Through the first two articles of this series, I assembled the bulk of the Strider Mini Quad frame, installed the propulsion system, and configured the flight controller. This time around, I will concentrate on the components of the First Person View (FPV) system, as well as the camera used to record in-flight videos.
The components that I chose for the Strider’s FPV system are quite common. The camera is a PZ0420 with a 2.8mm lens and IR filter. It mounts directly to the camera mounting plate that is provided in the Strider kit. The mounting plate is then sandwiched between the center plate and top plate of the frame. Since the center plate of the Strider frame features an integrated Power Distribution Board (PDB) there are 5-volt and 12-volt power taps for the camera located directly behind the camera mount. There are also inputs for the video and audio (if your camera has it) signal wires from the camera.
The camera I used does not have audio capability. It includes a 3-wire pigtail for power, ground, and the video signal. I shortened the pigtail considerably to reduce unnecessary wire on the airframe. The camera can accept 5-17 volts, so I plugged the pigtail into the 12-volt tap of the Strider.
My video transmitter (VTX) is a TS832 5.8GHz 600mW unit. Like most VTXs for FPV, it requires a FCC amateur radio license to operate. I attached the VTX to the bottom side of the top plate using self-adhesive Velcro. The rear end of the Strider center plate includes another set of power taps and nodes for connecting the video and audio signals. I again used the 12-volt tap and video signal.
I upgraded the stock VTX antenna with a circular polarized model. I also added a 7cm long extension between the VTX and antenna. The extension provides a flexible link between the antenna and its mount on the VTX. This isolates the VTX from the hard knocks that the protruding antenna is bound to endure.
When you are shopping for VTXs, antennae, and accessories, be sure to pay close attention to the gender of the connectors. Some components use standard SMA connectors, while others use reverse polarity (RP-SMA) connectors. You want your equipment to have the minimum number of connections and adapters, so get equipment with compatible connectors from the start.
At the recent TED conference, computer vision expert Fei-Fei Li explains how she and the researchers at the Stanford Computer Vision Lab are developing ways to teach computers visual perception by studying human vision. Her breakthrough a decade ago: instead of simply improving object recognition algorithms, her team increased the quantity and quality of input fed to the program to the tune of 15 million photos.
This post was done in partnership with The Wirecutter, a list of the best technology to buy. Read the full article below at TheWirecutter.com
If you regularly travel with devices needing Wi-Fi, get Verizon's Jetpack MiFi 6620L. Its battery life is among the best we’ve seen in hotspots, it runs on the largest and fastest U.S. LTE network, and its pricing is competitive.
Just about every smartphone can act as a hotspot, sharing its connection over Wi-Fi with tablets or laptop. But if you work on the road a lot, a hotspot offers a more reliable data connection than your phone and will run for much longer on a charge than a phone in tethering mode. Think two full days of work versus five hours.
AT&T, however, isn’t far behind and in parts of the U.S. beats Verizon. It also ended an advertising scheme to track subscribers’ unencrypted Internet use, while Verizon took until January to announce an opt-out.
The LTE networks of T-Mobile and Sprint, even after recent progress, can’t match the big two’s rural coverage--important in a device used often on the road. (For more on this, check out our guide to the best wireless carriers.)
Android tablets are going through an interesting transition right now. We're seeing the first few hints of 64-bit support, 4:3 screens, and some powerful gaming features. However, these products are still imperfect. I don't think there's such a thing as the perfect Android tablet for everyone right now, but there are a few good ones that might work well for you.
Let's check out all the top tablets on the market and see what they all have going for them.
If you like having access to the latest software and dig the 4:3 form factor, the Nexus 9 might be an appealing option. This tablet runs on a Denver dual-core Nvidia Tegra K1 chip with 2GB of RAM and 16-32GB of storage. The centerpiece is clearly the screen, which is above average compared to most Android tablets. It's an 8.9-inch LCD with a resolution of 2048x1536, just like the iPad. At 8.9-inches, a widescreen tablet would be awkward to use in portrait orientation, but the the N9 is quite comfy.
The Nexus 9 runs Android 5.0/5.1 Lollipop without any OEM junk added. This is Android as Google intended with updates more or less guaranteed for at least two years. The Nexus 9 might fall back to second priority in a year or so when new devices come out, but you won't be left to rot on an old version of Android within the expected life of this tablet. There are also full system images for the Nexus 9 and an unlockable bootloader, making for easy modding (and fixing your mistakes so you don't end up with a brick).
I think the biggest knock against the Nexus 9 is that the build quality simply isn't where it needs to be for a $400 and up tablet. The buttons are a little mushy, the soft touch plastic feels a little cheap, and it's slightly heavy. More recent production runs of the Nexus 9 are much more solid. It still takes a weirdly long time to charge, though.
More problematic is the state of the Nexus 9's software. It's overall a better experience than many Android tablets, but the N9 still stutters and hangs more than it should. Nvidia's Denver CPU core has a lot of power, but it seems like it's not being fully harnessed in the N9. Hopefully a future software update gives this tablet the extra boost it needs to be a better experience.
