.  A typical mosfet transistor nowadays has a gate 3 atoms thick, which means a whopping 3 atoms separate the plus and minus terminals inside the CPU from shorting out. .And a typical system-on-a-chip has 100 million mosfets in it; all of them practically shorted out.  

So these cases you buy--heat sinks, fans, etc., are the result of somebody figuring out how to dissipate a miniature Hiroshima bomb, and fanning that heat out to keep that heat inside down to a cool 70 degrees Celsius.  We basically short out the power supply to make it run fast, then leave it to somebody else to deal with the heat. 

So--those guys who do earn their keep.   

Currently CPU is 35c

GPU is 40c

Each of the 6 physical drives 30c

Not sure what the PSU is...but the fan's asleep [as usual]....

Sounds like a real winner!  

The temperature sensor is usually not the hottest part of the chip.  It's probably in the core, and even parts of the core are hotter than others:

https://www.researchgate.net/figure/The-thermal-map-of-an-eight-core-CMP-similar-to-Intel-Haswell-architecture_fig2_333361319

The sensor readings you're getting are still useful, though, because a higher reading still means a hotter chip, and a lower one means a cooler one.  And with Intel, AMD, Nvidia et al in competition, there's not much incentive for them to calibrate the temperature sensors to report hotter temperatures.  No one wants to be perceived as running hot. 

Probably true...except for those who had to work to earn the money to buy their machines.

Paradox warning: Clearly a want/need contradiction. People might want to eat ice cream, but realize that they need to eat healthy to control their blood sugar.

Who wants to hear a fire alarm in the middle of a great movie? No one, but people learn to prefer not being burned.

Maybe I didn't properly explain:   different chips have different acceptable ranges of temperatures that they can operate in.  Or more specifically, they have different ranges of temperatures that they can operate in *according to their own temperature sensor*.   All chips basically use the same materials:   polysilicon, phosphorous, boron, aluminum, etc..   They all have the same specific heat, etc. for the same material.   It's not like Intel can legitimately say they have superior phosphorous atoms to AMD's phosphorous atoms (although I have no doubt they would try).   

But the temperature sensor is a design.  They rely on the same universal laws of physics, but the designs are specific to the company.  Usually it comprises of a voltage divider, reed sensor, etc. of some sort, and that analog voltage gets fed to an ADC (analog-to-digital converter).   The sensor causes something to output a different voltage depending on what temperature it is, and that voltage gets converted into a digital number.   The company can calibrate that any way they want.  They can convert 35 millivolts into a 0 or a 16376, and they can interpret that 16376 to be any temperature they want (although if they ever interpreted it to be a negative Kelvin temperature, I think it's safe to say that design is definitively wrong).   The most logical choice is to calibrate it to output the temperature of THAT temperature sensor itself.   Which probably runs a good 17C cooler than the cores are running at that very moment.   So you may be reading 30C on the console, but the part of the chip you actually care about is running at 47C.    But that's okay--you already have an acceptable range in your head that the computer can reliably run at, and that is based on the temperature you are reading.  Your upper limit may be 90C, when we actually validate the CPU to run at 110C in-house (110C is a very common upper limit for non-military grade).

Personally, I think an accurate chip temperature sensor needs to reflect the WORST-CASE temperature on the CHIP.  But that is bad marketing.  If your AMD computer reads 30C and their Intel computer reads 50C, everyone's going to buy the AMD computer.  Even though they are running at the same temperature.

I understood you, tetleytea. The more successful marketing takes advantage of uninformed/noncritical thinking. A subtle and rarely considered fact as we rarely are intellectually honest enough to consider what we don't know and haven't learned.

Cool case. Now you just need the upcoming 5090 gpu to test its limits 😅

I have an EVA RTX3070 sitting here that'll fit - once I space out the GPU cover plate as I did for the first one.

I'm going to try to use what I have 'laying around' to keep things 'affordable'.  When the first one was done it was pretty much the top of the specs possible and its cost reflected it.... eg the first 32gig of ram was $1020 AUD ... but recently I doubled it for $256 AUD  [Corsair Dominator Platinum [4x8G] 2666MHz DDR4].  Yes, the board has 8 ram slots ...