Today marks the launch of NZXT’s newest case, the mini-ITX Manta. NZXT’s design approach to the Manta immediately draws attention for its rounded, Ferrari-inspired aesthetic and flashy-but-minimalistic layout. The enclosure fronts an all-steel exterior for the paneling – a hardened shell – and stands as a fusion of NZXT’s “old” and “new” case design philosophy.
The enclosure includes 2x 120mm front intake fans and 1x 120mm rear exhaust fan in its default configuration, but allows for up to 2x 140/120mm top fans or an upgrade to 2x 140mm front fans. Radiator positions are also open to user preference: Closed-loop liquid cooler (CLC) radiators up to 280mm can be installed in the top, the front, and 120mm radiators can be installed in the rear of the case. This opens the field of play considerably – radiator placement and case fan orientation (exhaust vs. intake, top vs. front, and so forth) will seriously impact CPU and GPU thermals. It’s important to look at your system configuration and implement an optimal setup that takes CPU and GPU thermals into account. That’s what this guide is for.
In my separate review of the Manta at GamersNexus, we published the most detailed cooling benchmarks the site has ever conducted on a PC case. We wanted to share this data with NZXT’s buyers, who can hopefully deploy the Manta fan and radiator positional testing for improved cooling in the ITX box. Let this benchmark serve as a guide for optimal cooling in the NZXT Manta and, if you’re a buyer who also hopes to use a CLC, be sure to pay careful attention to our radiator position results.
And just for clarity, before going to deep, “we” or “our” will refer to my team at GN. NZXT asked us to help with testing and data presentation for this Manta benchmark.
The full test methodology is available in the GamersNexus review and was used throughout this cooling optimization test. We’ll recap some of the most critical aspects here:
Thermal tests were conducted using an Intel Core i7-6700K Skylake CPU and Z170 mITX platform, parts listed below. A 120mm liquid cooler was used for the CPU. The GPU (GTX 980 Ti) was using an NVIDIA reference design cooler (blower fan), which will yield different results than aftermarket coolers that may use single- or dual-push fan setups.
Ambient was carefully measured using thermistors (air temp) and lasers (table temp). Ambient was left to run at a constant 21C. Minor fluctuations in ambient were accounted for by outputting a Delta T over Ambient value for all of our tests (including the below charts), which means that we’re subtracting ambient from the thermal diode readings to produce a delta value. If you’re curious what the temperatures would be in your home – although plenty of environmental factors can affect this data, too – add your ambient (room) temperature back into the values. For example, 40C becomes ~61C in a home that is kept at around 70-71F.
Case fans are manually configured to maximum speed using a custom BIOS profile. The liquid cooling pump for the CPU is placed on a custom fan curve for realistic workload increments that are reproduce-able in tests.
The parts used are as follows:
|CPU||Intel Core i7-6700K (Stock)|
|Memory||16GB ADATA XPG DDR4-2400|
|Motherboard||MSI Z170I Gaming Pro AC|
|Video Card||NVIDIA GTX 980 Ti Reference|
|Cooler||120mm Closed-Loop Liquid Cooler
|PSU||750W 80 Plus Gold Active PSU|
The NZXT Manta case testing was conducted based upon a matrix of “go-to” configurations. The below table shows what our technicians referenced when installing the fans and radiator in various locations in the case, and represents what we believe to be common, go-to cooling setups from PC DIY enthusiasts.
|Front||Top BACK||Top FRONT||Rear||Total Fan #|
|Stock – +1 Rear Rad||2x 120 intake||1x 120 exhaust||0||1x 120 W/RAD
|Stock – Rear Replace||2x 120 intake||0||0||1x 120 W/RAD
|Stock – +1 Top Rad||2x 120 intake||1x 120 W/RAD
|0||1x 120 exhaust||4|
|Stock – Rear Replace
Raised Front Fans
|2x 120 intake||0||0||1x 120 W/RAD
|Front Radiator||1x 120 intake
1x 120 W/RAD (front-top)
|1x 120 exhaust||0||1x 120 exhaust||4|
|Add 2x 140mm
|1x 120 intake
1x 120 W/RAD (front-top)
|1x 140mm exhaust||1x 140mm exhaust||1x 120 exhaust||5|
|Add 1x 140mm
|1x 120 intake
1x 120 W/RAD (front-top)
|0||1x 140mm exhaust||1x 120 exhaust||4|
|Add 2x 140mm
|2x 120 intake||1x 140mm intake||1x 140mm intake||1x 120 W/RAD
We’ve got two primary modes of data presentation for this test. “Equilibrium” charts show thermals once the CPU and GPU have reached their stable, peak temperature and have ceased fluctuating / rising. “Thermals over Time” charts show the rise and fall of thermals as the system balances its internal temperature. Because our tests are automated, all the data points start and end at the exact same time. This means that we get a clean “spike” in the Thermal over Time charts, showing precisely when load testing started (following a two-minute idle temperature logging period).
There’s a huge amount of data to convey for each configuration, which makes naming things a bit messy. The format is: +Fan Count (how many fans were added? what size?), [E] or [I] for exhaust or intake, and then Rear/Front radiator mounting. The final item is a parenthetical count of total system fans (3F, 4F, 5F).
