After my last crash the shaft on my rear tricopter motor was tweaked.
Fortunately I gave each motor a spin manually before powering things on and noticed so I had opportunity to fix it. The motor was clearly rubbing and I imagine would have promptly self destructed had I turned it on.
The motor is a NTM Prop Drive Series 28-30A 750kv / 140w though I would suspect this approach would apply to other motors including those used on quadcopters y6’s octocopters and other miscellaneous multicopters.
In essence, I removed a snap ring from the base of the motor which allowed me to separate the housing from the interior. I then mounted the housing with shaft in my lathe and using a dial indicator centered it up roughly (via nocking) in the chuck.
Once reasonably centered I ran the dial indiactor on the far end of the motor shaft. It was out by nearly 10 thousandths of an inch. I located the high spot and pushed on the shaft to nock it back to 5 thousandths. The process of measuring total outage, finding high point, halving, was repeated a couple times till it was under a quarter thou out.
Remounting and replacing the snap ring I found the motor now sounded healthy. Powering it up verified all was well; disaster averted!
As an aside; the earlier part of the flight actually went quite well. Trying it in a fairly open field made flying the tricopter a bit easier and I actually managed a pretty neat steeply angled circle around 1:36 into the video.
To configure the ardupilot/arduflyer for a tricopter you first need to load the tricopter firmware on the board (mine came with the quad firmware out of the box).
In theory, you can use mission planner to do this but for my board at least I consistently received an error when trying to load the firmware that way. I fell back to using the Arduino IDE to compile and load the firmware. This guide covers the process nicely. Basically you:
You should now be able to arm your motors. Detail instructions are here but essentially, hold the left stick to the bottom right for 4-6 seconds (the red LED switches from flashing to solid).
You can now add props and arm the tricopter. Turn on very light throttle and verify all of the props are moving air in the correct direction. If any motor is spinning the wrong direction swap two (of the three) power leads going to the motor to reverse its direction.
Now that the props spin the right way; with very light throttle try rotating the tricopter to verify the yaw servo moves the correct way. On my tricopter it didn’t so I reversed the movement by setting RC_7REV to -1 under the advanced configuration settings of mission planner.
I have the tricopter fairly together electromechanicaly but still need to tune the firmware.
At present the main issue is arming the motors; I have to do it via a laptop, using the controller it refuses to arm. Regardless, I did manage a quick test flight and it seemed to work great. Getting it to hover only took around 70% throttle so it should be peppy and have room to carry some camera gear. Looks like I can improve my landings a fair bit though 🙂
The hobby king KK2 board looks like a good low cost option at only $30. If you just want something cost effective and simple this seems a solid route to go.
Longer term though; I like the idea of autonomous flight capabilities and, even when manually controlled, the ability to automatically return to the launch site should something go awry.
To achieve these goals, the Ardupilot system looks quite good. Direct from DIY Drones, the creator, it costs $180 + shipping. RCTimer sells a copy called the ArduFlyer for $89.99 they also sell the u-Blox CN-06 GPS Receiver V2.0 for $26.99 ($116.09 + shipping total). Their version seems well reviewed so I opted for it given the savings.
You will need to use some jumper wires to connect the GPS to the TTL/USB adapter as the pinouts aren’t quite the same. Also, note you want to connect TX -> RX and RX -> TX for them to communicate properly. Going TX to TX and RX to RX will not work.
I found loading the settings to the GPS would make it ~3/4 of the way through and then error out. I believe this is due to the baud rate of the GPS changing partway through the config loading. To get around the problem I initially set the u-center software to 9600 baud and loaded the config as far as it would go. I then swapped to 38400 baud and re-loaded the full config; this second runs seemed to complete without issue.
I found this guide quite helpful for getting the arducopter mission planner software installed and setup. I attempted to install the latest tricopter firmware on my ArduFlyer before connecting to it as they instruct. I found the auto com port detect seemed flakey but after manually selecting the correct port (9 in my case) everything seems solid.
A number of people recommend the Spektrum DX6i transmitter for use with tricopters. It appears to go for around $110 shipped and, as the name implies, has 6 channels.
Further research indicated the Turnigy 9x is a better buy. It is only $54 from hobby king and comes with an 8 channel receiver. This transmitter utilizes an atmel microcontroller and has aftermarket er9x firmware available. Additionally, it is very popular to mod (backlight, better radio module, etc.)
Hobby King also provides a newer Turnigy 9xr. This model aesthetically didn’t appeal to me as much. Apparently it also has less internal space for putting in mods and isn’t quite as ergonomic.
The RCExplorer Tricopter V2.5 appears a strong option. Unfortunately, when I was ordering bits the motors they recommend were not in stock. Not wanting to munge my motor/esc/batter mix, risk being under/over powered or otherwise messing up, I decided to skip this design.
Upon further research the INFINITY T3 – DIY Carbon Fibre Tricopter looks like a great option. It appears to have excellent 12-15 minutes of battery life and looks to be an excellent platform for putting a GoPro on.
The official design is made primarily out of carbon fiber. My initial go of it will be out of wood to reduce expense when I inevitably crash it. Once my flying skills improve I’ll upgrade the frame.
The T shape appears to provide better visual orientation than the Y designs which is a plus.