SENTRY: Console-sized gaming PC case project

JUNEK9 did you replace just the heatsink? As in, did you put the vapor chamber heatsink inside the aero shroud? Or did you use the whole FE shroud and heatsink? It's also possible the card has a custom MSI bios that is more likely to ramp the fan rpm up at lower temperatures because the original Aero cooler is crap. Do you use any OC/tuning software like Afterburner? Maybe try making a custom fan curve?

In either case, I definitely recommend trying to seal the gpu to the side of the case. I got this adhesive weatherstripping from a hardware store for just a few bucks. It should be tall enough to meet the side of the case. So, I guess 1/2-5/8in (15-20mm). I think I posted something on this many pages ago. Here's some more pictures.

x6Ghibq.jpg

This ensures the gpu can only draw in cool, fresh air and not any hot recycled air. Pretty simple. Avoid getting the seal too close to the fan or it will actually make more noise.

LPkw8nJ.jpg


I also have 40mm fans on the side of the gpu chamber but honestly they don't do very much. Though I highly recommend getting fan filters if you don't already have them. These are generic magnetic ones. I did close off half the gpu chamber too as you can see towards the top of the picture, this increased pressure inside the case. Which works well when combined with a sealed cpu fan to greatly increase pressure inside the case. Below is the sealed cpu chamber which took quite a bit of trial and error to get the seal close but not touching the face of the fan.

qRNojUj.jpg


Hot air is ejected out the sides of the case pretty forcefully. Without this seal the cpu fan is totally passive, not really causing a net change in pressure, so it can't bring in as much fresh air or force out much hot air. With the seal, the cpu fan becomes a high pressure inward facing case fan. So additionally sealing half the gpu compartment, and some of the cpu side (obviously not where it would be choking the cpu or psu) can greatly increase positive pressure.

I've done a few tests with and without seals I think it genuinely makes a difference. But as always, your mileage may vary. If someone has a better solution or a tweak to anything I've mentioned here, definitely share.
 
Apologies if this has been asked recently already:

Any ideas on a retail date for these? (Seems like most backers have theirs already from the original run??) Super keen lol.
 
Apologies if this has been asked recently already:

Any ideas on a retail date for these? (Seems like most backers have theirs already from the original run??) Super keen lol.

We're preparing for something small (like 100 pcs) to test our new transport company (DHL). After the campaign the main problem were long delivery times which very often were longer than the numbers claimed by the shipping company. Before we will go forward we need to make some improvements in this field. In a week or two we will release a newsletter about it.
 
We're preparing for something small (like 100 pcs) to test our new transport company (DHL). After the campaign the main problem were long delivery times which very often were longer than the numbers claimed by the shipping company. Before we will go forward we need to make some improvements in this field. In a week or two we will release a newsletter about it.

Awesome, thanks for your reply! My ML03 Silverstone giant desperately needs replacing as a HTPC :p. Will keep an eye out! Keep up the great work!
 
We're preparing for something small (like 100 pcs) to test our new transport company (DHL). After the campaign the main problem were long delivery times which very often were longer than the numbers claimed by the shipping company. Before we will go forward we need to make some improvements in this field. In a week or two we will release a newsletter about it.
Where do I sign up for this?
 
View attachment 50204
Just installed 3D printed velocity stack on L9i decreasing idle temp and noise.. Max load still hot with slighty lower noise.

I've been so very tempted to obtain a 3d printer for this very purpose. Shrouds, ducts, cowling, seals etc, are very useful for sff pc's. They're almost necessary due to the lack of case fans that normally bring in fresh air. I'm very curious to see more info on this.
 
  • Like
Reactions: RosaJ
like this
I've been so very tempted to obtain a 3d printer for this very purpose. Shrouds, ducts, cowling, seals etc, are very useful for sff pc's. They're almost necessary due to the lack of case fans that normally bring in fresh air. I'm very curious to see more info on this.
Not sure where you live, but in the states most cities/towns have a couple libraries with 3D printers now. All for a few dollars if that, depending. I’ve wanted to make ducks for my Sentry and even a case for my rasPi for awhile. Just suck at the designing part.

Also ditto on the temps!
 
  • Like
Reactions: Ej24
like this
Not sure where you live, but in the states most cities/towns have a couple libraries with 3D printers now. All for a few dollars if that, depending. I’ve wanted to make ducks for my Sentry and even a case for my rasPi for awhile. Just suck at the designing part.

Also ditto on the temps!

