DJI Phantom and Wind

This week I took my drone to video Initial Point in Idaho. It’s a rocky outcrop about 20 miles south of Boise that was chosen as the initial survey point for the Idaho Territory in 1867. It has a concrete platform installed at the top with a survey marker embedded. There’s a rocky trail that can be driven to the top if your vehicle has a high ground clearance. I chose to walk to the top.

DCIM100MEDIADJI_0177.JPG

Initial Point Idaho

My car was parked at the base, I climbed up about 130 feet to the top. The wind was occasionally gusting at the base. It was constant at the top, with much higher gusts.

I carried the drone in my backpack to the top, and launched it from the platform. I manually flew it around the point, but was feeling extremely nervous doing so. I only flew about 36 feet above the platform during the entire flight.

The concrete platform has a metal railing surrounding it, and I didn’t trust the drone to return safely to land without hitting the railing so I manually had it land nearby. As I was hovering the drone before landing, it was holding a fairly constant 20° lean because of the wind.

After hiking back to my car I still had plenty of battery for another flight. From near my car I launched the drone and flew vertically to about 200 feet, centered the drone over the survey marker, and used the point of interest feature to create a video circling the point.

I don’t believe that the wind was any less on my second flight than it was on my first flight. The fact that I was in the wind on the first flight had me feeling significantly more nervous while flying than when I was out of the wind on the second flight. Watching the video from the first and second flight doesn’t appear significantly different. I would have liked the point of interest video slightly more if I’d been on top of the point the entire time, but I was too nervous to do it all while I was sitting in the wind myself.

DJI Battery Monitoring

DJI charges a significant amount of money for their batteries and calls them Smart Batteries. I’ve seen the statistics reported of how many times the battery has been power cycled, as well as details of how much flight time is available. This was a new and interesting feature to me.

Screenshot_20160502-075201

I had left the drone in the back of my car overnight. The temperature had probably gotten into the high 30s, and was still in the mid 50s with the sun shining. The warning message “Battery Temperature Too Low. Warm battery to at least 15 degrees Celcius before flying” came up on my screen and would not let me initiate a takeoff.

I have used batteries in cold climates in the past. I know how temperature affects both current output from batteries and future usability. I’m impressed that DJI has built in this feature to their firmware.

DJI Phantom 3 Standard

DJI has been the standard drone that people visualize when they think of consumer drones for the past several years. It’s a white quadcopter that comes with a white remote control.

This drone is the camera platform I was hoping for when I got into drones three years ago. The flights are very automated, the camera controls are fully integrated, and I can start flying in a new location with very little setup time. The picture below was taken to demonstrate how close the drone came to returning to its launch location and automatically landing after flying over a thousand feet away from where I was standing. The drone had taken off from on top of the ring, and it landed less than 4 feet from the same location.

WP_20160423_002

In early March DJI dropped the prices of their entire line, convincing me to buy the least expensive unit, now on sale for under $500. I ordered it on March 18th, received it on March 23rd, but didn’t successfully fly it until March 29th.

I live in downtown Seattle. I didn’t want my first flight to be anywhere that would draw excess attention. I’ve been flying drones for three years now, and have grown accustomed to things going wrong. I’d planned on flying at the RC field I regularly visit on the evening of  March 24th. When I got there, first I was not able to get the drone to power on, then when I figured out how to do that, I was forced to do a firmware update before I could fly. By the time I’d got all that out of the way I was running late for a meeting. I decided that the safe thing was to put everything away until my next opportunity.

The Phantom 3 Standard is delivered with two sets of propellers, the transmitter / remote control, a flight battery, an AC battery charger, and a few small extra parts related to the camera gymbal.  It is designed to work with an iPhone or Android phone to both control the unit and see what the camera is doing. I don’t have either of those phones, instead I’m using a Google Nexus 7 tablet. The device connects to the transmitter via WiFi. While my tablet is connected to the transmitter it is not able to connect to the internet.

My first flight with the drone was in a Seattle Park. The battery reports that it can fly up to 25 minutes. None of my previous drones would fly for more than about 13 minutes. The controller app on my tablet has plenty of feedback about the battery condition.

Normal flight mode for me has been that I tell the drone to take off via the app. It takes off and hovers about 4 feet in the air. Then I use the sticks to fly the drone to a location telling it to go up/down or rotate left/right with the left stick, and moving horizontally with the right stick. I usually enable the camera in movie mode before taking off, and only take it out of movie mode if I want to take still photos. When I’m ready to finish, I can either drive it back near myself, or toggle the left switch on the transmitter to cause it to go into return home mode. I’ve got it configured so that it will be at least 100 feet above the ground during return to home mode, which is good enough to clear most trees, but I make sure that’s true early in my flight just in case. If the drone loses contact with the transmitter for more than three seconds it will enter return to home mode.

