Rrrracer

Good work man. We were almost in your neck of the woods, Michigan's UP and were thinking about heading west but decided to go down state instead because of the weather Hope you get 'er dialed in before the snow hits!

Jesse99James

Well, I have an almost perfectly stable idle up until hot restart. It'll start pretty well but still falls on it's face hard when in transitions from cranking to closed loop idling. I extended that transition which helps but there's still an AFR oscillation that I can't figure out. AFR in general is damn inconsistent. One day I'll spend an hour tuning idleVE near perfection, the next day after WUE ends and its fully warm it'll show too rich (~13 AFR). If I tune idle again that day, the next it's too lean and unstable. It was running really well this morning after another hour tuning session in the garage. Idle was being caught smoothly 99% of the time with almost no droop so I took it out for top down cruise to run VE analyzer. I set it on Hard mode with limited authority to limit changes. I parked and went into a store, came back out and did the same on the way home. Both times there were smaller changes leaning out my cruise area cells. Great but the third time I took it out VEAL added almost 10% back into some of the same cells it had leaned out.

I'm starting to wonder if there is something wrong with my car. The Hydra Nemesis was also very inconsistent concerning AFRs which is why I replaced it. I got tired of idling at 11-12 AFR and figured it was a track focused tune, never fully polished for street use. Since the forecast looks like tomorrow is the last day of the season I'm going to focus on getting it into storage mode. Mainly fuel stabilizer, cleaning it, and trying out my newer, used wheel dollies. Next Spring I'll probably pay for tuning assistance as I really don't know what else to do...I've gone through almost three tanks of gas. I will have to get a better laptop if I want to do live tuning sessions. I can't see pairing anything to my phone for street tuning as I only have 5GB data limit but who knows, maybe remote tuning could still be done through a tethered data connection and text messages or phone calls for instructions. Maybe live tuning will have to be focused on starting, warm up, and idle tuning since I still have Wi-Fi in the garage. The rest might have to be done from analyzing logs, loading a tweaked tune and more data logging. Thankfully I found some lower risk, lower traffic highways where I think I can fairly safely record 4th gear WOT runs up near redline with less risk of a reckless ticket. There are so many traffic cameras around here!

I edited and added the last tune file of today as well as a log of three hot start attempts where the first two stalled. The third start I held RPM with the throttle and it calmed down to a great idle.

Rrrracer

Your ignition timing is all over the place on that log. You have Idle Advance set to On; as soon as PID engages, the Idle Advance drags timing back to ~6*, then your Idle RPM Timing Correction curve is fighting to bring it back up. Try turning off Idle Advance, but leave Idle RPM Timing Correction on.

You might also consider smoothing out your Idle RPM Timing Correction curve; currently it's set to add or subtract 10 degrees of timing at +/- 100 RPM, which is not much of a curve and is probably too aggressive.

There's other problems, but let's see what this accomplishes.

Jesse99James

That +/-10 Idle RPM Timing Correction was a recommendation by Curly and I'd seen it elsewhere but didn't know at that time where to try it. It made a huge difference back then and I recently tried a more gradual curve but switched back yesterday. I'll try less aggressive settings today. I also changed the Idle Advance to be less aggressive over the last few days. Originally any load caused idle destabilzation or near immediate stalling but my fuel map was quite bad at that time.

Last night I downloaded the TruboKitty MSM tune and compared each and every difference using my 60" HTPC to view that tune with mine on the laptop. Against better judgement I made lots of little changes hoping it'll get me closer to where my map needs to be. Looking back, I must have had an older TruboKitty tune saved or am only now recognizing just how close it is to where my map is headed...much better than the base map that came on the MS3.

Thanks for the guidance, I greatly appreciate it! I'll try the tweaks today and thankfully the extended forecast says I might get a few more driving days after work later this week into next week. It's been dipping into the very low 30s at night and we're lucky to see 40-45F this time of year. I need to start travelling on my annual leave to find where I want to move when I retire. Like you, I'm headed further south. I don't mind snow, just not 5-6 months of it.

curly

iTrader: (25)

I'll agree with idle advance needing to be off, however I've never had an issue with that timing correction setup, unless something else isn't tuned properly. Although your ASE taper time is sky high, you're not using much ASE itself, try increasing that, maybe never go below 24%. And under cranking settings, change ASE units from cycles to 0.1 seconds, 499 cycles is over in ~30 seconds, less if your start anything higher than your target idle. 499*.1 seconds is 50 seconds.

Jesse99James

Ok Rrrracer, that seemed to help now that my fuel map is close. I don't have Tuner Studio Ultra which allows custom dash setups so I left a few tables open and watched Idle Advance and confirmed what you said. I noticed it in the logs but didn't understand why it was dropping to 6 degrees. That 6-8-10-12 degrees was the base map and I think Idle Advance was ON but Idle RPM Timing Correction was OFF.

