I'll try and summarize what has been discussed to and from regarding the nissan trigger system.
First, we need to understand how the current trigger works.
It's an optical system located on the cam. There are two sensors. One of them records the engine speed using 360 small slots. The other one triggers the ignition events using different sized slots. It knows which cylinder is 'active' by counting the number of small slots while the larger slot is open. See [this image] and reread the paragraph and it will make sense.
![[this image]](http://200sx.jagenheim.se/images/trigger.png){kind=link}
This image is for the CA18DET engine. The SR20DET, RB26DETT and the other nissan engines works in exactly the same way, but differs on where the slots are placed and the number of cylinders.
So, how do we support this in VEMS?
360 tooth is a bit many for interrupt processing. So 2 ways are possible
- HW counter. Might be posible using a freed IO if PS2 keyboard is not connected. But firmware mods also needed. Honestly not top priority at the moment.
- AudiTrigger helper board
And of course the HW solutions
- replace toothwheel
- add extra sender (with necessary magnets)
Two ways have emerged. Either replace the trigger disk with a custom 60-2 disk. It will trigger every 15 (60/4) events and resync whenever we miss the two teeth.
Manufacturing of this disk may be a problem though. Also, it's not easy to replace the original disk.
Using the original disk we can time the slots to find out our relative position. This requires 720 crank-degrees to sync. We can't just look at one of the slots and see how large it is; since we don't know how fast the disk is spinning. We don't have the hardware to count the small slots either.
If we know the sequence of the slots; we can make sure that the next slot is always longer than previous. If this test fails; we know that we are back at the shortest slot again (#1 cylinder). This can also be used to sync the engine while running.
I'm not sure how accurate RPM readings will be from this, as we will just get two events per crankrotation.
Before we move on; we need to decide which of these approaches are the most feasible.