When I have Megasquirt set to fixed timing mode (not running from table) I witness the timing advance by 5-10deg when looking at the pulley with a timing light. My timing light seems to be fine because when I hook it up to my Miata (also running MS2) in fixed timing mode I don't notice the timing drift issue.

I have a 13b running on MS2 DIYPNP. I'm using the factory CAS for ignition pickup. I've verified that I have the VR signal polarities correct. I gave Arron Cakes guides a good read multiple times. I can't seem to figure out what I'm missing. People on the Megasquirt forums say this can happen if your polarities are flipped due to the nature of the VR sensor's voltage curve.

I've done a small bit of bench testing with the CAS connected to an oscope. Signals look clean and consistent when spun up with a drill.

I've tried changing the input capture setting from rising to falling... didn't seem to help. Any other ideas or settings to mess with?

 

How much slop is there in the gears from the eshaft to the CAS?

 

Having a spark hardware latency set when one may not be needed would cause this. Make sure it's zeroed out

 

I do have hardware latency zerod out.

I haven't had the chance to check the CAS backlash yet

 

Buying the FFE trigger kit was the best EFI-related decision I ever made. Fixes a lot of these issues, including noise (Hall-effect sensor)

 

I was just about to come post a similar question. You should log this event and make sure you don't have any sync issues. Also if you look at timing err, if your having the same issue as me there will be no sync errors but you will see the problem in the timing error log.

I started noticing detonation at 4k rpm and above at wot and dug though the logs and noticed timing error got worse as rpm rise. +- 10 degrees is what I was seeing.

I suspect it's noise related. My question was going to be about using megasquirt filters.

 

Timing error could be a false trail. I'll do some trigger logging today and see if I can't figure it out.

 

So in my case my trigger looks clean to me, at least when free reving.
The weather is crap so I can do any actual pulls.
After some reading the Timeing error thing is from something called 1st derivative predictive timing something or the other. Im seeing 10% +- so iess less than a 3 degree error.
Ill need to ask someone more knowledge questions about that.

I uploaded the log in case you want to see what (i assume) a normal log would look like.

 

I had my cas wired up wrong when I first started. the car ran but like total ****. I wish i could remember the symptoms.
Do you have any other odd thigns happening besides the timing drifting about?

Is it possible your timing is moving around because of a different table interacting? Im not sure if locked timing bypasses the other things like startup/idle maps

 

mikey D, it's good to hear that I'm not alone. Thanks for the tooth log. When I would log mine at idle, it looks exactly like your's. I never did any logging at speed to see what it looks like however.

Since my last post, I picked up the Full Function Engineering 36-1 hall effect kit. I installed it today, and unfortunately I'm still seeing the same symptom! This makes me think it's not related to the CAS...

I'm starting to question if it's something with my tune or wiring. Going through the tune last night I couldn't see anything strange (I've attached my MSQ file).

 

After a few months of tweaking things here and there, I've finally figured it out!!!!!

It turns out I had the wrong Megasquirt ignition setting. I needed Spark Ouput: Going High.



Note that I have MS2 Extra 3.4.2 firmware installed on a Microsquirt module (DIYPNP), so I'm not sure if this setting should be like this for everyone. It seems like all the guides I've been able to find say to use going low. Aaron's guide says to put spark out to "going low". DIY Auto Tune's guide also says to put "going low" for MS2 firmware.

This all seems to make sense now because I did have to replace my trailing coils last year. They probably burned up from being over dwell'd with the inverted signal. In the summer I was also have problems where the car would stall, and then start back up again after sitting for a few minutes. The coil were probably overheated, and after cooling off for a while they were ready to go again.

What really keyed me in to this being the issue, was when I hooked up my oscilloscope and noticed that the trailing coiling select signal also gets inverted when you change the "spark output" setting. I'm not certain I agree with the firmware developers on that one, but this also explains why I always needed to have my trailing spark plug leads flipped.

My timing is DEAD on now! There is no movement when I have the Megasquirt in fixed timing mode.

 

Your settings still don't look right to me.

 

What in particular doesn't look right?

 

Are you using the stock co cas? I'd expect rising edge and dual wheel

 

When I started this thread I was, but now I'm using the Full Function Engineering 36-1 trigger wheel.

Had I been able to figure out my spark output was flipped before picking up the FFE kit; the stock CAS probably would have been fine.