While in a holding pattern, waiting for power supply deliveries for both the PoE and Pi4B, I started some parallel work that was proposed in the RoadTest application. Since the specification indicated the PoE Hat was compatible with the Pi3B+, I figure I could start laying out a RoadTest Review framework using the Pi3B+. When the Pi4B is ready I will run the same tests.
Some observation were discovered while Installing the PoEHat on the Pi3B+.
- Raised heat sinks on a Pi may conflict for space. The heat sink (purple ellipse) had to be installed off center of the chip, in order to leave sufficient clearance for the PoE Hat transformer. Moving the heat sink to the left put it dangerously close to an SMD component (green ellipse) on the Pi motherboard. To alleviate both issues the heat sink is installed shifted down and to the left, resulting in it not completely covering the chip.
- The PoEHat completely consumes all the GPIO pins leaving them in accessible. The GPIO pins (red ellipse) are not long enough to extend through the PoEHat.
I have ordered (MORE SHIPPING COSTS!) a stacking header to extend the pins on the Pi. I'm not confident the existing space is sufficient for this to work. The plan to load up the Pi with a LED array detailed in The Road to Raspberry Pi4B/ PoE Hat RoadTest Review (Testing) will not be possible if the GPIO pin cannot be extended. Let's see what the stacking header does.
- PoEHat and Pi3B+ combination exceeds the size of a standard case.
I carved a hole in a plastic case I had available to enable the paired assembly to fit.
Let's test:
With the PoEHat installed and a new SD card loaded with a current Raspbian image I used a Pi power supply to fire up the combo. Wait, can you run a Pi with a PoE Hat using the Pi power supply? A look at the documentation on the raspberrypi.org site for the PoE Hat wasn't helpful. The documentation is more of a travel brochure with pretty picture than a manual that has useful information.
Truth be told, Yes Dorothy you can run a PoEHat'd Pi with a local power supply! Well it did in this example. I would assume the answer to the question you would find in the documentation. Not the case.
With a working Pi3B+ & PoEHat combo a stress test was used to establish a load on the devices. I followed the instructions provided in this link https://core-electronics.com.au/tutorials/stress-testing-your-raspberry-pi.html to establish a load and take measurements.
On the left there is a terminal output with the stress application running along with an htop output displayed on the bottom. On the right wthere is the terminal output of cpuburn application running along with an htop output displayed on the bottom. Both tests drive the CPU's, the cpuburn application claim to fame is to really heat up the CPU's. The test setup and performance matched what was in the documentation. There was some concern about the link be relevant because of date but that didn't prove to be a problem.
Discovery:
- PoE Hat fan turns on at when the screen displays a temperature of approximately 50 Degrees Celsius.
- The video in the link suggested the fan sound changed as cpuburn was engaged indicating an increased load on the power supply. The power supply voltage on GPIO pin 2, drop from 5.11 to 4.78-4.83 volts DC when the cpuburn test was running.
The voltage change using the cpuburn test makes it an ideal application to use for further testing as described in The Road to Raspberry Pi4B/ PoE Hat RoadTest Review (Testing) . It provides a setup that will put a strain on the PoE Hat power. I especially want to do this when the combo is attached to a maximum length of CAT 5.
I'm limited on further testing that can be performed at the moment, while I await parts delivery. This initial round of tests using a Pi3B+ has definitely provided some insights that will prove useful going forward. My updated plan is to confirm if the discoveries that were made using a Pi3B+ exist in the Pi4B. I will use the official RoadTest Review posting to document those findings.
