Tronbike 1.2.2 Video, 7/3/2008

Wednesday, November 18, 2009

Introducing EVroom - Motor Sounds for Electric Vehicles!

After a very sobering meeting with Mr. Native Cycles himself, I pushed a bit this weekend to get "EVroom" up and running. Its basically a server-side extension to my current ARM "TBM2" codebase that controls playback of sound-effects stored as various MP3 files, depending on the state of the controller and the computer's connectivity to it.

For example, on bootup, the bike announces itself and then informs of its connectivity to the controller, announces when communications is activated, and when/if its lost again. Its a great safety feature, I notice now all the time the controller is NOT connected and I don't know the immediate status - I need to tighten up latencies. Its proving a very useful companion for debugging some niggling hardware issues - namely, the wifi is unstable and seems to go out when firing up the solenoid. I have volume linked to amps, when at 0 volume is the loudest, under load the audio quiets down. I really need to get reliable speed/motion detection going too for volume or frequency control of the playback.

Further inspired by my meeting, I'm hunting for cheaper boards and high "NIT" displays, even big bold LEDs to serve as the dash. From my experiments so far, audio should really assist any simplified visuals.

In related news, my TTXGP DVD and fan-boy T-Shirt arrived. Went through every minute of the inaugural electric motorcycle race coverage, some great full length (25-35 minute!) POV camera sequences were included. Its clear I need to make more time for my tinkering.

Wednesday, October 28, 2009

TBM2 - Finally Some Good Data

The data logger finally worked well enough to plot some data of a commute leg.



The GPS is *not* the way to go for tracking speed and distance, but here it worked a little better than normal. Notice the GPS didn't sync up until 200 seconds into the ride.

In other news, the speakers in the computer are about to be used for something other than ballast. Audio indication of status like "controller on" or "power high" will be there, as will a background sound when at/near idle. In a test, I had the throttle controlling volume of an mp3 playback. Lots of cross compiling has been ongoing, the MMC card has been formatted for Linux ext2 for lots of space.

 A very busy chart of the same data, on a single scale/axis.

Wednesday, October 14, 2009

Arduino Tachometer and Temp Gauge

I dug into the ZX-11's cluster to see what  I could drive from the Arduino.

The temperature gauge seems to be straight voltmeter and easy to actuate with a little Arduino PWM Voltage out, the tach was harder, got no motion with voltage or pulses. Turns out tachs often assume wire to the ignition, a complicated approach.

SO I tore into the cluster and found a little board of electronics that seem to connect to a DC voltmeter for the tach. Snip snip, and into the breadboard we go!



The PWM (pulse-width-modulation) off the Arduino is a bit noisy, both the gauges hum a bit as they make their sweeps.

Friday, October 9, 2009

K.I.S.S. (keep it simple stupid)

Took the bike out with working GPS and new center iPhone mount. It all worked after 5 minutes of futzing around waiting for networking, waiting for
GPS syncing, waiting to go back inside to get the Bluetooth GPS, and then barely being able to see the iPhone display in the brightish overcast sky. Getting a 10K+ reading on the odometer and loosing the "persistent" settings did not turn me on either.

SO its cool it all worked as a technical marvel, but its just too much. Wifi, Bluetooth, Satellites, iPhone, Arm computers, but the silly Cateye bicycle computer is the only one I didn't need to fawn over or reboot, or even recharge - it just worked. I miss the big round gauges and analog needles that don't fade in glare. I hate that the iPhone doesn't work without taking off my gloves. I impatiently waiting for the GPS to be found and make sure the Wifi is up to talk to the pouty iPhone.

Back to the drawing board! Goal now is to get the Arduino driving analog gauges and replace the ARM computer completely. Maybe in stages, but its just too much right now.

Wednesday, October 7, 2009

Reflashed and Re-Installed!

Back on track. I have 3 backdoors onto the ARM computer, and a new mount for the iPhone installed. Seems the problem was only waiting for 2 seconds rather than 5 between USB WIFI startup and device/IP initialization. I removed the handlebar mount after placing a windshield mount on the new dash hole where the iPaq used to be.


I also installed 2 little sony speakers that have been bouncing around my collection for years, mounted to the inside of the front fairing - time for some audio effects. Huge MP3 collection on the SD card is avail in the meantime, its pretty loud in the garage. Boodler time? I previously ported this code to iPaqs, why not ARM too?

The Arduino awaits too - opto-isolators arrived today, hello voltage-meter-speedo city!

Tuesday, October 6, 2009

Back in...Reflashing the ARM Board

Well, I couldn't hack my way in, I left no easy holes. So I faced the dreaded decision to reflash the ARM board as per the ancient instructions, with examples in Chinese.

Using the wife's Vista host, I managed to connect the serial dongle for console and another USB cable for transmission, using a program called DNW (popular in the Q2440 ARM board circles, thank you FriendlyARM.net), I was able to follow the Chinese enough to reflash the OS parts one by one. Not very tough, I feel powerful now and free to make mistakes.

The TBM software distro SHOULD be fully compatible with the OS, and even provide the nasty kernel modules I put together myself to support bluetooth and the WIFI hub.

Lots of backdoors will be enabled before we re-install in Tronbike...

Monday, October 5, 2009

Locked Myself Out of TBMNet

I managed to lock myself out of the ARM computer. I was busy installing and tweaking and the WIFI failed to allow me to log into the machine. Removing the maching and attempting console fails because the machine is set up to aggressively use the serial port(and console) as the com route to the controller. Rule #1 is ALWAYS LEAVE A BACK DOOR.

Reviewing the startup script tweak, it should not be failing/behaving any differently than the last known good config, but I hope some sleep and some time in the shower will knock this loose. Geez.

In other news, I have an "Arduino" board in house - yet another computer for TBMNet. I am expecting some opto-isolators to create some battery gauges with this. I also want to make a speedometer/odometer too, maybe even talk to the controller over serial. At least it would be too dumb to lock me out.