HTPC upgrade? New mobo + CPU + RAM only

[Stereodude]

SD, are you messing with madVR with Kodi or still just MPC?

Just MPC-HC. I thought I read the developer of DSPlayer was taking a break which leaves using madVR in Kodi in limbo.

 

[Mercutio]

It's several versions out of date, which is one of the reasons I asked.

 

[Stereodude]

Well, I did some testing with different fans in different locations. I can only say that PC cooling is not even close to intuitive. :dunno:

There are spots for two 140mm fans in the front of the 600Q. There's one spot in the back on the bottom for a fan. The best GPU temps are achieved with one fan on the top in front blowing into the case and one in the back (bottom) blowing out. Adding a 2nd front fan so there is both a top and bottom fan in the front blowing into the case in addition to the single lower rear fan blowing out yields higher GPU temps than a single fan front fan blowing in + lower rear fan blowing out. Two fans in the front blowing into the case and none in the back runs the GPU hotter yet. FWIW, SPCR found the same behavior. Specifically that adding a 2nd front fan blowing into the case hurt GPU temps.

Where it starts to get really weird is that there seems to be a sweet spot for the amount of airflow moved out of the case by the lower rear fan. No fan in the rear blowing out = highest temps, but Fan A spinning at say 600RPM gives better GPU temps than a Fan B spinning at 950RPM in the lower rear spot, despite the fact that Fan B moves more air. The grill in the case the fan blows through is pretty open so I don't really see this as an issue of static pressure. I'm waiting on a little more hardware to arrive so I can test this better. Fan B lacks PWM control so I can't currently slow it down to the same RPM as fan A, and fan A is getting the same PWM signal as the CPU fan, so I can't speed it up in isolation. Right now I'm :scratch:

 

[Handruin]

Well, I did some testing with different fans in different locations. I can only say that PC cooling is not even close to intuitive. :dunno:

There are spots for two 140mm fans in the front of the 600Q. There's one spot in the back on the bottom for a fan. The best GPU temps are achieved with one fan on the top in front blowing into the case and one in the back (bottom) blowing out. Adding a 2nd front fan so there is both a top and bottom fan in the front blowing into the case in addition to the single lower rear fan blowing out yields higher GPU temps than a single fan front fan blowing in + lower rear fan blowing out. Two fans in the front blowing into the case and none in the back runs the GPU hotter yet. FWIW, SPCR found the same behavior. Specifically that adding a 2nd front fan blowing into the case hurt GPU temps.

Where it starts to get really weird is that there seems to be a sweet spot for the amount of airflow moved out of the case by the lower rear fan. No fan in the rear blowing out = highest temps, but Fan A spinning at say 600RPM gives better GPU temps than a Fan B spinning at 950RPM in the lower rear spot, despite the fact that Fan B moves more air. The grill in the case the fan blows through is pretty open so I don't really see this as an issue of static pressure. I'm waiting on a little more hardware to arrive so I can test this better. Fan B lacks PWM control so I can't currently slow it down to the same RPM as fan A, and fan A is getting the same PWM signal as the CPU fan, so I can't speed it up in isolation. Right now I'm :scratch:

You might be able to explain some of the behavior if you wanted to get creative by burning a simple incense stick (or some harmless equivalent) and watching the actual airflow pattern. Maybe there are hot pockets or leaks causing an imbalance you're not expecting?

 

[Stereodude]

You might be able to explain some of the behavior if you wanted to get creative by burning a simple incense stick (or some harmless equivalent) and watching the actual airflow pattern. Maybe there are hot pockets or leaks causing an imbalance you're not expecting?

That might be a viable option if I had a clear side panel. Running it with the side open so I could see the smoke would totally change the results.

 

[Handruin]

That might be a viable option if I had a clear side panel. Running it with the side open so I could see the smoke would totally change the results.

Code the side with cellophane? :-D

 

[Stereodude]

Code the side with cellophane? :-D

Should I do that in C, Java, or Pascal?

 

[LunarMist]

You don't have a fogger for smoke studies? :lol:

 

[Stereodude]

You don't have a fogger for smoke studies? :lol:

No, I don't have a wind tunnel either.

 

[Handruin]

Should I do that in C, Java, or Pascal?

Definitely not Pascal. I'd go for Java.

Note sure why I wrote "code". I think I was distracted with the code I was working on when I wrote my reply. :scratch:

 

[mubs]

Well, I did some testing with different fans in different locations. I can only say that PC cooling is not even close to intuitive. :dunno:
Right now I'm :scratch:

Damn, this is insightful. Keep us posted, SD, I'll be waiting eagerly.

 

[Howell]

Don't forget about the PS fan as an exhaust port. Is the placement of the two front fans such that air could go from a front to back pair on not cross the area with the gpu?

 

[Stereodude]

Don't forget about the PS fan as an exhaust port. Is the placement of the two front fans such that air could go from a front to back pair on not cross the area with the gpu?

The backplate of the video card is just above the bottom of the top fan, so... The top one, no. The bottom one, probably.

