Stuttering/Hesitation after many, many changes.

Adrian Hayden

Adrian Hayden

Professional Overthinker
Greenie N00B Member
Greenie Member
Hey everyone. Recently, I went through my entire Mazdaspeed6 and did a ton of maintenance (walnut blasting, injectors/seals, new gapped plugs, etc.), as well as some power upgrades. While the car had a TurboSmart BPV, CS TMIC, test pipe and catback before, I went ahead and installed HPFP internals, a 3-port EBCS in interrupt setup, and the Tru-3.5 aluminum intake from JBR. To begin tuning, I loaded the Base 93oct High Load map with Interrupt plumbing from the guys at VersaTune. I also borrowed the MAFCal from here. I was able to drive almost 75 miles without issues, and then I decided to dial in my MAFCal.

After making a few revisions, I was a little miffed: parts of the table from the forums were originally off by over 20% at idle, which seemed like far more than the variance you'd expect between different MAFs. Last night, the car developed a nasty stumble and stutters during moderate acceleration (I can consistently produce it with anything over 33% throttle in 3rd at 3000rpm, and it comes and goes throughout the midrange in 1st-3rd. Luckily, there's no knock). I suspected that the sensor was bad, and I changed it out for another MAF that I borrowed from a friend. I also tried loading the MAFCal that "worked" for almost 100 miles. While LTFTs were much better, the stumble persisted. The fact that the hesitation is a new phenomenon leads me to believe a sensor is failing, or there is a new mechanical issue.

I've attached a log with as many relevant parameters as I can think of. While I know that the poll rate isn't the highest on this one, I have many other logs with fewer parameters. If anyone can point me in the right direction and/or has suggestions as to what I should log, I would be very grateful.

Edit: Failing throttle body. Thanks to everyone for the help, and hopefully this post can be useful to others in the future. :)
 

Attachments

Last edited:
Fuel Pressure appears fine, everything looks good but if you pay attention to the load and the Mass Airflow you'll see it drops off a bit right around where you say it has issues. This could be a tune issue (overload correction; boost cut); this would require a retune if that's the case.
 
Enki said:
Fuel Pressure appears fine, everything looks good...
Click to expand...
I logged fuel pressure there to check exactly that :)
Enki said:
This could be a tune issue (overload correction; boost cut); this would require a retune if that's the case.
Click to expand...
Because I drove for almost 100 miles without so much as a hiccup, I am doubtful that this is a tune issue on its face. The weather has been the same, and this gas is from the same fill-up. I don't want to accidentally end up tuning for a mechanical/sensor issue.

I verified my EBCS is clicking with throttle and the key in the on position. After that, I zeroed out my WGDC Base table and logged boost related params (see attached). I noticed that the car still sees 20% DC and builds a couple psi past spring pressure. This doesn't exactly make sense to me, as all of the tables which modify appear to be multipliers, not additive. Do you think that all is well with the solenoid, or should I investigate that avenue further? I may swap a known working MAP as well, if only to rule something else out.
Enki said:
if you pay attention to the load and the Mass Airflow you'll see it drops off a bit right around where you say it has issues.
Click to expand...
I'm also potentially concerned about the VTCS sticking. Not sure if it could be catching and suddenly releasing, causing the "instantaneous" absolute load changes.

Thank you, again. Goes without saying how much I appreciate your help.
 

Attachments

Adrian Hayden said:
Because I drove for almost 100 miles without so much as a hiccup, I am doubtful that this is a tune issue on its face. The weather has been the same, and this gas is from the same fill-up. I don't want to accidentally end up tuning for a mechanical/sensor issue
Click to expand...

What you DON'T see is how adjustments are done to WGDC with load correction as a factor. Generally speaking, WGDC has several semi-persistent tables that operate exactly like fuel trims, and these work on a per gear basis, but only matter when load tuning (like all VT tunes are).

Take a look at the log you've posted, right around where you think the issue kicks in and look at the load, mass airflow, and WGDC row by row in that area; you should clearly see what I saw.


Adrian Hayden said:
I noticed that the car still sees 20% DC and builds a couple psi past spring pressure. This doesn't exactly make sense to me, as all of the tables which modify appear to be multipliers, not additive.
Click to expand...
First off, don't fall into the "spring pressure" trap. The wastegate spring doesn't really control the pressure you see at the boost sensor, more the amount of energy passing through to the compressor wheel, which while *should* be close to what you're expecting, it doesn't really work that way. Depending on the setup, you may need in excess of 40% WGDC to even see a change in boost pressure beyond margin of error.

If you're self tuning, then I can tell you for certain that if you're seeing a WGDC% output other than 0 on a zeroed out table that you've missed something relevant to actually getting "spring pressure."

