• Save 15% on ALL SFF Network merch, until Dec 31st! Use code SFF2024 at checkout. Click here!

Streacom DB4 build: "The Zen of Yin"

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
For two weeks ago, I started a build that I thought I would share with you.

  • Streacom DB4 - black version
  • Ryzen 1700
  • Biostar X370GTN
  • G.Skill 3200Mhz@cl14
  • Samsung 960 pro
  • MSI RX 550 ITX
  • Corsair SFX 600
  • HDPlex VGA cooler kit
  • HDPlex CPU cooler kit

I wanted to create a total black build by spray painting all the parts to give it a theme. I wanted to do this with a RX 470, but the miners did beat me to the punch every time they hit the stores.



This is my first test run just to verify everything was working correctly.




Here I have painted the frame black. This was in silver from the box.

The next images is a bit fast forward, since I didn't take any photos of the actual painting job... Wish I had now while creating this post.



Here we can see the PSU, videocard and backside of the motherboard.



This is from the other side.



I did make a hole in the top plate to fit a slim 120mm fan to suck air out of the case. I did also cover the top by cutting up a ladies thighs made from merino fabric and glued it there. I will have to redo this. Find a better way to do this with more grace.



Here you can see where I am today. I have not painted the heatpipes, but I will when I get the new ones I have ordered. You can also see that the graphic card is not cooled by the case. The computer refused to boot with the heatsink from HDPlex attached to the graphic card. I didn't bother debugging it since I am waiting for an RX 570 mini.


Now let us discuss thermal performance. As you can see I have 4 heatpipes connected to one panel and 4 other to another panel. One would believe that this could handle an Ryzen 1700, but it does not. Aida64 (CPU, FPU, Cache, Memory - all at once) brings the cpu to it knees after 30 minutes - it starts to throttle. This is very, very surprising to me. By my calculations those to panels with 4 heatpipes on each should be able to handle that easily. They don't - as of now.

My research, or should I say debugging, on this is that I have made two huge mistakes.

  1. The heatpipes is not aligned horizontally. I have come to my attention that this is very important for max performance. I have instead done all of them vertically.
  2. The bends... It is really hard to make those bends good without the heatpipes collapsing inwards. I have done my best, but there are some points where it looks ok-ish, but after some research I found out that those really hinders performance badly. I read somewhere that with a strait heatpipe carrying 100% and with a perfect 90 degree (molded) bend, it will bring it down to about 93%. I don't remember the math for it, but you added all the degrees of bends and divided that on some value created from the length, diameter and effect from the heatpipe.

With my new knowledge, I have contacted Streacom and ordered new heatpipes and heatsinks. I will tame that Ryzen 1700.



Here are my desk:


 
Last edited:

msystems

King of Cable Management
Apr 28, 2017
804
1,405
Yes, if you can redo those heatpipes so they only turn horizontally and/or up, it should help lower the thermal resistance. A pipe bender can help to do the bends smoothly so they don't kink.

Now let us discuss thermal performance. As you can see I have 4 heatpipes connected to one panel and 4 other to another panel. One would believe that this could handle an Ryzen 1700, but it does not. Aida64 (CPU, FPU, Cache, Memory - all at once) brings the cpu to it knees after 30 minutes - it starts to throttle. This is very, very surprising to me. By my calculations those to panels with 4 heatpipes on each should be able to handle that easily. They don't - as of now.

Pretty strange. Hopefully this is just a heat transfer problem and not the thermal limit of the cooling surface area. That size of area should be able to handle a 65 watt processor. Can you see if both panels feel evenly hot? Might give you some insight on if one side is under performing on the thermal transfer.

What are you using in the picture, is it the stock heatpipe kit? Or do you have the LH6 kit also?
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
@msystems I will look up a pipe bender and see if that would help. I would really like to get them as good as possible.

I tried it first with my 6700k (delidded) with only four heatpipes. Then I was quite impressed by the cooling performance for a 65w panel on a 95w cpu. I can say that when touching the panels, the one with the longest heatpipes(those who came with the case) gets much hotter than the other panel. I am not sure if this is a heat transfer problem or not. The one that is colder, only uses those small heatsinks that came with the LH6 kit. You can feel that it is really hot just where those are connected to the panel, but that heat does not spread like it does on the other panel.

I have however ordered new heatpipes and an extra large heatsink (like the one with the case) to spread it even more. Those two panels should be more than enough.
 