I extended the highest load out a bit more after turning on headlights with the blower motor on high. Of course it jumps way out when A/C is engaged but I set that highest load setting just above where it settled. More importantly I changed the lower loads to 10 degrees so the Idle RPM Timing Correction Curve is no longer fighting to drag it back up so far when idle RPM dropped. Since my base idle timing is set to 12 maybe that'd be the best setting so it's not retarding timing at all, only helping assist with higher loads.

I did more VEAL tuning and unlocked the idle cells leaving IdleVE engaged after saving the tune as a new name. I transfered the slightly leaner numbers back into IdleVE but haven't driven yet with those numbers...should be fine. Slowly, as the fuel map is leaned out I'm adding a bit more WUE to compensate but still saw 12 AFR with the old IdleVE numbers. I guess better rich than lean as it warms up, right?

Hot restart was better, only stalled the first time. The second time it settled like I'd expect but that was only one hot restart test. Thanks for more input Curly. I'll try turning Idle Advance off. After today's drive and log I again tweaked the Idle RPM TIming Correction Curve so it's a little more aggressive as RPM drops. With headlights on and blower motor on high I saw RPM drop down near 600-650 RPM but it quickly recovered without oscillating. I also changed the so it ramps up a bit faster but not as aggressively as +/- 10.

If you look at the newest tune I added quite a bit more ASE after comparing to the TruboKitty base tune. Thanks for the math explanation of taper cycles/seconds. I knew it ended quickly but didn't realize how quickly. I've read about Miata hot restart issues and IAT sensor location and your changes to increase 170 and 180F ASE taper match what was in the base map that came installed. I'll play around with cycles/seconds but hopefully based on my two hot restarts after it heat soaked while I was in the store, I think I'm on the right track to what my engine needs. Unfortunately, I forgot to log that first hot restart where it stalled. I did capture the second restart which was successful and felt/sounded normal.

Rrrracer

Curly, interesting! I've never tried running a car with the timing correction set that way, always used a more linear curve. I'll give it a shot sometime to see how it works. I also use seconds instead of cycles in ASE, makes it that much easier to look at a log and determine how long it needs to run.

Rrrracer

Unfortunately your log cut off just as the WBo2 kicked in lol. Idle timing correction is still fighting Idle Advance when PID kicks in, just not as much this time (Idle Advance was set at 10, spark correction was +6 deg)

You might consider re-binning your CLI Initial Values to include a row at higher restart temps so you have more accurate control over the IAC valve on heat soaked restarts (I noticed your top bin was 140*F)

Try this cal... all I did was turn off your idle advance, make your spark correction more linear and flatten the base spark table idle cells to add stability. Worst case, if it sucks you can put it back

Jesse99James

Thanks, I was going to try disabling idle advance next but wanted to see what happened with 10 degrees in the lower cells.I was thinking Idle Advance and Idle RPM Timing Correction worked together but now I see how they can fight or exacerbate instability. I am still too often perplexed looking at logs and being a bit overwhelmed with all the on/off states shown below the gauges. Upgrading to Ultra seems like a good choice so I can create custom tuning views (when I know what I needed to be looking at).

I didn't get a chance to look at that log and stopped it quickly after the engine started the second attempt with a stable idle because somebody wanted my parking spot. I'm looking at your changes on my 60" HTPC. My car idles around 32-33 kPa and I had a bit of advance in the lower RPM and low load cells were to help rebound or avoid RPM sagging with the 10# flywheel. I also still need to tune CLI PID and the sensitivity slider, last night I read about how tune it.

Your Idle Timing Correction "curve" looks similar to the one that was in the basemap. Maybe it'll work fine now that I fixed the RPM columns for Ignition and Fuel and my fuel VE cells are close. Idle has been much more stable. I had noticed the 140F in CLI Initial Values and was wondering...those are still the base map temperature values. I also didn't notice the active cells being highlighted during cold starts (45-50F) like you see in WUE as the highlighted oval moves through the active cells. I'll have to pay more attention when I rebin. Only when it was warm was I seeing which cells were active. I think taking a break over Winter will let what I've learned sink in and I can sleep on what I still need to try/do.

I think I finally got the FM reroute and crossflow radiator leak free. Took THREE attempts with the gasket at the back of the engine, the first attemp was my fault not letting the silicone set/cure. Second attempt was a dry gasket, still seeped through. Third attempt was a NOS gasket where the adhesive almost fell off so I used silicone on both sides. I had also removed the Koyo drain plug initially and struggled to get it back in...think I finally got it tight enough. I don't remember leaving it that damn loose and hope it wasn't cross-threaded which would mean draining yet again and taking it to a radiator repair shop. I wish I'd stuck with my low mileage MSM radiator with new hoses. The next project immediately after it's parked is removing the differential for 3.63 gears. New axle seals, input shaft seal, and pinion lock nut are on the way so I can take everything to a highly rated local driveline shop. A little progress will help balance this last month of tuning issues.

Thanks again!

