im having an issue where my car is running perfect at 14.5 afr, then suddenly pegs rich for a couple seconds, then recovers. it also happens while driving and feels like a moment of hesitatio. nothing in the log shows anything wrong. i noticed it with my diypnp, but it is more pronounced on ms3. probably due to better idle control with ve table.

if a coil dropped out, would it be enough to send afr to full rich on wbo2? does coil trigger record in datalog? i want to see if it happens on the same trigger every time.

It does sound like a coil going bad. I was able to find which one was bad on mine using the 4 widebands. 2 & 3 were going full rich when it happened.

Pull a plug, see if it does the same thing, my guess is not. My car will damn near stall if it looses 2 cylinders. You'd know for sure.

(Posting from my phone.)

Post the log. There's no cause that I can find in the megasquirt. The map rises a little just before the event. Then the rpm droops as the airs dump to 10:1.

Is it constant? I wonder off you can suit on the coil bench test mode for a few minutes and see if you can pinpoint a coil that's missing.

It seems kinda random. It's very obvious when it happens at idle, and I cn feel it when it happens while driving. I've got a few spare cops. Before I change the plugs I'm going to let it idle and replace one coil at a time to see if it disappears. I'll mess with the coil bench test mode and see what I see. The log is too big to post, I'll try to take another.

I changed the COPs one by one, and still get the issue. I did the trigger log,and when it happened the y axis increased significantly. I'm not really sure what I am looking at though. I've uploaded a datalog that shows it happening a few times.

Update: I replaced the spark plugs with NGK 4091 (resistor-less). Unfortunately this random rich condition is still there.

I used the realtime logger and did notice an increase in the pulsewidth when this happened.

At first I was thinking that maybe the fuelpump was receiving a voltage spike. However, this would not be logged in tunerstudio as an increase n pulsewith.

Possible culprits:

CAS: not sure how, but it is responsible for synchronization of fuel and spark. Hopefully one of the smarter individuals will chime in on this. I have a trigger wheel to install in the CAS, but I'm unaware if this is a solution for a worn or failing CAS since I am unsure about what actually fails when they go bad. I do have a spare CAS though. If this sounds like it could be a culprit, then please let me know.

Software: Something in software not configured properly. Always a possibility, but I have gone over everything pretty carefully looking for something that seemed like it could be a possibility.

Firmware: I think this is a possibility. maybe the firmware didn't write 100% correctly. Although the issue seems random. Today it seemed like it followed the same pattern as it did the day before. I'm concerned that something is going on behind the scenes that isn't showing up on screen. For instance. My CL idle dash light isn't on all the time while idling, yet all the parameters suggest that it should be on. It turns on, stays lit for a while, then turns off all while the idle remains constant. This makes me wonder if ASE or WUE is kicking in randomly for a second and then turning off. this would easily cause a momentary rich condition.

Hardware: Not likely. I don't see what could randomly send a signal to momentarily increase injector pulsewidth and then go away.

Ignitor: spark could be dropping out, but I would expect PW to stay the same. Realtime display suggest the pulswidth increased by ~ 12% going form 1.4ms to 1.6ms

I think I am going to upgrade to the pre alpha 1.1 version 14. I think there is enough difference in the software that if it were firmware related, then it would probably be removed. I guess I could reflash the current firmware, but where is the fun in that. Part of the reason I upgraded to MS3 is because I am a firmware junky.

I'm going to hold off doing anything until I get a few responses/suggestions

The log below, the PW follows the GVE (fuel map VE %) as the MAP increases. I dont think it's enough to be causing the dip, the dip happens a few moments after the fact.

really seems like the coils are dropping out.

there's a couple of instances on your log file where the AFRs dont drop rich, but the RPMs and MAP become really unstable.



https://www.miataturbo.net/attachmen...1&d=1316474510

??? SD card??? Then what?

one thing I find odd the the idle valve seems to be doing nothing... the entire log it's just slowly opening more and more.

If that is the case, then what are your thoughts about the CAS?

its not losing sync.

Bad idle valve. I found it odd that my car wasn't in closed loop idle while idling.

About to log to SD card

sending you a log. Go to 175s. Good pic of what's happening. Looks like I have been chasing a ghost. Something else going on.

I think I have to accept the fact that my car has a HG leak. Too much water is coming of the tailpipe while idling. I end up with a small puddle, and just before I finished I started to see little puffs of white smoke out of the tailpipe. I think the HG stop leak I had in there keeps it at bay if I am driving it regularly and circulating the coolant. For kicks I decide to run it a 4k rpm for a while to circulate the coolant. Low and behold that mysterious issue I had disappeared.

Hopefully it won't take too long to change the HG.

After a lot of googling it seems that this water issue is not uncommon during periods with a lot of moisture in the air and idling near stoich. I don't know. I'm going to button everything up and just start driving it. To me, the instability in the log would not lead to such a rich afr.

Doesn't seem like anything on the log would be triggering it but are the accel enrichments randomly kicking in, maybe?

(edit) Nevermind. Just reviewed your log and the accels never kicked in.