Thermals Over Time
Pasted from my review, here’s what we learn from these charts:
“The above charts show thermals over time – a bit harder to read than the equilibrium charts, but useful. It’s apparent that the ‘green’ configuration (add two 140mm top fans as exhaust) is the worst possible configuration for GPU cooling. This makes sense, too, since those fans are effectively siphoning cool air away from the internal components faster than it can cool them. This same ‘green’ configuration works reasonably for the CPU because of the front-mounted radiator, a configuration that dominates CPU performance charts for the NZXT Manta. This configuration pushes a bit more warm air into the case, but proper fan configuration will ensure the heated intake doesn’t negatively impact the GPU. This configuration is not ‘proper,’ as the triple-exhaust setup ensures all the air gets pulled ever upward and out – not ideal for a blower-style GPU fan.
The lighter blue configuration isn’t ideal, either, and it’s for the same reason: An exhaust fan positioned in the top-rear is pulling air away from other components. In this configuration, we also had the radiator rear-mounted, and so a top-rear exhaust fan is siphoning cool air away from the CLC before it ever reaches the radiator. In this way, the top-rear fan is parasitic in nature and detrimental to overall cooling performance. If configuring the system with a rear radiator, we can definitively state that a top-rear exhaust fan is a configuration that is best avoided.
An ideal configuration involves a front-mounted radiator for optimal CPU cooling, but it needs to be mounted in a way that doesn’t negatively influence GPU thermals. We recommend installing the front radiator in the front-top location, then leaving the rear 120mm exhaust fan where it was pre-installed. This ensures a clear path-to-exit for the warmed CPU air, but doesn’t siphon the bottom intake fan’s air from the GPU.”
Let’s move on to some more legible charts.
Averaged Peak Thermals
Also pasted from the review:
“The above charts represent thermals at equilibrium, or once temperatures have ceased their rising pattern and hit a stable number. We’ve got idle and load metrics available, though the various configurations are generally inconsequential to idle performance (but not always).
This data set is the same as what the thermals-over-time metrics used, just visualized in a different way.
For CPU temperatures, again, the triple-exhaust configuration produces the lowest readings at 34C, though idle is noticeably higher at 15.2C. It does not necessarily follow that the coolest load configuration will produce the coolest idle configuration – this is something we’ve seen in a few other cases, too. Correlating the triple-exhaust setup to the GPU chart, it becomes apparent that this is, in fact, one of the worst configurations possible. Pushing the GPU into the 71.4C delta T over ambient range means that, factoring-in an assumed ambient of about 20C, we’ve now got a GPU that is thermally throttling at 90C (absolute temperature). That’s one’s no good, then.
Conversely, the GPU is coolest with the 3-fan stock setup on the contingency that the radiator is not front-mounted. The 63.2C delta T value is achievable because of two factors: (1) The front intake is unobstructed and does not pass through a radiator; (2) the exhaust is a single-fan setup and not overpowering the blower fan’s ability to cool itself. This setup, though, is the worst for the CPU by several degrees.
Strictly looking at the charts, the best mix of CPU & GPU thermals would appear to be produced by the dual 140mm top intake. We’ve seen this in cases before. The only reason the effect isn’t more exaggerated is because of the Manta’s solid top panel, which aids in reducing noise emissions, but does trade-off some intake potential (exhaust is less of a concern – it’s easier to get rid of air). Unfortunately, this five-fan setup isn’t the best value and pulls only slightly ahead of the next configuration’s thermal readings.
The best setup is to run +1x120mm top-rear fan and a front radiator. This keeps the GPU temperatures at a superb (for the reference cooler) 63.3C, and only raises the CPU to 37.08C. Not bad at all for either, when considering the case’s focus on silence and discretion. Since most radiators include a fan, it’s likely that CLC users will be running four fans anyway.”
We also tested a top-mounted radiator configuration this morning (just prior to review publication), but haven’t processed the data into the GamersNexus charts just yet. A preview can be provided, though: Generally, our findings indicate that a rear-mounted or front-mounted radiator will outperform a top-mounted, 120mm radiator (assuming identical components, anyway). This seems to be a result of the ‘breathability’ of exhaust – the top uses thick, steel paneling and a small mesh to help control noise emissions (and very successfully contains most of the fan noise). This does have a slight trade-off for thermals, and is circumvented by mounting the radiator in the rear (where it’s fully breathable) or front (intake).
A general rule seems to be that system builders should avoid top-mounted exhaust fans, as those may actually end up “stealing” cool intake from both the CPU and GPU, assuming a tower cooler or rear-mounted radiator.
Assuming a 120mm radiator, our recommendation is to run NZXT’s stock configuration (2x 120mm intake, 1x 120mm exhaust) with no additional top exhaust fans, as those may hinder GPU cooling potential. The rear exhaust fan will receive the front intake from the radiator and exhaust its slightly warmed air, leaving the GPU to intake from the lower front fan. For larger radiator setups, the cooling potential of a 240/280mm CLC is so great that a top-mount – although it will yield warmer temperatures than a front-mount – will ensure the GPU doesn’t get suffocated for cool air. We would recommend mounting larger radiators in the top as push (exhaust) setups.
And, of course, we could go into far more depth than this – but that’s enough for today. Keep in mind that tower coolers are affected differently from CLCs, push/blower GPU fans have different cooling demands, and replacing the front intake with 140mm fans would also impact cooling. There’s plenty more room to play with the Manta’s cooling.