Noctua NH U9B Cooler SE 2 with Silverstone Sugo SG07

1. Introduction2. The Noctua NH U9B SE 2 Cooler3. The Installation4. Testing Methodology5. Noctua NH D9B SE2 Testing (full speed)6. Noctua NH D9B SE2 Testing (low noise settings)7. Noctua NH D9B SE2 Testing (ultra low noise settings)8. Noctua NH D9B SE2 Testing (fan disabled)9. Closing Thoughts10. View All Pages

Our recent review of the Silverstone Sugo SG07 stirred a lot of interest with our readers and I received emails asking if it would be possible to passively cool a high performance processor in this chassis. KitGuru seemingly has a core audience of readers who loathe noise, with a vengeance.

In our continued efforts to focus on cooling we decided to revisit this brilliant chassis and to figure out if this would be possible, without too many compromises.

In our original Silverstone Sugo SG07 review we focused on high end video cards and overclocked our Core i5 processor and today we are going to find out if it is possible to remove as many fans as we can and still have a stellar media PC with gaming capabilities.

As we have had so much success with Noctua coolers in the past, such as their class leading Noctua NH D14 (dubbed by KitGuru as the ‘Austrian Sandwich') we decided to give Noctua a call to find out if they had a cooler that would fit in this chassis and something more substantial than the Thermaltake Slim X3 we used before. The Silverstone Sugo SG07 is a reasonably sized media chassis, but we found that most of the bigger coolers won't fit into it with a 117mm height limitation and we also found that even graphics cards such as Sapphires fabulous HD5670 Ultimate Edition were too tall also.

Noctua offered their new NH U9B SE2 which is a revamped follow up to the hugely successful original cooler which received over 100 awards when it was originally released. The new SE2 version is designed with HTPC's in mind and adds support for 1156 and 1366 motherboards with the new SecuFirm 2 multi socket mounting system. The unit is also supplied with two high quality 92mm NF 89 premium fans.

The Noctua NH U9B SE2 is 25% narrower and 30% shorter than the U12P – we reviewed its cousin, the C12p SE14 on Kitguru a short while ago.

The NH U9B SE2 arrives in a traditionally styled Noctua box – which we love.

Inside there are two boxes, one with the heatsink sandwiched between protective cardboard, and another box which holds the fitting kits and extra 92mm fan which is supplied by Noctua. The same gatefold style manual package is supplied with left side for Intel and right side for AMD fitting. There are also adapter cables to slow down fan speeds for reduced noise levels. More on this later.

The heatsink itself is very attractive, as are all Noctua products – the engineering quality is exceptionally high and it shows.

The heatsink is shipped with a plastic protective cover on the base which helps to stop any marking during rough shipping. We have never had a Noctua protect arrived damaged yet, and they all are shipped internationally from several thousand miles away.

The top of the heatsink has an embossed Noctua OWL logo and the heatpipes extrude from here to dispel heat.

We are using the KitGuru award winning Silverstone Sugo SG07 for the system build today and we dismantled the system which we are currently using as a media center in one of our rooms.

The image above shows the excellent Zotac H55 ITX motherboard and the Thermaltake Slim X3 inside the chassis before we dismantled it. We previously had a GTX460 in the system but today we are going to see if it is possible to passively cool the system.

The Noctua U9B SE2 is supplied with three sets of plastic bags. One contains Intel specific components, another contains AMD specific components and the last has ‘shared' items between the system builds.

Both Intel and AMD builds use the mounting brackets seen above, these actually mount to the upper side of the heatsink and extend to allow mounting to the motherboard kit. This process takes only a few minutes.

The backplate is fitted in exactly the same manner as the NH D14 which can be seen here.

The center bolt on each side then screws into the motherboard mounting plates. It is worth noting however that the mounting position above will not allow graphics cards with any rear mounted passive heatsinks to be fitted. it is a few millimeters from the PCI express slot.

We therefore need to rotate the heatsink 90 degrees to allow for more room next to the Zotac H55 ITX PCI express slot. This obviously would be the best position anyway for fan mounting as the case has airflow holes on the side panels, not the rear.

