Stalled NCASE Project Sidearm

[Sean Crees]

I never really understood the use of this design as a good aesthetic choice. I can see why it works in the DAN case, to lower overall volume. I just don't know how many people are going to have their computers positioned in such a way that they will be able to see both sides of the case. I feel like if aesthetics are your primary focus, that both GPU and CPU should be visible from the same perspective, be it a side, top, or front.

 

[Necere]

I never really understood the use of this design as a good aesthetic choice. I can see why it works in the DAN case, to lower overall volume. I just don't know how many people are going to have their computers positioned in such a way that they will be able to see both sides of the case. I feel like if aesthetics are your primary focus, that both GPU and CPU should be visible from the same perspective, be it a side, top, or front.

Again, the right side of the case would not have a window (and would only show the back of the GPU if it did), so not being able to see both sides simultaneously isn't a valid criticism.

The GPU being obscured is valid, but that's a concession to other priorities, as I've explained above. I suppose the question should be, do you agree with those priorities or not? This design is really centered around the airflow concept, with the fans feeding fresh air directly to the key components, while still retaining the convenience of dust filtration and a relatively small volume.

 

[jeshikat]

As for SFX support, it's something I'm still considering. Oriented lengthwise in the case, it would preclude 3.5" HDD support and make running the tubing for watercooling trickier. I thought about supporting a cross-wise orientation as well, but that obstructs longer GPUs and would require lengthening the case by 10mm just to support 10.5" long cards:

Being able to sacrifice water cooling and 3.5" support in order to ditch the FlexATX for SFX/SFX-L seems almost necessary to me. Only having the option for FlexATX seems like it'll be too much of an uphill battle in marketing.

And there's no getting around physics, a 40mm at high speed will always be loud. So even if FlexATX really improves in idle and low-medium load noise with the upcoming models, noise at full load will still be a problem.

There's a structural beam running the length of the case at the top, which the handle bolts into, that could also serve as a conduit for the power cable:

Maybe add some cutouts in the flanges on the side of that beam for zip-ties?

At the moment I've got one of these luggage handles in as a placeholder. They look a bit cheap, and I'm not too sure about durability, but I like that it folds down pretty flat.

Hopefully you can finder a nicer solution, because those look cheap and don't match the otherwise beautiful design.

A machined handle might be nice, but it wouldn't be as low profile and the cost would be higher.

Yup, that's why I was thinking of the machined handle as an option, not the default.

 

[Necere]

And there's no getting around physics, a 40mm at high speed will always be loud. So even if FlexATX really improves in idle and low-medium load noise with the upcoming models, noise at full load will still be a problem.

Well, will it though? Consider a hypothetical 650W Platinum rated flexATX unit, running a load of, say, 300W. That's around, or even above, what a system running a GTX1080 and 7700K draws, right? So >90% efficient at <50% load, dissipating maybe 30W... would the fan really need to be spinning that fast? That's the question.

Maybe add some cutouts in the flanges on the side of that beam for zip-ties?

Yeah, that was the idea.

 

[jeshikat]

That's a good point, 30W waste heat isn't all that much.

Even so, you'll probably hear nothing but "FlexATX is a server PSU form factor, you'll go deaf with one of those in your PC!" if the case only supports it

 

[Necere]

Yeah, even if it were quiet under load, we'd definitely have our work cut out for us trying to convince people of it.

 

[K888D]

I love the case design and the thought you are putting into the airflow efficiency.

I agree with what you're saying here, but would the panel on the left layout not be vented? Because if it is that will play a bigger roll than fan clearance alone. But if the panel is not ventilated then I can see why you are going the route of layout on the right.

For example:

• 140mm fans need 35mm clearance
• 120mm fans need 30mm clearance
• 92mm fans need 23mm clearance

Its funny you should post this now, because I've just finished up writing an article this evening about this very subject! Article is HERE.

My investigations tie up pretty close with what your theories and conclusions.

Here is some of the data from the experiments, this first chart shows Wind Speed (km/h) of a few different fan sizes plotted at different distances from a flat surface on the intake side:

This chart shows the fan efficiency in percentage terms at a given distance from a flat surface:

 

[Necere]

I agree with what you're saying here, but would the panel on the left layout not be vented? Because if it is that will play a bigger roll than fan clearance alone. But if the panel is not ventilated then I can see why you are going the route of layout on the right.

No, the side panels are solid. That's one of the major points of the design; ventilating the side panels would defeat the positive pressure/forced airflow scheme. If you don't care about dust control, then w360's design (which draws air in through the sides and exhausts out the top, and is therefore negative pressure) is just as good or better in terms of cooling.

Its funny you should post this now, because I've just finished up writing an article this evening about this very subject! Article is HERE.

My investigations tie up pretty close with what your theories and conclusions.

