Flosstyle Build
I hadn’t a specific frame in mind when I started designing the TPU stack. I get the Flostyle frame from a local pilot for half the price, along with a Runcam Eagle. The stack turned out to fit pretty nicely in the frame, with the MT30 aligned with the arms. I just had to make some room for the arms nuts and for the lipo strap, and to remove the central standoff. Antennas and the XT30 connector are secured at the back by 3D printed support.
The Flostyle is a very compact and low profile frame, maybe a bit too much.. The runcam Micro Eagle barely fit in the front, and depending on the angle you setup, the back of the camera will touch the bottom plate. I filed the carbon to get more clearance around the camera.
For the power train, I picked up the XING-E 2208 (yes, partly to stay within the red and black color theme x) paired with the Nazgul 5140 props. From there, the build was very straightforward. After soldering the MT30 connectors to the motor wires, the quad is assembled in minutes.
Flight Performance
Flashed in BetaFlight 4.1, this quad is now my main freestyle rig. It weights 356g, 550g with a 6S 1000mah battery. Control and flight performance in general are excellent, as expected for a modern quad. It’s also quite efficient, I get up to 5min of cruising and light freestyle.
Video range is not so great despite the 600mW output power. Antenna mount is far from ideal for RF performance, the antenna is too horizontal and too close to the frame. But it’s very durable, as the antenna doesn’t stick out.
Incident
Althougth I’m very happy with this build, flight experience showed some weakness and limitations of the TPU stack design.
Overheating
As some people pointed out to me, encasing all the electronics in a sealed enclosure leaded to some thermal issues. After some minutes, even when flying, Betaflight warning will report a core temperature over 70°C.
When looking at the quad ready to take off with the battery strapped, the explanation is quite obvious: Both heatsink are covered. There is no airflow at all on Vtx heatsink, as it sit just below the battery. ESC heatsink is more exposed to airflow, but there is still the battery strap covering it.
Anyway, this overheating situation is not really a deal breaker for using the quad as is. The warning usually shows up mid-flight, and the highest temperature I got at landing was 77°C. Electronics components will just work fine at this temperature, I just have to wait between each pack for the quad to cool down.
ESC/Receiver collision
One day, as I was practicing fast and low flying, I hit a metallic fence. Battery died, as it took most of the impact energy. But the quad itself seemed fine. I then plugged my next pack, but there were no more Rx link. Both video and ESC were working, my motor beeped at power-up. So heading back home, and after taking apart the quad on my bench, here is was I found (after cleaning up all the mess) :
4 of the ESC decoupling capacitor were just gone, entirely burned. Even the copper of the pads area had melted.
What happen is that the impact compressed the whole TPU stack. Because the FC was mounted with soft standoff, it pushed the R-XSR against the ESC. The R-XSR crushed some decoupling capacitor on the ESC. And ceramic capacitors turn into a short circuit when physically damaged. In this case, a short.. directly on the battery.
Hopefully, the short probably didn’t last very long, just the time for the capacitors to be destroyed. But it must have been very hot inside for a moment, enough to melt copper..
Now, for the damage assessment: the R-XSR had it’s 3V3 LDO broken, but it was easily replaced. Surprisingly, the ESC itself works okay ! After cleaning and inspecting it, turns out that only the first copper layer around the capacitors was affected. I replaced it anyway, as I wouldn’t trust an ESC that went throught a fire in my quad.
The mistake here is pretty clear : the TPU stack must not be used as a structural part of the frame. Here, to accomodate the stack in the frame, I removed the four central standoff, and used the stack hardware instead. I knew it was a bad idea, but I didn’t have much choice if I wanted to use this frame. And I didn’t anticipated the consequence it would have.
Conclusion
The TPU stack concept turned out to be less versatile as I imagined it. Integration into the frame is very important to maintain good thermal performance and mechanical integrity, and therefore drastically reduce the number of frame usable with it.
Apart from the short-circuit issue, this stack proved to be very reliable, and I want to push the concept toward it’s original purpose: making a very durable quad. I will re-design it entirely to improve thermal and buildability, and develop custom hardware around it for an optimal integration:
- F7 flight controller with high end Vtx and enhanced cooling capabilitie
- Tailored frame to carry the stack