Mega Tuning problems
 

I posted this thread last year: https://www.miataturbo.net/forum/t27138/ asking why I couldn't seem to tune the miata N/A for some sort of fuel economy. Some of you helped me sort out a couple problems but I'm still at a complete dead end.

What I've done:

Cured Misfire w/cops using a 15,000mfd 110 volt Cap
Re-wired the grounds on LC1 to front of motor
Started from scratch building a new .msq
Double checked all settings
Tested every VE map posted on this board I could find...

Problem is, I've used maps from Hustler, Paul, Brain, etc... With your VE tables the car won't even start. I've generated a new map about 10 - 15 different times using megatune, car won't start. Anything under like 50 in the first four cells the car won't fire up.

Doesn't matter whatever map I load and datalog, anywhere from 20 minutes to 5 hours of logging yields the same two results. 1. Car is so freaking rich that it gets 8mpg, but runs like a rocket. 2. Car is so lean it won't even move. It goes back and forth log after log after log, no matter what target table I use in megalogviewer doesn't seem to make any difference.

What the heck am I doing wrong? This is super frustrating and I'm completely out of ideas on what could be wrong? :vash:

I'm sure you've heard this one, but:

Have you trying reflashing the firmware?

The AFR target table is not your problem, so just pick a reasonable one and stick with it. For simplicity I just have MLV use the same 8x8 table from the .msq.

It could be a simple case of MLV adjusting too much and overshooting the mark. If you have it set to "easy" then bump it up to "normal," if "normal" then go up to "hard," etc. The better the car runs, the higher that setting needs to be. It didn't take very long on "normal" before I had to bump it up due to MLV overshooting.

FWIW, I've had MLV overshoot a little even when set on "very hard." I wound up manually editing the proposed changes to the VE table such that I cut MLV's suggestions in half. As it got really dialed in, I'd manually edit the changes to allow no more than 2 points richer or 1 point leaner at a time.

whats your setup like; injectors, tps, etc?

post your msq.

Won't start up at all? When you're turning the key, the startup fueling is controlled by your cranking PWs and those are absolute values in milliseconds. The VE table has nothing to do with that part. If you loaded a .msq from turbo guys who run big injectors then the cranking PWs will be way too low if you've got stock injectors.

I see you've got an LC-1 and COPs but what's the rest of your setup? Put some details in your sig like we tell the noobs to do. ;) :laugh: I've got my shit dialed in reasonably well if you've got a 1.8 and are still N/A with stock injectors...

Shit, I totally forgot to include the details.

Its a 1995.5 1.8, V2.2, MS1, 029y4, MAF delete, Stock Injectors, Toyota COPs, LC1.

This is pretty much the MSQ I've been using only changing the VE table. The very first thing I tried after the last problems getting this thing running properly it was suggested I scale the VE table without boosted cells. That was the very first thing I Tried.

Going into generate a table, enter some made up values which sounded good enough for a ve table; 1800cc, 900rpm idle @24kpa, 110lb tq @ 5000 100kpa, 115hp @ 5600 100kpa, 7000 redline 100kpa. The table that gives me won't even get the engine started, and if I burn that to the CPU while the engine is running it just dies. Has the idle cells at 22 and my car seems to require 60 in all four of those to even get an idle and read 20:1 according to the LC1. I've also re calibrated the LC-1 multiple times and its a pretty new o2 sensor. Checked required fuel several times and always come back with like 11.3.

Loading the VE table from anyone else also won't even let the car start. Someone mentioned in my last thread that my values need to be extremely high in order to make the car run worth a damn. Several of them exceed over 100 in value especially in lower rpm high map portions of the table. Completely baffled on whats going on here.

The MSQ's attached:

https://www.miataturbo.net/forum/ima...attach/msq.gif Scott-NA-Tune-10-4-2008-4.msq (35.3 KB) <--- The car runs really well for the most part, but at cruise the AFR shows 7:1 to 9:1, same under WOT.

https://www.miataturbo.net/forum/ima...attach/msq.gif Scott-NA-Tune-7-18-2009-Final.msq (35.2 KB) <--- Car won't even drive. Idles fine at a good AFR, pretty smooth, but lacks fuel to even accelerate.

that fuel table just impregnated me.

i bet this works great. should be safe to burn as is.

Looking forward to getting home and testing that out.

