close/open loop issue

[thecrash91]

so one last question, if you guys all think this is normal, is it ok to run like this? i havent had any issues other than fuel cut. i do have a knock light, although its not been calibrated i havent heard any knocking and the light hasnt gone off except for a couple times when coasting. The car pulls nicely if i punch it at 3k, the afrs stay lean for a second, timing also feels like its being retarded a little extra, but it pulls decently once i spool up a bit. once i get an EMS how will i tune around this? If i tell it that it is less air than it is will it react like the stock turbo and go rich instantly? or do I have to add fuel? if i add fuel will the timing still stay retarded a tad? i am planning on obtaining a split second ftc or a megasquirt (prebuilt) within the next month or so. if i can get these answers i will feel a lot more comfortable accepting the fact that this is just the way the car is, lol.

 

[StealthWyvern]

I didnt really read anything other then the first post but that is normal for the stock tune

Get on it in 1st or 2nd gear and it will go pig rich at any sign of boost for awhile. However if you cruse in the lower rpms and get on it it will run 14-15 afr tell the trigger for open loop at aroudn 4k rpms. This is why you never want to boost in 5th gear!!!!

 

[thecrash91]

I didnt really read anything other then the first post but that is normal for the stock tune

Get on it in 1st or 2nd gear and it will go pig rich at any sign of boost for awhile. However if you cruse in the lower rpms and get on it it will run 14-15 afr tell the trigger for open loop at aroudn 4k rpms. This is why you never want to boost in 5th gear!!!!

i wholeheartedly agree, i never boost in 5th, but i have bigger issues now, those of the transmission variety.. damn things are made of pressed sugar, i swear. its never a broken gear though..

 

[Installshield 2]

just to try to clarify some of this: Everything i've read with your issue so far is deemed 'normal' for a stock MSP ecu...which is FAR from ideal...

There is a lot of misconceptions about these computers...most of which is still theory after 10 years...its very hard to get concrete evidence of exactly how these computers work...but so far, this is what i can confirm...

We have very 'closed loop biased' computers...to the degree that factory service manuals don't even call it closed or open loop, and certain OBD-II readers don't even know whats going on with fuel status...

The MSP computer attempts to run in closed loop as long as possible...this is seemingly independent of load, and only influenced by rpm, as far as mandated mixture is set...but ignition timing will be all over the place in this condition...which is largely why you're saying 'closed loop feels slow'...its true that a lean mixture, for the most part, burns hot and powerful...but only if ignition is set to take advantage of it...Mazda kept the tune safe by pulling ignition back with a near stoich closed loop protocol up to about 4k revs...bad for power, VERY bad for increasing boost or other mods, but good on gas and relatively safe for an unmodded car...closed loop, by definition, is a latent tuning protocol that reads exhaust gas contents and makes adjustments on the fly based on them...

open loop for an MSP is when the computer stops listening to the O2 sensor and runs off of preset maps...Mazda doesn't call it open loop, its just 'enrichment mode'...on an MSP, unlike MP3's and normal proteges, hit open loop closer to 4k...while an Mp3 can dig into open loop closer to 3200 at certain load levels...again, the MSP computer doesn't seem interested in load for whatever reason...

so, you hit about 4k...and everything is fine...under that, especially with mods, you're dangerously lean...sadly, this is completely normal for these computers haha.

also...i haven't been able to fully confirm it, as i'm on an MP3 computer...but there definitely DOES seem to be some sort of limp mode, at least in my case...My have twiggy cams that will often pop a random misfire code during cold starts (about once every 10 cold starts)...its only happening because of the lope these cams create...but, if i don't clear the code with my scanner...i'm definitely logging some pretty retarded ignition timing...i can't say i definitely feel it in engine response...but logs in my android app clearly show it...as soon as i clear the code, ignition timing gets much more aggressive again...through all of this though, mandated afr's and closed/open loop switching remain uneffected...

 

[thecrash91]

Thats good to know, It definitely makes sense. I just have one more question though, the stock turbo never acted like this, is this because of the less air flow? or was it because I had a split second unit forcing it into "enrichment mode"? if i get another EMS will it be able to force the car into enrichment mode earlier and thus eliminate the massive timing retard?

 

[f1yen]

I have made a discovery, and it relates to having the car go into enrichment RIGHT as you pass 0 psi.

