NefMoto

Please excuse me should this sound like any sort of demand, I just had an idea I was hoping to get feedback and gauge interest on.

What do you think of an opensource handheld flasher that would provide essentially the functionality of APR/GIAC/etc. ODB 'tune' selectors?  It would need help from a lot of different people but my initial thought would be to base it around an existing raspberry pi/beaglebone development environment.  Is anyone currently developing software that might lend itself to this function, run on linux?

Thoughts?
Thanks :)

I would definitely participate in this project.

btw it would be nice if it included some basic checksumming support...

Id contribute as much as i could also  ;D even a andriod app would be cool have all your bins in your phone

I suppose that would work utilizing Bluetooth but you'd absolutely need some checksumming capability in that case.  Not that you wouldn't want it if you could.

I was thinking storage on a microSD card with usb connectivity with a 7-segment type basic text display and a couple buttons.  Much like your average odb scanner.  Interfacing with the applications we currently use via GUI or commandline would take some scripting due to the limited interface.

Even better would be if this tool could include an FTDI chip and function as a standard scan tool/flasher with PC connectivity.

I was referring to the motronic checksumming, not data integrity between handheld device and the server supplying the file... if it is linux there are an infinite number of ways to do that (md5, bzip, etc).

I'm all in, although I don't have a smartphone... I don't believe in that crap ;)

It'd certainly be nice just to have something like Nefmoto being open source so we could integrate support for new ecus as well as ME7.

Wish List

1) CANBUS support!
2) Ability to change a variable stored persistently in the e2p. Then other routines inside the ecu can check this variable. Voilà - tune switching :)
3) Basic diagnostic routines, at least able to request fault codes from ecu.

Implement a KWP2000 engine. Then put applications on top of that. Layered approach also means the lower layers can be swapped around - K Line <-> CANBUS.

I can contribute towards the CANBUS and KWP2000 aspects.

:)

Just to clarify my original idea, not that new ideas aren't appreciated, I think an embedded solution might be the best approach versus a iphone/android 'app' because of the amount of development involved in that approach.  Besides the amount of fracturing in the android ecosystem between versions of android and countless phones so there would have to be near from scratch development of android/ios software for this purpose, a second interface would need to be developed either USB/Bluetooth so the phone could access the ECU.

An embedded microcontroller/nanoPC solution using beaglebone or Rasberry Pi might be able to run any existing Linux compatible tools with little/no modification.  An interface to ODB would need to be developed though this should be minimal as both these previously mentioned  platforms support GPIO and flashing the ECU is largely a serial communication via the K line.  So that would require little development.  The end result I'm hoping would be a 100 to 150 dollar unit that you could keep in your glovebox and swap between tunes on the go.  This is a great community but lets face it there aren't a lot of contributing devs out there so keeping overhead low would be essential to accomplishing this goal.  Once that task is completed there's room for added features like maybe DTC logging/clearing and maybe loading a bin file via USBOTG so you could retrieve a flash via email or the web and get it on your car in a pinch using a phone.

Can anyone suggest utilities that run on linux?  If there's a developer out there with something written that would be interested in working with me to get something running on a beaglebone or raspberry pi I will order one up and see what we have to do to make this reality.  The software running wouldn't necessarily have to be 'open' and could even be charged for.  I'd just like to see the hardware be open.  Shit I don't see any reason a thing like this couldn't be a kickstarter or could already be.  An ODB 'sheild' for a linux platform.  Or if the linux beaglebone/raspberry pi base was thrown out the window for a cheaper (to consumer, more expensive in development) embedded solution and you could have a programmable ODB computer, anything you can write, load or log.  Mechanics could log a clients car cheaply as they drive around for the week before the appointment.

I believe these are for sale, this has potential but is lacking in at least two ways we'd find useful.  There's no User I/O, buttons or native display and it'd be nice to embed an FT232 chip for doubling as an on the fly programmable OBD scan tool.

http://arduinodev.com/hardware/

Also for an 'embedded' solution it's fairly expensive however that's due to 'off the shelf' solutions utilized by the product, arduino base and bluetooth package.

Really interesting idea, I love the idea of some open source flashing tools - if only so when I get a weird error I can go and see what causes that bit of code to execute.

I would say that for future proofing, cost effectiveness, etc, something based on a cheap android tablet would be the best - maybe something like a nexus 7 - even if we restrict it to a single device. I always thought that porting stuff between linux and android was relatively straightforward - I'll look into this and see how bad it looks. There should be KWP2000 libraries availible for android that work with the bluetooth dongles, I cant believe that every app developer has written that from scratch.

FWIW, I ran up my eeprom data tool on an old phone:
(http://)

Woah that's impressive, i'm just cautious of these 'cheap' bluetooth dongles considering how crucial the task though simple it may.  In this case you may want to look into that OBD tool linked above, as it's open you'd have more access to things you may need though I have no idea what that'd be.

I kinda cheated, my tool is written in python (because I'm super lazy), so it works on anything with a python interpreter - I do most of my dev on OSX. So I just installed a python interpreter on my phone and away we went - going to have a play and see how easy it is to wrap an interface around it thou as its an interesting challenge.

I'll take a look at the odb tool linked above, itd be pretty cool to get something like the me7 logger working as a trial. Agree about the criticality of the task, it'd need a lot of testing on the bench!

I know BBB and RBPi will run Python, not certain how that is implemented but search either with 'python' and you'll find loads of results.  Even a book on learning python with the RBPi

It'll just be a standard python interpreter on either, although please dont think that I'm advocating Python for this project - although I'm interested to hear what people think is a sensible language to use? I guess it depends a lot on what library support there is already for the chosen platform.

Along the way I found this thread: http://nefariousmotorsports.com/forum/index.php?topic=4213.0 which looks kinda interesting, I might have a go tomorrow and see if I can get that working on the bench.

Incidentally, if you dont have one already a RBPi is an excellent toy and well worth the $30.

There is absolutely no reason we can't port python code to something more native-y

I am definitely willing to spend time on it.

Well I went big and ordered a BeagleBone Black Rev C today.  I liked the small form factor and I don't need all the I/O the RBPi provides.

this seems pretty cool
have somthing to change tunes on the go
i know my buddy uses a windows tablet for vagcom if thats of any help

I'm inclined to say we should make it less nativey, I like the idea of having something I can run on linux, osx, windows, android without having to port!

add raspberry pi.

linux/windows/android we can do with java

There is nothing we can do for all 5.