Completed The MI-6 Microtower Case: Performance in 6.7L

[whatanoob]

Were there any changes made to the second batch that would prevent the NH-L12S from fitting this time around?

Also, regarding the bottom fan, I was eyeing the NF-A9x14, however it says in the website that I can fit a fan up to 20mm thick since I'm going with the minimalistic front, so maybe there's a better one for my case out there.

Does anyone know about any good 92mm fans that are 17-20mm thick?

 

[dcbn]

Good News Everyone! I dropped my temps 10-20c!

Tonight I swapped out my Noctua NH-L12 for the Noctua NH-L12S. For those that don’t know, the L12 uses a 92mm fan underneath it to cool the CPU, while the L12S uses a thin 120mm fan. The L12S is technically out of spec for this case by a few mm, but does fit. It does touch the case side.

Both coolers had the fans in a pull config; pulling air in from outside the case.

My CPU is an i7-8700 non-k. At stock it’s a 65 watt TDP CPU. However, at stock settings it will clock down pretty hard in Blender and Prime 95 AVX. As such, I set the long term TDP to 110 watts. This locks it in at 4,3ghz on all 6 cores.

To compensate for the increased heat, I undervolt the CPU. I set the BIOS to adaptive VCORE at a -0.110. This is Prime 95 AVX stable 24-7.

My temps with the L12 were in the mid 90C range in Prime and Blender. With the L12S I’m now in the mid 80s in Prime and Blender. Gaming in BF1 got the biggest drop of nearly 20C. I believe that this is because the side of the case is now essentially part of the heatsink, and the larger fanmoving more air. The side of the case now gets quite warm where as before it was cool.

Additionally, the CPU fan now blows directly onto the M2 socket which dropped its temps by 15-20C.

Overall, the L12S just rocks for this case.

If possible, please, could you share some pictures?

Thank you for your attention!
I'm sorry for any inconvenience.

 

[firewolfy]

Were there any changes made to the second batch that would prevent the NH-L12S from fitting this time around?

Also, regarding the bottom fan, I was eyeing the NF-A9x14, however it says in the website that I can fit a fan up to 20mm thick since I'm going with the minimalistic front, so maybe there's a better one for my case out there.

Does anyone know about any good 92mm fans that are 17-20mm thick?

There was no change on the second batch that affects the cpu cooler height.

I have not yet found 92mm slim fans around 20mm thickness that have variable speed, unfortunately.

 

[NoSk0pz]

for those using custom length PSU cables: what size? been thinking about picking up the sleeved kit from amazon for the sf450, but i've been reading they're about the same length as the originals. at that point i might try to go shorter (unless the sleeved cables won't fill that space as much).

(i don't really NEED shorter cables, but i'm looking for any excuse to tinker more)

 

[firewolfy]

for those using custom length PSU cables: what size? been thinking about picking up the sleeved kit from amazon for the sf450, but i've been reading they're about the same length as the originals. at that point i might try to go shorter (unless the sleeved cables won't fill that space as much).

(i don't really NEED shorter cables, but i'm looking for any excuse to tinker more)

This is the same as I put on the webpage--
You should measure for your system, but typical cable lengths needed are: MB 24-pin: 250mm, EPS 8-pin: 250, PCIE 8-pin: 300, SATA: varies.

@rfarmer got some custom cables from Aliexpress HERE. Maybe he will jump in and give some thoughts.

 

[rfarmer]

for those using custom length PSU cables: what size? been thinking about picking up the sleeved kit from amazon for the sf450, but i've been reading they're about the same length as the originals. at that point i might try to go shorter (unless the sleeved cables won't fill that space as much).

(i don't really NEED shorter cables, but i'm looking for any excuse to tinker more)

I used 200mm for both the 24pin and 8pin EPS, 300mm for the PCI 6+2. SATA power are also 300mm which was the shortest I could get from Aliexpress.

 

[Engr62]

I currently have an i7-4770 set-up with a Cryorig C7 cooler in my MI-6. I recently acquired an ASRock AB350 Gaming ITX/ac motherboard and a Ryzen 7 1800X CPU that I'm thinking about putting in the MI-6. It looks like many use the Scythe Big Shuriken Rev. B for Intel CPUs, but I'm not sure what the go-to cooler is for Ryzen CPUs in the MI-6. I'm afraid the NH-L9a is not enough for the 1800X.