Some guidance:

Boost in ANY gear for 12-15 seconds: (be nice to your car for this... boost lightly like 2 or 3 psi)
1)WITHOUT punching the throttle to WOT (WOT for 5 seconds will go into open loop )
2)You must also stay away from lifting the throttle to hit fuel cut on decel
3)You must stay away from the open loop onset by 4K rpm

To achieve this easily, I will normally roll on a side street in 2nd or 3rd, and boost (2-3psi) and drag some brakes to keep me away from accelerating past 4k rpm. You can do this in as low of an rpm as you want (will be easier with less brake drag)

Once you are done with this, test my theory, and go past zero psi and see if you dip down into the enrichment... IN ANY RPM for the rest of the time until you turn off the car.

Now doing this is a pain, and I am trying to make a PLC mimic some signals to hopefully accomplish this without dragging brakes etc.
Please report back if this works... has worked on my car for over 2 weeks.... I can post a video if needed.

 

[thecrash91]

i feel like this is more of an issue with the gt28rs than with the stock turbo. lol. it makes total sense because me and f1yen both have this exact issue and both lack an EMS, lol. i cant exactly try this out right now though cuz my tranny is shot, and the one im putting in for the time being is probably none too reliable either.

 

[dskelly37]

Just for reference, every car goes into "fuel enrichement mode" when you go WOT, it's not just a Mazda thing. Most cars have a % of TPS vs. RPM table they reference to determine when the vehicle is WOT (100% TPS doesn't always mean WOT to the PCM). So yeah, I'd say you just need to tune the damn thing before anything else.

Also post your LTFT and STFT for the lulz.

 

[f1yen]

I am aware of the slight "tip-in" fuel enrichment when the throttle is mashed.
I guess I'm just confused as where 100% load (ref. from torque app) will only sometimes trigger the WOT mega fuel dump in lower rpms. (same fuel dump that happens at the normal 4k rpms)

for lulz
I adjusted with the SSFT, for trims in vacuum. so I am +/-3% ST and LTFT .
The problem that I see, is that IF the car stays in closed loop when in boost (obviously below 4k rpms) my trims un-altered would sweep way negative -18 or something. I do not know if this has to do with relocating the MAF.

I ruled out MAP for forcing open loop, the MAP (or EGR boos sensor) only ratchets open the solenoid (to see manifold press/vac) in the beginning of driving, right as the car gets to operating temp. any boost input at these times delays the EGR calculation and it will keep attempting to see vacuum. (on a side note, I may be able to make the micro-controller mimick this for EGR deletes )

I will keep working on getting the car to consistently transition to open loop upon boost,(and will probably be the death of me) as I think this is necessary to tune properly.

Would it help if the IAT were near the MAF??
I was under assumption that the iat only controls some of the base timing maps.

 

[dskelly37]

I suppose IAT and MAF should be close to each other since both values are used together to determine air density. If the MAF is way up front registering air flow but the IAT is way down the intake stream and measuring hot ass air, I could see the PCM possibly calculating leaner fuel trims thinking the air is less dense than it really is (this is pure theory craft of course!)

What about your igntion timing, is it advancing like it should or is the PCM pulling it back before 4k? When my Subaru had a melted knock sensor, it would pull timing back intermittently (sometimes even negative lol) and it would just fall flat on it's face. You'd sit there with the pedal mashed and slowly creep up to 3.5k RPM, then suddenly it'd take off like a bat out of hell.

 

[f1yen]

I have seen some clever ways to get the IAT on the hot pipe, I will have to invest in that.

Yes, I have some horrible timing being pulled... sometimes into the negatives... I thought that the sensor was the problem a while ago, (looked like I had the death goo coming out of it) so I replaced it with a spare I had, not much change, if at all.
I will take some logs and post them.

Maybe I should think about buying a new one knock sensor.... ?
I hate throwing parts at this car, just as good as throwing crap in the air, useless and you just feel like you get crapped on anyways when it falls on its face.

 

[dskelly37]

Not sure with Mazda, but some brands are extremely particular about the angle to centerline and torque spec of the knock sensor.

I'd be focusing your efforts on what's keeping the ignition timing back since that's what is holding up the revs. The only possible way I could see fuel trims causing timing to pull back is if it was so lean it really was knocking, but I kinda doubt it is.

How's engine vacuum at idle? I had a MR2 once that was timed a tooth off on the intake camshaft. Ran and idled pretty good actually, but it felt like it had turbo lag even though it was NA haha. Didn't think anything of it though till I popped a vacuum line off for s**** and giggles and it wasn't sucking at all.

