Production SENTRY 2.0: Evolution of console-sized gaming PC case

[NeroFX]

An update on my stripped screw guys, and a happy 2021 to you all!
I ended up buying a combi drill as I don't have one and I'm sure it'll come in handy for years to come.
I also bought a screw extractor that went down to a T10 size, tiny!
I got the thing out, I'm so relieved as I want to send back my Zotac 3060 Ti to overclockers as the fan noise is far too high, and there's an LED that can't be switched off.
The last thing is I'm still wondering how this happened, but you can see in one picture the hole alignment is off, and I wondered if that has something to do with it?
Perhaps the power cables are pushing the case out causing force on the screws? I'm looking into some braided cables now, before I attempt putting on the lid again.

I also have some black acrylic paint Dr Zaber sent me ( I had a slight defect on my case), so I can touch all the scratches back up.

 

[tja4430]

Wanted to talk about the 5600x performance and thermals in the Sentry 2.0 case. I've had my build in my Sentry 2.0 running since early in August this year with the following:

Ryzen 5 3600
Noctua L9A Chromax
Gigabyte B550I AORUS PRO AX ITX Motherboard (Bios - Gigabyte Re-released F11 Bios, Final version in Tweaktown Beta BIOS thread)
WD SN750 500 GB PCIe NVME
SK Hynix S31 Gold 1 TB Sata SSD
HyperX 16 GB (2x8 GB sticks) DDR4-3200 CL16 (XMP was set for all testing)
Nvidia 2070 Super FE Graphics Card
Corsair SF600 Platinum.

I upgraded to a 5600x I was able to snag 2 weeks ago.

I did a bunch of testing today with the 5600x at both stock settings, and some mild all-core OC with fixed Vcore, and wanted to share my findings.

The first big difference between the 3600 and the 5600x, is that the 5600x has a single CCD and CCX. The 5600x has a lower PPT of 76W vs. the 3600's 88W PPT.

When doing stress testing with the 3600 at stock with Prime95 (smallest FFTs), my temps instantly rose into the 90s, quickly hitting 95 degrees C, which is the AMD rated max temp. When doing stress testing with the 5600x at stock (smallest FFTs - ran for 2 hours), the maximum temperature I hit was 91 degrees C, with an average of roughly 88-89 degrees C. This is significantly improved from the 3600 at stock settings. The average Vcore read by HWInfo was roughly 1.031 V.

Its important to note here based on these results that running the 3600 and 5600x in Prime95 in this case with a small fan cooler like the L9A at a fixed voltage with an all-core OC will probably run you instantly at 95 degrees C, and probably more. The AMD internal chip voltage and current optimization is very smart with stock settings, and manually setting a vcore of even 1.1 V will cause the temperatures to spike considerably.

I also tested the 5600x with Cinebench R20. After a few runs, I noticed max temps of 81 degrees C, with a pretty consistent score of 4080 +/- 15. Note that my memory is not necessarily the fastest kit, and obviously, there are multiple 5600x's that have achieved scores in Cinebench much higher. Compared to the 3600 with the same setup, the 3600 scored roughly 3550 or so at stock several months ago when I tested it, but with temperatures of 87-88 degrees at stock.

I also tried running Cinebench R20 with a mild all-core OC (again DO NOT run a fixed VCore all-core OC with Prime 95 in this case), of 4.0 GHz at 1.14 VCore, and for the 5600x temps settled around 79-80 degrees C. I repeated this several times, and got similar results. With the same voltage set for the 3600 testing I did a few months ago, the 3600 temps settled around 85-86 degrees C.

In terms of raw performance, the 3600 and 5600x wont provide much difference for the average user. But, the 5600x does have a significant improvement with thermals at stock, and does provide better out of the box performance based on benchmarks, and games (in terms of FPS) for that matter.

I have ordered the JHack 10 mm shroud (for 92-mm Noctua fan), so I am curious at how much the thermals improve for the 5600x at stock.

In conclusion, I would say the 5600x is a CPU that fits right at home in the Sentry 2.0 case, and has better thermal performance at stock over the 3600.

UPDATE: I had the incorrect temps for the Cinebench R20 all-core 4.0 GHz tests. Fixed those. Not as drastic as the previous incorrect numbers, but still a fairly significant reduction in temps for the 5600x.

