The first ECU was created

The initial tests they were realized using a PIC 18F452 working in 40Mhz, but to working with the new flywheel of 60 teeth and motivated to learn about the ARM architecture, was created a new system using a ARM Cortex M4 and I created my own RTOS to schedule the temporal and angular tasks that It need to do for generate the spark ignition.

This system was debugged and validated, it worked very well in a flywheel simulator that I did, however along the practical tests, I don´t had success, the cause is very simple, my system did not had a efficient instrumentation, the optical sensor did not worked correctly because the alignment between flywheel and optical sensor varies depending on the engine vibration...

In my tests I found that is very critical this instrumentation and is very hard to develop it using a hand made solution (without specific resources), this problem forced me seek new solutions for my project.

When I found this difficulty to use the flywheel with 60 teeth, inspired in a motorcycle solution that I observed, I tried to create a new reference system composed of 12 teeth, for this, I glued small metal parts on the magneto with epoxi resin and used the hall sensor again to index the crankshaft.

Again I had problem with the index signal generated… At this time, I can´t explain the reasons to did not works, but I believe is critical that dimension of metal parts and the existence of magnets inside the magneto, the screws that hold the magnets generate interference in the signal.

I stopped with this approach after some attempts using different sensors, settings...

My last tried was install a automotive VRS sensor ,  but the problem was the same with the hall sensor, however with this type of sensor I could observe a significant quality improved in the signal, when I used hall sensor there was a distortion in tooth duration time then I can not explain, I believe this happens because the magnets inside the magneto (magneto´s construction ), maybe the hall sensor is not a correct sensor that have defined to this kind of application...

Initiating the activities

My first idea was create a Igniton System using only one point as crankshaft index for the calculations, but in a short time I noticed it is not a good idea, it is very hard to predict revolution time with calculation. For this experience was created a support for a automotive hall sensor (flywheel sensor for VW GOL application), the aligment is the index on the TDC. I will show you bellow:

This is the first aquisition using a digital oscilloscope:

We must have in mind what the speed of crankshaft is not constant during the time and this characteristic  is fundamental to understand the needs of control solution, because we need estimate the next period.

Is not trivial to do a predictive calculation and my experience it was not the best, I understand that you need more point to use as reference, specially in my case, where I set my index in TDC…

For more accuracy, was created a flywheel with 60-2 teeth , it was sculped using a rotary tool as show below: 

This is a comparison balance (2 tooth) for static balance of the flywheel

Here, I tried to align the flywheel

Optical sensor support

Initial test with a hammer drill fixed in a table and a optical sensor

My father as technical support... : )

After finished the first test and validate the concept, I assembled my creation in the engine:

JD37 - The powerfull 37cc

My biggest challenge to start this project, was gathering of information about this engine and its peculiarities…

There is not a lot information on the internet about him, it was produced to Brazilian factory called Hatsuta Industrial S.A and his model is JD37.

                                      

It has a 37cc displacement and 1hp at 7500rpm, it’s a elementary two stroke engine (similar a used in a grass cutting machine) and its used in a vehicle called Walk Machine (a success vehicle in the 1980s was created to transport and to amuse children and adults, similar a small motorcycle).

The JD37 has a original ignition system using a contact breaker (first version of this engine, but the newest version use a Capacitor Discharger Ignition System and don´t needed a contact breaker, only a high inductance coil), it does not have ignition timing compensation through load, engine temperature, I have a conjecture that this ignition timing curve is not opymized for this engine, because in Brazil this CDI is applied in other engines that have totally diferent characteristcs, biggest displacement, other geometria, his never works optimized…

The main characteristic I desired improve with the advent of my new project was a possibility to map the ignition timing angle, because in my analysis its the best way to get better combustion efficiency and optimize for all engine speed, engine temperature, load that works (facility to crank, more stable in Idle and fast response when desired), other important consequence is increase the rpm, this way increase the top speed.

PS: My engine has a exhaustion system and air filter different the first figure, it is like the picture above (old version)!

Blog´s objective

My objective to publicate this personal research is discuss about Ignition System design, calibration and all the things envolved. This theme is very complex and there are lot of things to be considered to succeed in applying.
I will updating the blog as the project evolves.
If you have any doubt, curiosity or suggestion please send me a message or leave a message here, this is a channel for discussion about this theme.