Am slightly concerned about some errors it got, 6 to be exact. Tried to attach the audit.log file but it doesn't like the .log ext and the file is too large for .txt. If its needed I can zip it and attach, but this is the important part here I think:


C:\Documents and Settings\Administrator\Desktop\megasquirt200907211 238.msq:Open
WARNING(01): Signatures don't match.
Found in msq "MSnS-extra format 029q *********"
Expected from ini "MS1/Extra format 029y3 *********"

WARNING(04): Constant "BoostMaxKn" units mismatch, "PSI" in msq, expected "KPa" from ini.
WARNING(04): Constant "KnockBoost" units mismatch, "PSI" in msq, expected "KPa" from ini.
WARNING(04): Constant "OvrBKpa" units mismatch, "PSI" in msq, expected "KPa" from ini.
WARNING(04): Constant "StartWIKPa" units mismatch, "PSI" in msq, expected "KPa" from ini.
WARNING(06): Constant named "latency" remains unaltered.

Warning Details
---------------

WARNING(01) means that the file you read was not written from the
configuration that you are currently using. If the
two configurations were similar, you should see few
warnings and most settings will carry over successfully.
If they were quite different, you may have lots of
problems getting things to work, but you need to make
the judgement call on this by examining the other
warnings above.

WARNING(04) means that you have different units set in the msq and
your current configuration. This almost always means
that the value of the constant is incorrect and you must
manually correct the situation. The easiest way might
be to reset your configuration's settings, say from
Celsius to Fahrenheit, then re-reading the msq.

WARNING(06) means that the constant found in the current configuration
was not found in the file, and thus has not been altered by
the file read. Make sure that the value of the constant
makes sense in the context of the other changes.

C:\Documents and Settings\Administrator\Desktop\megasquirt200907211 238.msq:Close

ignore it. Load it locally, and save it back out, then open it; all the errors should be gone.

Opened, saved. Re-opened without errors.

Reflashed 029y4 to the car and sent the msq. Cranked and it hardly started. When it would start, I had to modulate the throttle to keep it running. Tinkered with the warmup wizard with no avail. Once it completed warmup I attempted to drive it. Nope. Completely and totally impossible to drive. Car bucks and bounces, o2 pegged at 20:1.

Drove it anyway logging for ~ 15 minutes. Almost hit some jerkoff walking his dog down the middle of a street instead of the sidewalks. :rolleyes: After about 3 ~15 minute logging sessions it started to improve in a few places, especially 2nd gear between 2500 - 7500 (oops) and pulled harder than it has ever before. That was the end of improvement, every logging session that was run through megalogviewer just went down hill from there. Progressed back to the point where it was nearly impossible to drive again. :facepalm:

The table is back to huge ass numbers again. Am I doing something wrong in log viewer? Are my injectors bad? Fuel pressure maybe? Can't believe I'm having so much trouble with this shit when it really should be a piece of cake.

Attached is the very last log, and the msq resulting from it. Had to change a few things... over-run settings, taco output...

Didn't mention I'm running the LC-1 DEFAULT settings inside the LC1, is that a problem? Checked my ini file and its configured for the default LC-1 settings..

something is wrong with your shit then.

Maybe. When you are tuning the VE table in MLV, how big are the changes you are making? Does the average cell change by 5, 10, 20?

Does the car drive fine on the stock ECU? If it does, you can probably eliminate injectors or fuel pressure as the root cause of your problems.

Do you have a gauge for the LC-1, and does it agree with the gauge in megatune when the car is running?

how old is your fuel pump?
how old is your fuel filter?

Mine did weird shit like this when I replaced it and all problems went away.

Don't have an LC1 guage in the car but I did hook a second laptop up to the lc1 and observe the output from that and it matches up with Megasquirt's display. Think i've changed too much to put the car's stock ECU back in but I'll look into that. ALL emissions controls under the hood are gone... Actually think if i just put the MAF Back in I'd just have to change the values for the air temp sensor since I'm using the GM temp sensor, it might work... *sigh*

Fuel filter is less than 2 years old, has maybe 4000 miles on it TOPS.

Fuel pump, rail, pressure reg, injectors are all 14 years old. OEM. I'd really hate to replace any of that shit and have it turn out not the problem.

so have we pretty much ruled out 100% that megasquirt would be the problem??? Whatever I need to replace in order to get this pile of shit running proper i'll do... Just need to know where to start. :(

Also reset the tps values, I didn't mention that before... not like it made any difference at all.

if its not idling with that map i gave you something is fubared. get it to idle and start pulling the injectors and spark plugs one by one.

