Mech5700

I believe that is how it is. Because I had no problem running them on the stock ECU.

FAB actually suggesting raising the cranking dwell to 8ms, up from the 6 that I had.

------------------------------------------------------------------------

And as for the falling v rising edge, the starting was great when I left work, about 105* ECT. Definitely felt some slight hesitation while accelerating, drove for a few miles and it felt sluggish. Pulled over, set back to rising edge, reset base timing, gave me a slight hard time cranking, and drove off. Felt normal again. Had to turn of the laptop, was running out of battery.

The start:

Nice and clean, no sync error.

You can see the RPM get kinda fuzzy when accelerating. I think this is what I was feeling:


The log kept failing to upload, had to post the screen shots...

I still have yet to swap back to stock coils and see what happens. But I'm running out of ideas... I came home and switched back to falling edge, reset base timing again, and then it gave me a hard time starting just like rising edge did! I have no idea why things are so inconsistent.

Every log I take of a rising edge start, there are sync losses (first uploaded log)... 2nd is a falling edge example, no sync losses.

When I was on the side of the road and switched back to rising edge, I had to give it 20* of offset to meet the fixed 10* base timing. Mind you, earlier today I had pulled #1 plug out, verified TDC, and remarked the crank pulley @ the TDC mark. When I set base timing originally almost 2 weeks ago, I had it set at 9* of offset to line up when I turned on the fixed 10* timing. Earlier today, when I set it to falling edge, I had to set the offset to 18* to get it to match up and it started fantastic. Came home and verified TDC AGAIN, it was still lined up at this point on the pulley. And right now as it sits, set to rising edge, there is only 4* of offset to set base timing. And out of a few start attempts, I get that kickback like I'm way too advanced, and my crank advance is 0*...

W


T


F

Mech5700

This morning I swapped back to the OE coilpack setup after another failed attempt to a good start. It made no difference so I feel that is not a factor in this. Log #1 (COP) and #2 (OE) show this. Both have 5 sync losses during cranking.

After I found it made no difference, I put the COP set back in, set back to falling edge, reset base timing, and started nearly perfect. Log #3

I don't know how I ended up with 18* offset yesterday, but I am pretty sure that's why acceleration did not feel right. As of now, it is at 4* offset. I just drove it and it feels normal. I still feel a slight ripple in RPM under light load, so there are still adjustments I need to make. It does smooth out as things get warmed up. Log #4

Once in maybe 10 starts I will get the crank "lockup" and a sync loss while it is set to falling edge, When I do, it looks like the timing is advancing too early. Slightly raising the crank rpm helped this I believe. Log #5

It always wants to start better when it is warmed up, so I will have to verify this current configuration tonight after work.

Side note: In old threads, I keep seeing an earlier setting for ignition that said Miata 99-00, and now it is only Miata 99-05. How could the setting be different a few years ago and now they are classified as the same?

Mech5700

Pretty much got it dialed in now

The last thing that I can't quite figure out yet, is the spark advance when it hits maybe 200ish rpm. In the log you can see as soon as spark advance hits its peak, the rpm's dip for a split second and jump back up while it is still cranking. On the cold start, the cranking RPM is slower and the effect is more dramatic, and will stop the cranking for a moment. If i let off the key and start again, it starts almost immediately. Even with "cranking rpm" set to 400 and the "cranking advance" set to 0*, the ignition table is being used too early. The workaround I found so far is to put the advance to 0* in the timing table on the 700rpm, 70-100kpa cells. Right when it catches, the cell pointer will jump up to the next cell over and start advancing the timing. It's not perfect, but it works so far.

hector

I wonder what is different about your car that requires such off-the-path settings to work?

About the only real difference between you and me is VVT and ecu, and my car runs off the settings evreyone else uses.

Braineack

iTrader: (62)

the spark advance on that table is fine, it's going through a spot that's demanding 27°. Open the tune in MLV along with the log to see--looks like 1500RPM to me; the car is running at that point, why would you put 0° in the cells where you did? I dont see this on the log posted. The ramp up in spark angle is fine.

Mech5700

Why is it trying to hit 27* before it's even done cranking? It immediately stops the rotation of the crank when it goes from 0 to 27 at 200 rpm. It thinks it's running before it even gets over 200...