That said, because boost pressure is tied directly to the ratio of exhaust pressure to intake pressure, you will literally always see a boost curve that loosely follows your torque curve, as engine efficiency changes across RPM (volumetric efficiency). As the engine flows more freely for a given RPM, that feeds back into the turbo that makes the black magic happen. The only way around this is to feed the wastegate with static pressure air (like they do in drag racing where the feedback loops result in instant compressor/exhaust wheel destruction on spool) or by somehow controlling the amount of air going into the engine to the point where airflow per cylinder is identical across all RPMs (so basically, not a fucking chance).

Adrian Hayden said:
Do you think that all is well with the solenoid, or should I investigate that avenue further? I may swap a known working MAP as well, if only to rule something else out.
Click to expand...
Leave them alone, they are fine. The solenoid only takes instructions from the ECU, and cannot report back a %, so any % WGDC you see is because the tune is commanding it.


Adrian Hayden said:
I'm also potentially concerned about the VTCS sticking. Not sure if it could be catching and suddenly releasing, causing the "instantaneous" absolute load changes.
Click to expand...
Not happening. You're assuming that the VCTS will be preventing air from entering the engine when at boost, and that's simply not going to happen. If it did, it's something you'd feel immediately, and the boost cut bucking (what I assume you feel) is something that triggers only when you're over load target for a period of time.
 
Enki said:
WGDC has several semi-persistent tables.
Click to expand...
Is there a reference to any literature for these tables? I know I'm missing logic here, but I really want to know *how* they interact with the WGDC. I come from HPTuners, and it's pretty trivial to get information about the ECU logic. If there's any write-ups on it, I would love to give them a read.
Enki said:
Take a look at the log you've posted, right around where you think the issue kicks in and look at the load, mass airflow, and WGDC row by row in that area; you should clearly see what I saw.
Click to expand...
Definitely. I only know to identify that as the time of the event because of the dip in load and mass airflow.
Enki said:
If you're self tuning, then I can tell you for certain that if you're seeing a WGDC% output other than 0 on a zeroed out table that you've missed something relevant to actually getting "spring pressure."
Click to expand...
I agree 100%, I'm curious as to what I'm missing in the first place. I'm sure it's simply within tables in the scope of the load targeting, since I didn't make changes there.
Enki said:
The solenoid only takes instructions from the ECU.
Click to expand...
This is extremely helpful to know.
Enki said:
If it did, it's something you'd feel immediately, and the boost cut bucking (what I assume you feel) is something that triggers only when you're over load target for a period of time.
Click to expand...
Looking at the second log I sent you, boost builds to 12.5psi at ~3350rpm, and it holds within 1psi until I stop accelerating. As soon as absolute load eclipses 1.3, stuttering begins. Isn't that basically immediate? Looking at my Load Targeting tables, as well as limits for Overboosting and absolute load/boost, I don't see a way I could overboost/overload at 1.3 load, <13psi, 3350rpm being at sea level. If I'm wrong, don't hesitate to correct me.

I cannot get my head around why my current load-based tune would work for almost 100 miles and then suddenly shit the bed. Without weather/gas changes that cause me to access different parts of the table, wouldn't I have to have a mechanical issue? I honestly don't believe it's the VTCS, but I'm wondering if something somewhere has/is letting go (VTCS, injector seals, and really anything is the air, fuel, or spark department is suspect to me).
 
Adrian Hayden said:
Is there a reference to any literature for these tables? I know I'm missing logic here, but I really want to know *how* they interact with the WGDC. I come from HPTuners, and it's pretty trivial to get information about the ECU logic. If there's any write-ups on it, I would love to give them a read.
Click to expand...
Nope. Pretty much all tribal knowledge. They clear when you flash a tune or pull the power, and cannot be directly edited. That's about all you really need to know about them.

Adrian Hayden said:
I agree 100%, I'm curious as to what I'm missing in the first place. I'm sure it's simply within tables in the scope of the load targeting, since I didn't make changes there.
Click to expand...
There's the WGDC table which is used as the base, obviously. Then there are tables for coarse and fine corrections, tables for pedal and throttle positions IIRC, maybe one or two for barometric correction, and transitory tables between open and closed loop, which have multiple triggers for transition. Use the table descriptions, as they are generally pretty good, but may not be 100%.

Adrian Hayden said:
Looking at the second log I sent you, boost builds to 12.5psi at ~3350rpm, and it holds within 1psi until I stop accelerating. As soon as absolute load eclipses 1.3, stuttering begins. Isn't that basically immediate? Looking at my Load Targeting tables, as well as limits for Overboosting and absolute load/boost, I don't see a way I could overboost/overload at 1.3 load, <13psi, 3350rpm being at sea level. If I'm wrong, don't hesitate to correct me.
Click to expand...
No, because the overload condition may be just south of 1.3 load, depending on what table is being used for limiting (hint: it's pretty much always the lowest value for the condition).