 

[Dax911]

Wanted to talk about the 5600x performance and thermals in the Sentry 2.0 case. I've had my build in my Sentry 2.0 running since early in August this year with the following:

Ryzen 5 3600
Noctua L9A Chromax
Gigabyte B550I AORUS PRO AX ITX Motherboard (Bios - Gigabyte Re-released F11 Bios, Final version in Tweaktown Beta BIOS thread)
WD SN750 500 GB PCIe NVME
SK Hynix S31 Gold 1 TB Sata SSD
HyperX 16 GB (2x8 GB sticks) DDR4-3200 CL16 (XMP was set for all testing)
Nvidia 2070 Super FE Graphics Card
Corsair SF600 Platinum.

I upgraded to a 5600x I was able to snag 2 weeks ago.

I did a bunch of testing today with the 5600x at both stock settings, and some mild all-core OC with fixed Vcore, and wanted to share my findings.

The first big difference between the 3600 and the 5600x, is that the 5600x has a single CCD and CCX. The 5600x has a lower PPT of 76W vs. the 3600's 88W PPT.

When doing stress testing with the 3600 at stock with Prime95 (smallest FFTs), my temps instantly rose into the 90s, quickly hitting 95 degrees C, which is the AMD rated max temp. When doing stress testing with the 5600x at stock (smallest FFTs - ran for 2 hours), the maximum temperature I hit was 91 degrees C, with an average of roughly 88-89 degrees C. This is significantly improved from the 3600 at stock settings. The average Vcore read by HWInfo was roughly 1.031 V.

Its important to note here based on these results that running the 3600 and 5600x in Prime95 in this case with a small fan cooler like the L9A at a fixed voltage with an all-core OC will probably run you instantly at 95 degrees C, and probably more. The AMD internal chip voltage and current optimization is very smart with stock settings, and manually setting a vcore of even 1.1 V will cause the temperatures to spike considerably.

I also tested the 5600x with Cinebench R20. After a few runs, I noticed max temps of 81 degrees C, with a pretty consistent score of 4080 +/- 15. Note that my memory is not necessarily the fastest kit, and obviously, there are multiple 5600x's that have achieved scores in Cinebench much higher. Compared to the 3600 with the same setup, the 3600 scored roughly 3550 or so at stock several months ago when I tested it, but with temperatures of 87-88 degrees at stock.

I also tried running Cinebench R20 with a mild all-core OC (again DO NOT run a fixed VCore all-core OC with Prime 95 in this case), of 4.0 GHz at 1.14 VCore, and for the 5600x temps settled around 79-80 degrees C. I repeated this several times, and got similar results. With the same voltage set for the 3600 testing I did a few months ago, the 3600 temps settled around 85-86 degrees C.

In terms of raw performance, the 3600 and 5600x wont provide much difference for the average user. But, the 5600x does have a significant improvement with thermals at stock, and does provide better out of the box performance based on benchmarks, and games (in terms of FPS) for that matter.

I have ordered the JHack 10 mm shroud (for 92-mm Noctua fan), so I am curious at how much the thermals improve for the 5600x at stock.

In conclusion, I would say the 5600x is a CPU that fits right at home in the Sentry 2.0 case, and has better thermal performance at stock over the 3600.

UPDATE: I had the incorrect temps for the Cinebench R20 all-core 4.0 GHz tests. Fixed those. Not as drastic as the previous incorrect numbers, but still a fairly significant reduction in temps for the 5600x.

Thank you for this feedback , it is very appreciated. Can you also share your thermal with the shroud when you receive it please?

 

[aaaadrian]

Hey guys, I bought the new 5cm LINKUP PCIE 4.0 riser cable and managed to fit it into Sentry 2.0 properly. I can't find this kind of information anywhere on the internet thus sharing my experience so you guys are informed.

If you want to try this, you will need to cut off the PCB on one side and PCIE slot locking tab on the other side in order to allow the black sentry 2.0 riser piece to bite into the riser properly. Besides cutting, another major challenge is bending riser. Even at 5cm the riser is still too long and you'll need to bend and tuck away the excessive cable. I ended up bending it inwards so it doesn't touch side panel when I close it. Bending took me a lot of time.