I think I found the culprit last night.

I think it is the VICS vacuum lines. The VICS solenoid was bypassed on the dyno because the tuner said it actually hurt performance when they were activated at 5200 rpm. The lines going into the solenoid were plumbed into a t fitting with the 3rd end capped. This results in the manifold being directly tied to the vacuum pod behind the throttle body.

It appears that the vac lines are slightly too big for the ports. When enough vaccuum is pulled, air is able to momentarily leak past the rubber tube, but then once the vac is reduced the seal is sufficient to discontinue the leak. Vac builds back up and it happens again. However, there doesn't seem to be an exact amount of vac that causes the leak so it appears random. I'm going to get new vac lines today, and hook up VICS again. Maybe I didn't have them working right on the dyno. Peopel seem pretty confident that it helps low end torque a little. Not sure if my cam gear has a negative effect on the VICS.

thats because VICs should be ON until 5300 or so.

settings should be 0 for start, then 1 for active at the desired rpm?

should be:

trigger 1 | 0

rpm > 5300 hysterisis 100

so solenoid is powered below 5300.

Brain,

When performing the coolant pressure test while running the car, are you just checking to see what pressure it maintains? If so, what is reasonable? I'm going to perform the "running" test tonight.

IIRC, you're supposed to pump it to around 15psi or so, then start the car, if it drops and/or pulsates theres a problem.

Well, I started the car with it pumped to 18 psi. It never dropped or flickered when I started the car. However, after about 10 minutes I looked at the gauge again and it was at 25psi. The shenanigans also started again when the pressure was high. Once I removed the tester and replaced the radiator cap it did not hicuup again.

Seems like it is pointing to a HG leak.Damn I don't want to replace the HG.

Well, should I buy this for confirmation.

https://www.napaonline.com/Catalog/C...006_0282509199

Or just put the money towards a head gasket???

if your using the tester that hooks up to the radiator cap it will get above 18psi when its hot because 18psi is when the radiator cap opens and lets the coolant into the overflow

I guess I am going to find a leakdown tester. The HF cheapo model seems to get mixed reviews. However some sound like they aren't using it correctly. When I turned the car off it did hold 25 psi for what that is worth.

I think the problem that I am dealing with is a firmware issue. I decided to take it for a drive tonight (no log) However the drop out occurred and it is too rhythmic to not be something hardware or firmware controlled. Unless the two capacitors in my COP harness good be affecting the voltage getting to the cops, then I think it is firmware related.

I've started to notice a pattern in some of the events. I've seen in multiple instances that one pules will occur 25s after previous pulse. Here are some logs for kicks.

https://www.miataturbo.net/picture.p...pictureid=1814
https://www.miataturbo.net/picture.p...pictureid=1813
https://www.miataturbo.net/picture.p...pictureid=1812

here's your MS running on my STIM with the VR input for crank and MS3x input for Cam.

hopefully we don't see these gaps anymore.

https://www.miataturbo.net/attachmen...1&d=1317570810

Very nice

I put my MS3 back in after having a new circuit built that uses the MS3x board. The rich spike still occurs, but I never have a problem with the the CAS signal dropping out.

I decided to pull the head. I've been getting water out the tail pipe for some time, but my water level never seems to go down. When I pulled the head I noticed that all 4 cylinders had a clean strip that started at the edge of the cylinder and moved in towards the center. I think this would indicate coolant since it kept the piston type nicely cleaned in this area.

https://www.miataturbo.net/picture.p...pictureid=1837

The nice part about pulling the head is that I have been able to look into my fears of det with my OP gauge flickering. There didn't look like there was a single pit in any piston. They were perfect. I'm going to use the opportunity of the head being off to add an aftermarket OP gauge since the sending unot is easily accessible at the moment.

https://www.miataturbo.net/picture.p...pictureid=1838

I didn't want to wipe the pistons down before I took the picture. The coolant makes them look a lot worse than they were. I wiped them with papertowl,and removed a very thin film of gunk.

SD logs
 

Here are some clips of my SD logs converted to .msl files

This was posted over on the msextra forum:

Re: MS3x random rich condition
by wrenchdad » Thu Oct 27, 2011 2:23 pm

I had this same problem, could see the rich spikes at idle and had like constant missing while cruising PLUS sync lost. When I rewired my condensor fan output from using INJ F to an unused VVT output, no more misses, idle smoothed out but sync lost when down to 1-3 per hour of driving before it was 30+ in 10 minutes.

hope this helps
later wd

Then try moving the wire on p4 on the MS3x DB37 to pin 27, copy the settings of INJ F to VVT under outputs.

Good news! I'm going to leave everything alone and test alpha 15 when it comes out.

EDIT: There is a sync loss thread over on the Msextra site as well, and it was confirmed that the error can corrupt the PWs. It might be noise from the INJ outputs that is fed back into the system. I'm just glad I wasn't crazy. I guess I got the experience of changing a head gasket :)

tech speak

Solved! Alpha 15 code solved the issue

great, glad to hear.