CamdenBoss Hex – Raspberry Pi 4 Case

Having previously reviewed the CamdenBoss Cube case, it’s time to look at another unique design from this UK manufacturer. The CamdenBoss Hex Raspberry Pi 4 case.

Construction

This is another 2 piece design Raspberry Pi 4 enclosure, where the two halves are bolted together providing a rigid and specifically with case provides an added degree of security not found in many enclosures.

• 4 x 12mm plastic standoffs with brass inserts provide a secure and raised mounting for the Raspberry Pi 4. The cross-head screws provided are annoyingly non-magnetic.
• 3 x T10 security Torx bolts secure the two halves of the case together.
• The case was supplied with 4 push-fit rubber feet, which I found a little difficult to install manually, without damaging them.
• 2 x 30mm fan mounting points, pre-drilled knockouts are provided. (Fan’s were not provided).
• 2 x mounting points on the rear, 100mm apart, with slots to allow fixings to be with the case at a number of angles. Interestingly 100mm is also one of the VESA mounting sizes, so perhaps mounting on the rear of a display was one of the design considerations?

Design

• Unlike the CamdenBoss Cube case, the Hex case does have ventilation provided by the two 30mm fan positions and some diagonal slots at the rear.
• Supporting two optional fan mounting points is useful, I doubt there are many scenarios that would benefit from both positions being used.
• The case prevents access to many of the connectors on the Raspberry Pi 4. Leaving only the front USB and Ethernet port accessible once the case is sealed.
• There is a slot towards the rear of the case which allows for cables to be installed before the case is bolted together. Though as can be seen, the standard USB-C power connector is a bit tight and would be better suited to a right-angled connector
• The SD card isn’t accessible, again this could be a security benefit.
• The status lights on the rear of the Pi are visible by looking through the rear vent slots or the rear cable opening.
• Both the Hex and Cube cases are larger than is typically found, but this Hex cases size does provide some additional security which may be important for some scenarios.

CamdenBoss provide offer a number of customisations to the design, including different colours and materials, digital printing directly on to the enclosure, internal RFI/EMC coatings, see more at https://www.camdenboss.com/camden-boss/cbrpi4b-hx-bk-hex-raspberry-pi-4b-enclosure/c-23/p-24094

Thermal Testing

Passive Cooling

As the Hex case wasn’t supplied with any Fans, initial testing is performed without any active cooling.

Without active cooling, throttling begins to occur at 50% load as the 2 core stress hits temperatures of 84°C resulting in frequent throttling down to 1000MHz. The 75% load test sees more continuous throttling, with the test running at 1000MHz with frequent drops to 750MHz and several drops to 600MHz. In the 100% load scenario, the system reaches temperatures of 87°C and suffers from extensive throttling throughout the test. The majority of the test sees the CPU core clock running between 750MHz and 600MHz but with frequent dips to 500MHz.

Active Cooling

Whilst the Hex case provides mounting points for up to two fans. I elected to test with just a single 30mm fan installed, on the side nearest the GPIO header pins.

For this test, I used 3 different Sunon MagLev fans which I sourced from reichelt.de. Tests were performed running the fans at their rated 5 Volts and at a reduced level of 3.3 Volts, which reduces the speed and overall noise. The manufacturer specifications for the fans tested are:

I would have liked to have obtained the MF30100V3 a 30x30x10mm fan running at just 6000rpm and claimed noise of just 10.2 dBA, however, I wasn’t able to find anyone currently stocking this variant.

Thermal Results

The graphs below show the temperatures under the 100% load scenario with the fans running at 3.3V and 5V. The passively cooled results are also included to show the impact.
The CPU temperature plotted is the temperature above the ambient room temperature.

With a single fan running either at 5V or 3.3V, the CPU no longer throttled under any of the load tests. The performance of any of the tested fans is more than sufficient, but it does come at a cost.

Noise Testing

The hollow construction of the CamdenBoss Hex case appears to act as a sound amplifier for the noise generated by the Fans tested.

Running any of the fans tested here, at any voltage produces an unacceptably (my subjective opinion) loud droning sound which would be unsuitable to installation in a quiet home or working environment. If the case is to be used in a noisier environment, then the impact should be less annoying.

As noted earlier, there is a lower speed, quieter 30mm MagLev fan available, the MF30100V3, however, at present I’m unable to source this to see it can provide an acceptable level of noise whilst providing adequate cooling.

Conclusion

Like the Cube case, there is an element of form over function, with key capabilities of the Raspberry Pi 4 restricted from being used. This is listed on the site as being ideal for smart home IOT devices, I’m not convinced due to the poor thermal results, but maybe if under light load or where the throttling isn’t going to be noticed maybe it would do, but it would seem wise to choose something which didn’t hamper performance.

• The CamdennBoss Hex enclosure requires active cooling if being used for anything other than the lightest loads.
• Unfortunately, acceptable noise levels weren’t possible with the fans tested so far.
• Restricted access to internal headers will be a limiting factor for some.
• No access to the SD-Card, USB-C, and the Micro HDMI connectors once the case is secured is also a significant limitation for many scenarios.

Parts Tested

• Raspberry Pi 4 (Amazon) (Pimoroni)
• CamdenBoss Raspberry Pi 4 Hex Case
• Fans:
  • Sunon MF30060V1 (reichelt.de) (mouser) (digi-key)
  • Sunon MF30060V2 (reichelt.de) (mouser) (digi-key)
  • Sunon MF30100V2 (reichelt.de) (mouser) (digi-key)