After flying the QAV250 micro quadcopter for several months I was lusting after something bigger and more stable. I settled on GetFPV‘s QAV400 since I’ve had such a good experience with the GetFPV store and with the QAV250 quadcopter frame.
Instead of using the OpenCC3D flight controller for my QAV400 (which I’ve used and loved on my 250), I opted to set the 400 up with a 3DRobotics ArduPilot Mega (APM) 2.6. I had the APM sitting on my shelf after using it for fixed-wing planes about 6 months ago, and really wanted to try an autopilot with a quadcopter.
I’ve spent the last week building and tuning my QAV400 with the APM 2.6, and it’s really stable now.
Today I flew my first autopilot mission, controlled almost exclusively using a Samsung Galaxy S3 phone (wow!). Check it out:
Unfortunately I lost my first QAV250 quadcopter in some mountainous terrain last month. I was able to snag a used one online, so we’re back in action! This is one of my first test flights — now using a GoPro Hero 3 Black for production video.
A few weeks ago I flew my Lumenier QAV250 mini quadcopter around City Park. I threw the video into iMovie and did some quick cuts, transitions, and audio edits. I’m hoping to do a lot more of these in pretty places over the next year.
TL;DR (the short version)
For under $500 in parts I put together a remote controlled plane that I can fly without direct line of sight. The plane has a camera that broadcasts its image in realtime to a 7″ LCD mounted to my remote control. It has about a 1-mile range and 20-25 minutes of flight time.
Here’s my friend Jimmy flying it for his first time:
When I was six years old I watched Home Alone 3. The boy in the movie stuck a camcorder on a remote control car and drove it by watching the video feed on a TV in his attic. He used it to spy on the bad guys. My god, that was so cool.
Ever since I was six I’ve been trying to re-create that setup. I’m 22 now, so it’s been a few years. In my teens (2002-2009) I obsessed about the project, poring over parts on eBay, but the gear was just too expensive. Today, as result of the increase in RC hobby popularity, the prices are affordable now.
Over the past 6 months I learned how to pilot RC planes entirely by flying and breaking them. I’ve gone through 4 planes, and I don’t regret it one bit. In the long process of developing an autonomous UAV (unmanned aerial vehicle), I’ve built a plane that I can pilot by watching a live video broadcast from a camera on the plane via an LCD screen on the ground. Yes, it’s a drone. In the remote control enthusiasts world, this is called First Person View (FPV). The entire bill of materials for my first FPV system is under $500:
Plane: $60.02 – HobbyKing Bixler 1.1 – Link
Speed Controller: – $7.18 HobbyKing 20A ESC – Link
Remote Control: – $69.97 Turnigy 9X 2.4GHz 9-channel transmitter – Link
Camera: $109.99 – Boscam Explorer HD19 – Link
LCD Monitor: $30.65 - 7″ Car LCD from Amazon – Link
Video Radios: $54.99 - Boscam 5.8GHz Video transmitter + receiver – Link
Batteries for Plane and Video Ground Station: $7.89×2 - Turnigy 2200mAh Li-Po – Link
Battery Charger: $24.07 - Turnigy 50W Balancer/Charger – Link
Building the system
In the world of RC this is a pretty simple setup. The plane is a cheap foam model, which some people may scoff at, but I love. The reason: when this thing crashes marvelously nose-first into the ground, the foam body absorbs the impact. You’ll be gluing and duct-taping this plane body back together many times, but you’ll never have to replace the electronic parts. In the old days our parents built their RC planes out of balsa wood (or some other ancient material like it). They spent 20hrs building the plane so that it was just right. Then it crashed into a tree and took 30hrs to repair. I’d rather tape-up and keep flying.
Most of the RC gear is plug-and-play, though you’ll have to solder some connectors to get the speed controller plugged into the plane. If you’re not comfortable soldering, take all of the parts to your local hobby store and they’ll solder it for you. By now I’ve learned how to solder connectors for every component in my system, but when I got my first order from HobbyKing I went straight to the local hobby shop. They soldered all of my connectors for me, gave me clear directions on flight prep, and even showed me how to work the transmitter.