Okay, if I burn the exact unmodified MSQ you posted, the car idles like complete shit at 1200rpm and a 56kpa map.. If I burn one of the modified msq's that I got lastnight, like the second or third attempt it idles smooth at 900 and 23kpa map. WTF is going on with that?

I swapped out the injectors with some other stock miata injectors I had laying around, still does the exact same bullshit. I did find though if I squirt some starting fluid on the injectors the engine speeds up a bit, so I have a vacuum leak on the injectors (I hope) or the intake manifold.


Have another update. Plugged in the factory ECU and MAF, put the MAF inline with the intake. Car started right up, immediately. Idled perfectly. Left the LC1 putting out the wideband signal to the ECU even though it has no idea what the signal is trying to tell it IIRC. Car runs like garbage below 4000. Above 4k it pulls like a striped ass ape. Shit, forgot to hook that stupid solenoid up between the intake manifold and the Fuel pressure regulator. Hooked that gizmo back up, made no difference. Had LogWorks open logging my LC1's output and when you're under load under 4000rpm it was 20:1, once it hit 4000 and switched to open loop it started giving it fuel and the car started to pull.

Not sure where to go from here. That kinda sounds like an indication of fuel pressure maybe? Shit, I don't know.... The check engine light came on right as I pulled in the driveway but I haven't pulled that code yet to see what it could be crying about. The EGR valve is disconnected from the Electronics, and the two solenoids are missing along with the charcoal can. Its likely one of those three.

Did somehow I mix up the fuel lines when I put the motor back in 2007? Maybe have the feed line hooked to the pressure regulator or something? Would that make the car run this way?

Programmed the LC1 output to emulate a narrowband O2 sensor. Car runs like a freaking champ with the factory ECU. 100%.

Problem is 100% megasquirt related, but the question is... What is it??

I've reflashed the firmware multiple times, used Scott's MSQ with his settings that should work, no positive results.

Just got to thinking back to https://www.miataturbo.net/forum/t14615/ when I dropped some solder inside my megasquirt with it powered on. That fried the processor, but it was replaced and seemed to be working since. But maybe not?

Could that have damaged U7, Q2 or Q7 at all? I think the solder piece grounded out those outputs or maybe hit +12 volts from the power supply on that side (Pin 21 and 22 from the Proc) and fried one of the injector outputs. Could that have damaged the U7 injector FET driver?

thats why i suggested getting it to idle and then pulling the plugs and injectors one by one, seeing if one or two or many might not be firing/squirting.

If it runs like a raped ape above 4k, then you have enough fuel. I don't want to sound like I'm ignoring a potential mechanical problem, but it really just sounds to me like your shit just needs to be tuned.

The base map from DIY had my car running really lean below 4k. Your MLV-adjusted VE tables make baby jesus cry, and the msq Brain posted would have my car running lean as hell, first because the VE table values look a bit low, but also because reqfuel of 10.2 is too low for a stock 1.8 & stock injectors.

The displacement of a stock 1.8 is 1839 cc, your stock 94-97 injectors are 265 cc/min, and if you plug in an AFR of 13:1 into the reqfuel calculator it will spit out a reqfuel value of 13.1. Even if you typed in 14.7 as your target AFR, the reqfuel calculator would return a value of 11.6. These msqs with reqfuel of 10.2 must be from 1.6 cars (?). With reqfuel set that low it's no wonder you have to run what appear to be really high values in the VE table.

I'm gonna post my msq, but keep in mind there are a couple differences in our cars (I don't have COPs for example, and I run parallel with the stock ECU controlling idle & fans) so you might not be able to just load it and run it. You could modify it for your dwell and whatever else you need to change, or if you want to just extract the meat & potatoes, set your reqfuel to 13.1 and import my VE, spark and AFR target tables. The spark is from DIY's turbo base map and is safe. My VE table is not tuned much in the high rpm & low MAP cells (lower right area of the table) but it works great for both street driving and on track.

I run hi-res, which I did have to retune a little bit after upgrading from 029v, so it won't be perfect on 029xx but at least it's reasonably smooth and shouldn't be off-the-charts lean.

EDIT: I started typing all this out before reading that you may have screwed up your MS. But some of it still applies and it can't hurt to try a msq that will feed in a little more fuel than what you've been using.

Well I did that with the coils and one by one noticed a chance in engine behavior. They're all sparking. Injectors on the other hand I didn't have a good enough idle to tell for sure. Do need to replace my injector harness though because I noticed a few wires were exposed on #3's connector, the jacket was pulled back on them or something? Still seemed to make connection though. Hard to tell. Will try that again tomorrow with a pig rich ass map and see if that makes any difference or not I suppose.