Any suggestions?

 

[rfarmer]

I currently have an i7-4770 set-up with a Cryorig C7 cooler in my MI-6. I recently acquired an ASRock AB350 Gaming ITX/ac motherboard and a Ryzen 7 1800X CPU that I'm thinking about putting in the MI-6. It looks like many use the Scythe Big Shuriken Rev. B for Intel CPUs, but I'm not sure what the go-to cooler is for Ryzen CPUs in the MI-6. I'm afraid the NH-L9a is not enough for the 1800X.

Any suggestions?

You can buy the Noctua NH-L12 again, if you aren't in the US check Amazon. You could also try the Noctua NH-L12S, even though it is listed as too tall a couple people have fit them in the MI-6.

 

[Engr62]

You can buy the Noctua NH-L12 again, if you aren't in the US check Amazon. You could also try the Noctua NH-L12S, even though it is listed as too tall a couple people have fit them in the MI-6.

Thank you for the suggestion. However, on the newegg page you linked, it says that cooler is 93mm tall with the fan. The suggested cooler height is 65mm, correct?

 

[chyll2]

It is 93mm with the standard 120mm on top. When they recommend L12, they meant only using the 92mm fan at the bottom which is around 66mm height

 

[Engr62]

It is 93mm with the standard 120mm on top. When they recommend L12, they meant only using the 92mm fan at the bottom which is around 66mm height

Thanks, I understand now.

 

[rfarmer]

It is 93mm with the standard 120mm on top. When they recommend L12, they meant only using the 92mm fan at the bottom which is around 66mm height

Thanks @chyll2 forgot to mention that.

 

[Engr62]

You can buy the Noctua NH-L12 again, if you aren't in the US check Amazon. You could also try the Noctua NH-L12S, even though it is listed as too tall a couple people have fit them in the MI-6.

It is 93mm with the standard 120mm on top. When they recommend L12, they meant only using the 92mm fan at the bottom which is around 66mm height

I may not be able to put the 1800x setup in my MI-6 after all. I just don't think I'm going to be able to cool it well enough with a top-down cooler.

I put the 1800x system together on my bench top (i.e., my desk) with the only cooler I have available right now with an AM4 mounting solution--my Noctua NH-D9L. It's definitely a no-go for the MI-6 since it's a 110mm-tall dual tower, but I wanted to get an idea of how cool it would keep the CPU under full load. While trans-coding a 1080p movie (2-pass 3000 kbit/s high profile) with Handbrake, the CPU temperature peaked at 71C. That's only 4C of headroom before throttling occurs. By the way, the 1800x achieved 80.7 fps on the the 1st pass and 87.1 fps on the second pass--so, I'm very pleased with those results.

I have an AM4 mounting kit on order for my Cryorig C7 that I was going to test with the 1800x while it's still on my bench top before I spent $60 on the NH-L12, but I doubt the C7 will perform well enough considering the D9L didn't have much headroom.

I've thought about getting the loft for the MI-6 when it's available and going with an AIO solution, but I'd really like to stay on air if possible since I may need to travel with this PC.

 

[dcbn]

I have an AM4 mounting kit on order for my Cryorig C7 that I was going to test with the 1800x while it's still on my bench top before I spent $60 on the NH-L12, but I doubt the C7 will perform well enough considering the D9L didn't have much headroom.

In my opinion, It makes more sense on acquiring the NH-L12s, It's considerably better than C7 and fits in the MI-6 according to @rfarmer.

Also, using an AIO solution probably won't give you that much performance when compared to the NH-L12s and you'll end up "adding some liters on your build".

Jay Mattison (@Tek Everything) used the NH-L12s (on an i7 8700K) in his build as well.

Link to the video.
Link to the article.

Best of luck!

 

[Jello]

Could try undervolting as well.

 

[K888D]

While trans-coding a 1080p movie (2-pass 3000 kbit/s high profile) with Handbrake, the CPU temperature peaked at 71C. That's only 4C of headroom before throttling occurs.