Here is some of the data from the experiments, this first chart shows Wind Speed (km/h) of a few different fan sizes plotted at different distances from a flat surface on the intake side:

This chart shows the fan efficiency in percentage terms at a given distance from a flat surface:

Great, real data! I'm actually surprised though at just how much airflow drops off, even as far out as 50mm+. It's a wonder those M1 builds with the bottom rads/slim fans pulling through 10mm of ground clearance accomplish anything at all.

Going by your chart, the fans in my design would be running at about 50% of their max airflow (ignoring the effects of other restriction). That's pretty significant, but perhaps only just puts it on par with designs that have the fans drawing through restrictive vents instead, like Lian Li's 3mm holes on the M1 and A4 (~40% open area pattern).

 

[alamilla]

I've been keeping tabs on the this on [H].
Great to see you bringing this over to SFF.

Love the design - particularly support for water-cooling.
As many have expressed, my only concern is that of PSU support as I would prefer SFX/SFX-L.
Alternatively, you could go the HDPLEX 300W route!

 

[Kwirek]

"Data & Science"

I'm just curious, would it be possible for you to test say one fan mounted with the same early distances (say up to 8 cm since the plots plane out after that) but with intake from the table? That would allow for a clear intake of air from all sides (instead of 3 as it is now). I don't really think it will change anything, but it could strengthen the data you've gathered.

 

[K888D]

I'm just curious, would it be possible for you to test say one fan mounted with the same early distances (say up to 8 cm since the plots plane out after that) but with intake from the table? That would allow for a clear intake of air from all sides (instead of 3 as it is now). I don't really think it will change anything, but it could strengthen the data you've gathered.

I see what your saying about the 3 sides may be skewing the results slightly. If I get chance I'll see what I can do.

One thing I did notice during the testing is that vent design and restriction has an affect on the direction of the air being pushed out the fan, especially perforated style vents including dust filters.

Fans generate most of their pressure around the outer most edge of the fan, so what you get is almost a ring of air being pushed out with a dead zone in the middle of the ring next to the fan. As this air gets further away from the fan it spreads out and merges into the breeze you can feel with your hand for example.

But, what I found is that if you restrict the fan with a perforated style restriction the air is pushed out the fan sideways at something like a 60-70 degree angle around the fans circumference, there is a massive dead zone in the middle of the fan where you can actually put your fist in front of the 140mm fan and not feel any air blowing as its passing around your fist. I am going to put some diagrams together to try and explain better in the next part of the article.

 

[Phuncz]

Amazing work @Necere and @K888D with the research, valuable information for anyone trying to understand more about SFF design !

The handle does look good in its resting position but I'm worried it's not going to be the same experience the case offers in quality. Unless it's really well made. I like that FlexATX is choosen because it is a better fit for this project but indeed, an SFX fall-back will be something to seriously consider. I remember at the inception of the M1 the flak SFX received over ATX. While I expect it would be the opposite now. But I expect the same issue with FlexATX: unknown and unloved, now. But that doesn't have to mean you have to listen, I'm still glad you made SFX the primary PSU for the M1, probably the entire SFF community is thankful for this.

 

[3lfk1ng]

Ok, even short rubber feet would make sense.

 

[EdZ]

So for this style of layout, where the GPU and motherboard are sandwiched together, putting the GPU behind the motherboard actually makes better use of space than the reverse. That's partly because we have a large area where the GPU extends past the motherboard, but also because we can use the space that's required, but otherwise unused, for the motherboards standoffs/backside clearance:

As is plainly apparent, the layout on the right provides considerably more intake clearance for the GPU fans. Now, depending on what drives are mounted or whether we end up putting in an SFX PSU, that area towards the front may be considerably reduced, but it's still going to be better than with the GPU facing the side of the case.

One other layout would be a 'facing opposing' layout, with the backsides of both the motherboard and GPU attached to opposing case walls, and with both sharing a single 'plenum chamber'. Maintenance is complicated in that removing the case side means having the GPU temporarily retained by the flex riser and power cables (though once detached you have more room to work inside), but fan restriction is minimised and a short flex riser can be used.

But, what I found is that if you restrict the fan with a perforated style restriction the air is pushed out the fan sideways at something like a 60-70 degree angle around the fans circumference, there is a massive dead zone in the middle of the fan where you can actually put your fist in front of the 140mm fan and not feel any air blowing as its passing around your fist. I am going to put some diagrams together to try and explain better in the next part of the article.

I believe the fine-pitch exit stators on Silverstone's Air Penetrator line were intended to counteract this effect. Sadly there are no 92mm or thin 120mm variants. A flow collimator could be retrofitted to existing fans though, either of the stator design like in the Air Penetrators, or a more common packed-tube design (think taking a bundle of aligned drinking starws, then cutting them to the desired length).

