Production FormD T1 Classic (READ FIRST POST)

[hetvler]

mid 70's is absolutely fine for the system.
I'd personally advise to keep the nvme drives as cool as possible because those will degrade under high temps. Below 50 would be ideal.

degrade meaning performance (read/write) or like... physically degrade and shorten lifespan.

 

[Navic]

Is there a point to having temps as low as absolutely possible (ignore boost clocks) or can i safely run my system at slower fan speeds and noise levels and sit in the mid 70s to minimize noise without endangering my components?

Where is the threshold? I don't mind dropping clock speed since the performance hit will be very minimal but I do care about preventing damage to my parts.

For modern GPUs, high 70s is as high as I would go
For modern CPUs, mid 90s is as high as I would go for intel CPUs, high 80s for AMD CPUs
SSDs like to have the NAND run warm, but not super hot. The controller for the SSD (the temp that is shown in monitoring software) should stay under 75c.

degrade meaning performance (read/write) or like... physically degrade and shorten lifespan.

It can shorten the lifespan of the drive, but not by a huge margin. But if you want the SSD to last as long as possible, refer to above.

 

[hetvler]

For modern GPUs, high 70s is as high as I would go
For modern CPUs, mid 90s is as high as I would go for intel CPUs, high 80s for AMD CPUs
SSDs like to have the NAND run warm, but not super hot. The controller for the SSD (the temp that is shown in monitoring software) should stay under 75c.


It can shorten the lifespan of the drive, but not by a huge margin. But if you want the SSD to last as long as possible, refer to above.

Word, thanks. I'll see which ceiling i hit first regarding CPU/GPU/SSD thermals and as silent as I can go on the exhaust fans.

The GIGABYTE B550I AORUS PRO comes with a heatsink (and heatpipe!) so i think i'll be alright there or at least have more headroom than I would without it.

Thanks for the guidelines, looking forward to tuning my system all the way in early january when i get the case

 

[PKAWA]

degrade meaning performance (read/write) or like... physically degrade and shorten lifespan.

Both actually.
To counter the rear nvme running hot I've opted to ad a 30cm extension cable to it and routed it to the top just behind the radiator.
Stays well below 50c now and that was a 20 euro cable.

For the front, I've added a 40mm fan and have it running at 60%. my board doesn't have a fan but does have the chipset right under the m.2 connection so it tends to get hot without the fan.

The Gigabyte x570 board i'm interested in has a fan already build in and should perform well if I can believe the reviews.
I had hoped Fabio could shime in on fitment with the LT240 and temps.

 

[PKAWA]

Just found the measurements for the new AMD cards on overclock.net:

We just learned the dimensions of all 3 cards thanks to SAPPHIRE, they list the products already, they're selling Reference design;

RX 6800 XT and RX 6900 XT
Dimension: 266.7(L)X 119.75(W)X 49.75 (H)mm
267 x 120 x 50mm (2.50 Slot)

RX 6800
Dimension: 266.7(L)X 119.75(W)X 39.75 (H)mm
267 x 120 x 40mm (2.00 Slot)

So for people contemplating using the dc-lt40 this might be useful info.

 

[Ray2r]

nVidia RTX 2060 waiting for a 3080 that is dual slot or more, depends on what I find

I am happy to report that all 3080s are dual slot or more [=

Founders edition 2.0 Slot, EVGA XC 2.2 Slot. Everything else around 2.5 Slot or more.

I myself am waiting for a 2.0 Slot 6800XT so I can use a bifurcation riser.

IDK what this poll for "More GPU Slot Options for T1?" means. Does it mean that there will be a side panel that extrudes past the case just for the GPU side so u can potentially have 3.5 or 3.75 PCIe slots? Because that would work awesomely for me to use a bifurcation and a >2.0 slot GPU.

 

[PKAWA]

It means that if you have a 2.5 slot card, ideally we would like to be able to use the lost space on the gpu side on the cpu side by having a 2.5 slot mode.
That way lpr would not be necessary in most watercooled cases and routing tubes would be easier.
I think this could be easily implemented by grinding away the notch between the two cavities for the centre beam, tapping screwholes and supplying a alumimum part that fills the left over hole in any setup.
Problem is probably cost of changing production, so I don't see that happening any time soon.
But it could be done at home too.
(Actually I thought about it some more, and it cannot because the backplate would need to be redesigned too.)
Another idea I have is replacing the psu like 1cm further from the front wall and thus creating a gap so that the tubes of the radiator could come straight down right behind the front panel.
Offcourse this would only work with a flipped psu and a J-hack unit because there would be near zero space between the psu and the motherboard.
Benefit would be unhindered airflow for the psu outlet and intake of the fan, easier tube routing with less bends and a cleaner look.
the way I see it there would only need to be tapped another two screw holes in the centre beam.

