Heatsink with ducting

- Home
- My CV/Contact
- Astronomy

- Home
- My CV/Contact
- Hamradio
- Old systems
- Current antennas/systems
- SatPC32 to ACLog interface
- Tiltplate for beam antenna
- Greencube Terminal Program
- IC-9700 PTT Expander
- RotorController
- Flatpack2 HE interface
- PA overcurrent protection
- Heatsink with ducting
- REPAM - Remote PA Monitor
- Dual RF Head
- 70CM 500W Amplifier
- 10 MHz Reference clock distribution
- SkyScanner
- SimpleCalc

- Astronomy
- RC planes

Interests and experiences

While designing a couple of Power amplifiers (for HF, 70cm and 23cm) the need for some serious cooling came. I wanted to find some easy to obtain heatsink, one that would not be too expensive (this is always relative), and one that would allow me to remove a lot of heat efficiently.

I choose a heatsink from Wakefield and two fans from Papst.

Now, the heatsink is rectangular at the end while the fans are circular, not something that mates too well. In order to make the airflow as efficient as possible, I designed a duct that would connect the fans to the heatsink and at the same time, would make sure the air passes as perfect as possible from the fans to the heatsink.

The duct can be 3D printed, the STL file for this is available for download here.

The heatsink (12 inch long), Wakefield-Vette partnum. 122258, can be ordered from Mouser from this link

You can get the heatsink in different lengths, 3,6,9 and 12 inch. Scale the numbers below accordingly.

The heatsink measures 304.8 x 187.3 x 79.7 mm. When entering the data from the datasheet for the heatsink in a "heatsink calculator", we can get some estimated numbers.

There are a number of online calculators, I used this one

When using the sizes of the heatsink, I get the following numbers (this is for 3 inch (76.2 mm) of the profile)

From this we can see that at 400 LFM airflow, the thermal resistance of the heatsink is calculated to 0.19 deg/W. This matches the datasheet for the heatsink pretty well (orange arrow, close to 0.2 deg/W for a 3 inch length). Also the thermal resistance at 800 LFM is shown in datasheet to be around 0.15 deg/W and the calculation says 0.14 deg/W.

Both are close enough to be trusted :)

Next up, we enter the full length (305 mm) of the heatsink that will be used. We get the following thermal resistances and pressure drops at different airflows. The pressure drop is needed when we are going to select a suitable fan(s).

Now that we know some details about the heatsink selected, next job is to find a suitable fan (in this case, two fans). Every fan has a specified airflow rate, this number looks nice and large once you find them, but be VERY careful!

The published airflow rates for fans, are typically specified at 0 Pa pressure drop! Once you attach a fan to a heatsink, the fan will see some amount of pressure drop because of resistance to the airflow thru the heatsink.

Each fan type handles this differently. The bigger (higher power) the fan motor is, typically the better it handles a large pressure drop. It is very important to study the datasheet for the fan you have in mind!

I searched for some time, and found an reasonable candidate from Papst, the 3412NG fan (12VDC, 1.9W, 2700 RPM, 32 dB(A)). The size is 92 x 92 x 25 mm.

This fan is available from mouser: Papst 3412NG

The datasheet for the fan can be downloaded here

The most important thing is the airflow it can provide at different back pressures. According to the datasheet, the 3412NG follows curve (3) on this graph:

From the curve (3) above, we can see that with a back pressure between 5 and 12 Pa the fan can deliver between 30 m3/h (200 LFM) and 75 m3/h (506 LFM).

We now know the capability of a single fan when attached to the heatsink!

The fan can deliver 506 LFM at a back pressure of 12 Pa.

Looking at the calculations for the heatsink, we see that at 500 LFM it will give a back pressure of 11.2 Pa

These two numbers means the we should be able to get the 500 LFM from the fan.

But, we have two fans! So the airflow is actually double that amount!

From the calculated data for the heatsink, this means a thermal resistance of 0.05 degC/W.

If we for example need to remove 500 Watt in heat, this means a temperature rise of 500 x 0.05 = 25 degrees!

This is a very good number, even if we dissipate 800 Watt in the heatsink, this will only give a temperature rise of 40 degrees.

Please note that this is with the two fans running at full speed!

In order to direct the air from the two fans as optimal as possible, I designed a 3D printable duct. The two fans mounts to this duct, and the duct is mounted to the heatsink using 4 screws.

Please note that the center divider on the outlet between the two fans, must be up tight on the heatsink, use some strips of soft rubber etc. around the frame of the duct to make a tight fitting (so air don't escape/short it self).

The duct can be 3D printed, the STL file for this is available for download here.

Example with a BUD alubox and a W6PQL 70cm PA and LPF filter

I used the heatsink/fan combination in my 70cm 500W amplifier. I did some 1 minute long full power transmission to test the cooling capability of the combo. Seems to do a decent job. The amplifier was running at 55% efficiency at 480W output, so around 400 Watt was dissipated by the heatsink.

The ambient temperarure during the test was 22 degC.

After 10 transmit/receive sequences, the LDMOS temperature stabilizes around 40 degC. This means that the heatsink/LDMOS was at 18 degC above ambient.

This fits wonderfully with the calculations above of 25 Deg. rise with 500W dissipation. In the data above, 400W was dissipated at 50% duty cycle (1 min TX, 1 min RX).