Last edited:

TinyAudio

Cable-Tie Ninja
Jan 9, 2017
188
132
just spitballing: Could the fact you are using forced extraction on a passive case be cooling in some way the heat pipes? limiting the transfer of heat to the panel?

This case looks awesome in black!
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
@TinyAudio That could be the case. I haven't done any extensive testing with/without the fan, since my focus has been on the heatpipes themselves. In some tests I have seen a drop by 5 degree Celsius. This was done with stress testing. However, a test I did when playing Battlefield 1 - with the fan I hit 73 degree Celsius, but without I hit 72 degree. I haven't taken those numbers seriously yet. There are many factors like room temperature and minutes played in one session. But, it did surprise me.

The fan is 1200 rpm fan, but I have ordered a 2000 rpm one. I hope to see a more consistent result with that fan on 100%. My goal is to have the fan only kicking in when I am gaming to help get those extra fps.
 

zovc

King of Cable Management
Jan 5, 2017
852
603
Regarding bending the heatpipes, I'd check with HDPLEX. If they have a non-liquid filling (i.e. foam), I'm under the impression you could damage the material on the inside. Even if the pipes themselves don't rupture, breaking down the foam or whatever else is in there would almost certainly affect the conduction properties of the pipes.
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
@msystems I just bought a pipe bender. Turns out they had it at my local hardware supplier. It was quite cheap as well.



The result is amazing. Actually so amazing that I believe that without knowing one can't tell apart a bend from this tool and a bend from the factory. Lets give it a go. In the next image there are two bends, one from the tool and one from the factory. Can someone see the difference?



In short, thanks for the tips @msystems! I give you 10 internet points.
 
  • Like
Reactions: Julio Espinosa

msystems

King of Cable Management
Apr 28, 2017
804
1,405
@msystems I just bought a pipe bender. Turns out they had it at my local hardware supplier. It was quite cheap as well.



The result is amazing. Actually so amazing that I believe that without knowing one can't tell apart a bend from this tool and a bend from the factory. Lets give it a go. In the next image there are two bends, one from the tool and one from the factory. Can someone see the difference?



In short, thanks for the tips @msystems! I give you 10 internet points.

Cool, glad that helped.

I suspect you have a transfer problem on one side. If the heat transfer is working optimally then both sides should heat up and reach equilibrium. Lets hope the new heatpipes fix it.
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
Time to post some number on the effect on the fan mounted on the top.

For the testing, I removed the top mesh so that I got optimal airflow.



I now have to type of fans. (http://www.scythe-eu.com/en/products/fans/slip-stream-120-mm-slim.html)

I have the 1200 - and the 2000 rpm model and done some testing with both, with both of them in two variations: pulling and pushing air.
The test was simple in the sense that I let the CPU get down to 45 degree Celsius and the start a Aida64 stress test (CPU, FPU, Cache, Memory) and did let it run for 10 minutes exactly with fan on 100%. Did the same for all the runs.

  • 2000 rpm fan (pulling air out) 74 degree Celsius
  • 2000 rpm fan (pushing air in) 75 degree Celsius
  • 1200 rpm fan (pulling air out) 76 degree Celsius
  • 1200 rpm fan (pushing air in) 77 degree Celsius
  • Reference (no fan) 82 degree Celsius

The difference between best case and worse case is in this test 8 degree Celsius. I also did a test with Battlefield 1 after I removed the mesh on top. About 1 hour of playing with the mesh and with a fan, gave me 73 degree Celsius. 1 hour of playing without the mesh, but with a fan, gave me 61 degree Celsius. That is a 12 degree Celsius difference! That gives me some hope that I can squeeze out extra bit of performance of an RX 570 when that comes around. Compared to the other user on this forum doing this build with an RX 570.

It also says that the mesh I had on top must be stretch thinner so that air more easily can pass through.


Last, a warning, don't ever buy this:



I hoped that it would be able to transform 0 - 100% PWM to 0 - 12v DC, but that is still a dream. I haven't measured it with a voltmeter, but I guess it goes from 5v (or 7v) to 12v. This is not good enough for me. I want to shut the fan completely down when not gaming and only begin when I stress the system. I will have to figure out something different since the motherboard don't support 3pin control.
 
  • Like
Reactions: Phuncz and zovc

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
@zovc I just got an reply from HDPlex saying that the heatpipes are bendable, but they can't be cut. That would render them useless. For future reference.
 