At what fan speed/power though? Because if the fan was for instance operating at only 20% power then thats a pretty impressive result, but if the fan was flat out at 100% then yeah it's quite high for open an open bench test.

You may only 4C headroom, but how much headroom have you got left in the fan speed?

 

[Gautam]

the CPU temperature peaked at 71C. That's only 4C of headroom before throttling occurs.

I don't know much about AMD, but 75C sounds awfully low for a throttle temperature. It seems like its actually 95C.

Unpopular opinion but nearly none of these coolers are well equipped to deal with a real 100W load. Handbrake/video encoding is a completely different ballgame from gaming, in many cases it'll have the CPU dissipating double the heat. 71C for a real 95W load isn't bad at all on a D9L. An L12 will probably be barely enough, but you'll have to stomach 80C+ temps.

 

[Engr62]

At what fan speed/power though? Because if the fan was for instance operating at only 20% power then thats a pretty impressive result, but if the fan was flat out at 100% then yeah it's quite high for open an open bench test.

You may only 4C headroom, but how much headroom have you got left in the fan speed?

I need to check to see what the fan speed is while trans-coding. I didn't change anything in the BIOS, so I'm assuming it was at the "standard" fan control setting and not "quiet" or "performance."

I don't know much about AMD, but 75C sounds awfully low for a throttle temperature. It seems like its actually 95C.

Unpopular opinion but nearly none of these coolers are well equipped to deal with a real 100W load. Handbrake/video encoding is a completely different ballgame from gaming, in many cases it'll have the CPU dissipating double the heat. 71C for a real 95W load isn't bad at all on a D9L. An L12 will probably be barely enough, but you'll have to stomach 80C+ temps.

From what I've read, there is a 20C offset for the Ryzen X CPUs (at least for the 1st generation). It's my understanding that the chip was actually running at 55C when it is reporting to the monitoring software that it's running at 75C. It had something to do with having consistent fan curves for the complete line of Ryzen CPUs (including the non-X). There is lots of conflicting information about this.

 

[Thehack]

I need to check to see what the fan speed is while trans-coding. I didn't change anything in the BIOS, so I'm assuming it was at the "standard" fan control setting and not "quiet" or "performance."



From what I've read, there is a 20C offset for the Ryzen X CPUs (at least for the 1st generation). It's my understanding that the chip was actually running at 55C when it is reporting to the monitoring software that it's running at 75C. It had something to do with having consistent fan curves for the complete line of Ryzen CPUs (including the non-X). There is lots of conflicting information about this.

Ryzen master should give you the actual reading. Most of that silliness is solved in second Gen and later bios update. They were afraid motherboard manufacturers would read the much cooler temp of Ryzen (soldered) in comparison to Intel and would have a bad default fan curve.

 

[Engr62]

Ryzen master should give you the actual reading. Most of that silliness is solved in second Gen and later bios update. They were afraid motherboard manufacturers would read the much cooler temp of Ryzen (soldered) in comparison to Intel and would have a bad default fan curve.

Thanks for the explanation--best I've seen. I checked the temps in Ryzen Master tonight. It is giving the same temperature readings as I was getting with CPUID HWMonitor.

At what fan speed/power though? Because if the fan was for instance operating at only 20% power then thats a pretty impressive result, but if the fan was flat out at 100% then yeah it's quite high for open an open bench test.

You may only 4C headroom, but how much headroom have you got left in the fan speed?

I was correct in my assumption... the CPU fan profile was set to "Standard" in the BIOS. I set it to "Maximum" and re-ran the trans-coding test. The peak temperature dropped to 62C with the fan running at 2077 RPM. I set it back to "Standard" and re-ran the test so I could see what speed the fan is running since I failed to record that earlier. Set to "standard," the fan peaked at 1849 RPM while achieving a peak CPU temperature of 70C.

I found this article on ITX coolers for Ryzen. My results with the NH-D9L jive with the results presented in the article when the author tests with AIDI64 FPU. I guess I'll go ahead and test with the Cryorig C7 when the AM4 adapter arrives since that is my only option of coolers I currently have that will fit in the MI-6, but I don't expect it to be sufficient.

My next step will be to fork over $60 when I can afford it to get the NH-L12. Either that, or just run this system in my NCASE M1. But, I prefer the CCD MI-6.