The Nexus 9 is a good tablet, but it's pricey. If you can find one on sale, it might be a good buy. Even if you can't the form factor makes it worth considering.
Will's on a quest to to find a new car, and is considering an all-electric vehicle. This week, he test drives the BMW i3, a unique hatchback that can run for 80 miles on a full charge. We take the i3 on the freeway, on San Francisco's steepest hill, and test its self-parking feature.
Artist Paul Michael Johnson finally finished the TIE Fighter animated short film that first made the rounds several years ago. His seven-minute story, told from the perspective of the Empire, is animated in a wonderfully nostalgic 80's anime style.
In the first installment of this series, I assembled most of the frame components of the Strider Mini Quad. I also installed and soldered the motors and ESCs. Although the flight controller was installed in the frame, it still required attachment of the various wires and configuration of the firmware within. Let’s focus on those tasks and keep moving.
The flight controller is the nerve center of any multi-rotor. It takes your control inputs and the data from its onboard sensors and translates it all into commands for each of the ESCs. There are several different brands of flight controllers. Considering all that they do, most of these units are incredibly small. The flight controller I chose is the OpenPilot CC3D (CopterControl 3D), which fits perfectly on the Strider’s stock flight controller mount.
From a wiring standpoint, the flight controller is situated between the radio receiver and the quad’s ESCs. First I attached the ESCs to the CC3D. The CC3D has a bank of pins that accept the standard receiver plugs found on most consumer RC equipment. The quad’s motors are numbered sequentially as you go clockwise, with the #1 motor being the front left. I attached the plug from this motor’s ESC to the #1 pins on the CC3D and then followed suit with the other ESCs.
To connect the CC3D to my Futaba R617FS receiver, I used the 8-wire harness included with the flight controller. The colors of my wires didn’t match those on the OpenPilot diagram, so I just referenced the pin order. The first two pins are negative and positive power. The remaining pins are signals for channels 1-6 respectively.
PPM (Pulse Position Modulation) receivers like the FrSky model shown in the Strider manual, and Sbus receivers like some Futaba models require only one signal wire for all of the channels. The R617FS is a standard PWM (Pulse Width Modulation) receiver. As such, it has three pins for each channel: positive, negative and signal. The positive and negative connections from the CC3D can be connected to any channel on the receiver. The signal wires must be connected to their assigned channels.
In addition to the connections to the CC3D, I made a signal wire connection from the receiver to a pin on the Strider’s frame. This wire allows a 3-position switch on the transmitter to operate the Strider’s built-in LED lights, lost model alarm, and also toggle crosshairs in the OSD function. I used channel 7 for this, since it was already mapped to a 3-position switch on my Futaba 7C transmitter. By going this route, I did not need the channel 6 signal wire from the CC3D.
While it was somewhat overshadowed by the announcement of Inspire 1 quadcopter last year, DJI also released an SDK for its Phantom line of consumer quads. This was a big deal--the SDK allows developers to tap into the data feed and capabilities of the Phantoms, including video streams, camera controls, flight telemetry, and most interestingly, flight control. It meant that devs could make apps to serve as alternatives to DJI's own Vision flight app, or apps with specialized capabilities to serve specific user needs. Notable apps that have come out of this program include autonomous mapping and photogrammetry from Pix4D, as well as multiple UAV fleet control from PixiePath. Today, a startup called Autoflight Logic has released its own app using the DJI SDK--one that gives the Phantom the ability to autonomously follow and film a moving subject.
We've discussed this idea on the podcast before--the Phantom technically should have enough information in its telemetry to know where it is relative to any fixed target. It's just geometry: you can use altitude (height) and lateral flight distance (length) information to calculate not only the Phantom's absolute distance (hypotenuse) from you, but the angle at which it would need to aim its camera to center you in its sights. That kind of autonomous tracking gets more complicated for moving subjects, but an autopilot app could tap into the relatively precise GPS information provided by a phone or cellular-enabled tablet. The quad knows where it is, it can know where you are, the rest is math.
Of course, implementing such a system isn't really as simple as that. There are so many factors to consider: the accuracy of the GPS, how often data is sent between Phantom and app, limitations of the SDK, failsafes, etc. There's also the consideration of quadcopter as a cinematography tool--something we've had a little experience with. Automated camera control needs to simulate the steady and graceful pans of manual control, or at least produce footage in predictable way that can be edited later. The video in this promo for Autoflight Logic's Autopilot app ($20) looks promising:
Autoflight Logic claims to have solved for many of these problems, and it's the first third-party autonomous flight app approved by Apple's App Store review team. We were given access to the final build of the app ahead of its release this morning, and spent an afternoon testing it in San Francisco's Golden Gate Park with the help of our friend (and experienced Phantom pilot) Jeremy Williams. Some of our flight footage from the test is embedded below.