Here's a short update on ideas for the octocopter I'm going to make. I haven't spent too much time thinking of ideas recently as I've been too busy with basic electronics instead.
So I'll start with the mechanical ideas. The plate that is called on previous posts, the sprung plate, I think might need to be made of aluminium. This is due to the shear and axial forces it will be put under with the weight of any payload being bolted onto this while its only sprung to the lifting skeleton. I think these springs that hold the plate on, should also be adjustable as the payload may vary (different cameras etc). These could also be electronically adjusted so that it could change on the move or automatically.
I have also found out why (well I think why), ducted fans aren't used as much as they could be on multicopters. This is due to ducted fans introducing weight and high speed drag penalties.
My next idea is something which I think will improve efficiency and make the use of something for one job, do another for free. Basically as I've found out I need an inverter to convert from DC power to AC power for the brushless motors that power the props, I need FETs that will be switching relatively fast. Due to this the FETs can get quite warm or hot. So the idea is, to make the stators out of aluminium. These would then obviously be cooled by the air passing from the rotors. This should hopefully improve cooling characteristics. As the frequency increases the switching of the FETs, they heat up. This though, also increases the fluid flow over the stators and thus taking away more heat.
Next RF or wireless communications. So for RF communications for the multicopter, I think I will use two channels, one for control and one for streaming video or other data. This is because the lower ISM bands are narrower than the higher ones and so the bandwidth of video can't be transmitted. On the other hand the lower frequencies travel further so are much better for control. Specially seeing that video could be stored on on-board SD cards. So in Australia, there is no 400MHz and the 900MHz band is smaller than elsewhere in the world. The 2.4GHz is as normal. So I think I will have to buy the chipsets and set them up specifically for the 900MHz band of 918 MHz to 926 MHz and use standard chipsets for the video streaming on the 2.4GHz range. Alternatively use HF radio for controlling.
With the 2.4GHz communications, it would be great to make the aerial directional by possibly using tracking with the help of GPS, accelerometer and/or a gyroscope. I also thought maybe put these in the controller so these can be relayed back for interpreting of where to angle the aerial.
So when sending data over these RF communications, I need to validate the data, specially the control data so to avoid malfunctioning or hacking. To do this I think it would be good to use cyclic redundancy checks on all data and then authentication and or encryption.
I think I might be able to make the PCB stackable. This would be good as I can seperate different functions out and make it sort of plug and play and much easier to upgrade individual parts without having to upgrade the whole lot. These could communicate via I2C, SPI or CAN.
Having a separate video board would be good. It would mean that processing of video could be done independently and it wouldn't use all of the processing power of the main controller. Hardware encoding/transcoding would be good too as this is a very computational intensive process.
I have also had a look into sensors for video. They are quite hard to find in my opinion, but I managed to find some on DigiKey. I dislike DigiKey as their website is horrible and its just much harder to use than Element14/Farnell. Thankfully these sensors seem to support what is called Camera Interface. This is supported on most of the STM32 Chips (called DCIM) and likely on other brands so should be relatively easy to setup. The OmniVision sensors come in lots of different resolutions upto sixteen mega pixels.
Finally, I think it would be good to implement a safety (and security) measure of using RFID cards to validate and use as a sort of start key. In essence, without the RFID card in the copter, the copter won't start.
Well thats all for this post. I should have some programming tutorials up this week. Next week I will be back at uni but hopefully will have some time to spare to play round with making PCBs.