I verified that I'm now able to enable PCIE 4.0 without problem. The main reason I did this is to avoid taking apart my wife's Velka 3 (trust me this is not fun) and borrow her graphics card every time I need to upgrade my bios, since I have a 5600x and bios updates are very frequent. This is definitely not ideal the new riser is much harder to take out or put back in, but it'll work till (hopefully) someday new sentry-compatible gen 4 riser card comes out.

I have taken some pictures for your reference:

 

[Alster370]

Wanted to talk about the 5600x performance and thermals in the Sentry 2.0 case. I've had my build in my Sentry 2.0 running since early in August this year with the following:

Ryzen 5 3600
Noctua L9A Chromax
Gigabyte B550I AORUS PRO AX ITX Motherboard (Bios - Gigabyte Re-released F11 Bios, Final version in Tweaktown Beta BIOS thread)
WD SN750 500 GB PCIe NVME
SK Hynix S31 Gold 1 TB Sata SSD
HyperX 16 GB (2x8 GB sticks) DDR4-3200 CL16 (XMP was set for all testing)
Nvidia 2070 Super FE Graphics Card
Corsair SF600 Platinum.

I upgraded to a 5600x I was able to snag 2 weeks ago.

I did a bunch of testing today with the 5600x at both stock settings, and some mild all-core OC with fixed Vcore, and wanted to share my findings.

The first big difference between the 3600 and the 5600x, is that the 5600x has a single CCD and CCX. The 5600x has a lower PPT of 76W vs. the 3600's 88W PPT.

When doing stress testing with the 3600 at stock with Prime95 (smallest FFTs), my temps instantly rose into the 90s, quickly hitting 95 degrees C, which is the AMD rated max temp. When doing stress testing with the 5600x at stock (smallest FFTs - ran for 2 hours), the maximum temperature I hit was 91 degrees C, with an average of roughly 88-89 degrees C. This is significantly improved from the 3600 at stock settings. The average Vcore read by HWInfo was roughly 1.031 V.

Its important to note here based on these results that running the 3600 and 5600x in Prime95 in this case with a small fan cooler like the L9A at a fixed voltage with an all-core OC will probably run you instantly at 95 degrees C, and probably more. The AMD internal chip voltage and current optimization is very smart with stock settings, and manually setting a vcore of even 1.1 V will cause the temperatures to spike considerably.

I also tested the 5600x with Cinebench R20. After a few runs, I noticed max temps of 81 degrees C, with a pretty consistent score of 4080 +/- 15. Note that my memory is not necessarily the fastest kit, and obviously, there are multiple 5600x's that have achieved scores in Cinebench much higher. Compared to the 3600 with the same setup, the 3600 scored roughly 3550 or so at stock several months ago when I tested it, but with temperatures of 87-88 degrees at stock.

I also tried running Cinebench R20 with a mild all-core OC (again DO NOT run a fixed VCore all-core OC with Prime 95 in this case), of 4.0 GHz at 1.14 VCore, and for the 5600x temps settled around 79-80 degrees C. I repeated this several times, and got similar results. With the same voltage set for the 3600 testing I did a few months ago, the 3600 temps settled around 85-86 degrees C.

In terms of raw performance, the 3600 and 5600x wont provide much difference for the average user. But, the 5600x does have a significant improvement with thermals at stock, and does provide better out of the box performance based on benchmarks, and games (in terms of FPS) for that matter.

I have ordered the JHack 10 mm shroud (for 92-mm Noctua fan), so I am curious at how much the thermals improve for the 5600x at stock.

In conclusion, I would say the 5600x is a CPU that fits right at home in the Sentry 2.0 case, and has better thermal performance at stock over the 3600.

UPDATE: I had the incorrect temps for the Cinebench R20 all-core 4.0 GHz tests. Fixed those. Not as drastic as the previous incorrect numbers, but still a fairly significant reduction in temps for the 5600x.

Ive recently upgraded from a i5 4690k to a 5800x, which is in the sentry 2.0 with a noctua NHli9. Idle stabilizes at 42c. In games I get 4.5~4.6ghz boosting across all cores, package temperature holds at about 87C, processor power seemed to be capped at 65W even though its not in eco mode (5800x is 105w cpu?) unless it was limiting power based on thermals.

