Alexj84nl

Thanks again for all the replies and help, really appreciated.
Without this, I would have still been stuck way further down the line.
What I did
- gave the FPR it's own vaccuum line (oem)
- connected MAP hose from MS3 to the cruise with T splitter
- gave the VE map lots of additional (+10 and later on +20) fuel in all regions except idle (which is pretty rich at the moment, around 13 / 13.5)

I didn't warmup the engine, just took it for a short trip around the block to see if the changes would have made any effect,
At first, I thought it was just a bit smoother on the throttle. At least from idle it felt that way.
But when driving around it turned out it wasn't and after 2 minutes I parked, I quickly pushed throttle, and it stalled.


Screenshot below shows the stall from idle.
AFR 13.7 and RPM 1000
Throttle from 0 to 3.2%
Immediately, RPM goes down and AFR goes sky high.





Screenshot below is me driving around and applying throttle.
Car starts shaking (too lean I guess) so you see multiple green spikes (my foot hitting throttle). Notice the lean AFR when applying throttle.


As a reference my previous VE table (below), and my VE table (above).
I did try some values in between.





So I'm not sure what to do at this point.
Maybe someone notices something in the screenshots that can help, like weird MAP values? Is the line broken/squeezed somewhere?
Otherwise I will have to switch back to OEM ecu and see what happens. (not looking forward, it's cold/raining and I don't have a shed/storage place. And, it ran great before conversion so not sure if it will help.

Panici

Fuel:
Change that left column to 500rpm.
Fatten up that idle, and make it a progressive increase from 15kpa to 45kpa instead of blocked off like you have now.

You can idle it down all the way to 11ish AFR before it starts rich misfiring.

Can always lean it back out later once drivability is there.


Spark:
Post your spark map.
Make sure your ignition table is also binned down to 500rpm on the left, and pick up the timing a few degrees in the columns to the left (relative to idle @1000rpm).
I would expect idle timing in the 15-17deg range if the base timing is set properly (with your 800rpm and 500rpm rows timed a little higher then the 1000rpm column as mentioned)


Other:
Oh, and brace the right side of your foot against the transmission tunnel (assuming your car is LHD), and roll it onto the gas pedal. That's an old 4x4 trick to prevent bouncing the throttle inputs when in the rocks.

Make sure you warm up the engine before your testing. Otherwise you're fighting Warmup Enrichment (WUE) as well, further complicating things.

Alexj84nl

Thanks for the help.



Both my ign and my ve table where more progressive increasing when I started out.
Especially the ign table is pretty increased, I was at 16 before for the idle region.
Earlier tips/feedback made me change the idle VE and idle IGN regions to equal numbers.

For now I did change it back based on your comments (see attached tune), but I won't think it will make any difference.
The problem is not in the idle region, but in transients from hitting the throttle.

Also, I tried to disable Acceleration Enrichment, not sure how, so I put the (accel tpsdot) threshold on 1000. I will not hit that (with the crappy performance).

Thanks for the other tips, will try to practice with the foot/tunnel.
About warming up the engine, first time I didn't, I hoped I fixed the problem. (problem is that noticeable, it doesn't matter if WUE is active or not)

Because Im going over all things again, trying to find the issue.
I had a hard time setting the base timing because there was no clear marking.
I had to put the 'trigger wizard' for timing at '-0.5' degrees in tunerstudio to get the results posted below.
Wonder if you think this is correct.
Video: Photo/screenshot from video:

Nethawk

Nothing stands out in your logs or tune.

Some thoughts:

• we can trust the wideband sensor is working properly, because the stall/stumble correlates with the lean reading
• lean reading means O2 in the exhaust, so either not enough fuel, or a misfire
• your latest VE table already has all of the relevant cells a fairly high values, and there is no other corrections pulling fuel. This leads me to believe the MS is commanding plenty of fuel, but it may not being delivered physically.
• Thinking about misfire, there is no sync loss, and spark advance shows reasonable values throughout your data. This doesn't rule out coil or other ignition hardware issue though.