The video gear should also be relatively straight forward to set up. Plug the video receiver into your 7″ LCD and fire that puppy up! Plug the video transmitter into your Boscam and fire both up! I power my video tx (transmitter) and camera from the same battery that the plane is powered from. I created an extension cable for the battery pack and added a JST connector to it that powers the video gear. You can do this, or you could power the video components from their own battery to keep things clean.
I’ve tried various methods for mounting the camera. Using a rigid mounting method like popsicle sticks or a plastic mount often results in “jelly” video like this. I’ve found that the cheapest, most flexible, and highest resulting quality method is to use lots of small strips of Gorilla tape:
Building your first FPV system is all about experimentation and learning. For your first system get some inexpensive gear and see if you really enjoy it. Then, start upgrading parts as you go. The setup that I’ve detailed here is certainly not professional-grade, but it’s a testament to how inexpensive the hobby has become and how much fun you can have at such a low cost!
I was playing with this system back in September 2013, and since then I’ve dug much further into the hobby of amateur FPV and drone control. Today I’m testing full autopilot systems that control the plane autonomously, and for FPV I’ve set my 7″ monitor aside and now use a set of goggles with LCDs built-in. I’ll be documenting more in due time, but I wanted to get a post up about my initial setup. I think some of those first FPV flights are what fully sucked me into the hobby, and I’m so glad they did.
Here’s one of the most beautiful flights we’ve had. My friend Jimmy is flying over Horsetooth Reservoir in Fort Collins, CO:
I’ve been bringing up self-driving vehicles in conversations with friends lately. There are some interesting potential benefits that would arise if self-driving cars were universally adopted. Here are some that I find fascinating:
Valets, shuttles, taxis, and parking be damned!
With a self-driving car there’s no reason to park the thing yourself. You also won’t have to find parking in a crowded area (e.g. San Francisco). The car can drop you off at the front door and go find its own spot to park. Then you can beckon it with your smartphone once you’re ready to leave (read: Uber integration). No more need for fancy valet services. Airport shuttles will make little sense if your car can drive you to the airport and then park itself back at home. Taxi drivers also seem unnecessary.
One car is better than two, or three
A self-driving car can drop you off at work, drop your wife off at work, drop the kids off at school, and then park somewhere equidistant to all three locations. Unless there’s a long-distance trip that only one person is going on, a self-driving car can do the work of several cars. Most families could probably get by with just one.
Never again: “can you give me a ride?”
My mom spent all of my childhood driving me to friend’s houses, or picking up friends and bringing them to our house. I know, I am super lucky. However, a self-driving car could have done that and would have taken zero of her time.
No more traffic lights. No more pedestrian walk signals
Google’s and Stanford’s self-driving cars already have thousands, if not millions, of real live miles logged in California. The cars collect an incredible amount of data through multiple sensors, and they’re programmed for safety. On a street with only self-driving cars, there’s no need for a pedestrian crosswalk. Just walk or bike across the road and all of the cars will stop for you automatically.
Goodbye jams, goodbye road rage, and goodbye ‘thank you wave’
Computers are typically optimized for efficiency much better than humans . A highway full of self driving cars that can communicate with each other could solve a significant portion of traffic jams. No more need for the ‘thank you wave’ from Seinfeld. If all cars on the road know that you’re taking the next exit, they’ll make room for your car when it’s time to get off.
I think that I’m an OK driver — but I’m positive that a computer collecting terabytes of data per second about the surrounding environment is better. I have to look in my rear-view mirrors to see what’s behind me, which takes my attention away from what’s in front of my car. That’s not so with a computer.
I’m skeptical, sure, but honestly I can’t wait to safely text and drive. Or eat and drive. Or _____ and drive.
Just uninstall (delete) the Android File Transfer tool.
This was my first year attending Y-Combinator’s Startup School. It won’t be my last. The line-up of speakers had a lot of hot names, but every talk was interesting. Some were inspiring. Here are my favorite snippets from each speaker:
- Phil Libin (Evernote) — Surround yourself with people that are smarter than you. Persist.