After owning the drone for one month, I also purchased the DJI backpack specifically designed to carry the drone. It has been what really makes the drone fun for me because I can store the drone in the backpack, knowing it’s fully protected, and easily grab the entire thing and throw it in my car to go somewhere that might have interesting things to photograph.

 

FrSky X4rSB Receiver controlling Eachine Racer 250

I wanted to use S.Bus communication to connect my receiver to my new Eachine 250 
Racer drone. I wanted to be able to remotely control the lights on the drone. I wanted to bring back telemetry from the drone to my transmitter. All of those are possible with the X series receivers from FrSky.

20160513_093703

I’ve used the X8R receivers in the past, but I bought an X4R-SB receiver specifically for this drone.

In it’s default configuration it will output PWM on the pins for Channels 1 through 3 and S.Bus on the 4th connector. It has a separate input connection on the side for S.Port telemetry devices and an analog data line.

By putting a jumper on the signal pins, as shown in the picture above, during the binding process, the output is changed so that CPPM channels 1-8 are on the first port, PWM Channel 9 and 10 on the second and third ports, and S.Bus on the fourth port.

I’ve got the S.Bus connection going to the main port on my CC3D flight controller, a cable to control the lights on the third port (channel 10) and the first two ports remain unused. I have the momentary switch SH on my Taranis configured to control Channel 10. Toggling it cycles the LEDs on my Eachine Racer through the  off and on colored states.

The telemetry cable is connected first to a FrSky SP-GPS – Smart Port GPS Sensor (GPS-V2) and daisy chain connected to a FrSky SP-FLVS – Smart Port Lipo Voltage Sensor.

20160513_10061520160513_10062720160513_10162020160513_101628

The GPS Sensor is a new item in my arsenal. I’ve used GPS chips connected to my flight controllers in the past, which allow the flight controller to direct its flight position using GPS. Since this is not connected into the flight controller, it’s purely a toy, though it may help finding the drone if I get confused and crash it away from myself. The last data received in my transmitter remains on the display, which I could then use to assist my search for the drone.

The voltage sensor connects to the balance plug on the battery, allowing me to monitor the state of the individual battery cells during flight, as well as having low battery alerts reported on the overall voltage.

For installation of this receiver in my drone, I’d come across a 3d model for a holder. I had never used a 3d printer before, though I’ve been fascinated with them for several years. I realized that Windows 10 has a 3d Modeling program installed by default, 3D Builder, and that it can print using an on-line service.  Because the service accepts the file and tells you how much it is going to cost delivered, this was an easy first try. I downloaded files for both a battery protector tray and the previously mentioned receiver tray, merged them into a single model file, and had them delivered for $27.15. Other than the time involved for the delivery I was happy with the result. I submitted the order on 4/26/2016. I received notice that it shipped on 5/10/2016. It finally arrived on 5/13/2016.

20160517_17394420160517_17375520160517_174014

The design here has the receiver extending in the model over the flight controller and under the video transmitter. I’m not positive that I’m going to use the battery tray. If I don’t use it, I just need to get screw extenders to install the receiver directly above the flight controller.

Eachine Racer 250

I decided to get into the FPV Racing Drone scene and found this unit available in an almost ready to fly version for $140.

20160513_101937

Almost ready to fly means that I need to add my own transmitter and receiver. Because it’s FPV, it also means that I need my own video receiver and display.

I have an FrSky Taranis transmitter, several compatible receivers, and a set of Fatshark Attitude V2 goggles,  so when the model arrived, I expected all I would need to do was charge the battery and it was ready to fly.

It turned out it was slightly more complicated, but the past three years of playing with drones meant that it wasn’t too involved.

First I found a useful wiki page dedicated to the racer, with plenty of information. It allowed me to understand that I’d got version 5 of the racer, with the difference from version 4 being that they’d removed a safety tray that kept the battery from crushing the flight controller.

I figured out that to use S.Bus to communicate between my receiver and the flight controller, I’d need to download software to my PC, connect the PC to the flight controller with a USB cable, and configure it’s inputs. The instructions I found all referred to OpenPilot. OpenPilot is an open source project that is now defunct, including the domain name itself not going anywhere. The replacement project that is functionally similar, is LibrePilot. I believe that http://opwiki.readthedocs.io/en/latest/user_manual/index.html is the old documentation from the OpenPilot project, while the LibrePilot home page has links to all of the new project.

After downloading LibrePilot and installing the software on my PC, I couldn’t change the settings on the firmware that had shipped on my drone without upgrading to new firmware and resetting all of the configuration.  After some hesitation, I allowed it to upgrade, then followed the wizard to configure the new firmware. I was impressed at the ease of setup, and later when I flew the device, it proved that the defaults worked nicely out of the box.

