MicroSquirt.com read only forum (Jan.15,2018)

I am using MicroSquirt for a dual spark application on a 4 cylinder GSXR 600 with an 8tooth crank trigger (VR sensor) and a 1 tooth cam trigger (VR sensor). After a lot of trial and error (and a lot of thought/documentation reading), I was able to get the engine to start with the following settings:

Base Ignition (MegaTune 2.25 p3)...
Trigger Offset: -45
Skip Pulses: 2
Alpha: 85
Beta: 22
Gamma: 0
Ignition Input Capture: Falling Edge
Cranking Trigger: Calculated

Advanced Ignition...
Trigger Wheel Teeth: 8
Missing Teeth: 0
Skip Teeth: 4
Delay Teeth: 3
Dual Spark Options: "M-0 Wheel w/ Rising Crank Sync"
Offset (advance) for Output #2: 0.0

The rest of the values (fuel, etc.) were kept the same from a MicroSquirt configuration I had reliably working with a 12-1 wheel in another GSXR 600 engine.

I've monitored the ECU output using both an oscilloscope (both VRs and both Ignition ouputs) and using a strobe on the crank/TDC mark. The biggest issue I had was that the Trigger Offset value wouldn't reliably produce spark for most values. For other values it would produce spark but wouldn't vary it's location or would put it at a non-intuitive angle within the crank cycle (e.g. +90 when we had entered -30, though these values are just a general example... I tried so many that it's hard to remember the actual observations). The delay teeth parameter also produced curious results, in that it would 'more reliably' produce spark when configured for Trigger Offset: -45 deg and Delay Teeth: 3 compared to a more intuitive setting of Trigger Offset: 0.0 deg and Delay Teeth: 2 using the engine manual. With either of these settings, if I change the Trigger Offset value by say 10 deg, the ECU wont produce spark. The Alpha Beta Gamma parameters seem to affect whether I get spark or not as well.

What I'd like to know is why the Trigger Offset/Delay Teeth parameters don't always produce the desired output. Also, if you have any suggestions regarding the above configuration, I'd greatly appreciate it (especially good Alpha Beta Gamma parameters).

The worst case for timing accuracy is at low rpm and having only 8 teeth doesn't help matters. I have found that there is not a lot of improvement with alpha-beta-gamma. Your best bet during cranking is to use trigger rise rather than calculated - there should be very little error there because it fires instantly when it gets the tooth pulse - it doesn't do any prediction.

As far as trigger offset and delay teeth, these are limited by the fact that the dwell + advance begins at one tach pulse and must complete by not the next one, but the one after that. That is you have 2 tach intervals. If you keep moving trigger offset or delay so that the spark goes over the end of the dual tach interval, then you will spark at that point regardless of the advance. Likewise if the dwell has to happen too early, then you might get a short dwell because it won't move the location of the spark, for obvious reasons. Absent these cases, changing trigger offset should affect the location of the spark, but it can be hard to tell during cranking if the engine is jumping all over and you are trying to predict where it is going to jump to.

Judging by your description of the engines mechanical setup the dual spark configuration should be 'Rising or falling cam synch with tach or wheel'.

Paul,
I believe that 'Rising or falling cam synch with tach or wheel' is used only when the number of crank trigger wheel teeth M = #cylinders/2. For our case that would require a 2 tooth crank trigger wheel but we have an 8 tooth crank trigger wheel. We did try this setting and it would constantly lose sync. The M-0 wheel with Falling Cam signal worked for us.


Grippo,
I understand that the timing is difficult to control precisely during cranking because the engine speed varies alot. Our engine starts easily and it seems like the spark plug a few degrees after top dead center, I would like to have it 5-12deg BTDC but if that is not possible it doesn't seem to make much difference.

We are using the CAN bus to control ignition retard (to be implemented with a traction control system in the future) while the engine is running and that seems to be working as expected.

bigbrotherjohn, I meant to respond sooner with this screenshot, but laptop elsewhere.Paul_H is right with his statement Rising Cam Sync with Tach or Wheel. You need to have two tach events per crank rotation, so skipping 4 teeth will do that.
This is a screenshot from my Thunderace settings...

The rising/falling option can be applied to both sensors whenever you have specified 2 sensors. The falling/ rising input capture is for the first sensor and the 2 pairs of Dual Spark Options: Rising/Falling cam synch and M-0/rising/falling crank synch options take care of the second sensor. In your case, you said you have a cam synch, so the appropriate choice should have been crank with rising/falling cam synch. This pair of options can be used with or without missing teeth - just set Miss_Teeth = 0. In the last 2 options zero missing teeth is specified because it doesn't make sense to have a separate CRANK sync and missing teeth on the wheel as well - which one should the processor synch on. It seems as though you should be seeing the wrong rpm and it should be missing synch since with a cam synch it expects it to occur once every 2 revs. Or is this a 2-stroke ?

I guess he set the skip teeth to 4 to have his second channel 180°.
The correct setting for your application should be skip teeth = 8 witch will make two 8 teeth tach completing 720 ° rotation every 2 tach, right now both tach are made in a single revolution in a wasted spark manner, making cam synch somehow useless.
The 180° offset for cyl 2-3 may be set trough "offset (advance) for output 2", either to 180° or -180°
This may trigger some unwanted events like I had, I could put more offset and spark will quit or jump farr off, or hit a wall making further advance impossible, this was due I think to the tach event too short for the dwell and spark time.
The cam VR sensor would quit working after a while and be somehow inconsistent, I swapped it for a Hall and hall was well