WTF Tuning Paraphernalia - Android Dashboard Project

[Heide264]

I've been wanting to do this for a long time. I plan on making an 'android dashboard' for my car. Ambitious project for me. I know a lot of people on this board have experience with programming and electronic design, and I will need all the help and suggestions I can get.

Main Goal: To display real time tuning information. The first main benchmark will be to display various thermocouple outputs along with a few pressure sensors. I want to be able to calculate both turbo and intercooler efficiency - experimentally (not just theoretically). Down the road there are several 'dreams' or optional benchmarks - not excluding reading information directly off of the CAN bus in the car.

Background: I want to be able to collect some data that isn't commonly taken with these cars. I feel that the data should be taken to help both the power and long term reliability of our cars. When I was on the track last time, I noticed how much my car slowed down after the first few laps. After cooling off, the cars power returned. I personally believe that the droop in performance is due to under bonnet pressure that is reducing the effectiveness of the top mount intercooler. Instead of throwing money at the car, I want to take some measurements and prove my theory. Alternatively, it is a great opportunity to learn about the physics behind the car. If I get the display to the point where it is recording CAN bus information, real time display could be an extremely useful tuning tool.

Hardware: My Nexus 7 and an Arduino ADK on the processing & display end. The controller contains the USB hardware already, along with plenty of I/O. For those familiar with the Arduino line, it is basically an Arduino Mega with USB hardware added on. I should be able to use this to both charge and connect to my tablet quite easily.

As far as the sensors, I will have to deal with the hardware design myself. I plan on contacting a few vendors to get some pressure sensors and thermocouple odds and ends. Hardware design with thermocouples is quite challenging (especially under the hood of a car) and should provide a good opportunity to learn a lot.

Schedule: Whatever I want it to be. I will set goals as I can, but I have other things to do in my life as well.

My Experience: I'm an Electrical Engineer by my undergrad. While I am fairly competent in hardware design, I have little software experience that would be relevant (I should be okay with C code to program the controller with). I have spent a few months reading some Java programming books as well as a few android application books.

[Heide264]

Current Snapshot & Next Goal:

Got two way communication going!



Next up I have a few bug fixes dealing with the usb connection (only connects when usb debugging mode is NOT enabled - sadly... and doesn't reconnect without closing the program) and some form of fragment based interface.

[Heide264]

Initial Impressions with the Arduino ADK:

The controller was oddly friendly. In my past experience, it has taken a few hours if not days to get the hardware version of "hello world" running (A blinking LED). It took me under an hour to get all of the required software downloaded, some sample code loaded up that flashed an LED, and one way serial communication into the laptop over putty of both an analog and digital signal. While it was sample code for the most part, it is a way friendly setup than I have dealt with in the past with controllers while I was in school.

The other end of things, the android application, will be a much larger hurdle.

[Rambo]

Subscribed! Perfect timing, I was just looking into a project similar to this for my rooted Nook Color. I plan on using the IOIO to make an integrated HU, and maybe piggyback off this project. The IOIO might be better for a more permanent install once you've gotten the prototype ADK setup how you like it.

[Heide264]

Subscribed! Perfect timing, I was just looking into a project similar to this for my rooted Nook Color. I plan on using the IOIO to make an integrated HU, and maybe piggyback off this project. The IOIO might be better for a more permanent install once you've gotten the prototype ADK setup how you like it.

Nice looking little toy there. Haven't seen that one before. I may look into that down the road a bit. Considering I already have the controller, I am going to go with that one for now. It has a lotta nice perks for prototyping. If the project works out nicely, I may do some more hardware work as opposed to just using a development board.

That being said, I don't know how much the android application code would change. I bet both devices are picked up as an accessory (2.0 or whatever). Will see. I will definitely keep posting up on my results.

[shadowpr]

this would be great.

[Heide264]

this would be great.

Thanks.

Update - spent most of my available time for this updating my android development stuff. Was back in early ice cream sammmich software. Those APIs take longer to download than I remember. Hoping to get the bare bones application up and moving tonight. Doubtful that I will get to the communication tonight, but will see.

Also ordered a 12VDC 2A wall wart. The controller can require up to 1.5A to charge the device as well as supply enough current to the I/O.

[Heide264]

Update: Got hello world running on both the phone and tablet (and hence adb for both devices)

Started briefly adding some stuff. Got a basic interface activity with a title/text box/button that works. The button does log correctly on the laptop as well.

Still got a ways to go in order to receive any serial communication by the end of the week.

[Terminal Kuz]

Could you give me a really brief dumbed down version of the main goal? Is this to tune on a tablet? or more like a gauge screen like boost, oil psi, temp etc?

[Heide264]

Update: Spent some time on it before getting knocked on my butt from a cigar.

There are two main ways to interface the controller to the android OS it seems. One being a 'pre-built' set of stuff, and one just using cut and dry eclipse & arduino. I plan on doing it without the interfacing program. I believe I got a good chunk of the java done on the android side (basically a toggle button that should control an LED on the arduino). I should throw up a digital indication on there as well (fake light).

Tonight will attempt to get the arduino half done and connect the two. I know I have to import a few libraries into the arduino software, but it may go smoother than planned.

Could you give me a really brief dumbed down version of the main goal? Is this to tune on a tablet? or more like a gauge screen like boost, oil psi, temp etc?

Just a gauge display for now. More geared towards non-ecu determined parameters at first... So anything you would use a gauge for normally. The ECU interface may happen down the road a ways.

[Heide264]

Update: Got a few hours in. Realized I was far from the amount of code required on the android side of things.

Found this fantastic site: Android ADK: How To Blink an LED Code and Explanation

It has a good bit of deprecated (old) code that I'll need to look through and fix up a lot - which is hard considering this stuff is a good bit over my head. I'm hoping I should be able to get it working though.

Good news being the arduino code looks fairly simple on the receiving side.

[Heide264]

Update: Didn't get much time over the weekend sadly. I did find a good book that has a chapter on what I need to get moving - Professional Android Sensor Programming (Wrox Programmer to Programmer): Greg Milette,Adam Stroud: 9781118183489: Amazon.com: Books

I picked it up. The digital preview is pretty nice. I also think most of the information in this book will be applicable to my project.

[Rummie]

Fascinating stuff here, enjoying the steps as much as wanting to see the end result. Good luck.

[Heide264]

Fascinating stuff here, enjoying the steps as much as wanting to see the end result. Good luck.

Thanks...

Update: Having issues. It appears that there are various ways to implement the Android Open Accessory stuff. Most guides I have seen use a different way than google recommends with the 2012 release. I'm trying to find out what the recommend... but I can't get git/repo working to get their source files. No idea why they can't just post them for download.

Got git to work... then realize they require this 'repo' tool... now I can't get the repo tool to save anything but an .xml file. I just want my alarm clock code =P

[zax]

Whatevs.