 

[Arc_Lag]

It means that if you have a 2.5 slot card, ideally we would like to be able to use the lost space on the gpu side on the cpu side by having a 2.5 slot mode.
That way lpr would not be necessary in most watercooled cases and routing tubes would be easier.
I think this could be easily implemented by grinding away the notch between the two cavities for the centre beam, tapping screwholes and supplying a alumimum part that fills the left over hole in any setup.
Problem is probably cost of changing production, so I don't see that happening any time soon.
But it could be done at home too.
(Actually I thought about it some more, and it cannot because the backplate would need to be redesigned too.)
Another idea I have is replacing the psu like 1cm further from the front wall and thus creating a gap so that the tubes of the radiator could come straight down right behind the front panel.
Offcourse this would only work with a flipped psu and a J-hack unit because there would be near zero space between the psu and the motherboard.
Benefit would be unhindered airflow for the psu outlet and intake of the fan, easier tube routing with less bends and a cleaner look.
the way I see it there would only need to be tapped another two screw holes in the centre beam.
@Wahaha360 Is this a mod you'd be willing to implement?

I'm also curious! What is the status of the variable slot thingy? Do you have a specific release date in mind? @Wahaha360

 

[PKAWA]

I think at this point this was only an idea from somebody in this thread, but not actually something they are actively working towards implementing.

I will dry test the second setup when I'm building in the new case by just using a appropriate stand off between the front and psu mount.
If it's feasible I might tap screwholes myself.

 

[biopunk]

Just found the measurements for the new AMD cards on overclock.net:

We just learned the dimensions of all 3 cards thanks to SAPPHIRE, they list the products already, they're selling Reference design;

RX 6800 XT and RX 6900 XT
Dimension: 266.7(L)X 119.75(W)X 49.75 (H)mm
267 x 120 x 50mm (2.50 Slot)

RX 6800
Dimension: 266.7(L)X 119.75(W)X 39.75 (H)mm
267 x 120 x 40mm (2.00 Slot)

So for people contemplating using the dc-lt40 this might be useful info.

Thanks, so a DDC pump with a 25mm top like Magic-Cube or Eisdecke will also fit.

 

[PKAWA]

Well we now know that 313mm gpu's are supported and that leaves 46mm behind the 6900xt.
For me the height was something I was looking for, my current 3080 is 127mm so I'd have 7mm more room to route tubes under the gpu.

 

[grinny]

Thanks, so a DDC pump with a 25mm top like Magic-Cube or Eisdecke will also fit.

Exactly what I was thinking. I need to measure what is the thickness of EK 3.2 PWM pump + Magic Cube sideways.

 

[rfarmer]

Ali tests water cooled 3090 with different radiators and looks like 10900k and RTX 3090 is doable with single 240mm radiator. He will be installing it into his T1 in the next video.

 

[SoSquidTaste]

Ali tests water cooled 3090 with different radiators and looks like 10900k and RTX 3090 is doable with single 240mm radiator. He will be installing it into his T1 in the next video.

Only OT can make me think to myself that I should really jump on a PNY 3090 hahahah

Anyone happen to do a ST30 (as tested in this open bench) vs. the ST25s that we run in our builds? IIRC it was noted that the core thickness between the older ST30 and current ST25 is minimal, so hopefully performance maps between the two?

This probably better fits the subreddit. Moving there!

 

[SquareCubedBox]

Anyone happen to do a ST30 (as tested in this open bench) vs. the ST25s that we run in our builds? IIRC it was noted that the core thickness between the older ST30 and current ST25 is minimal, so hopefully performance maps between the two?

I think radiator area seems to be a more important factor rather than radiator thickness when it comes to cooling potential (source: )

 

[PKAWA]

Okay so decided to open up the case and testfit the psu closer towards the motherboard using the screw holes already in the center beam.
It does not provide enough clearance for the tubes with springs, which I already suspected.
Without springs it does work barely but given the angle they would need to make underneath the psu the tube walls would collapse.
So all in all not a viable option.
The distance to the motherboard is not enough to tap a third set of screw holes unfortunately.

I've meanwhile seen that the Asus x570-i would be the better board over the Gigabyte but I'm having trouble coming up with a viable solution for the tubes in 3 slot mode and high profile ram.

 

[T_Dawg]

I think radiator area seems to be a more important factor rather than radiator thickness when it comes to cooling potential

It absolutely does, my Bachelor's Degree is in Chemical Engineering and one of the basic formula for designing heat exchangers is

q=UAdT

where:

q is the rate of heat transfer (W, J/s, BTU/hr, etc),
U is a Coefficient of heat transfer determine from a variety of factors including materials of construction, fluid properties, etc,
dT is the temperature difference of the cooling fluid entering and exiting the cooler (slightly more complicated than that, but that the gist) note that in our cases this is the change in air temperature
and
A being the surface area of contact. <-----

This is balanced against the desired amount of heat transfer from the fluid, which is a similar, but different equation

q=mcdt

where

q in this case is still rate of heat transfer
m is the mass flow rate of the fluid (equivalent in our case to say, pump speed)
c is the heat transfer coefficient of the liquid being cooled
and dT again being temperature difference of the liquid exiting the cooler vs its inlet temperature

Generally when you're designing heat exchangers, you're trying to figure out how big they need to be, so you're solving for A to determine the number of tubes, fins, etc. So yes, thickness of these finned coolers does directly impact that surface area, A, but so does fin design, density, etc.