FirmwareMismatch2016-04-20 (1)2016-04-20 (3)2016-04-20 (4)

The original firmware had a rose icon, while the new firmware shows a warning icon. That could mean something important, but I never found any mention and the system seems to be working correctly. You can see that the drone shipped with a firmware dated 2015-03-12 and the new version is dated 2015-10-21.

Connecting the receiver to the flight controller using the S.Bus connection required using the MainPort connection on the flight controller. An appropriate cable shipped with the device that included 4 wires, colored black, red, green, and yellow. For S.Bus operation, the green wire is not used. I removed it from the flight controller side of the cable. When connecting to the receiver S.Bus, Black = (-) Red = (+) Yellow = (Signal).

20160517_173734

My goggles only support 8 channels. The drone video transmitter supports 32. Finding a match required a bit of reading, and then deciding on what was least likely to cause conflicts with other people near the field I regularly fly at. I chose D7 on my transmitter and CH6 on my goggles, which worked out to 5840MHz.  The newer version of my goggles supports 32 channels. (8 channels, on each of 4 bands.)

I found that http://www.rcgroups.com/forums/showthread.php?t=2266883 has a very nice explanation of the frequency bands used for FPV including details of how they overlap and recommendations for which frequencies to chose for the least interference between racers. It includes a google docs spreadsheet that’s been color coded to have the frequencies sorted in ascending order and make the bands more visible.

I got the video transmitter and receiver in sync by selecting the channel on my goggles, then cycling through all of the possible channels on the transmitter until I got the clearest video picture. My goggles auto select NTSC or PAL depending on the signal they receive. A friend using a full sized monitor wasn’t as lucky. That’s how I figured that the Eachine Racer 250 ships with a PAL camera.

The Eachine Racer 250 has a pair of bright white LEDs on the front, one on either side of the camera, and a LED lightbar on the back. The LED light bar on the back can be cycled through a series of colors by sliding a power switch located on the left side of the main board that turns all the LEDs off and on. There is also a two pin cable connector that can be plugged in to a receiver PWM output allowing remote light control.  To use remote light control, the local switch on the drone must be in the OFF position.  I’ll write more about this in a separate post.

After going through the LibrePilot setup wizard and putting the appropriate propellers on each motor, the racer flew completely as expected. It’s very responsive, and also very resilient to the basic crashes I’ve had so far.  The biggest learning experience for me has been to add throttle when I think I’m going to crash on the ground. With my larger drones, I’ve wanted to stop and recover the drone when it hits the ground. With this drone, it is much more likely to bounce and be able to recover itself if I can get it off the ground.  I recommend this drone as a good entry into the FPV racing drone scene. I’m sure that there are plenty of drones that are more resilient or responsive, but there’s also plenty more that can be spent than I did on this.

85% Processor Usage while idle

My Microsoft Surface Pro 4 has recently started showing high CPU usage even while I’ve got nothing running in the foreground.

It’s annoying because I’m used to the idea of lower CPU usage being related to extended battery life, and if the CPU is being used by an unnecessary program, I’ll stop using that program.

2016-05-15 (1)

The other interesting thing is that Microsoft Edge is showing as a significant user of both CPU and Memory. I’ve not yet launched Microsoft Edge since I told the operating system to restart.

Is this some placeholder in the Task Manager?

If I run the Sysinternals Process Explorer side by side, it shows the machine being very lightly used, which is closer to what I expect to see from task manager. If I want real information on what’s going on with my computer I’m more likely to use the Sysinternals tool, but the task manager is already installed on all windows machines, and has some simple graphs to look at, including watching network traffic.

2016-05-15 (2)

SpeedTest.Net results from different devices

I’m visiting my parents today, and one of the normal things I run a check on is the condition of their internet.

I’ve got the speedtest.net app installed on my iPad. Running it produced acceptable results. 17Mb/s is not great, but it should be good enough to stream HD video, and that’s the main thing I want to just work when I’m not visiting.

SpeedTestBoiseiOS

I brought up the website in my browser on my Microsoft Surface tablet and received significantly better results.

SpeedTestBoiseSurface

71Mb/s download is almost comparable to what I’m getting at home. At home I’ve got symmetric bandwidth, so my upload speeds are often better than my download speeds.

Both of these tests were run through an old Cisco RV110W Wireless-N gateway that only runs on 2.4GHz frequencies.

I’ve registered significantly higher speed transfers on my iPad in the past.

Is the iPad limited in it’s transfer speed when running 2.4GHz? It’s possible that the higher speed transfers in my iPad history were all when I was connected to my home router running 5GHz.