Bi-copter racer

[Piotr Szalczynski]

New design of small Bi – copter racing drone based on my previous small builds. Soon to be CNC-ed.

• This topic was modified 4 years, 10 months ago by Piotr Szalczynski.
• This topic was modified 4 years, 10 months ago by Piotr Szalczynski.

Attachments:

1. 

   1.jpg
   

2. 

   2.jpg
   

3. 

   1-1.jpg
   

4. 

   2-1.jpg
   

[Kevin_Erik]

Very clean design!

Quick question: Are you using the header pins to attach the Servo to the FC (quick disconnect) or is everything soldered in? Reason I ask is that replacing Servos has always been a pain to do. Not because of the soldering (I’m 2M Solder Qualified via the U.S. Naval Lab) but because I often have to take apart a large portion of the Copter to effect the swap.

Cheers!

Kevin

[Piotr Szalczynski]

So far I was soldering a cable with connector (female plug) to FC and the connect to servo (male). So I didnt need to cut servo cables. In my future desing I plan to use a new flight controller which has pins from PDB so no need to solder anything.

• This reply was modified 4 years, 10 months ago by Piotr Szalczynski.

[thorkan]

Hi Piotr,
congrats, light weight construction. How do you attach the servo? I think if you support it by the upper plate its a stiff solution?!

What a great and detailed 3d model you did! What Program do you use therefore? Did you build all the 3d parts by yourself? Must have cost days to do that detail level;-)
Keep on the good work!

[jihlein]

Very well done!

[Kevin_Erik]

Any reason for the 2 or 3” props & High KV motors?

[Piotr Szalczynski]

Actually the 1106 6000 Kv are there only cause I did not wanted to spent time on modeling the motor I want to finally use. I was thinking about 1206 4500 Kv or 1306 2700 Kv / 4000 Kv. The reson begind 2-3″ props is simply the size. I dont want to intemidate new Bi-copter users with a dangerous machie. We all klnow this configuration can behave pretty rough so using 4-5″ is too scary especially indoors.

[Piotr Szalczynski]

I use Fusion 360, Solidworks, CATIA and NX . This model was made in Fusion360. The servos will be attached by zip-ties which I didnt modeled in CAD. and servo itself is locked in the cutout in carbon frame. I think its still the lightest and cheapest solution. So far worked very good in all 3D printed prototypes. https://www.youtube.com/watch?v=O5djisDuUdQ

[Kevin_Erik]

Quick question: Which ESC protocal and Servo frequency have you been using?

Kevin

[Kevin_Erik]

How are you progressing with the project, any pic’s ?

[Bellah1z]

wow looks great, thanks for the future expansion of our Bicopter Universe !

[Piotr Szalczynski]

https://youtu.be/GJbuMYEdd9A I tried to reduce wobble with rapid changes in flight direction before I will manufacture racer. This is what I achieved so far. What do you think about this reactions for rough stopping and change of flight direction?

[Kevin_Erik]

Looks very manuverable and fast. Did hear quite a bit of warbling during the turns. Thinking that your Pitch gains are a bit high and causing the oscillations turning turns. Question are you using D-Term? I ask cause in dRonin and Beta flight, D causes Servo delay / lag that lead to oscilations. Only Roll is immune to this due to the high reation speed of the motors.

[Piotr Szalczynski]

My PID are quite low P 22 I 20 D 20. Do you think I should try make D 0 ? I can send you the log from this flight so you can see what was happening during it.

[Kevin_Erik]

Its pretty much agreed upon that D-Term doesn’t work well for any Servo controlled axis as Servos are 100 times slower compared to modern brush-less motors. Now as previously stated, the delay D-Term adds appears to cause each axis to oscillate more easily. If you look at some of my Bicopter flight videos you’ll see oscillations during reverse flight due to Yaw and Roll fighting each other (read my example) or near constant oscillations from Yaw being over driven / having D-Term active. Try this and see for yourself, should be easy enough to do.

Example 1: During Roll, the motors induce a small amount of Yaw that the Gyros detect but the Servos are too slow to respond to effectively. Thus during this Yaw, the change in blade orientation (think gyro) causes a small amount of Roll. The motors while quick to respond, once again induce Yaw, causing the Servos to again react. For Roll / Yaw PID’s that are too strong, Yaw with D-Term or reverse flight, this can create oscillations that easily become a self sustaining loop.

The “why” of reverse flight Oscillation is pretty strait forward but not everyone agrees with it. IMO It all comes down to the propeller Aerodynamics…

Example 2: During forward flight, the tips of the props are generating the most lift when they’re the furthest way from the copter. (assumes left prop CW and Right prop CCW rotation) This is also the point at which the prop experiences the least turbulence. During reverse flight, the point of greatest lift is now the point closest to the copters body and also the point of greatest turbulence. This turbulence comes from the prop vertices / thrust interacting with the frame, carbon arms and Servos. Knowing all this its easy to see why forward flight can be stable while, reverse flight oscillates for the same PID’s. What’s interesting about this is that reversing the props and motor spin direction will make reverse stable and forward flight oscillate! IMO, propellers with an even number of blades should work best. Reason being is that the both the lowest and highest points of turbulence are crossed simultaneously. Thus canceling each other out or at the very least minimizing the effects of said turbulence. For your size of copter, four bladed props would work best. They’d fly smoother, quieter but your not going to see efficiency gains. Which is not something I’d really worry about for such a small copter.

Beyond all that I’ve only recently managed to achieve stable flight. This was done by keeping D-Terms off Yaw / Pitch, while keeping Roll / Yaw gains low. Then Tuning Pitch below the point it starts self oscillation. Keep in mind that Roll and Yaw are both very powerful Axis and thus should be kept low. While Pitch needs a bit of help to be able to manage the copters orientation for smooth forward / reverse flight.

My remaining issues are:

* Self-leveling modes are misbehaving, causing roll to one side during longer flights.
* During sharp turns (with both Roll and Yaw) the copters nose can drop suddenly. This is either a case where the minimum motor speed is too low or Leveling is having problems maintaining pitch as the Copter Rolls past a certain point.

-Kevin

[Author]

Posts