In this position there will not be enough room for extended heatspreader memory fitting – but our tried and trusted 1600mhz DDR3 Kingston ram fits without a problem.

The images above show one 92mm fitted to the cooler, even though two fans are supplied the fitting of the other one would completely block the PCIe slot on the right. This board does have onboard graphics but as this is going to be used as a 1080p media center we are aiming to include a high quality discrete solution.

We are also going to try this system in a fanless state, although whether the Core i5 CPU can operate without airflow is a question yet unanswered.

The Zotac motherboard installed back in the Silverstone Sugo SG07 chassis. You will see the Sapphire HD5670 Ultimate Edition also installed. It fits barely, as the heatsink of this graphics card extends to the rear of the PCB and also increases the breadth of the card from the slot.

Regular readers will remember that in our original review of the Silverstone Sugo SG07 that there is a 180mm AP181 Air Penetrator fan installed, which we reviewed a short while ago.

This fan however has to be removed for two reasons. Firstly – it takes up a large portion of the inside chassis space meaning neither the Noctua NH U9B or the Sapphire 5670 Ultimate Edition can be installed. Removal is straightforward – three screws hold it in place at the top of the chassis and two tiny screws at the rear of the chassis hold the ‘low' and ‘high' fan speed controller in place.

We can now seal up the chassis and fire the system up.

This review is focused on monitoring cooling performance and noise levels as we lower fan speeds with the goal of trying to remove them completely from the main system build. Ideally we would have liked to remove the Silverstone power supply and use a silent passively cooled Zalman unit, but samples are limited and the Silverstone power supply has been specifically designed for this chassis with custom cabling lengths. As it is downward firing, and filtered – we also have never noticed any noise emitting from it in the past and will put it to the test today.

Our ultimate goal is to get a reading of under 20dBa via our noise meter in our specially created sound room. More on this later.

Media Center – Minimal Noise build

Chassis: Silverstone Sugo SG07
CPU: Intel Core i5 655k Engineering Sample @ 3.2ghz
Graphics Card: Sapphire HD5670 Ultimate Edition (Single and Crossfire X reviews) – passively cooled.
Motherboard: Zotac H55 Mini ITX
Thermal Paste: Noctua NH H1 Thermal Compound
Memory: Kingston Hyperx 8GB DD3 1600mhz
Cooler: Noctua NH D9B SE2
Hard Drive: Crucial 256GB RealSSD
Monitor: Panasonic 42 inch NeoPDP 600hz Plasma

SkyTronic DSL 2 Digital Sound Level Meter (6-130dBa)
Thermal Diodes
Cinebench R10
Furmark

Recently we have changed our method of measuring noise levels. For most reviews we have built a system inside a Lian Li chassis with no case fans and have used a fanless cooler on our CPU. We are using a heatpipe based passive power supply and an Intel SSD to keep noise levels to a minimum. The motherboard is  passively cooled and we use a Sapphire HD5670 Ultimate Edition graphics card which is also passively cooled. Ambient noise in the room is kept as low as possible. We measure from a distance of around 1 meter from the chassis and 4 foot from the ground to mirror a real world situation.

Why do this? Well this means we can eliminate secondary noise pollution in the test room and concentrate on only the components we are testing. It also brings us slightly closer to industry standards, such as DIN 45635.

To test today we are using the Silverstone SG07 chassis with the supplied 600W PSU, we will be using the Noctua NH D9B SE2 in various configurations as explained on each page. We normally use air conditioning in our testing environment which generates noise levels around 21dBa. With air conditioning disabled our meter reads 16.2dBa ambient.

The room environment held fairly stable between 24c and 26c throughout all our testing.