ScottFW: I'll defiantly give that one a shot too, but it makes no sense why my car needs so damn much fuel to run even remotely well while boosted cars use the same IN BOOST as my map gives N/A!

Maybe I'm not understanding what you're saying so pardon me if you already know this, but you cannot just look at the raw numbers in your VE table and assume you're dumping more fuel into your motor than somebody else whose VE table numbers are lower. MS takes reqfuel into account when determining how much fuel to inject. If someone else has a VE table for a 1.6 (which it must be if it has a reqfuel of 10.2) then your VEs will have to be a bunch higher on a 1.8, because I see from the .msqs you posted that you haven't changed the reqfuel value.

With your smaller stock injectors you could also have higher VE table numbers but actually be injecting less fuel than someone with 460s or 550s with lower VE numbers, depending how reqfuel was scaled. Bottom line, there are other factors involved and the raw VE numbers don't tell you how much fuel you're actually injecting. They can give you an indication of relative fueling across the range of operation on your engine, but there are a few variables that confound any direct comparison between cars of the VE numbers alone.

Now if you were comparing pulse widths with another similar car (same engine setup, injector flow rate, fuel pressure, etc) and you needed much bigger pulse widths at 70 kPa than the other guy at 150 kPa while turning the same rpm at the same IAT, AFR, etc, then I'd be mighty surprised.

Not sure how much tuning you've done before, but my experience is that it took a couple nights of driving, logging and MLV'ing to have the VE table roughed in and the car running merely "acceptably." It just doesn't happen after two or three logging sessions. I've probably got 40-50 (maybe more?) cycles of logging & tuning between street and track to get mine where it is now, so that I don't feel compelled to bring the laptop with me every time I go for a drive.

I assumed Scott set my required fuel before he sent me that msq, I did notice that my required fuel was different. Tried that tonight, changing it to like 12.3 for reqfuel and made zero difference. About the only change it made which was readily obvious the idle map went from 54 to 33 using his msq. (Edit: I just remembered that I never set the required fuel BEFORE I tried to tune with log viewer..bet that makes a difference..) I know the raw fuel numbers in teh VE table don't indicate the exact amount of fuel going in, but I do know that its been pretty normal if one of those values on my table says like, you know, 120 I expect the AFR to be pegged at 7:1...

From what I recall however, I should be able to use another miata guy's fuel map without much trouble doing nothing more than changing required fuel and doing a little bit of tuning to compensate for differences in motors, turbos, etc...

http://jugrnot.com/tuning.jpg

I've been fucking with this since Jan 2008 with very little success. There HAS to be something broken with this piece of shit. Thats less than 1/3rd of my logs I've taken and run through megalogviewer, and not a single one of those is less than 10mb in size. Still hasn't produced SHIT worth using. Its either PIG ass rich or so lean it won't make the car go.

Please don't take this as a rude reply or anything, I'm just completely and totally out of ideas and openly admitting I've no idea what the hell I'm doing I guess. Really I don't care anymore, just want the damn thing to freaking work. Even approaching just paying someone to do this shit for me because its starting to get pretty obvious I can't do it myself. Really am wondering if there is any hardware damage to my megasquirt left over from dropping that piece of solder inside..

Swear to god, I have the worst luck with electronics. Just recently spent 2 weeks dicking with this 2 way radio I pieced together and programmed. Had it working like a champ, got a remote control unit cable for it and smoked the piece of shit within an hour. The remote cable has 2 power inputs for +12 volts, but if its in dash mount the rear of teh radio has a +12 input from ignition.. that ignition wire came off and scraped across the chassis when I was moving it off the desk to finish programming. After it quit working I pulled the radio apart and found there had been FLAME inside the radio. There went $160 bucks and 2 weeks of my life.

I might cut out of work early tomorrow and dick with this thing more. First thing I'll do is try loading your VE table, target table, spark table, and make sure I set my required fuel properly then see what happens. Will also make sure I check the injectors to be sure they're all firing.

Need some education on a few different things:

1. My LC1 is set for default output 0-5v .5-1.5 lambda. Is this optimal? I've noticed some people change it but I'm not sure why and if I might benefit from it?
2. I have MegaTune Configuration set for Default LC-1 settings
3. That being said, Are the settings I have in MegaTune correct for EGO Correction? EGO Switch Point (V): 2.352 volts, Ignition Events: 24, Controller Step: 1, Controller Authority: 10, Active above: 160, Active above: 1500rpm, EGO Correction: Ign Pulses
4. In MegaLogViewer: Calculated Fields I have Wideband O2-AFR set for "Innovate LC1 Default" and Wideband O2-Lambda also set for "Innovate LC-1 Default" is that correct? One or the other?
5. Do you use the MegaTune Target AFR Table in MegaLogViewer, or a larger table like some other guys in place of the "Default Target" table?
6. When is it appropriate to change from Easy, Normal, Hard, Very Hard when running the Analysis?

