Added mapping tables for the intake/exhaust valves #3
Conversation
|
|
|
|
||
| // Control constants. | ||
| const int DEG_PER_MAGNET = 6; // Number of degrees for per magnet. | ||
| const int ROTATION_TWO = 760/DEG_PER_MAGNET; // Number of interrupts in the full cycle. |
There was a problem hiding this comment.
This should probably be 720 / DEG__PER_MAGNET. However, a 60-2 wheel has 58 trigger edges over 360 degrees. Shouldn't there 58*2 -> 116 interrupts per cycle, not 120?
There was a problem hiding this comment.
Yep, I totally messed up that maths. Why is it 60-2? One for the missing tooth and one for zero based array?
There was a problem hiding this comment.
I think I made the assumption that he had made a standard 60-2 (60 teeth, 2 missing teeth) wheel. Re-watching the video it almost looks like a 60-1 wheel. Either way, it might be easier to define all of these in terms of teeth:
static constexpr int NUM_TEETH = 60;
static constexpr int NUM_MISSING_TEETH = 2;
static constexpr int NUM_TRIGGER_EDGES = NUM_TEETH - NUM_MISSING_TEETH;
static constexpr int DEG_PER_MAGNET = 360 / NUM_TEETH; // Number of degrees for per magnet.
static constexpr int TRIGGERS_PER_CYCLE = NUM_TRIGGER_EDGES * 2; // Number of interrupts in the full cycle.
static constexpr int TRIGGERS_PER_ROTATION = NUM_TRIGGER_EDGES; // Number of interrupts in a half cycle
```.
This way if a change to the trigger wheel design is made you can just update the NUM_TEETH and NUM_MISSING_TEETH. This would make the intake/exhaust maps a little more complicated. This could be overcome by creating the maps on an arbitrary grid and then interpolating for the current calculated crank angle.
There was a problem hiding this comment.
Agreed, rather than teeth I've used STEPS incase there is ever a different way to generate position data. It follows the conventions of stepper motors which implement a similar positioning index system.
I didn't use constexpr as I'm not sure it is a standard for the Arduino?
|
From what I understand, you try to map the amount of valve opening (not just 0-1, open-closed) to the shaft rotation using the I believe that just focusing on opening and closing the valve at the right time might be easier to implement and almost as efficient as being able to partially opening of the valve. Looks like people successfully implemented pressure control using an electromagnet: https://ntrs.nasa.gov/citations/20180006153 |
|
@valeriu-balaban My understanding is that 0-1, binary open-close is the benefit of this approach. After watching https://www.youtube.com/watch?v=WwlNqaz9q_0 it was clear to me that the binary open/close is why it's more performant. |
|
I think I've removed the need for any time related code. Using a second interrupt to detect each rotation it now has a clear and reliable check point. |
|
That is a good idea. Dual wheel optical cam sensors should be able to be interfaced with the crank with a little bit of work. I've used DSM cam sensors with a trigger wheel like this in the past: Also I'm currently using a AEM EPM on an Alfa Romeo ITB/EFI project. It has 2 outputs: one for cam sync and one for a crank signal. I've adapted it to a bosch distributor, but it could also be adapted to work on the 212 crank for testing: My adapter for the AEM EPM Its been a while since I've looked over dealing with nesting/concurrent interrupts on an AVR. Just as a precaution I would set the single trigger interrupt to a a higher priority and keep it as tight as possible. As long as the single trigger tooth is offset from the N-tooth trigger wheel, it shouldn't be a problem. AVR ISR priorities are described briefly here: |
|
Really interested in the project, just wanted to throw in an idea about the benefit of partial valve opening: it would be interesting if partial valve opening did work with the freevalve type tech because it might remove the need for a throttle, and might actually bring similar performance benefits as ITBs, as well as potentially improving low down torque (?) |
|
Could it be possible to reach a similar effect by varying how long the valves open, in place of how far the valves open? It seems easier to implement here, though variable lift and duration would clearly be ideal. |
Sure, I think so in general. Being specific though, it depends how much control you get over the valve opening and closing time, I made an issue to mention about that, it seems like there will be some slowness no matter what, There are faster solenoids but the price starts ramping up. So if you have less control over the minimum opening and closing times, it would be good to also control how far they open |
6 participants
Based on the last video I've added mapping tables for the intake/exhaust valves. Going to grab a Harbour Freight engine today so I can start testing.
Notes;