“A journey of a thousand miles begins with a single step” Lao Tzu
So it is with this project. It’s arguable that we took those first steps a long time ago, but we’ve hit a milestone in ordering much of the internal electronics needed for the drone. I wanted to do a write up on the parts that were ordered and our logic behind them. There is a considerable amount of time spent behind the scenes doing research before leaping into any of the purchases, and we could probably spend more time doing research, but at some point you have to take that first step.
To start off our budget for the project is $1,100. This is the amount of money that my employer is able to put into the project. I will purchase some of the cheaper stuff myself, in the interest of me seeing the project through, but the major components I’ll have my employer purchase. That budget may seem like alot, but you see how it quickly gets eaten up.
The Flight Controller
For this build I decided to go with the Pixhawk 2.1 otherwise known as the Cube. Pixhawk flight controllers are well known and have quite extensive industry support. This is important as if I run into issues then there is likely others who have run into the same issue, and can help me work through it. I could have gone with the original Pixhawk, or a clone, and saved a few bucks, but I decided to get the Cube for it’s redundant features.
I fly over tails ponds, active mining pits and other potentially hazardous environments. I would rather have the security in knowing that if something goes awry with the controller that there is a good chance I’ll still be able to bring it in. We also fly in cold weather, and having heated electronics, will extend the operational range of the drone to be able to fly on those cold windless days.
The typical base price I have found for these controllers is $235. They can be sourced from a variety of places, I purchased mine in a package that included the GPS, and the Radio Telemetry package.
GPS
When it comes to electronics there is a benefit to going with the same manufacturer. Typically all of the components speak the same native language, or the manufacturer has figure out how to get component 1 to speak to component 2. So I was incentivized to stick with ProflicNC who makes the Pixhawk 2.1 for the GPS sensor.
With our Ebee we used ground control points (GCP) that were surveyed in by our mine surveyor. This saved the extra expense of the most costly RTK version of the Ebee.
Over the last three years, I’ve built up an impressive catalog of GCP’s across our mining site. So when I go to conduct a flight, I spend a few minutes to freshen up the points with a little paint. Usually I can conduct several months worth of flights before needing to go back out and do GCP maintenance. The nice thing about doing aerial surveying is I can tell when the GCP’s have been disturbed or are in need of paint from the comfort of my desk.
RTK GPS sensors have come down in price, but I decided not to get one at this time. It may be an upgrade in the future but to prove the concept we wanted to match the Ebee performance and not exceed it. The Here+ kit that I saw was ~$600 so six times the price of an Here 2.
Again I bought this as part of a kit, but I am sure the kit price is similar to the individual price.
Radio Link – Telemetry
Can you pre program a mission, launch the drone, and forget about it until it comes back? I believe so, but that is definitely not the way you want to go. Most of the time there isn’t an issue, the drone takes off, does it’s thing and comes back when the battery gets low, or there is an issue. (The Ebee was really good about this) Sometimes Murphy wakes ups, and the mission suddenly does not go as planned.
In these incidents you need to know where the drone is, call it back, adjust the mission plan or know it’s last known location. For this you have to have some kind of radio link, some way to talk to the drone. Most of my flights have mixed terrain of large flat areas with excellent line of sight to large mountains which obstruct the line of sight. Since we are in the desert, there is very little in the way of foliage that will get in the way of a good radio link.
The biggest challenge I have is good landing areas. The ground is covered in sage brush, and sagebrush can do an impressive amount of damage to a drone. Sure we can retrieve it, but I’ve had it break spars, props, and poke all kinds of holes through the foam wings. The roads seem like an obvious choice, but they act like one big long strip of sand paper. For these reasons I try and find something soft to land in, like a field of cheatgrass, or on a strip of vinyl liner.
For this reason the drone often has a bit of flying to do to get to the mission area. We need the radio link to cover the entire mission area and the journey to and from the mission area. From the home point to the furthest point being photographed. This can sometimes be a little over a mile.
As with the GPS we looked for a Tx/Rx combo that would talk with the Pixhawk 2.1, and work with Mission Planner. Being that this is our primary communication device with the drone, we want it to just work, with minimal effort.
The RDF900 seems to be a popular choice, and one that is offered by ProficNC for use with the Pixhawk 2.1. The Tx/Rx bundle is offered in a few different configurations for $218 to $250 bucks depending on the configuration and who it is purchased through.
The Motor and ESC
The design I downloaded from Flite Test, and I relied on them pretty heavily when looking at what to get in terms of the motor and ESC. They recommend their Power Pack C. This includes the motor, servos, mounting hardware and ESC. I decided not to go with the whole pack, but did go with the recommended motor and ESC.