The university I'm at has several 3d printers. I even worked in a biomedical engineering lab with one but the scientists that used it were insanely protective of it, I didn't even have the passcode to get in the room so I didn't even bother asking to tinker with it.

Designing is the hardest part. I used to be fluent with Solid Works when I was younger but since moving to biomedical research I've forgotten everything I knew. I've tried messing around with FreeCAD and others. But it's frustrating. If I ever get back in to it I'd like to supply custom pieces and parts for modders, especially for projects like Sentry. But according to my girlfriend I already have too many hobbies. :D
 
  • Like
Reactions: RosaJ
like this
So I can't open my Sentry anymore. For some reason, after a few weeks of usage the screws were almost impossible to unscrew. I was careful, but the attached key destroyed the head of one screw (they are VERY soft) after just one attempt of unscrewing it - great. What do I do now? Should I just get a dremel and cut through the middle of the screw's head to be able to unscrew it with a flat screwdriver?

Also, what precision torx screwdriver should I use with the Sentry screws? I'm not really a handy-man so the sizes/symbols used are just confusing.
 
Last edited:
So I can't open my Sentry anymore. For some reason, after a few weeks of usage the screws were almost impossible to unscrew. I was careful, but the attached key destroyed the head of one screw (they are VERY soft) after just one attempt of unscrewing it - great. What do I do now? Should I just get a dremel and cut through the middle of the screw's head to be able to unscrew it with a flat screwdriver?

Also, what precision torx screwdriver should I use with the Sentry screws? I'm not really a handy-man so the sizes/symbols used are just confusing.

I recently had this issue myself. I ended up taking a dremel tool to the screw head, cutting a slot into it, and using a regular flathead screwdriver to remove the screw.
 
  • Like
Reactions: kyjol
like this
After a long process and a PSU/MB failure, I have almost finished my Sentry, still on the hunt for the best card and now that prices are crazy, I'm stuck waiting. I might wait for Volta/Ampere at this point.

Don't miss the benchmark/temps results at the END of this post!

Here's my sentry build: https://imgur.com/a/zeU8a

ASRock Fatal1ty Z370 Gaming-ITX/ac
Intel 8700 @ stock
Corsair Vengeance LPX 16GB (2x8GB) DDR4 DRAM 2666MHz (PC4 21300) C16
Thermolab LP53 Slim & Quiet CPU Cooler Intel 1155/1156
Noctua 92 x 14 mm Low-Profile Cooling Fan with A-Series Blades (NF-A9x14) attached w/ Silicone Cooking Bands
Samsung 960 EVO Series - 500GB NVMe - M.2 Internal SSD (MZ-V6E500BW) /w JP-P600-05 JunPus JP-P600 High Performance Thermal Pad 6W/mK (100x100mm) 0.5 mm on the front of the stick.
Seagate FireCuda 2TB SSHD
Corsair SF600 600W PSU - stock cables, would go for custom cables if someone else has the measurements for the board.
Demcifilters - Sentry in vertical position.

Future:
Nvidia gfx / BFGD monitor
5tb+ 2.5 storage drive


The termal pad helped temps out of the case for the 960 evo, 8C lower. It's the same temp inside the case with the thermalpad vs out of case without thermal pad, which is saying something.

I first had Kyronaut, but I found it really hard to apply. The paste is very dry and it sticks more to the spatula tool provided than anything else. The 10g tube looks to have a better applicator, I had the 1g tube.
The TIM was sliding around on the processor and the lp53.
I used PassMark Performance Test 9.0 on open air bench, my max cpu temps were between 80-85C, depending on the core with Kyronaut. I then decided to try Conductonaut, which got me a max of 80C, that was consistent across all cores. I then started to read a few stories from people having their conductonaut leaking and others warning it shouldn't be used between the cpu and heatsink. When I applied the conductonaut, I was trying to be careful, however, I didn't realize there was a small tip that I needed to attach. So I ended up with a ball instead of a pinhead. After spreading it, there was a small puddle aka lake effect. I knew I had too much. So I got paranoid and tore it down again, removed the conductonaut with q-tips and ISO. I went to reapply the LM, but it turns I used the whole tube in 1 use :(.

I reapplied the kryonaut but got worse temps, applied it a few times and still poor temps compared to the first paste. I finally decided to live with it and close up the case. When I powered it on, it died almost immediately. Pressing the power button barely spun the fan and got nothing. I tore it down again and flipped the case upside down.....a screw fell out. I never had this happen, I didn't believe it, I flipped it over before I closed the case.