  • Like
Reactions: zovc

msystems

King of Cable Management
Apr 28, 2017
804
1,405
Last, a warning, don't ever buy this:



I hoped that it would be able to transform 0 - 100% PWM to 0 - 12v DC, but that is still a dream. I haven't measured it with a voltmeter, but I guess it goes from 5v (or 7v) to 12v. This is not good enough for me. I want to shut the fan completely down when not gaming and only begin when I stress the system. I will have to figure out something different since the motherboard don't support 3pin control.

Can you explain in more detail what happened... I assume it was still spinning the fan at the minimum setting? Was it at a low/quiet rpm? Most 12v fans won't spin at 5v, they could spin at 7v though. So if that is the case, the minimum voltage by this adapter is probably around 7v, still enough to spin the fan.

Anyway you should be able to just buy a good 120mm PWM fan and get rid of that adapter. Make sure it is a true PWM 4 pin fan.
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
This adapter, well, how can I put it nicely, but still get the message through?
For the fan that runs at 2000rpm at max, this adapter brought it down to 1800rpm when I forced PWM signal to zero. I confirmed that the right signal was sent with an PWM fan and everything was according to my expectations for that. On my old(er) pc, I have this manual fan control with just a single knob to rotate that controls a 3 pin fan. Tried it with the 2000rpm fan, while it can't bring fans to zero that one, it took the rpm down to the 920rpm range.

If the adapter could scale from the 1000rpm to 2000rpm range, I could talk about it with a good taste in my mouth. Well, even for a range between 1300/1400rpm to 2000rpm would make sense, but for a range of 200 rpm, I really can't see the use anywhere for this adapter. It does something, I'll give it that, but that is also it.

On a more hopeful note, I have done some research and have bitten the bullet to buy a more expensive unit, in hopes that the NZXT Grid+ V2 fan controller will suit my needs. It is still in the mail, so haven't had the opportunity to play with it yet, but it has 0rpm mode, fan curve settings and connects to the internal usb header pins. AND supports full control of both 4pins and 3pins fans. I liked that, seems like a good solution and brings with it a opportunity that I will try since it can control more than one fan:
- A small 40mm fan on the VRM's as well. Damn that was hot when doing the old measurement trick by pressing my finger up against it while the system was under heavy load. No overclock or anything, but that was much hotter than anything I have felt on a computer so far. So it needs some work. I am thinking that when I kick in the fan on the top of the case, I can might as well kick in the 40mm fan at the same time since I then change lands from passive to active, so then noise is no longer a concern. My gaming needs/wishes are however then.

I am sure that the NZXT Grid+ V2 will have something unexpected about it that makes less than my ideal, like a "coil-whine" or something, but I will make a post about when I have gained some experience with it.

P.S
I need to use this Schythe Slip Stream 120 mm Slim because of clearance issues between the top and the internals in the case. I guess I could squeeze in the new Noctua slim fan with a tiny modification on the motherboard mounting cover, since that supports PWM, but a black fan wins in this build and spray-painting a Noctua fan that they have used years on perfecting the balance on the fins seems to me like not the right solution.
 
Last edited:

msystems

King of Cable Management
Apr 28, 2017
804
1,405
This adapter, well, how can I put it nicely, but still get the message through?
For the fan that runs at 2000rpm at max, this adapter brought it down to 1800rpm when I forced PWM signal to zero. I confirmed that the right signal was sent with an PWM fan and everything was according to my expectations for that. On my old(er) pc, I have this manual fan control with just a single knob to rotate that controls a 3 pin fan. Tried it with the 2000rpm fan, while it can't bring fans to zero that one, it took the rpm down to the 920rpm range.

If the adapter could scale from the 1000rpm to 2000rpm range, I could talk about it with a good taste in my mouth. Well, even for a range between 1300/1400rpm to 2000rpm would make sense, but for a range of 200 rpm, I really can't see the use anywhere for this adapter. It does something, I'll give it that, but that is also it.

