Many G1/ADP1 owners have been using the app Tetherbot to get internet access on their laptop via USB to the phone’s data connection. The app relied on the Android Debug Bridge to forward ports. It worked, but people wanted a solution better than a SOCKS proxy. The community figured out a way to create a properly NAT’d connection using iptables and then [moussam] rolled them up into easy to use applications. There’s one for setting up a PAN device on Bluetooth and another for adhoc WiFi networking. It requires you to have root on your phone, but hopefully you’ve achieved that and are already running the latest community firmware.
When we first saw [Jeffrey Nelson]’s G1 based robot we immediately wondered what the transport for the controls was. The G1‘s hardware supports USB On-The-Go, but it’s not implemented in Android yet. It turns out he’s actually sending commands by using DTMF tones through the headphone adapter. The audio jack is connected to a DTMF decoder that sends signals to the bot’s Arduino. He wrote client/server code in Java to issue commands to the robot. You can find that code plus a simple schematic on his site. A video of the bot is embedded below.
Continue reading “Forknife, Android G1 controlled robot”
[ghostwalker] dropped in on our previous Debian Android post to let us know that he had streamlined the install process. The first time around, it quickly became difficult to complete the process because firmware updates had taken away root access. Hackers have since figured out how to downgrade from RC30 and install BusyBox. All you need to do to put Debian on your phone is download the package from [ghostwalker] and then run the installer script. This isn’t technically a port since Debian already has ARM EABI support. What would you run on your phone if you had access to the entire Debian package tree? A video of Debian starting up is embedded below.
Continue reading “Debian on the G1 once again”
[Luke Hutchison] has come up with a rather clever hack to get multitouch support on the G1. He wrote a patch against the Synaptics touchscreen driver. When two fingers are placed, the driver reports the x/y of the midpoint and a radius for the size field. If only one finger is used, the size is reported as zero. The nice thing about this approach is that it’s backwards compatible; the extra data will be ignored by current apps. Unfortunately, Google’s Android team says that if multitouch is ever added, it would identify individual fingers and definitely not using this method.
It’s the season of gift giving. Did you get anything interesting/hackable? What will you work on next?
We gave ourselves an Android Dev Phone 1 (ADP1). We hadn’t really considered getting a G1 until the ADP1 was announced… It’s actually a lot of fun to use as our primary phone. Our favorite app so far is connectbot, the SSH client. The interface is really smart, way better than all of the iPhone clients.
What did you get?
In our Dev Phone 1 excitement last week, we somehow overlooked phoneWreck’s teardown of the T-Mobile G1. The complex slider mechanism is certainly worth looking out. One of the major oddities they point out is the inclusion of two vibration motors. One is mounted next to the SIM on the mainboard. While the other is mounted in the frame next to the earpiece. We wonder what was gained/solved by using two. The phone also includes a digital compass module. We’d like a more detailed explanation of how the Xilinx CPLD is used. From this article in 2006, it seems HTC uses them to generate custom clock signals and switching off devices for power management.