Adrian Hayden said:
I cannot get my head around why my current load-based tune would work for almost 100 miles and then suddenly shit the bed. Without weather/gas changes that cause me to access different parts of the table, wouldn't I have to have a mechanical issue? I honestly don't believe it's the VTCS, but I'm wondering if something somewhere has/is letting go (VTCS, injector seals, and really anything is the air, fuel, or spark department is suspect to me).
Click to expand...
There's a table that sets the limit to corrections on the top and bottom end, and by default it's +20% -100% but that can be changed to a point where you can overload and the ECU *won't* pull more DC%. I've been there; it's gnarly.

If you ever want to make the car run but undrivable, that's a good way to do it, lol
 
Enki said:
There's a table that sets the limit to corrections on the top and bottom end, and by default it's +20% -100% but that can be changed to a point where you can overload and the ECU *won't* pull more DC%. I've been there; it's gnarly.

If you ever want to make the car run but undrivable, that's a good way to do it, lol
Click to expand...
That just sounds like an over-all bad time haha.

Since I recently reflashed the same tune that I drove all those miles without so much as a misstep, doesn't it concern you that the tune is suddenly no longer driveable? If I had waited weeks, or made physical changes in between, it would irk me a lot less. I'm fairly worried that I have a fundamental issue with one of the base three (air, fuel, spark) that has suddenly reared its head. I'm going to do a compression test to sort of indirectly peek at the seals, but beyond that plug gap is the only other thing that jumps out to me to investigate. Since the stumble has spread to 2k rpm at somewhat lower loads, I'm worried it's more of a gradual failure like leaking injector seals.

If it ran like shit within the first few miles, I'd have no apprehension changing load tables. But it made it through two full days, so I just don't see how it couldn't be some sort of mechanical/electrical change. After I'm convinced everything is in proper working order, I'll make those adjustments.
 
If you're going to recheck anything, make it the plugs.

As for being worried: Nope. Not even a little. Next time you cold start the car, when the idle comes back down, check your STFT and LTFT. When you get back from whatever trip you went on in it (hopefully longer than 10 miles or 30 minutes), check them again and compare the numbers. LTFT might not have changed, but STFT should be pretty different.

Any small environmental change can have a big impact on tune adjustments, even stuff you normally wouldn't figure like barometric pressure and humidity.
 
Enki said:
If you're going to recheck anything, make it the plugs.
Click to expand...
Plugs are one step colder, gapped to 0.028, for the record.

Enki said:
Any small environmental change can have a big impact on tune adjustments, even stuff you normally wouldn't figure like barometric pressure and humidity.
Click to expand...

If I accept the premise that what is basically Versa's high load tune doesn't work at sea level in the low to mid 80s with the added airflow of the TIP and test pipe, that doesn't exactly explain why it ran without a single stumble for so long. I can appreciate the small variances making big changes, but I just wouldn't expect it on a tune that's so conservative. You definitely put me at ease a bit though with your certainty.

Do you have suggestions as to what I should be looking to adjust first? This all came to a head because I cannot even begin to dial in the MAFCal without encountering pretty bad stumble (and it still bugs me that it's off by over 20% from what exists on the forums). Since airflow calculations affect basically everything, I don't want to put it off.
 
I hope you're running straight pump gas and have a big turbo for those plugs...Did you just change to them?
Also does your intake have an airflow straightener? If the filter is old or super dirty on one side it can throw things off too.
 
Enki said:
I hope you're running straight pump gas and have a big turbo for those plugs...Did you just change to them?
Also does your intake have an airflow straightener? If the filter is old or super dirty on one side it can throw things off too.
Click to expand...

I'm just running 93 from the pump. Haven't changed out the k04 yet, thought it would be best to learn a new tuning application with a relatively safe turbo. Filter is brand new, it is quite literally the 3.5 Aluminum intake that JBR sells on their website. As far as the plugs, they are what I saw recommended in MSF, here, and by a bunch of the tuners. 0.028 was on the high end of the gap spectrum compared to most, and I never saw anyone experience issues with that heat range. They are NGK 6510s. Changed to them during all of the other maintenance, plugs that came out were downright awful. Let me know if I should be looking into something else
 
Enki said:
That's what I'm thinking too, especially if step colder came out of it and looked nasty too.
Click to expand...
You misunderstand; I put step colder plugs in the car. What came out was probably the plugs from when the car was assembled. The new plugs have 100 miles on them, and looked fine when I pulled them yesterday
 
You must log in or register to reply here.
Back
Top