 

[Stereodude]

So I got my 3-pin fan speed controller yesterday along with a few more 140mm fans. I did a little more testing late last night. With a different fan blowing into the case in the top front position I found no appreciable difference in the GPU temperature with the lower rear exhaust fan running at the minimum speed the controller allows, the maximum speed the controller allows, and more of a middle speed. :scratch:

 

[ddrueding]

To me that indicates that the fan is mainly preventing air from flowing the opposite direction. Try just blocking off that part of the case and see what happens.

 

[Stereodude]

To me that indicates that the fan is mainly preventing air from flowing the opposite direction. Try just blocking off that part of the case and see what happens.

I swapped the fans and did a little more testing. Some of the fans seem to have much more focused airflow. When using the fan with the most focused airflow up top in the front blowing in the GPU temps barely seem affected by the speed of the lower exhaust fan (no more than a 1C swing from minimum speed to maximum speed of the lower rear exhaust fan). When using one of the fans with the more diffuse airflow in the front up top blowing in the speed of the lower exhaust fan does seem to matter much more (like 3C swing from min to max speed). I attempt to reproduce the results I observed earlier and was unable. The GPU temps were lowest when the lower rear exhaust fan was spinning the fastest. Slowing it down did not lower GPU temps.

Also, despite having the air conditioning on in the house, I'm beginning to think the ambient air temperature by the system is changing enough to impact the changes I'm seeing in the data I'm collecting. I'm now running an oscillating household fan to circulate air around the area of the house where the system to try to keep the corner where the system is from warming up and impacting the data. The fan doesn't blow on the system though.

 

[time]

You need to get reasonable airflow across the GPU subsystem, but that case is making it nearly impossible. Your only useful exhaust is the PSU - the holes in the back above the video card are probably a thermal short circuit to the PSU and not helping your cause. The lower exhaust fan will help the CPU a little, but as you discovered, can easily divert air from the marginal low pressure area created by the high-efficiency PSU with a fan that spends most of its time idling.

Fans need to be close to where you need to create cross-flow. It's extra difficult when you're running the fans slow to keep noise down, and especially when the case is over-sized.

There's no point worrying about case pressure unless you can seal up the hundreds of holes that the case manufacturer has helpfully punched into it.

You could try positioning the third fan as an intake on the bottom towards the front. It *might* encourage circulation around the video card, but it's a long shot.

Any chance you could squeeze an 80mm fan onto the back near the top?

 

[ddrueding]

The advantage of oversized cases is that you can position fans inside the chassis right next to the components that you care about. I hope to be revising my personal rig today (if the kid lets me). I'll post some pics when I can.

 

[Stereodude]

You need to get reasonable airflow across the GPU subsystem, but that case is making it nearly impossible.

Uh... This is one of the best quiet cases for GPU thermals out there per SPCR. It was the best they tested that had optical drive bays, which I need. It's measurably better than the Antec P180 I started with and the GPU doesn't bake the CPU. I'm keeping the GPU under the throttling threshold of 82C while keeping the system virtually inaudible. That was the goal. I'm just doing a final bit of tweaking to find the best fans for each spot to keep it as cool as possible while being virtually inaudible.

It's not really oversized IMO. It's wide, but it's not particularly deep or tall. Wide is good because I have a rather tall CPU heatsink. The Asus Strix GTX 1080 is quite long. There's about 3" between the front of the 140mm case fan and the 1080.

Ultimately, I got to where I needed to be. It was more challenging than I expected, but it's suitably quiet. It's not quite inaudible. With the TV off, the receiver and HT system off, the fridge not running, and the HVAC off you can tell that it's on if you know what to listen for. Mainly that the background noise floor in the room is slightly higher with it running, but there's no real pitch or distinguishing noises emanating from it.

 

[time]

Uh... SPCR rated this case third out of the group of 6 it chose to compare it to. In terms of GPU thermals, it came fifth. Four of the six support an optical drive, including the runaway leader, the Silverstone FT05.

If you look further into SPCR tests, you would find the Fractal Design Core 500, also with optical drive support and superior thermal but fairly comparable acoustic performance.

If you want quiet, check out the SPCR review of the BitFenix Prodigy. All of these tests use the same video card.

I'm glad you're now happy with your setup and I hope you get great enjoyment from it.

 

[Stereodude]

So you recommend me two mini-ITX cases that won't hold my gear. Perfect. Lets ignore that for a minute though. The Fractal Design Core 500 was slightly noisier, the SSHD and MB were considerably hotter, the GPU was 1C hotter, but spun the fan slower. The BitFenix Prodigy was slightly quieter, the MB was considerably hotter, the GPU was hotter by 5C, but spun it's fan slower. It's impossible to say how they would compare if the GPU target temp was 80C.

I'm also not sure what you're talking about with the SPCR results... The only two cases that best it in that review with respect to noise and other thermals since the GPU temp is held constant are the SilverStone FT05 and Antec S10. I'm real skeptical that 1120 vs 1170 RPM has any audible difference, especially considering the system was quieter, so I don't agree that it finished in 5th. Contrary to your claims, the FT05 has no 5.25" bays. The Antec S10 is outrageously expensive and also has no 5.25" drive bays. As reviewed the 600Q had less than stellar fans and still did quite well. I've replaced them for better ones.