- Dan Siroker (Optimizely) — Use the hell out of your customer’s feedback.
- Ron Conway (SV Angel) — Entrepreneurs have to mature at an alarming rate to keep up with their company. Learn quick.
- PG + Sam Altman office hours — Be open to letting your idea evolve.
- Chris Dixon (A16Z) — The best idea for a startup looks bad to everyone else. How ironic. Knowing a ‘secret’ is very important.
- Diane Green (VMWare) — Make people’s lives easier. They’ll love you for it.
- Balaji Srinivasan (Counsyl) — Change comes from voice within or voice by exit. Voice by exit is powerful en masse.
- Chase Adam (Watsi) — Work on something that you care about more than yourself.
- Mark Zuckerberg (Facebook) — Hire people that you would want to work for.
- Nathan Blecharczyk (Airbnb) — Pick ONE metric. Focus on it.
Heads up: I’m pretty much regurgitating what Paul Graham said in this interview: http://www.youtube.com/watch?v=BDA0t49AaZ4&t=7m37s. Go watch it, it’s phenomenal. This post is my own experiential take.
I’ve spent the last six years looking for coders to build my ideas. In 2009 I stole a friend’s business model for a Photoshop to HTML service. I even went as far as hiring the SAME web designer he used to build his site. That business failed because I was a bad friend and an idiot, but it was failing anyways because I couldn’t write a single line of HTML or CSS.
In 2011 I hired a few iOS programmers to help me build apps that I had ideas for. The simple apps got finished and did just okay. The complex app that we tried to build got abandoned and rejuvenated twice. It never shipped. That was 50% bad team leadership on my part and 100% me not knowing how to build anything for our app. I was the ‘visionary/marketer/team leader… yeah, right.
There are a handful of other projects where, if I knew how to code, I could have at least built a shitty prototype in a few days. Instead these projects took months, and the things that other people built for me were often shitty anyways. Full disclaimer: I still can’t code, but I’m learning Ruby on Rails and it feels incredible to know that I’ll be able to build from an infinite box of virtual legos soon. Here’s why I think every entrepreneur should learn how to code:
Being the “idea guy” is bullshit.
I’ve known one person that can pull this off. I can’t.
There are a million ideas out there. Yours might be great, but what really matters is execution. Great idea + bad execution = crap. Bad idea + great execution = gold (maybe) — because likely along the way of executing, your idea will evolve into something that’s actually valuable. Either way, ‘maybe gold’ is always better than ‘definite crap’.
Knowing how to build things greatly improves your chance of executing well. It also just makes you so much more valuable to the company.
Things you build yourself will always be better.
Your idea hopefully alleviates pain somewhere. Your intimate knowledge of that pain is your biggest asset. When you try to outsource the building of your pain killer, you have to describe the pain to whoever you’ve hired. They never experience the pain for themselves. How fucked up is that? The people building the solution don’t even have the problem. Yikes.
This is why we always hear that successful companies came out of tools that the founders built for themselves. When you build a solution for your own pain, you build it right and you build it fast. Then you make it better.
Coding is the fastest and cheapest way I know of to build magical things.
Building tangible magical things like electric cars and reusable rocketships takes a lot of people’s money and time, and a lot of bureaucratic paper movement. You can build great software by yourself in your basement with a mediocre computer.
Almost every day in 2013 I find a new plugin or website that makes me say “holy shit! I can do X ten times faster with this!” That’s valuable magic. As starting entrepreneurs ( ones without gobs of money) building software is the fastest and cheapest way to create real value for users.
Entrepreneurs are heralded as scrappy individuals with great ideas and great hustle. They bring people together to solve problems, and that’s fantastic. I’m just learning, over time, that an entrepreneur who knows how to build version 1 of his/her product will have a business with customers way faster than the entrepreneur that has to hire a web design firm to do the same thing.
As my friend Scott would say “A hacker that can hack code and can also hack business is dangerous!” I sure wouldn’t mind being dangerous.