 

[Pat-Roner]

Nice! Been waiting for something new from Ncase.
Would love if there where room to put sound dampening on the solid panels, since then you can "contain" the sound in the middle of the case. Would also be awesome if it where possible to orient the mobo and gpu towards fade each other (backside out). Would work poorly with windowed panels though...

@Necere Have you thought about TFX psu? Seasonic has a pretty nice one.

 

[Necere]

One other layout would be a 'facing opposing' layout, with the backsides of both the motherboard and GPU attached to opposing case walls, and with both sharing a single 'plenum chamber'. Maintenance is complicated in that removing the case side means having the GPU temporarily retained by the flex riser and power cables (though once detached you have more room to work inside), but fan restriction is minimised and a short flex riser can be used.

Would also be awesome if it where possible to orient the mobo and gpu towards fade each other (backside out). Would work poorly with windowed panels though...

A face to face layout actually requires a much longer riser:

I'd be concerned if the fans are too close that it will result in extra noise from turbulence. Also this layout precludes a window.

Would love if there where room to put sound dampening on the solid panels, since then you can "contain" the sound in the middle of the case.

I'm not sure how much value foam would have, since the top of the case has open vents. You could install it on the CPU side though, at the cost of some fan intake clearance for the CPU cooler

@Necere Have you thought about TFX psu? Seasonic has a pretty nice one.

Unfortunately TFX is a bit too long for this case (it's 25mm longer than flexATX), as well as thicker, which gets in the way of the radiator ports:

Also the Seasonic is only 350W. I feel like TFX is the neglected middle child here, that will likely get passed over in favor of its beefier bigger brother, SFX, and its lean younger sibling, flexATX. TFX just doesn't have a lot going for it in comparison.

 

[Arboreal]

I would be keen on a strong handle, not sure how strong the one you've shown is.
For my last 2 LAN builds I have used a re-purposed Panasonic Toughbook removeable alloy handle:

It is pretty light, but a little bulky for the Sidearm - BTW, what happened to the 'D' at the end that's on it's name on [H]?
A metal handle that drops flat when not being carried would be really neat, so long as there is room in the centre void for it to drop into.

 

[Necere]

I would be keen on a strong handle, not sure how strong the one you've shown is.
For my last 2 LAN builds I have used a re-purposed Panasonic Toughbook removeable alloy handle:

It is pretty light, but a little bulky for the Sidearm - BTW, what happened to the 'D' at the end that's on it's name on [H]?
A metal handle that drops flat when not being carried would be really neat, so long as there is room in the centre void for it to drop into.

Well, the handle is designed for luggage, so I would expect it to be good for at least 20-30lbs. It has an internal flexible steel band, which isn't likely to fail. But yeah, it remains to be seen what we can do for the handle. We may get one custom made that's a a bit higher quality.

"SideArm" was the original name I came up with years ago for a much earlier iteration (which you can see in the [H] thread), long before even the M1 project started. W360 added the "D," I guess because he was able to buy the domain for it. He went ahead and posted the thread without consulting with me first, so there you go. Personally I think it's a bit awkward with the D.

 

[Arboreal]

Well, the handle is designed for luggage, so I would expect it to be good for at least 20-30lbs. It has an internal flexible steel band, which isn't likely to fail. But yeah, it remains to be seen what we can do for the handle. We may get one custom made that's a a bit higher quality.

Fair enough, you had me at the steel band .
Whatever else you do with the design, I really like that it is braced on something structural, not the case roof like I did.

"SideArm" was the original name I came up with years ago for a much earlier iteration (which you can see in the [H] thread), long before even the M1 project started. W360 added the "D," I guess because he was able to buy the domain for it. He went ahead and posted the thread without consulting with me first, so there you go. Personally I think it's a bit awkward with the D.

Interesting, I hadn't seen the early stuff; I lurk on [H], but have never participated.
A SFF portable case is as sweet by any nameD!
Really great to see NCase in action again, especially as one of your other ideas for a 'flat' / Steam case seems to have been put into action all over the place.
Nice bit of development in a slightly different direction. Again, I like the shape better than the shoeboxes in the 10-12L size (says he with an SG05 and SG06 here...).
I'm watching with interest, envy at the design and creation skills, and trying not to drool.
I'm hoping the £ will recover some time, as kit from the US is looking waaaaaay expensive in this lat few months.
Bring on the sandblasted aluminium, and I'll be a dribbling wreck!

 

[Necere]

One advantage of this layout is it's able to fit some of the largest oversized video cards on the market:

This is in a relatively narrow case at 135mm, whereas a conventional layout where the card inserted directly into the PCIe slot on the motherboard would require a minimum case width of 175mm+ for the same height compatibility.