Despite my suspicions that the motor is oversized for our needs, I decided that in this case more power is not going to hurt. We’ll pay the weight penalty and use it as a starting point. Motors are inexpensive and if we want to downsize later, we have that option.
We also purchased the Flite Test ESC that is recommended for this motor.
I can appreciate companies like Flite Test. They do offer free plans for download, and sell everything to support their creations. I want to give them my business in order to support what they do.
Batteries
The eBee batteries have a custom connector, and I am not sure how to wire them to be used for our new drone. It is something I’ll likely try and do down the road. It’s a shame not to use them, they are expensive $100 per battery. Some of the batteries I have have only 10 or so flight hours on them, and I’d like to get my money’s worth out of them.
In looking at batteries I really debated on what I wanted to do. I ended up ordering the recommended batteries from Flite Test.
The battery provides the ballast to get the GC right on our aircraft. So we have to be cognizant of how large the battery is and where we place it on the aircraft. I could go with a larger battery, and according to the CAD model, it would be fine. However larger batteries are more expensive. (Overall dollar amount, but about the same price $/mAh)
With the eBee we were getting 24-26 minutes per flight on a calm no wind day with a 2200mAh battery. We can expect to get less that that with our drone. Our drone is larger, we are going to be heavier, and the power consumption of our individual components will likely be higher. However we do have enough room to fit two batteries, wiring them in parallel, if needed. So we have some options.
Batteries are one of those things we will play with once the drone is up and flying. It’s a variable that’s easy to change. Which is why after toying with the idea of getting 4S 3300mAh, I decided to go with the 3S 2200mAh batteries.
Camera
The core of a photogrammetry drone is the camera. Otherwise you just have something that can fly in cool patterns in the sky. With the Ebee we used the S.O.D.A camera system. However when we first bought the drone it came with a Sony DSC-WX220, which worked fine for over a hundred landings until the dusty environment of the desert finally did it in.
Looking at what has been done before has been one of the driving factors of this project. So I looked to see what I could find on the market that was inexpensive, and might be modified to achieve what I need.
Proper cameras are going to fall in the $400 to $500 range, once the camera and lens is purchased. I have several “Point and Shoot” consumer level camera that I can work on modifying and practice on. In my book it’s worth a shot, and I’ll likely learn something. If it doesn’t work we’re out $100 bucks. It’s not much of a gamble.
Other Sensors
The last sensor I needed, or think I need, is an air speed sensor. Air speed is important and the air moving over the wings provides lift, and the air moving over the control surfaces provide control of the aircraft. The flight controller needs to know how fast the drone is moving through the air, and adjust the control as needed. For this we need to have a sensor. Ground speed isn’t a good indicator of Airspeed.
The eBee used the same sensor chip for measuring differential pressure and had a plastic Pitot tube. The cost for this was $44 dollars.
The Last Expensive Piece
The last piece we bought was a Radio Controller. I do not have confidence that i can get the drone to land it self, and I am not entirely convinced that we want it to, given some of the challenges we have had with the Ebee. For this reason I elected to purchase a controller, and we’ll have to learn how to land this sucker.
Again I leaned on Flite Test’s recommendation for this purchase.
Even though we have an autopilot on the drone, having the ability to take manual control is going to help prevent some of the sticky situations we had with the eBee. Although it may also get us into some interesting situations too….so maybe it’s sixes?
Ebee will void the warranty if you take manual control to land the drone and damage results. It make sense, but it also discourages you from taking manual control when you probably should. Why risk having the warranty not cover the expensive repairs, just because you think you can do it better? The Tx/Rx was ~$200.
The Conclusion
The overall cost of the drone is more then electronics that goes into it. If I spend less than 100 hours getting this up and running, that will be a small miracle. I am likely to spend several hundred hours of my time, conducting flight tests, making adjustments, and repairing the drone. This amounts to several thousand dollars, likely more than what it would cost to have the Ebee replaced. At this time, we are willing to put that time in, and I am willing to do what it takes to learn and develop something. In all we squeezed in just under the $1100 budget.
There are other components out there, and if you are willing to piece together something from the cheapest components you can find, you can save a few hundred dollars. However I did not want to spend the majority of my time getting Component A to talk to Component B. I did want plug and play solutions as much as possible and that went into a lot of my decision making.
Hopefully if you are looking to follow in my footsteps on this project, this write up is helpful. Keep in mind as this project progresses we may find that we missed parts that we needed, or that one component might not work, and we may swap it out. So don’t consider this a final parts list.