On a more hopeful note, I have done some research and have bitten the bullet to buy a more expensive unit, in hopes that the NZXT Grid+ V2 fan controller will suit my needs. It is still in the mail, so haven't had the opportunity to play with it yet, but it has 0rpm mode, fan curve settings and connects to the internal usb header pins. AND supports full control of both 4pins and 3pins fans. I liked that, seems like a good solution and brings with it a opportunity that I will try since it can control more than one fan:
- A small 40mm fan on the VRM's as well. Damn that was hot when doing the old measurement trick by pressing my finger up against it while the system was under heavy load. No overclock or anything, but that was much hotter than anything I have felt on a computer so far. So it needs some work. I am thinking that when I kick in the fan on the top of the case, I can might as well kick in the 40mm fan at the same time since I then change lands from passive to active, so then noise is no longer a concern. My gaming needs/wishes are however then.

I am sure that the NZXT Grid+ V2 will have something unexpected about it that makes less than my ideal, like a "coil-whine" or something, but I will make a post about when I have gained some experience with it.

P.S
I need to use this Schythe Slip Stream 120 mm Slim because of clearance issues between the top and the internals in the case. I guess I could squeeze in the new Noctua slim fan with a tiny modification on the motherboard mounting cover, since that supports PWM, but a black fan wins in this build and spray-painting a Noctua fan that they have used years on perfecting the balance on the fins seems to me like not the right solution.

Thanks, you make it clear.

Look at this fan, it looks perfect for what you need:




http://www.prolimatech.com/en/products/detail.asp?id=2722


I even found a review, on SFF network:

https://smallformfactor.net/reviews/prolimatech-ultra-sleek-vortex-12-a-slim-competitor
 
  • Like
Reactions: rinoandrejohnsen

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
Wow, another solution! Wish I had researched myself to this one, but guess I got stuck in a one-track train of thought with the Schythe fans.

Under pro's for the fan you linked to: "A valid competitor to the Scythe Slim Silence, and superior when using on air coolers and behind fan filters."

Would have been perfect since I have not solved the problem with the fans losing massive effect under the mesh at top of the case. I will give this a serious thought if I can't find a solution for the effectiveness behind the filter.

Thanks!
 

rinoandrejohnsen

Chassis Packer
Original poster
Jul 23, 2017
18
19
Time to post a update on the mesh attached to the top.

I said in my first post that I would have to find a more graceful way to attach that mesh, so I have used a bit of time to think about it.
While this was boiling in the backside of my brain, I also found out that the fabric I had chosen didn't let enough air trough, so I had to find something different than wool. As I walked through a local curtain shop, a lady approached me and as it turn out, the knowledge from someone working with different fabrics every day was gold.

She told me to travel to a big fabric outlet and look for a plastic mesh that was commonly used to have as a inner curtain on bigger windows to limit the field of view a bit to create a more pleasant look.


It isn't perfect yet, my mounting mechanism, but by far outweighs my first attempt at this. I have been experimenting with a solder iron to create smooth seams. I haven't got good enough about it yet as the this plastic fabric melts fast, but I have made some good seams on a small piece of fabric. Hope I will be ready to some work on this plate after I get better at it.

I have used two layers. The first layer with the mesh rectangles orientated vertically and the other layer with them horizontally. I wasn't sold at the look of it first, but after tighten the layers around the plate, I realized that this could be good.


I must say that I am very happy with the look of this. Those screws on the edge are perfect for connecting the look of the case and the plate with the mesh. I believe that those screws in this scene can be described with what I believe is a technical term in psychology as a "transition object" between two different states. The lines on the case itself are contrasted very well against the complex forms of the plate. This black screws closest to the fan are what I am least happy about.






Now, down to numbers.

The performance of this fabric in terms of air flow is really good compared to the wool mesh I used first.

The old values without a mesh:
  • 2000 rpm fan (pulling air out) 74 degree Celsius
  • 2000 rpm fan (pushing air in) 75 degree Celsius
  • 1200 rpm fan (pulling air out) 76 degree Celsius
  • 1200 rpm fan (pushing air in) 77 degree Celsius
  • Reference (no fan) 82 degree Celsius
The new value with the mesh:
  • 2000 rpm fan (pulling air out, with mesh) 75 degree Celsius
I got this value by doing the same test as for the old values. I am very happy with the results for this. I can also feel the air be pushed out from the top with strength and when I place my hand on the side plates and bottom, I can feel air passing my fingers into the case.

Fun fact, the plates are made in such a way that you can see light pass through when holding it up against a strong light in fairly dark room, so air will pass through there as well when the fan is pulling in air which is really good for the temperature. Didn't know that at the start of this project.
 
Last edited: