I have recently been pushing the limits of the XBee Pro to see how far it can go.
Unfortunately my current plane (SkyFun) only has a 38 minute duration and a maximum travel distance of 20km, this means the furtherst I have been able to go is 10Km (a 20Km return trip).
I have a 14dB patch antenna which has still shown signal of 85%+ out at this distance.
I wanted to find out how far the XBee could really go.
This meant planning a one way mission.
I spent about a week looking for the perfect start and destination, plotting the course and arranging access to the landing location with a local farmer.
The start point was my normal local park and the landing location was a farm paddock 16Km away.
The flight was over sparsely populated farm land and at a constant incline meaning I could increase altitude during the flight in order to get greater range.
My plan was to launch and monitor the flight from the launch location until I lost contact with the plane via Telemetry. After this I would drive to the landing location where the plane would be circling and manually land.
I calculated that even if I maintained contact with my plane during the flight and it made it all the way that I would still have time to make the 12 minute drive without running out of battery power.
With a tail wind, the plane would average 60Km/h at 45% throttle.
With everything planned and 45 test flights already in the bag with this airframe I was ready.
I launched and the plane took off to the waypoint as planned.
I was hoping for at least 12Km as this was ideal conditions.
The point on the map above shows the point I lost contact which is 10.6Km from the launch location.
The XBee maintained a very strong signal up until the last Km and then it began to drop off very rapidly.
It begins to be unusable below about 40%.
I packed up the ground station and drove to the destination which took just 11 minutes.
The plane was circling perfectly as expected and I put it in FBWA and landed without incident (apart from the slightly startled flock of sheep).
It turns out that the 10Km I had been flying to is around the maximum range of an XBee.
I am working on a much bigger aircraft that will have a 1 hour plus duration at a higher speed so I am going to have to look into more long range options for telemetry. GPRS / cell is preferable as we have excellent coverage here and fully routable and cheap internet plans. However, for experimentation I have been surprised at the distances that can be achieved with an XBee with a cheap patch antenna added on.
Here are a couple of tips for people wanting to achieve maximum range..
- Minimize antenna cable length. My XBee is mounted on the back of my patch antenna and I run a long USB to my laptop. This maximizes gain and reduces lost signal.
- I use either the ground or my car to enhance the ground plane effect. Putting the antenna in front of my car door or right beside my car makes a big difference to the distance. Also keeping it as close to the ground as possible.
- Altitude is king, the higher you can fly the greater the achievable distance. Due to the fact my destination was on an incline, my end waypoint was 400m high relative to my starting point.
- Fly long straight missions so you can line up your antenna properly, I am building an antenna tracker that removes the need for this but Ardustation2 is currently broken since the last Mavlink updates.
Comments
So you've obviously disabled the RC failsafe RTL function. As I don't think anyone has set up an Xbee based failsafe, I have to think that flying in such a condition is inherently unsafe. Believe it or not, most of the UAV's operated in theater right now DO have a geek on an RC transmitter. Not the big boys, but the smaller ones (shadow, silverfox, raven, etc.).
But it brings up a really good thought: Xbee based failsafe. How hard would it be to implement a lost data-link failsafe? At least in the US, this is required to even be considered for a COA or experimental certificate. I'm not even saying that RTL would have to be the mode, just so that if all communication with the aircraft is lost, I KNOW what it is doing.
To me there are pros and cons associated with going cellular. The pros are the obvious lack of having to worry about frequencies and licensing and all of that...but to me a real drawback is that I REALLY like to fly far out away from people when I am testing / playing with this technology, and there may not be good coverage out in the middle of nowhere.
@Grips - If there IS good GPRS / Cellular coverage where you intend to fly it might be a good solution to your licensing / spectrum concerns.
@Simon
Yes, they are normally NAT'd but this can be a good thing...
If your APM is behind a NAT and your laptop has a data card on the same network, then all IP addresses behind the same NAT are routable. That means you don't actually need internet routable addresses. You can just connect to the device as you are both behind the same NAT.
If this doesn't work, then you can ring your service provider and ask to change the APN that your SIM is on to an Internet routable SIM. A NAT'd SIM is usually the default for simplicity but ALL providers have internet routable APN's available for customers to use. In some cases, you don't even need to ring your provider, you do a google search and look it up and change to using that APN and your in business.
There is usually no charge for this.
Here in NZ all our providers have fully routable non NAT'd plans by default.
I have past experience doing this in both the USA and Australia (for my real job which is GPRS telemetry related).
I really think GPRS is the way to go!!!
Great work, Toby. This is the type of article I love to see here. Well thought out, well planned, and equally well executed - yielding useful information and a great conversation. Keep it up. Have you considered the (relatively expensive) microHard or Digi 9Xtend modems?
Good work Can you test the range with 3dbi dipole and patch?
The big boys with the predictors and reapers land and take off locally because of the latency of the satellite link. Different cause, but same result, control lag or latency. All most guarantied to cause a PIO (pilot induced oscillation).
I would never give up the RC link. I would enhance it(dragonlink ?). Going thru 2 computers and Xbees also leaves a lot to go wrong as well as adds control latency. It's not the distance, it's the processing and Xbee. I would always want a hardware fail safe,but hey, that's just me.
I suppose it depends if you consider your project an RC plane with an onboard autopilot or a drone.
Personally, I have other models I fly for the joy or RC but for this project I consider RC as an unfortunate backup that must be used. The title of the site is "DIY drones" not "DIY RC models with autopilots" :)
If the autopilot is built with enough redundancy and smarts then you should never need to use RC. It's a safety net many people struggle to let go of but I have done over 40 flights only using the RC link to flare for landings. If we got sonar working it would not be needed at all.
@ Julian,
10 km / (299.8 km/ms) = 0.0343 ms.
You would never see it.
Does anyone know of a MS lag time for transmitting over a distance like that? Im working on a project thats looking to do processing at the ground station level and send flight commands back to the airplane, wanted to see if this is unrealistic or not...
-
1
-
2
-
3
-
4
of 4 Next