Due to the shortage of all new cards ive crammed my 980ti jetstream into the sentry. its a tall card, and its a 2.5+ slot card, its ridiculously tight it there, no spare room for power cables, about 0.5-1mm between gfx fans and the case, its pretty loud!

 

[Nomo]

Good news from EVGA.
I think that could be a good option for a AIO Cooling + GPU Combo.


I hope this helps.

 

[riba2233]

Good news from EVGA.
I think that could be a good option for a AIO Cooling + GPU Combo.


I hope this helps.

I think it is to long for that combo.

 

[tja4430]

As an update, I received my 92 mm fan shroud (10 mm - Tall RAM compatible version). It fits perfectly in the Sentry Case.

I started playing with PBO2 today and undervolting with the 5600x. I watched Optimum Tech's video above to learn a little more about PBO2 and curve optimizer. I put an all curve negative magnitude of 30, while keeping the stock power profile (PBO Limits -> disabled), similar to what Optimum Tech did in the above video.

I ran tests using Cinebench R23, and my motherboard is the B550i Aorus Pro AX - F11 Released BIOS with XMP Enabled (DDR4-3200 CL16). Stock PPT settings (76W) still applied, as well as TDC and EDC.

5600x at Stock, Cinebench R23 - 10,270; Temps Max at 91. Average all core frequency of 4.05 GHz

5600x With PBO2/Curve Optimizer (all core negative magnitude offset of 30), Cinebench R23 - 11,130; Temps Max at 86. Average all core frequency of 4.35 GHz

With PBO2/Curve Optimizer with 92mm fan shroud installed, Cinebench R23 - 11,198; Temps Max at 82. Average all core frequency of 4.39 GHz.

Keep in mind the F11 BIOS for my board has the "pre-released" PBO2/Curve Optimizer with AGESA 1.1.0.0d. The AMD final released AGESA with PBO2 will be 1.1.9.0.

In any case, this is very significant improvement with this feature, not only for average all core performance, but with thermals as well.

 

[Alster370]

5800x cooled by noctua nh-l9a , what temps/clocks should I expect. Ideally I want to hard lock it to 65w max, and then do the PBO2 above to improve efficiency.

Doing the BMW blender render my processor hits the 90C package temp limit, then ryzen starts lowering clocks. Starts at about 85w, finishes the test at about 60w @3.7ghz. 3m 21s
Is this expected? Or do I need to reapply thermal paste?

 

[SaperPL]

That's how it works. The turbo is pretty aggressive and those CPUs will run at the boundary of their temperature and power limits.
The big misleading thing with Zen2 launch was that we've got a lot of people getting 3700X with boxed Wraith Prism which is like rated for 135W TDP iirc and people reported good temps for their "65W" CPUs, but that's completely different than running it on an actual 65W TDP cooler.

I wouldn't worry to much about the temps if you're just reaching 90 degrees sometimes since AMD stated those temps are okay if they are below the max temp. It's the same as 80~85 degrees on GPUs. The issue is the noise and clock speed, whether your setup is crippled by SFF fan or not.

 

[Alster370]

That's how it works. The turbo is pretty aggressive and those CPUs will run at the boundary of their temperature and power limits.
The big misleading thing with Zen2 launch was that we've got a lot of people getting 3700X with boxed Wraith Prism which is like rated for 135W TDP iirc and people reported good temps for their "65W" CPUs, but that's completely different than running it on an actual 65W TDP cooler.

I wouldn't worry to much about the temps if you're just reaching 90 degrees sometimes since AMD stated those temps are okay if they are below the max temp. It's the same as 80~85 degrees on GPUs. The issue is the noise and clock speed, whether your setup is crippled by SFF fan or not.

That makes sense. Havent had any instability running it at 85-90c for extended periods, so il take AMDs word that its normal. Hopefully a 3d printed duct should drop temps 5c or so just to get it in a more comfortable range.

On another note, im running a 50mm thick 980ti in the sentry at the moment. temps are ok, about 80c at full load, but predictably its quite noisy, id say my tolerance is fans on 65%.

Im looking at a 6800/6800xt, im wondering, even though the 6800 is a 2 slot, and so its quieter because there is more room between the fans and the perforations, it runs about 5-8c hotter than the 6800xt as the cooler is physically smaller, which means it will need to spin the fans faster to hold the same temp as the 6800xt. So what im asking is which do people think will be louder during gaming?