 

[f1yen]

Here is some data from timing.
Warm Idle

2nd Gear WOT (full throttle at least!)
<a href="http://s1337.photobucket.com/user/f1yen/media/Screenshot_2013-07-08-17-29-30_zps089f6a54.png.html" target="_blank"><img src="http://i1337.photobucket.com/albums/o662/f1yen/Screenshot_2013-07-08-17-29-30_zps089f6a54.png" border="0" alt="2nd gear WOT photo Screenshot_2013-07-08-17-29-30_zps089f6a54.png"/></a>

2nd Gear again WOT
<a href="http://s1337.photobucket.com/user/f1yen/media/Screenshot_2013-07-08-17-24-54_zps7856bf1c.png.html" target="_blank"><img src="http://i1337.photobucket.com/albums/o662/f1yen/Screenshot_2013-07-08-17-24-54_zps7856bf1c.png" border="0" alt="2nd gear WOT2 photo Screenshot_2013-07-08-17-24-54_zps7856bf1c.png"/></a>

As seen from the 2 WOT runs, the jump in timing (different rpms in both runs) in lower rpms may be relating to timing being pulled.
The timing seems to progress fine once 4k is hit.

Vacuum is great, -22 or so made a hot pipe so wasegate is on there now. boost is 6-7psi.

 

[dskelly37]

Can you graph out short term fuel trim and O2 readings as well?

I actually went back and re-read through this thread, probably should have done that from the start lol... What I don't understand is why the PCM under closed loop can't compensate for the added boost by dumping in more fuel. It seems that it rolls into boost and then leans out hard because it wont enter fuel enrichment mode, so it begins to knock and pulls ignition timing back instead. Pretty derp of Mazda if you ask me.

 

[StealthWyvern]

How did u get that graph in torque!?!

 

[thecrash91]

really glad this thread picked up a little, cuz im still lost, but also without an engine management, lulz

 

[f1yen]

If you purchase the app ($5 or something) it has more functions and on of them is graphing. I would recommend.

I have SSFTC, so yes management. does anyone else have input on my timing? can they log? I want to know if i should be buying a new knock sensor.

In a real hot day yesterday with the AC on, I am having the hardest time driving the car, I know AC pulls alot from a 4 cyl.... but this is ridiculous. I almost instantly have to push the car into boost, (sees 100% load) and then the PCM will pull abuncha timing. (seen 100% load timing pull here http://www.mazdas247.com/forum/entry.php?799-Fixing-the-Mazdaspeed-Protege-01 )

Doing so I went into the "early open loop" real soon after driving (seeing open loop any time the car sees 100% load aka just past zero psi)

 

[thecrash91]

If you purchase the app ($5 or something) it has more functions and on of them is graphing. I would recommend.

I have SSFTC, so yes management. does anyone else have input on my timing? can they log? I want to know if i should be buying a new knock sensor.

In a real hot day yesterday with the AC on, I am having the hardest time driving the car, I know AC pulls alot from a 4 cyl.... but this is ridiculous. I almost instantly have to push the car into boost, (sees 100% load) and then the PCM will pull abuncha timing. (seen 100% load timing pull here http://www.mazdas247.com/forum/entry.php?799-Fixing-the-Mazdaspeed-Protege-01 )

Doing so I went into the "early open loop" real soon after driving (seeing open loop any time the car sees 100% load aka just past zero psi)

so you do have an engine management? that means buying an engine management wont solve this issue than.. damnit since so many people said that would work i thought that was it.

 

[f1yen]

My tuning is very mild, (well mostly because the only reputable area to play with the fuel is after 4k) and I refuse to use the "overboost" setting because it just outputs a garbage O2 swinging signal from .8 to .2 volts. The ecu reacts to it more on the lean side and starts increasing the trims....

In my case the car acts "fine" for the first few boost events (pulls timing at 100% load, stays around 14.7 untill 4k rpm, etc)... but something triggers the 10afr (open loop) in lower rpms (say 2000) at 100% load.(and it will do this the rest of the time the car stays on: any time the car goes into 100% load, the car goes into open loop and dumps the fuel in.
This is extremely bothersome when autoXing, starting off the car drives 1 way, and then 4 corners in, the car is acting different (especially if you are in the low revs in 2nd gear on a corner).

My goal is to find out what is forcing me into that "early open loop" , force it, and then tune with the SSFTC the whole boost range (because it will be in open loop).

So far, since it was easier for me to do this when it was hot out, I will be looking for several things... IAT, knock possibly, and time at 100% load.
Anyone want to donate a NA protege ecu to me for "science"?

 

[dskelly37]

What's funny is my current DD Subaru started acting up today, intermittent loss of power. Turned out to be a bad upstream O2 sensor. Try going for a drive and graph every PID you can. We have a few solid ideas now but need more data on your end lol.