Hello all noise phobics.
I avoid instruments with cooling fans if I can, but the RTB2004 tempted me, and after reading many comments on how quiet it was I decided to give it a try. How bad could it be? Well, I soon found out. After 2 minutes I turned it off in disgust, and realized I had to fix it or get rid of it.
To give some context: where I live there is no external noise pollution, my computers are all in Hdplex fanless cases and use SSDs for storage. The RTB became the loudest thing in my house. I could easily hear it from an adjacent room.
Initially I had hoped to reduce the noise by just replacing the fan with a better 80mm fan. So I tested with a Noctua NF8, but the difference was minimal. They both run at the same rpm, and the high speed of a small fan is the main contributor to the noise. I needed a larger fan, but the RTB chassis construction and the fan orientation makes it difficult to fit a larger fan inside the case. I did not want to modify the chassis to make space, but opted instead to modify only the plastic case
Before mods:
------------
Fan = Sunon HA80251 v4 , 1700 rpm @10.5 volt
Noise: rear= 40dBA , right= 34dBA , front= 35dBA , left= 34dBA ( measured at 30cm )
Internal heat sink temp: adc=34C , micro=50C
Exhaust temp: 32.5C - 40.7C Air speed=1.2 m/s
Peak temp pcb backside = 62.9C (pic 'before')
The RTB fan blows air over the top side of the pcb only. There is no air flow over the back side. The attached pic 'before.jpg' shows the temperature on the pcb back side after one hour. The hottest part at 62,9C was a component (MPKM 3620 246 ?) under the adc.
I decided to:
-use a larger slow-running fan placed outside the case.
-modify the case to make the best use of the air flow.
I tested several fans. Some did satisfy my noise requirements when running at reduced rpm, but showed starting problems when the speed was reduced too much. (The RTB monitors the fan, and displays a warning message if the fan stops). The fan I selected was Noctua NF S12A ULN
To minimize the needed air flow I decided to open up fan inlet so air can pass through the cavity under the pcb so bottom side acts as heat sink. To allow air to enter the space, the fan cutout must be placed as low as possible. Air exit was created by adding several vent holes ( pic 'bottomview' ).
The case was modified as follows:
-all vent holes on the right side blocked.
-the top three vent hole rows on the left side blocked.
-added vent holes to the bottom of the case (pic bottomView). Air exit from pcb back side.
-opened up the rear air intake to suit a 120mm fan.
-added a fan grill and cover (pic 'rearview')
After mods:
--------------------
Fan = Noctua NF S12A ULN 550 rpm @10.5 volt (with uln resistor)
Noise: rear= <30dBA right,front,left = <26dBA (below noise floor of meter)
Internal heat sink temp: adc=35C , micro=50C
Exhaust temp: 36.4C - 42C Air speed = 0.8 m/s
Peak temp pcb backside = 47.3C (pic 'after')
All measurements taken after one hour.
Ambient temp = 26C
R&S specifies noise as 28.3dBA at 0.8M. I used 0.3M since that is more like the actual distance when I use the scope.
The noise level values tell only part of the story. The greatest change with this modification was the noise character. It is now atonal (like pink noise) and it does not disturb me at all.