Rrrracer

Right on man, holler if you need some help

Jesse99James

Good news and bad news. I started getting sick last weekend on my long Fri-Sun off weekend, mostly a very bad headache that maximum doses of pain killers barely touched. I had minor sinus congestion and no fever, figured it was another sinus infection headache but suspected Covid. I worked Monday wearing my mask in the office (mostly on the street by myself all day) and felt worse. Tuesday I called in sick as soon as I realized I couldn't smell the coffee brewing nor the peanut butter I just spread onto my muffin. Got tested a few hours later and learned within 12-13 hours that I'm Covid-19 positive with a break through infection (Moderna).

So far symptoms have been fairly mild other than the headache. The loss of smell.taste was right on schedule around day 3-3.5. I'm not coughing much at all, more of the feeling and slight urge but I'm hoping now that my sinuses have cleared that it's not moving into my chest...won't know be sure for another day or two but I'm hopefuly the worst is behind me and the vaccine allowed my immune system to rev up faster. I've read a few studies, more point to loss of smell as being related to those that have more mild symptoms avoiding the hospital. Another that said the opposite stating the more intense the loss of smell/taste, the worse the symptoms seem to be. So, not surprisingly, nobody still seems to know and the data collection I've experienced didn't ask about symptom severity. Maybe it'd have been more helpful to rate each symptom on a scale of 1-10 instead of just clicking boxes stating you have/had that symptom!

Since I'm off work for 10 days after the onset of symptoms, and feeling better today (less headache, clear sinuses) I worked on tuning. My Miata idled ok with the minimal RPM timing correction but while tuning IdleVE yet again changing RPM to match the table and load, it was hunting more than it used to. Turns out mine likes something like +/- 5 degrees as the first step instead of a curve. I'll play around and see if I can't minimze the curve more as I dial in the tune.

I played around a bit with tuning EGO PID control. SImple EGO mode allows quite a bit of drift around target +/- two to three tenths AFR (seemed more easily read to type instead of +/- .2-.3, at least to my failing eyes). Anyway, I'll mess around more with tuning EGO PID later. I turned it off while tuning IdleVE. Although it's supposed to filter the EGO correction, seeing the gauge dithering (?) bothered me as I don't fully trust auto tuning idleVE cells. I'll play around with CLI PID tuning tomorrow if I'm still feeling well.

I found two resistors from a long dead LED turn signal project where I planned on drilling 10-12 mm holes for a row of bright LEDs in each side mirror. I'll take the inline controller I recently bought and add a terminating resistor on the controller side. Maybe that'll increase accuracy by decreasing error. I think I'll cut into the outer plastic bundle sheath so I can pull the two wires up/out slightly. Then stripping a little insulation to wrap the resistor wires around before soldering. Spacing out the connections will allow the resistor to be aligned with the bundle and I have a fresh bottle of liquid electrical tape that be easier than taping before tucking everything back into the sheeth. May I'll add a little more liguid tape before taping the bundle sheath. Since this weekend is warming up into the 50s, maybe I'll swap out the glove box gauge controller for the inline controller and just use the TunerStudio gauge. The gauge controller wires are overly long and the gauge is difficult to read with the top down. While nice to tuck away and close to box, not attracting attention I think the inline controller with short CAN wires and adding the terminating resistor will increase accuracy. ??

Ok, enough of my bored rambling. Hopefully I'll be able to update next Spring with a tuning success story or to get more guidance but if Covid takes me down, the engine starts and idles well enough that it can be driven to a new home. My mother and brother certainly won't want to mess with this project!

cerevo

Dude, THANK YOU, I have been following the OGPedXing videos to tune my idle in on my new MS3pro PNP on my 04 MSM and I keep getting idle dialed in so good, then I'll blip throttle and it'll start going dead lean on a few cells, and then just get worse and worse oscillation in it. I started just richening up all the cells it leaned on out, got that nipped, then it was doing it when turning on lights, a/c oddly enough had pretty good recovery.
I'd think I had it about 90% good, turn the car off for a few hours, get in it to go somewhere and all of sudden 2 or 3 cells in my idle VE are super lean again and causing oscillation... I went through this about 4 or 5 times, back and forth, it was too lean then it was too rich, I'd fix the oscillation and then the next start it was back.

I couldn't figure out what the heck was going on, I was debating making a post. I found Jesse99James thread here and started reading through it, seems like he went through this whole thing a few months before me, I followed your advice, as well as learned some things kind of along with him about the issue and in the end, I just turned off idle advance, and adjusted my rpm correction to be a little more aggressive, it idles almost perfectly no matter what changes. Which to me means I had the idle dialed, I just needed that last piece of the puzzle. I'm shocked this isn't mentioned in the OGPedXing video, but then again maybe the NA's need it on or something, I don't know.

I also agree that my base map seemed way further off than what others experience. I have a pretty much stock MSM other than the little enchilada kit so I would have thought it'd be pretty close.

Anyway, I just wanted to say thanks because I think your advice saved me a lot of the headache that Jesse99James went through