This rambling aside has been brought to you by an early 1990s degree that was never really used. There's a lot of gross simplification here, but this are the first principles.

 

[Navic]

It absolutely does, my Bachelor's Degree is in Chemical Engineering and one of the basic formula for designing heat exchangers is

q=UAdT

where:

q is the rate of heat transfer (W, J/s, BTU/hr, etc),
U is a Coefficient of heat transfer determine from a variety of factors including materials of construction, fluid properties, etc,
dT is the temperature difference of the cooling fluid entering and exiting the cooler (slightly more complicated than that, but that the gist) note that in our cases this is the change in air temperature
and
A being the surface area of contact. <-----

This is balanced against the desired amount of heat transfer from the fluid, which is a similar, but different equation

q=mcdt

where

q in this case is still rate of heat transfer
m is the mass flow rate of the fluid (equivalent in our case to say, pump speed)
c is the heat transfer coefficient of the liquid being cooled
and dT again being temperature difference of the liquid exiting the cooler vs its inlet temperature

Generally when you're designing heat exchangers, you're trying to figure out how big they need to be, so you're solving for A to determine the number of tubes, fins, etc. So yes, thickness of these finned coolers does directly impact that surface area, A, but so does fin design, density, etc.

This rambling aside has been brought to you by an early 1990s degree that was never really used. There's a lot of gross simplification here, but this are the first principles.

yeah i aint gonna read that. too many sciencey words and letters. it's too late in the day for me to read that stuff tbh
I'll just assume what you said was that while radiator thickness doers help, a bigger radiator will help more as it adds more surface area. did i get the answer right? do i get a gold star?!

 

[BOGLOAD]

Finally got my case and built the thing. Unfortunately it came with a few small finish imperfections, not sure if I should raise an issue with this for QC purposes?

Incredibly tough with the lack of decent instructions, had to watch a few videos/look at other builds to really get a grasp of what was going on. Currently running 3-slot mode in anticipation of a 3080 (or similar).

Some observations:

• Case is half the size of my previous InWin A1, and has much better thermals. Also near silent, except for the GPU which you can really only during gaming. Hooray! Not going to lie, it didn't look all that 'premium' to me in pictures, but it really does in real life. Beyond happy
• I had to file down a motherboard stand-off to stop it from hitting the riser. I used a nail grinder (for my dog, ha) to get it to a flush, acceptable level. Better than grinding into the riser, but seems to be a design oversight.
• In 3 slot mode, an EK-120 AIO will *just* fit. The block will literally touch the side mesh panel, without too much noticeable bulge. RGB diffusion through the mesh looks great. I am a stickler and can deal with this. I couldn't use the full length brackets, had to use the smaller ones so my AIO is sitting 'freely', but snugly.
• Using G.Skill Trident RGB RAM, best to mount the tubes on the mono shroud side, so they've got more room to travel above the block+ram-lock, and less 'bunching' occurring above the PSU. Mounting with tubes closest to the ram is tough.
  • Works well using the ROG Strix X570-I motherboard, but can get in the way of some of the fan headers if mounting shroud side.
• Current CPU temps are 30c idle from boot, sits around 35-45c most of the time after extended use
• Current GPU sits around 40-50c idle, and hits about 75-85c depending on game (Zotac 2080 Blower, rather hot)

The ASUS TUF 3080 seems to be the card to go for in terms of thermals/noise, but I am concerned about reports that its too tight that you need to snap the power connectors off - has anyone managed to get one in there without modding this?

 

[arthasdklol]

It absolutely does, my Bachelor's Degree is in Chemical Engineering and one of the basic formula for designing heat exchangers is

q=UAdT

where:

q is the rate of heat transfer (W, J/s, BTU/hr, etc),
U is a Coefficient of heat transfer determine from a variety of factors including materials of construction, fluid properties, etc,
dT is the temperature difference of the cooling fluid entering and exiting the cooler (slightly more complicated than that, but that the gist) note that in our cases this is the change in air temperature
and
A being the surface area of contact. <-----

This is balanced against the desired amount of heat transfer from the fluid, which is a similar, but different equation

q=mcdt

where

q in this case is still rate of heat transfer
m is the mass flow rate of the fluid (equivalent in our case to say, pump speed)
c is the heat transfer coefficient of the liquid being cooled
and dT again being temperature difference of the liquid exiting the cooler vs its inlet temperature

Generally when you're designing heat exchangers, you're trying to figure out how big they need to be, so you're solving for A to determine the number of tubes, fins, etc. So yes, thickness of these finned coolers does directly impact that surface area, A, but so does fin design, density, etc.

This rambling aside has been brought to you by an early 1990s degree that was never really used. There's a lot of gross simplification here, but this are the first principles.

Yeah, but that depends on your definition of thickness?
For example, I would expect 2x240 30mm to have the same performance as 1x240 60mm at the same number of FPI and "floors/levels/tiers" inside ?