Toyota Celica GT-Four -88 install
I'm installing a Genboard 3.2 in my Toyota.
The engine is a 3SGTE 4 cyl turbo.
I've sorted most of the board configuration, WBO2, EGT etc.
Now i've started thinking about the trigger options.
The engine has 2 trigger wheels with 3 VR sensors mounted in the distributor.
One wheel with a single tooth and 2 VR mounted 180 degs apart
- 1 VR timed at 30 deg BTDC for cyl 1
- 1 VR timed at 30 deg BTDC for cyl 4
Another wheel with 24 tooths even spaced, no missing tooth
- 1 VR timed with one tooth timed at 10 deg BTDC
The distributor is adjustable, atleast +-10 degs.
Multitooth option
- plan is to use one VR on each wheel, the single tooth VR is used as trigger-sync
- both LM1815 used (one for 24-even spaced teeth)
- the output of the two LM1815 combined to make a multitooth type pattern
Coil type option
- either both LM1815 used (for the 1-tooth wheels) and output connected or
- two VR signals connected in series, and one LM1815 used.
Connecting the LM1815 outputs is safe (open collector), but the LM1815 output we use is stupidly designed to be pulled to low most of the time, and only released (to the 5.6k pullup resistor) when the pulse comes (input passes 0 and falling). Check the LM1815 datasheet, there might be another output that we normally not use. Or diodes can be used to act gates.
At worst, some firmware mod is probably needed (or more HW hacking).
Here is a link with pictures of the wheel and VR sensor arrangement.
(Borrowed the link from EmilLarsson, hope you don't mind.)
The question is then what does it take to make this work with the Genboard?
Some general trigger questions - updated 2005-05-21
- How are the ignition events scheduled, are they timed to the trigger-sync or to the last tooth before the actual event?
- How often are the timerticks/crankdeg calculations done, once every trigger-sync or once every tooth?
- 2..3 times every crankrot, but lasttooth firmware is coming (slowly)
- 2..3 times/crankrot ok i assume it's done in "userspace" (normal prio eventque) then, timertickssampling when is it done ,randomly or synced to some tooth or sync?
- Sampling window for timerticks, is it ticks per tooth or per revolution?
- 2..3 times every crankrot, but lasttooth firmware is coming (slowly)
The scheme used above is very critical to cranking/starting advance acurracy, especially on a high-comp, low-inertia engine with drained battery. The crank speed varies very much during each revolution, if the sampling used for timerticks/crankdeg is done far away (in crankdegs) from the actual spark event, or if the sampling is done on a per/rev basis you end up with very bad acurracy. Maybe the problems described on Jose's page MembersPage/JoseLCortes/FullySequencial is somewhat related to varying crankspeed during cranking, in his case it messes up the trigger state-machine?
- "multitooth" means that spark is scheduled trigger_tooth (say 7) teeth after the first tooth after missing tooth, than at every another_trigger_tooth (say 20) teeth. The 7, 20 example means trigger is at teeth 7,27,47 (only makes sense for a 6cyl with 60-2 wheel, eg. 2.5L BMW M20B25 block)
- "coil" means that a spark event is scheduled at every trigger
Some notes about supply noise
- I noticed that the supply to the AD597 gets very noisy when the WBO2 heater is active, the result is a jumping EGT figure that averages about 3 degs higher than normal due to the noise.
(The docs aren't exactly for dummies... :-) )
There is an L4 and some decoupling caps for the EGT power that should. Can you identify if noise is in the GND or in the +supply?