Mech5700

Check this, it is from this morning.

Advance spikes to 19*, and the pointer is at 0* on the ign table.

Why.

Mech5700

Tired to set crank advance to -9*, to see if it jumps 19* and goes to 10* total advance.

Fail. Jumped a total of 28* from -9* retarded to 19* advance during cranking. Also had a sync loss AFTER IT STARTED, and it just fell on its face and died.

If you watch "Base advance", you see the spike to 19*. If you watch "Spark table1", you see it follows the table (0*)

Braineack

iTrader: (62)

because at 500rpm and 92kPa your map is demanding 19° timing.

when the rpms drop below the cranking rpm value, they drop back to -9°.

you're the only person I've ever worked with using values like that. I'l built hundreds of MSes and everyone is just using 10° cranking and my default timing map.

you get a sync loss cause the motor cut out, it appears to have run lean and died.

Mech5700

Where do you see that?



It sees 19* adv @ 900 rpm, and this is happening at 200 rpm.

I have set these values to try and find out why it's doing this.

And with crank advance at 10*, it jumps to 29* advance while its cranking!

hector

You have idle advance on and it has 19* although at 38% load. Can't say this is it but maybe turn it off and see.

Seems to me the commanded 19* is at above cranking speed. Like at 578rpm.

Braineack

iTrader: (62)

I see it now, i was using the wrong msq in MLV.


I think it's because idle advance is turned on.

turn it off and just enable idle rpm timing control -- that uses spark timing to for idle control; it's really good.

idle advance locks the timing based on the table; it's not so good.

Mech5700

I noticed that as well.

I set it to not use PID activation, and set the parameters so it wouldn't engage idle adv till a higher rpm and it still hung up.

I will give that a shot, I haven't tried it with idle adv completely off. But that is the only thing I see that requests 19* of timing.

Mech5700

Turned off idle advance, turned on idle RPM timing correction, and put the ignition table back to normal.

It got a little better, still hangs up hard. The RPM doesn't fall to 0, but you can see it drops significantly. In this log I kept the key held forward, and you can see the voltage drop really low while it tries to overcome that stroke.

Matt Cramer

I'm seeing a processor reset at the end of the log. Was the key switched off or power to the ECU interrupted, or did it just stall?

Mech5700

I hope this ordeal can help someone in the future, I really do...

Obtained a new igniter (more friendly with MS3) and a used crank pulley from Bryan @ FAB9 for a reasonable price. Installed this morning and here are my findings...

Rubber failure on the old crank pulley


This is the new pulley stacked on top of the old one. You can see how far the outer ring of the pulley walked. You can also see that the marks I painted on the pulley are 10* off as well. I think this is partially where the "it sounds too advanced while cranking" came from... When I originally found TDC, I marked where the timing cover 10 and the T corresponded on the pulley. THIS IS NOT TDC. When these marks line up, it is actually 10*BTDC!


Thanks to Matt @ DIY, he directed me to this
This is what you should see with your timing light when setting base timing with the fixed 10* advance on.

MegaSquirtPNP by DIYAutoTune.com

That woulda helped since the beginning...

So far so good, it got warmed up when I was setting base timing again and starts great. As always, only 1 shot a day at this so tomorrow morning will tell me the truth.

Mech5700

Key switched off if I remember correctly.

Mech5700

Here is the last start I did, I still don't see why it wants to advance the timing per the table, then thinks again and drops it down to crank advance, then back to the table commanded timing. Is this the effect of the crank RPM setting?

Mech5700

Experimented with crank RPM setting. I wanted to see why the crank would slow down/stop sometimes while starting, even after everything I just did.

I set crank RPM to 1000, just so it wouldn't be advancing any timing till the engine was actually running, and set crank advance to 0*

Still stopped when cranking.

Back to square one, scaled back cranking PW% about half as much as it was currently (180*=100%, to 180*=50%) and it started every time after that without the hesitation.

Matt Cramer

It uses the cranking advance setting below cranking RPM and the table above that. It also uses the table if the engine has been running for a while (30 seconds or more, IIRC) and drops below cranking RPM - this was for engines that operate at very low RPM like some off road vehicles.

Next time it won't start, try pulling a spark plug and see if it's wet with fuel (too much) or dry (too little).