 

[kai535]

Here's my recent build with the short evga 3060ti I placed a 92mm aio at the end to show how much space is left, I think if you installed the aio first and slid the aio down under closer towards the psu the tubing connectors that sick out would have enough space and clearance without and case mods but it would be then missing some of the ventilation and you would probably need custom cables as I have a ton of slack on my 8 pin cpu and and pcie cable.. Also im using the new J-hack M2426 that cleans up the the 24 pin atx cable and frees up a ton of space its also the 220mm length adapter too

Lastly I toss out the 3.0 usb cable a while back and used a 2.0 cable instead for less cables but the asrock z390 tb3 mobo has a weird placement for the USB connector and now I need to get a longer one. Or go back to a 3.0.. any idea on where to get one?

 

[SaperPL]

Here's my recent build with the short evga 3060ti I placed a 92mm aio at the end to show how much space is left, I think if you installed the aio first and slid the aio down under closer towards the psu the tubing connectors that sick out would have enough space and clearance without and case mods but it would be then missing some of the ventilation and you would probably need custom cables as I have a ton of slack on my 8 pin cpu and and pcie cable.. Also im using the new J-hack M2426 that cleans up the the 24 pin atx cable and frees up a ton of space its also the 220mm length adapter too

I'm curious about the performance of that thing. Running a 120 mm AIO didn't really help much with Zen2 temperature issue - the designs of the blocks might not be perfect for the chiplet layout until you go for those big triple fans rads that already got designs made for Threadripper.

Lastly I toss out the 3.0 usb cable a while back and used a 2.0 cable instead for less cables but the asrock z390 tb3 mobo has a weird placement for the USB connector and now I need to get a longer one. Or go back to a 3.0.. any idea on where to get one?

Here's a one that has precise specs: https://www.aliexpress.com/item/32947900751.html

Essentially you need to make sure you'll get those specs - merchant may just have the same photo without stating dimensions and will send you whatever version they have. It needs that 47 mm of distance between screw holes / 60 mm total width. I think this is a store somehow connected to our supplier since this detailed image popped up after we shipped first Sentry. Anyway if you have a store closer to you, ask for these dimensions before you place the order.

 

[Alster370]

@SaperPL Just wondering if you guys did any testing regarding blanking/closing the two open sections of the central support, using either plastic or rubber covers (the ones with the slits in for cable management). I know that the hole just opposite the PSU is mostly blocked off by wires anyway, but the area where the pci passes through is fairly open and would allow hot air to pass from gpu to cpu section and vice versa.

something like this for the area opposite the psu, plastic cover for the other, but would need to be installed after the riser has been fixed in place

 

[SaperPL]

We did some tests early in the design of 1.0, but the performance difference was negligible for the CPU while also it caused some issues with various cable layouts. We planned to use the area behind the central support wall for excess PSU cables and it worked pretty well for some of first SFX-L PSU cables. It was also limiting in desktop/horizontal orientation of the case for the gpu cooling.

Note however that we did tests this for 150W TDP cards as we assumed this to be a target for first Sentry and we ended up now with 300W+ TDP cards now and also we see that most users of Sentry do use it on a desk with vertical stand, so we might revisit the idea at some point.

 

[phinix]

Is there a temperature difference when using the case in horizontal position?
Any disadvantages, cons using it flat?

 

[SaperPL]

Is there a temperature difference when using the case in horizontal position?
Any disadvantages, cons using it flat?

In horizontal position the airflow to the GPU compartment intake is limited so it's better to use blower style cards in this configuration. Open air cooler cards will be at disadvantage in horizontal orientation.

 

[phinix]

What if I place it GPU up? So my 3070FE intake would be on top.
Of course I would add some tall feet.

Do you have any temp tests with this position with actual temp measures?

 

[SaperPL]

If you do that, you'll most likely throttle your CPU because it'll be like an oven since the CPU cooler won't be able to effectively push the air outside that will accumulate in motherboard area.

 

[phinix]

If you do that, you'll most likely throttle your CPU because it'll be like an oven since the CPU cooler won't be able to effectively push the air outside that will accumulate in motherboard area.

I thought this would happen, thanks for confirming.
So basically there is no point putting the case horizontally then, it will fail each way.