Honestly I would be looking for wiring, or other issues in the engine bay. Fuel pressure regulator could be bad? Your easiest sanity check is to see if the car runs fine on the original ECU.

Alexj84nl

I think it does, because it gave trustworthy values running connected to the OEM ECU.
Also, I'm running fully open loop (afaik), so the MS3 should be running solely based on all the maps (and not based on the wideband signal).

The FPR was certainly not bad before installing the MS3. I did remove the silicon hose from the FPR to insert the T splitter to connect the MAP hose.

I cannot think of other wiring issues in the engine bay.
This Miata was absolutely 100% stock and totally clean.
Sure, to install the MS3 I had to make changes, maybe I ruined something.

• remove the MAF, insert IAT
• remove TPS, insert VTPS
• remove O2 narrowband, insert Wideband O2.

Im not totally new to making changes under the hood and engine tuning.
Only thing I can think of is the Throttle Body gasket. (but, then again, would it give these kind of problems & wouldn't my 'brake cleaner test' have shown some RPM change)

Thanks for all the help, will convert back to stock ECU in the coming days.
Will connect it to my Wideband (as just before the MS3 install).

Picture of the main suspect attached, hope I got the issue fixed soon, miss driving around.

Alexj84nl

I think I have found the problem. Will require a bit more testing, but Im pretty hopefull this is the problem.
The T-splitter I used, might have been too restrictive to the airflow.

I disconnected the hose from the 'evap can' to the intake manifold. This gave me one free unused nipple, which I used for the MAP hose.
I took it for a quick spin and I didn't have any weird AFR lean spikes in my log.
More importantly, car felt pretty OK. Sure, way too rich, cold engine, etc etc, but nothing I can't fix with the tune.

When I look at the logfile, It seems like the MAP signal follows the TPS signal more closely, there seems to be less of a delay / discrepancy.

Tomorrow I will remove the evap system / bypass it, and roll back to a regular base map.
And then see what happens.

If everything works fine, I will eventually bring the evap system back with a proper T-splitter.

DNMakinson

iTrader: (3)

You still need to better understand the various parameters. For instance: TPSdot is a measure of how fast your throttle is opening, and is in no way dependent upon how your engine responds.

Unfortunately, we jumped into tuning when the problem was pneumatic.

TBH, we should have caught that there was a time lag in the MAP response shown in the very first graph you presented.

Glad you found the issue.

DNM

Reverant

iTrader: (10)

Yeah it makes sense if you look at it now. The ECU won't inject the required fuel in time if the MAP signal comes with such a delay.

It reminds me of another wild goose chase I had years ago, with a guy who had installed the wideband after the cat converter, and "the ECU was not responding to fuel changes fast enough".

Alexj84nl

I think we had a little misunderstanding, I think I do understand the TPSdot and MAPdot (not at the start of the thread, but right now). Not sure where the confusion started.
And by no means I want to give the impression I understand all the parameters, still learning a lot by reading various topics / manuals.
I'm really grateful for all the help and advice I got in this thread, amazing how you guys helped me out! Again (can't say it enough); thanks alot for all the help!

Still didn't had any time to move on with the project (busy december month at work), I will report back if the issue has indeed been resolved, hopefully this weekend.

Panici

Seems obvious now. Think I missed it as I typically keep TPS and MAP on the same plot.

Glad you found a root cause. Hopefully you picked up some tuning tips along the way.

Alexj84nl

Definitely, a lot! Thanks!

Problem solved. (it was that T-splitter, unbelievable, it's a standard automotive part I used)

Had a great run yesterday. Got the VE map roughly OK and it runs absolutely great.
Will have to dive into the rough idle and cold start issues, but that has been described plenty,
Thread can be closed.

Reverant

iTrader: (10)

Good to know.