Combine these settings and you have everything OK. Some earlier versions of MegaTune had problems with the LC-1 default file being off, but this is fixed now.

Looks OK.

This is correct.

Agree with Matt, that stuff looks fine to me.
I used a 12x12 AFR table for a while, but now I have MLV do its adjusting based on the 8x8 table from Megatune. There's nothing inherently wrong with using a 12x12 target table that you create in MLV, but if there's any discrepancy in the interpolation of values between the two tables, your EGO correction while you drive will be fighting the tuning you did with MLV. I experienced that once and the car drove like crap during cruise. It's simpler to just use the 8x8, and I don't need a larger table for AFR targets.

This is getting to the meat of what I think could be the problem, given that you've done a bunch of tuning and have gone from lean to pig rich between tuning sessions.

Stoichiometry can be touchy. Have you ever done titrations or pH'ed a solution in a chemistry class? You have a basic (as opposed to acidic) solution, and you try to get the pH more neutral by adding concentrated hydrochloric acid. You add quite a lot of acid and the basic solution slowly becomes more neutral, but when you're at the point of neutrality, just another drop or two of the acid will swing the pH way the hell off neutral very quickly. It is easy to overshoot the mark. It's the same with tuning your car, and my experience is that it's real easy to overshoot with MLV, so you go from pig rich to super lean in one logging/tuning session. The exact same thing happened to me when I was learning to tune.

If you have it on easy or normal MLV will often suggest huge changes and have you overshooting the mark. Just set it to very hard. But don't blindly accept the changes to the VE table that it suggests. If you mouse over a cell, it will tell you what the old value was so you can see how much it's changing. A change of 10 is HUGE. With my table pretty well dialed in, if I take a cell in the VE table and add 5 it will run really rich, and if I subtract 5 it will run very lean, so when MLV makes a change of 10 it can swing you from one extreme to the other.

Even on very hard, it still may suggest changes that are too large. Mouse over each changed cell and manually edit it so the change is no greater than 5 in either direction. Do a couple logs/tunes like that and you should be creeping towards your target and getting incrementally better, rather than overshooting back and forth. As it gets more dialed in, make sure it's only adjusting 2-3 points in areas of the map where the car already runs okay. Then as it gets really close, only allow 1-2 points at a time. I wish there was a way to define settings more stringent than "very hard" but I don't think there is.

Tuning of the part-throttle area of the map (40-75-ish kPa) will be more touchy than tuning the WOT (~100 kPa) cells. This is because at part throttle you've got less air coming through the throttle body, so the amount of fuel injected has a greater effect on the AFR than it does when you've got more air coming in. You can get the WOT cells tuned pretty quickly but spend weeks or months dicking around with the part throttle cells, a point here, a point there. Fun!

Scott, Scott.. Scott here. :giggle:


Okay I used Brain's MSQ as a template, double checked EGO settings, changed my required fuel to 13.1, Loaded ScottFW's VE, AFR, and Spark tables.

Pulled out the o2 sensor, reset the heater and free air calibration. Put it all back together and fired it up. Cold start was slightly rough, but I think it'll be pretty easy to smooth out in a couple places. Let the car come to full warmup and was completely and totally shocked to see the car idling at a near SOLID 16:1!! Went for a drive and was also very shocked to see how close the table was for my car, it runs great. Decided to tinker a little and loaded up Hustlers psychotic and scary spark table on a 90 degree ambient day http://smiliesftw.com/x/hide.gif. Car ran pretty well overall but theres a couple little hiccups in light load low mid RPM that I need to figure out. Noticed out of the corner of my eye the map goes slightly lean at the point when it 'bumps' or misfires..

Logged about 30 minutes of driving and came home to check on the log, then run through the viewer on Very hard setting. Max cell change was 7, so I looked it over a bit and decided to save it and we'll try it later. Kinda warm outside and I don't wanna go back out for a while now after spending all evening last night replacing my muffler and this morning tinkering.

Man all along... was such a simple problem! If I'm ever around Virginia will defiantly have to look you guys up and buy a few beers!