Based on the reviews and my requirements, I don't think I made some sort of insane idiotic choice like you're trying to suggest. Was it the absolute best if I was starting with a clean slate? Probably not, but I didn't have a clean slate.

 

[Clocker]

Sounds like this was a fun project and a learning experience!

I know this feedback is too late to make any difference but just wanted to chime in that for my 'server' that I keep in the basement, I use a Coolermaster HAF 912. The 2x 200mm fans provided a lot of air flow at a very low sound level. IIRC, I slowed them down just a bit with an inline resistor.

The case is also deep enough so that the side panel fan could be spaced away from the mesh opening in the side panel with a thin hollowed out case fan frame (fan removed) that I used as a shroud. Use of the shroud made the air flow almost inaudible even at a close distance and allowed outside air to be directly fed to the GTX970 video card I was Folding with at the time. If you are looking to quiet the system down a bit more or experiment with them, I suggest you give creating some shrouds (or buying them) a try. Getting the fan setup so it is not close to the mesh opening of the case really makes a difference sometimes (depending on the mesh etc.).

The case is ugly though! Would not want it out in plain sight. For $49 shipped free I could not pass it up though.

 

[time]

My post was in response to your snark. Everything in it is accurate, including the fact that the FT05 supports an internal optical drive.

The MiniITX case examples were an attempt to broaden your perspective, in response to "This is one of the best quiet cases for GPU thermals out there per SPCR". The Core 500 shows what an out-of-the-box solution can achieve when there's some good engineering behind it. If the default setup already delivers a relaxed GPU, there's scope to change the case fans for quieter models and otherwise tune the cooling/acoustical compromise.

 

[Stereodude]

My post was in response to your snark.

What did you think you were going to get when you drop in and try to tell me I'm doing it all wrong and am using the wrong case for the job? I didn't start a thread looking for blank slate case ideas, but I started one just for you. Here you go.

Everything in it is accurate, including the fact that the FT05 supports an internal optical drive.

Accurate, but totally non helpful. It still doesn't have any 5.25" drive bays to support the optical drives I already have.

The MiniITX case examples were an attempt to broaden your perspective, in response to "This is one of the best quiet cases for GPU thermals out there per SPCR". The Core 500 shows what an out-of-the-box solution can achieve when there's some good engineering behind it. If the default setup already delivers a relaxed GPU, there's scope to change the case fans for quieter models and otherwise tune the cooling/acoustical compromise.

And they still won't hold my motherboard, cooler, or the two 5.25" half height optical drives I have. The comment you're highlighting has an implied "that holds my gear" attached to it, but you're trying to treat it as an absolute statement when it's clearly made in the context of other posts and the thread as a whole.

 

[Mercutio]

Do you have any room in your chassis to experiment with ducts for managing airflow? I have no idea what it looks like in your new baby but a cardboard shroud doesn't seem like an unreasonable step in experimentation.

 

[Stereodude]

Do you have any room in your chassis to experiment with ducts for managing airflow? I have no idea what it looks like in your new baby but a cardboard shroud doesn't seem like an unreasonable step in experimentation.

That's a good idea I hadn't really considered before. Clocker suggested the same. Yes, there is some room for that in the case. I'm not entirely sure where to put it for the lower rear exhaust fan, but I have some ideas for the top front one.

 

[snowhiker]

Do you have any room in your chassis to experiment with ducts for managing airflow?

I was going to suggest this as well but after Time's response above, and you reply to it, I was getting my popcorn ready! LOL

 

[Clocker]

https://www.youtube.com/watch?v=MYbG-uv4Op4

Used to look at the stuff on this site for hours:

http://www.frozencpu.com/cat/l2/g30...ooling-PC_Water_Cooling_Radiator_Shrouds.html

http://www.frozencpu.com/cat/l3/g47...essories-Fan_Accessories-Fan_Ducts-Page1.html

 

[Howell]

Do you have any room in your chassis to experiment with ducts for managing airflow? I have no idea what it looks like in your new baby but a cardboard shroud doesn't seem like an unreasonable step in experimentation.

That's also what I was thinking with my last comment. Instead of cardboard, plastic wrap might be able to give a sense of pressure differential if it is not too small. Plastic wrap might be able to temporarily replace the side cover for a smoke test.

 

[Howell]

That's a good idea I hadn't really considered before. Clocker suggested the same. Yes, there is some room for that in the case. I'm not entirely sure where to put it for the lower rear exhaust fan, but I have some ideas for the top front one.

You could start with just a simple case divider to isolate the influences.

 

[Clocker]

You can get a big enough sheet of clear acrylic for $10 at Lowes. http://www.lowes.com/pd/DURAPLEX-0-08-in-x-18-in-x-24-in-Clear-Acrylic-Sheet/3143519