So I thought the MB was dead. I was able to replace the motherboard. I tested it open air bench and turned it on.....nothing, not even a fan spin. I was about to flip my shit when I had the idea to see if the PSU is bad, I bought a refurb unit. I luckily had a PSU for 10 years that I never used, and I was able to test it. Power On, PC boots. At this point, I have to assume the screw came from inside the PSU. So I had to RMA my Refurb SF600 that I picked up over Black Friday.

While I waited for a new PSU, I lapped the heatsink and the CPU with 1000 grid 10x in each direction, then again for 2000 grit. Spoiler alert, lapping doesn't do anything on this setup.

I tried MX-4 since I ran out of Kyronaut, it is easy to apply. I re-pasted twice and got temps that matched the initial kryonaut paste. It's really on the application rather the material.

The lapping made the copper shiny again after the liquid metal, but no observed improvement. I also tried the LP53 with fins parallel to the ram vs perpendicular. Temps were better with airflow, parallel to the ram, heat not trapped between VRM and RAM, but this does put pressure on the ram.

I spent a bit of time doing cable management the best I could with the stock SF600 cables. I closed it up and it works (phew).

I benched it again, using the same test. My temps' max now hits 99-100C :(
I'm worried that when I add a gfx card, it will heat the CPU even more. I'm not overclocked but turbo boost is on. The RAM is running at CPU stock (2666). I'm just at the thermal limits of the heatsink and the case. My max is now 4.4ghz, with 2 cores going to 4.5, compared to 4.6 out of the case. Anyone else with similar temps?

The Intel diagnostics tool heated up the CPU more than Performance Test or Prime95, it hit 100C out of the case and failed the test. With the MX-4, it barely passed but still hit 100C (probably a fluke, since the temp is checked every 2-3 seconds)

I never delidded and after seeing these stress temps, It's looking more likely I'll be willing to attempt it. Anyone have temp comparisons for a 8700 delid?

I'll be keeping the sentry vertical.

I still need to get a GFX card.

I was also in a unique position where I was able to capture benchmarks with various thermal pastes and I was able to test before and after Meltdown microcode updates.

https://imgur.com/a/XyLno#puDqlIi

Wow that's a lot of tinkering, trial and error, and headaches. Just to clarify, you are using the 65W tdp i7-8700? Non-k? If so, you definitely shouldn't get temps like that. I use the LP53 as well and never get temps like that. I was previously using a 4790S (65W tdp) in one of my Sentry pc's, which I upgraded to a 4790k (88W). With the 65W cpu, iirc, I would get temps up to ~75C when running Passmark. With the 4790k I'd reach maybe mid 80's to 90C. Then I delidded my cpu replaced the tim with liquid metal and temps dropped like crazy, I rarely see over 70C anymore during normal gaming use.

I don't know exactly why you're seeing such temps. You've been using decent thermal paste. LP53 is a good cooler.

You still have the fins parallel to the ram?
Are the screws holding the heatsink down fully tightened? Is pressure being applied evenly? When you remove the heatsink do you see a good spread of paste? Or is it only spreading in one direction?

The cpu power ribbon cable looks like it might be blocking some outward airflow. Probably negligible though.

Maybe try a different method of mounting the fan? I know some people here have come up with crafty solutions for the lp53. The rubber bands may be hindering air intake slightly.

Don't worry, we'll figure this out. Sentry is a small case but you still shouldn't get temps like that.

Edit:
did I see correctly that you're cpu vcore peaked at 1.375V? That's definitely not ok for stock coffee lake. Something is wrong. A 14nm Intel cpu like that shouldn't go over ~1.2v vcore unless you're manually tweaking voltage for OC purposes. My older 22nm Devils Canyon doesn't even go over 1.2V stock!

Check your bios, some boards may ship with a bios that is configured for some sort of Auto Overclocking that raises vcore to unacceptable levels. Something is wrong with your vcore.
 
Last edited:
Looks like AR11 will cool well but be really noisy like the C7 because of the 47mm height pushing the fan up against the case.
 
Looks like AR11 will cool well but be really noisy like the C7 because of the 47mm height pushing the fan up against the case.

With the 92mm x 14mm fan from Noctua it should be fine. I'll be curious to see how it performs. Though direct contact heatpipes are of dubious benefit in graphics cards due to the miniscule contact area, I think it will work alright for cpu's as they have an IHS to spread the heat.

We'll see how it compares. It's only 290g in weight. The NH-L9i is 345g. The LP53 is 410g. It's hard to compete with a larger mass of metal. Though fin arrangement and heatpipe configuration can make a big difference, so we'll wait and see how it performs.
 
Last edited:
With the 92mm x 14mm fan from Noctua it should be fine. I'll be curious to see how it performs. Though direct contact heatpipes are of dubious benefit in graphics cards due to the miniscule contact area, I think it will work alright for cpu's as they have an IHS to spread the heat.

We'll see how it compares. It's only 290g in weight. The NH-L9i is 345g. The LP53 is 410g. It's hard to compete with a larger mass of metal. Though fin arrangement and heatpipe configuration can make a big difference, so we'll wait and see how it performs.

The original fan is 15mm, so the noise will still be intense even with the Noctua fan.
 
The original fan is 15mm, so the noise will still be intense even with the Noctua fan.

Is it? I thought it was 20mm. Smallformfactor.net listed it as 20mm but Silverstone's website says 15mm so I guess it is 15mm.
 
Is it? I thought it was 20mm. Smallformfactor.net listed it as 20mm but Silverstone's website says 15mm so I guess it is 15mm.
The original prototype version that debuted back in Computex 2017 had a 20mm fan. It's pretty clear as day that it's a different fan in this picture:

https://smallformfactor.net/news/silverstones-new-argon-ar11-49mm-lp-cooler

Having said that, my guess that this will not be as good as the C7 based on direct contact heat pipes. No matter what marketing BS they feed you, direct contact heat pipes is not better than having a flat surface. This is strictly a cost cutting measure. Therefore, don't get your expectations up in matching the C7 let alone beating it.
 
Last edited:
Anyone remember the GPU cooler clearence height ? so I dont have to open it up again. I was considering putting the Raijintek MORPHEUS II CORE EDITION on my old Nano card (pretty sure it will mount to it) to silence it maybe see if I could jam a low profile 120mm in behind it.
 
Anyone remember the GPU cooler clearence height ?
There is ~48mm clearance for the GPU cooler.

Instead of fans underneath you could try a whole lot of 40mm fans along the side.
 
Last edited:
There is ~48mm clearance for the GPU cooler.

Instead of fans underneath you could try a whole lot of 40mm fans along the side.

Hmm great then it should fit with a few mm left. I was planning on having one 40mm intake right on the vrm of the card and since the print is so short on the nano i think a low profile 120mm would fit on to its rearside. Now im only missing the cryorig copper edition to free up my 4x40mm noctua fans from the cpu but I guess that that might never arrive just like the vega nano :(
 
Hmm great then it should fit with a few mm left. I was planning on having one 40mm intake right on the vrm of the card and since the print is so short on the nano i think a low profile 120mm would fit on to its rearside.

That cooler is 44mm thick. How are you going to fit a slim 120mm fan?
 
That cooler is 44mm thick. How are you going to fit a slim 120mm fan?

Yeah there wont be any room left against the chassis side, but look at how short the nano card is, I think there might be enough room "above" the cooler on the part that will stick out way beyond the pcb against the case side.
 
Thank you, I don't have it overclock, but you might be right. I'll check and post my settings.

Edit: Everything is set to auto.

I'm on MAX non-turbo performance with the ASRock board.
I agree with Ej24, that vcore seems wack! I’d start asking around here on the Intel forum and see if they may know. Or do some google searching as even those idle temps are high.
 
Thank you, I don't have it overclock, but you might be right. I'll check and post my settings.

Edit: Everything is set to auto.

I'm on MAX non-turbo performance with the ASRock board.

Everything is set to auto? Strange, your VID is still high under the cpu heading. Under motherboard voltages the vcore looks less scary, still 1.28v is high. I assume you've tried resetting to defaults in the bios? Its possible you have a bad cpu or possibly a bad motherboard. These temps even in Sentry are not normal. This is a 8700 or 8700k?
 
Ugh, so the price went up for the 100 unit limited preorder? I was expecting ~$235 and they are ~$290 instead. Shipping and fees brought it to about $360 shipped to the US.

I didn't run the conversion till after I placed the order (dumb). I'd read $235 so many times that I had no doubt.

Edit: Yes I see the 230 euro note. I've been waiting over a year to order one so I hurried thru the check-out process, as the internet has conditioned me to do for limited items. I was expecting to be in around $300 total so it sucks, but the kids won't starve or anything. I also found a statement by them in the SFF forums where they covered this, so it's working as intended.
 
Last edited:
USD to EUR crashed during previous month and a half from all time high to the lowest ratio during the last 3 years, thus it's more expensive in USD for you.
 
Back
Top