Motovation and Introduction

I recently got a Grinderino 3.5" frame as we saw it was very durable and the stack had a lot of room. I attempted to put a 5" stack rated for 8S capability, and this build was the final configuration after I originally started out with a Mamba F722 45A stack equipped with a Runcam Racer 3, TBS Nano32, and T Motor Pacer 1804s I had laying around. The original standoffs were too short so I had to get taller ones as the stack height I needed was up to 50mm. I wanted something with a lot of performance in a tight form factor and spooling at high RPMs. I have built and rebuilt around 9 quads from whoop size, up to a 9" tarot hex but mostly 3.5" c;ass.

DESIGN AND SPECS
The AUW with a 1050 8S R-Line was ~450g. I placed 2 caps as seen in the photo to reduce voltage spikes and the placement was fine with double sided mounting tape but they loosened in a crash so we fixed them in place with hot glue.
The battery connector was semi rigidly connected and assembled around the TPU antenna, as there were assembly issues with creating strain relief on a standoff.

1. Frame - The 3.5" Grinderino by BOTGRINDER and Ummagawd was selcted for its interior build space and its durability in a true X form factor
2. Stack - Browsing RDQ the newest capable stack with a 20x20mm configuration on 8S was the Sequre stack. I had a faulty MPU6000 Gyro but I was able to get a replacement that works great. For the E70 ESC 70A was plenty of headroom, and if this somehow burns they have an AM32 version
3. Motors - Originally starting out with T-Motor Pacers, they were great quality but we needed something with more thrust, and we looked toward Chris Rosser's motor testing videos and selected the Datura 2105.5 with a theoretical max amp draw of 62A and 1.2kg of thrust each on 6S
4. Propellers - I required a high pitch aggressive prop to achieve the maximum accelerations and the Scimitar's were durable and compatible with the Daturas. Although a lot of the prop blade was covered by the motor bell and coils at the tradeoff of increased airflow for cooling
5. Camera - I also wanted to experiment with low light capabilities, to compare my Walksnail 3.5 quad with the best analog competitor at the time (now its the new Caddx AI camera at the time of writing)
6. VTX - A reliable vtx was required, and high power was a nice to have. The Rush Tank Max Solo 2.5W was the best for penetration and a good competitor to the TBS Pro32 series. Digital systems were not selected due to mounting space and other requirements
7. Powerplant - I opted for dual 4S 600mAh in series or a single 8S 1000mAh. The highest performance option was the R-Line 8S 1050 mAh as it was also proved to work nicely in the Sequre stack review video, and by itself was lighter than dual 4S configuration.
8. GPS - I plan to add the Luminier Remote ID Module to record maximum speeds but at the time of this writing I am still on backorder so it was not included.
9. Reciever - I had great experience with the RP4TD true diversity reciever on my QX7/Ranger Micro Module ground station, and it was feature rich considering the scope of this build
10. Antenna - My main Video setup is Walksnail on LHCP, so to make everything compatible I stuck with the proven TrueRC Singularities on LHCP.

FLIGHT PERFORMANCE & OBSERVATIONS
During testing of this quad there were a few surprises that I did not expect. First Id like to give a couple pointers from the Sequre stack review video that is different from a typical 6S 5":

• Noise. An 8S freestyle build's noise is much higher pitch than a 6S with a notable tone difference and ear piercing when doing a close flyby
• Acceleration. A small blip of throttle on the low end sends you much faster and farther into the air, especially on lighter builds.

I expected these two pointers in the build and thery indeed translated over on a 3.5" class. Here are all of the things that stood out:

• Throttle. Even in the bottom 30% of the trottle range the quad was expremely responsive and the hover point was very low even for a quite heavy microquad past the 400g mark
• Noise. The motors spooled up very quickly because of the kV rating even on a LOS blip hover test in the backyard. I only ever tested them on the Scimitar, the best I can describe the sound is the sound of a whip before it cracks the air. That sharp whirp sound is loud and ear piercing, and I would reccomend ear protection for your VOs
• Motor heating. The large motors generated heat during flight and every flight. There was a lot of wasted energy going on just flying in the park.
• Battery heating. The battery was also hot after landing
• Frame arms heating. New one for me, there was so much heat generated on the motors it significatly raised all four arm temeratures to warm. It makes sense since theres not a lot of thermal mass contained in the frame and that it was also heat conductive.
• Motor mount pressure scarring. Since the frame was pretty light on usable before this build there was significant indentation of the motor standoff connection points on the carbon arms.
• No propwash. The responsiveless power and weight did not make this build floaty at all, but very precise and plenty of power.
• Overall frame wear. Becase of the higher standoffs and increased weight the main back plate was beginning to deliminate and crack, and the top plate was becoming more flexible.
• Bigger plug spark. Its fun to see a huge spark when plugging in and decreasing system power some time after battery is dosconnected due to the capacitors.
• Louder startup tone. The custom tones and beeps are louder on these motors and voltages when comared to my 4S AOS 3.5"
• Very short flight time. Flying aggressively put me into the sub 2min flight time with all the inefficiencies in mind.
• Really f**** fast

TUNING
As far as I am aware no other build on 3.5" has been attemped on 8S, so I was going into a grey area here. Nonetheless Rosser's turning and filter video was followed and I started off with the the defaults on Betaflight 4.5.0 and turned down the master multiplier to 0.3 resiulted in low throttle oscillations. Then I increased to 0.8 which led to trilling, then finally settled on 0.6. Although I wish I could have done more blackbox focused turning becase the latest flight that broke an arm showed the blackbox data observed oscillations on all axis but Roll axis the worst. This was I believe the cause of the motor desyncs that kept occuring. Gyro filter 2 was turned off for all of the tuning.

Because of the desync issues on BLHeli32 I looked at the guides and selected values that were on the safer margin for decreasing the chances and severity of motor desyncs, such as increasing motor timing and setting demag compensation to high.

Based on my calculations for the performance and weight of the build, With no RPM limit anf Motor Output limit, the build had an expected 15:1 thrust to weight ratio and would hover at around 6% throttle with no expo. Of course the 2650kV motors were not tested for 8S only 6S, so a theoretical increase from 63,000 RPM to ~84,000 RPM assuming no heat losses and motor physics isnt saturated would put this thrust to weight ratio a bit higher. Even with a safe valued tune the motor still became extremely hot because of these RPMs. It was a certainty that these motors would smoke and fry so before the maiden flight and onward a motor output of 75% was placed and still pulled a maximum 17.3 g's during a full throttle desync.

FUTURE WORK
I hope this serves as a good starting point for building an 8S on a microquad for those who also plan to build one as well. I had to make sure that my 8S stack would fit the frame I had, and I was also particularly concerned with voltage spikes from the motors. I would advise that low kV motors are a requirement, but if I were to try the build again I would go for a low pitch high kV, or a high pitch low as you can go kV. Also 8S batteries are expensive considering that one unlucky crash broke the arm and that same arm punctured cell 5,6,and 8 which means it had to be retired after 5 flights. Luckily the stack was undamaged and I planned to move everything to an AOS 5". For the price of the Sequre stacks it was a great premium value compared to other brands like T-Motor whose 8S ESC is considerably more expensive.