KitGuru noise guide
10dBA – Normal Breathing/Rustling Leaves
20-25dBA – Whisper
30dBA – High Quality Computer fan
40dBA – A Bubbling Brook, or a Refridgerator
50dBA – Normal Conversation
60dBA – Laughter
70dBA – Vacuum Cleaner or Hairdryer
80dBA – City Traffic or a Garbage Disposal
90dBA – Motorcycle or Lawnmower
100dBA – MP3 player at maximum output
110dBA – Orchestra
120dBA – Front row rock concert/Jet Engine
130dBA – Threshold of Pain
140dBA – Military Jet takeoff/Gunshot (close range)
160dBA – Instant Perforation of eardrum

Most Noctua coolers are supplied with two noise adapters – a black coloured ‘Low Noise' cable and a blue ‘Ultra Low Noise' cable. We will be using both today and lastly removing the fan completely to see if our system can be kept in check. We won't be underclocking either as we want the full performance from the 655k processor. The chassis will be closed and placed on a table, just as it would be used.

For the first series of tests we are running the system with the single 92mm Noctua fan on full speed to get a reference point from which to work from.

Temperatures on the processor never exceeded 50c which is a good starting point for the 655k processor. Considering there is no other airflow in the chassis this is impressive.

When the single 92mm fan is set to full speed we measured a total noise of 21.9dBa which was not noticeable by our real world ‘ear' testing. Next we will hook up the Black ‘low noise' adapter and measure performance levels.

The Noctua Black Adapter is attached to the NH D9B SE2 which reduces fan speeds to 1,300 rpm and a Noctua rated 13.1 dBa.

When using the low noise adapter processor temperatures rise by about 9-10c and system temperatures also increase slightly due to the reduced airflow which helps the chassis and heatsink driven Sapphire HD5670 to cool. They are still within safe parameters however.

Noise levels drop to 16.8 – this is extremely difficult to record as our breathing can even affect the result. We held our breath and measured the noise levels in the room.

For the final fan test we attached the blue Noctua ‘Ultra Low Noise' Adapter which Noctua rate at 7.9dBa for this fan. The fan slows down to around 1,000rpm.

As expected temperatures rise by another 10c or so on the processor while the Sapphire card temperatures increase by another 1c under load. These are still well within safe parameters and we can't even tell the system is on.

Again we had to hold our breath and maintain still movement to get our final measurement of 16.2dBa. Obviously the PSU and other ambient noise is adding a little to our reading but this was as low as our equipment was able to measure in this room. Even with our ear to the chassis we could hear nothing at all. The Silverstone PSU is clearly also very effective.

We disabled the single 92mm fan completely for these results and ran our tests again.

To measure these temperatures we left the system at idle for 90 minutes then tested again under load. The cores would hit 80c but even after 3-4 hours of testing never exceeded this point – the sheer mass of the heatsink was able to keep the 655k cool. It would be possible to run this system passively cooled long term as 79/80c is not dangerous, especially with default voltage. Whether we would however is another question.

Measuring noise levels with the single fan completely disabled did not lower our meter reading at all, we can therefore take an educated guess that ambient room noise and the noise of the PSU is generating the rating of 16.2dBa. We can also assume that leaving the fan on, with the blue, ultra low noise adapter is a better long term option as temperatures are lower and noise levels are not increased at all.

Once again Noctua have created a phenomenal cooler which is without a doubt one of the best solutions for this market we have yet to test on KitGuru.

Even at default fan speeds the Noctua NH D9B SE2 does not generate much noise at all, and to the human ear it is hard to notice from a distance of one meter. Our recording equipment did measure around 22dBa. Moving to low noise and ultra low noise adapters brought further reward with dBa recordings of 16.8dBa and 16.2dBa respectively – clearly our equipment and testing room would need further improvement to measure more accurate dynamics at these settings.

Removing the fan after using the ultra low noise adapter didn't help our recorded dBa figures at all so we would leave a single fan at this setting for a little airflow, especially as it helped drop temperatures by 8-10c.

The Silverstone Sugo SG07 still manages to impress us, especially as within this particular review we have managed to measure how quiet the power supply is with the right partnering equipment. ‘Out of the box' you are able to have a high performance gaming system built into this chassis and for people who want to build a quiet system you can easily remove the 180mm fan and upgrade the heatsink to a Noctua NH U9B SE2.

KitGuru says: Noctua never fail to impress us as a company, their heatsink solutions are without a doubt the finest on the market, bar none.

Become a Patron!