I mentioned a while back that I was going to design an electronic indicator panel for the Savage, and I have my 1st prototype ready to demonstrate.   For those of you who grok technical jargon...  I am using an Atmel Mega8 microcontroller.   This 8 bit controller has 8k of flash memory, 1k of RAM, 3 timers, 6 channels of 10 bit A to D conversion, and other nice features.   I wrote the program in assembly language and current project code is only about 1k binary.  My algorithm is fully interrupt event driven; i.e., all events are asynchronous and independent.  I am posting a picture and a short movie clip showing real time circuit operation.   I'm sorry about the movie quality, but the original movie turned out to be a 175meg AVI, and I had to strip it down to 10meg to make it a reasonable download, however, in doing so, it lost a good amount of video quality.   The temperature sensor I decided to use is a DS18B20 from Dallas Semiconductor.   This amazing device looks like a standard TO92 case 3 lead transistor, but is in fact, a tiny computer with its own memory and CPU and a single bi-directional data pin.   The 3 leads are +5V, ground, with a middle lead that implements a 1 wire network client, slaved to the Mega8 master - pretty danged amazing!   I designed the circuit so the temperature sensor is optional, and if omitted only 4 connections are required: power, ground, and the two wire rotor pulse signal.   The optional temperature sensor requires an additional 3 wires and can be placed anywhere you want to read temps on the bike.   Interestingly, since the sensor uses a shared single bi-directional wire, more sensors can be added and queried in sequence.  The only restriction is available space on the display.   In the finished product I plan to layout the circuit on a PC board and piggyback it onto the back of the LCD display board.   Then it's just a matter of packaging it into a small plastic box (any packaging designers out there?).  Of course, questions, suggestions, comments, criticisms, are welcomed.

Pretty slick. Are you going to make your own PCB or have them made? A guy on overclock.net made a pretty slick fan-control panel and had Silver Circuits make his PCB's for pretty cheap.

*edit* This is the post of his PCB's:
http://www.overclock.net/t/1248141/custom-fan-controller-build-now-with-backl...

Very nifty!

Just throwing stuff out...

If you could replicate the "connector" such as a pass-throughto the stock speedo cluster, then it should provide you with power and an easy method for users to hook it in to the savage system.

If you did so and since you would then be privey to all the data the stock cluster had  yo could:
replicate the neutral indicator
replicate the left and right turn indicators - also could "tune" for LED?

In this way, some one could replace the speedo and use your package to replicate all the missing cluster info. ( IE they could replace the tank ...lose the stock speedo but gain all other info back.

Sorry but it aint finished till I get a gas gauge!!! LOL

Other sensor ideas
vaccumn - this is the "economy" sensor
MAF- gawd would only know why
ignition ON idiot light  -- YES PLEASE!!!!!!!!!
clock
USB power out

Pine wrote on 08/11/12 at 10:57:42:


Thanks for your suggestions Pine.  I did consider some of these and other features that COULD be implemented, but the primary goals for the project are KISS and no mechanical mods to the bike.   I wanted a plug-n-play unit that would be the least problematic or intrusive to the bike, and that anyone could install.  In doing so, I thought it would have the broadest appeal.  Modding the speedo case or cable would provide extra features, but would be too intrusive for most.   I did consider a clock, and clocks are reasonably cheap and easy (i.e., Maxim 1307, for instance), but there are a couple of problems...   A) clocks require either a battery or continuous power from the 12V system, and B) they require setting, which means buttons for hour & minutes.

One feature I thought of was a security system using a single button.  Since I already have the rotor pulse signal, I could effectively disable it at startup, preventing the ignitor from firing.  This condition would persist until a unique button sequence was entered.   So for instance, the display might initially say "Suzuki Savage", and then wait for the a button sequence to be entered.  The sequence would be simple, perhaps 3 or 4 button pushes, but would consist of short (< 1 sec) and long (> 1sec) pushes. Example: puuuush-push-push-puuuush.  Once the sequence is entered (and you have x retries), the rotor signal is re-enabled, the display returned to normal, and the bike is permitted to start.  This feature would only add a single button to the unit (and the software written).   Note that electrically, the rotor pulse is a simple floating signal from the sensor coil and not referenced to 12V or ground.

I also considered, at least briefly, designing in a true cruise control.  In this case, when you hit "cruise", I would latch in the current rotor pulse count and attempt to keep it constant by altering the throttle using a servo motor attached via short cable to the carb.   In other words, in parallel to the regular hand throttle cable.  Just like in a car, I would disable the servo motor using either the brake light signal, or by hitting cruise button again.  Of course, this feature would violate my no mechanical modding rule, so I threw it out.   The only reason I even considered this feature is because my hands are VERY subject to handlebar vibration, thus it would be personally helpful.

CalisOsin wrote on 08/11/12 at 11:12:23:


Hi Caslis, I have thought about selling the unit either as a finished product or kit form (soldering required), but there would have to be enough demand to order the materials (otherwise, it will just be one unit for myself - LOL).   I haven't calculated an exact bill of materials, nor have I decided on an external packaging (housing) design, but a rough guess is the unit could be offered somewhere in the range of $30 .. $40.

Well, if this thread is about finding out if there is interest, count me in as well.

rfw2003 wrote on 08/12/12 at 10:24:10:


Yep, yep, yep, very tempting to throw the "kitchen sink" in here as there is so many other cool features I could add (and this MPU has plenty of horsepower left over to do them).   The problem is, each one comes at a price, and I'm not necessarily talking about $$$.   For instance, there is reliability, complexity, cost, real estate (space), heat dissipation  etc., to consider.  And that's only the internal aspects; i.e., you can add a single feature and lose half your audience, if it becomes too complicated or expensive.  In the industry we call it "feature creep" or "creeping elegance" - LOL.   In the worst cases, you can actually cast your project into a state of "improbability" - trust me, seen it!   This is why every good engineering project establishes a strict set of "requirements" at the very beginning.   When you've met them, you're done designing.  In engineering, you get paid by what you accomplished, not what could have been.  

I wonder if this project was completed.