Sale ends todayGet 30% off any course (excluding packages)

Ends in --- --- ---

wrx 'stumbling' issues

General Tuning Discussion

Forum Posts



Tech Articles

Discuss all things tuning in this section. News, products, problems and results. 

= Resolved threads



I have a 2017 WRX

It has an intake, fmic, catted j-pipe, tgv egr deletes, and ethanol sensor

Using Accesstuner, with the stage 2 Cobb ots map, so far the tgv valves are set to be always open and adjusted the ignition comps and avcs comps for the egr, turned off stratified start, and zeroed out the wastegate. It is also running in full time closed loop, but I think that's how it came from Cobb.

Here are some logs of light throttle, about 0-1 psi Log 0

and a WOT run, Log 1

There seems to be irregularities in each section, occasionally, log 0 will show, the afr's are not consistent with the cl target, however other times as you can see in the end the target is seemingly on point.

Log 1 seems decent for a wot, the afrs are targeting just fine, however at about 3600 rpm, the car stutters, you can feel it and see it in the graph.

What could cause this? Is there a table I might have not seen to either zero out or edit?

Any help is appreciated.


Attached Files

I'm sorry I did not take the time to look at the logs but have you tried decreasing your spark plug gap? Does this occur at peak boost? Misfires can show as lean in the pipe. So the misfire can cause the AFR error so to speak.... I don't know if you've ever seen carbon tracking on spark plug porcelain but that is where you have spark jumping down the outside of the plug laying carbon tracks which are conductive on the outside of the spark plug porcelain between the tip and the hex area.... Also make sure that you don't have any exhaust leaks because that can throw lambda readings off....

Thanks for the reply.

There are no indications of feedback knock or fine knock during the pull, so I dont think its misfiring.

I have a tune done by shop on the car and there are no issues with stumbling when that is flashed to the car, so I doubt the spark plug gap is really to blame here.

There is no stumbling when going slow in the car, just weird AFR values when cruising around.

During a WOT pull the AFRs are on point, even while the car jerks/stumbles the afrs are close to the targets.

The stumbling during wot happens slightly before peak boost as seen in the logs, as its on a 14lb wastegate it happened at about 11psi

Anything else I should look at/ log that would cause this? Do you have more experience with accesstuner an the listed mods?


Misfiring or delayed ignition is not the same as knocking. You asked for suggestions, one that is easy to try was presented, and you didn't even try it.

There are only three results from a change -- better, worse or the same. I suggest you confirm that decreasing the plug gap absolutely doesn't make a difference.

Thanks for the reply.

The spark plug suggestion was provided without the user even reviewing the logs, as they mentioned.

This issue does not occur when a different map is flashed to the car. It only happens with the map I have edited so far, which make me believe there is something about the map itself and not a physical issue with the car. Somewhat seems like a step back to make a physical change to the car when the issue is not present simply by changing the map.

I will try what is suggested and change the spark gap even when there have not been issues with the current spark gap and a better map.

Did you get a chance to look at the logs yourself? Would a misfire occur under low load in vacuum and cause AFR to be unstable and AF correction to appear as a square wave?

I can't view the files, but a common problem when approaching boost/torque peaks is an ignition breakdown because the voltage required to jump the gap increases with the cylinder pressure. This is why closing the gap was suggested - an easy way to check one variable to see if it had an affect.

That said, a mis-fire would be expected to show up as a lean mixture - unless the unburned fuel and oxygen is ignited in the exhaust manifold. You are reporting the AFR as being consistent, which would suggest the fuel is being burned and so you may be dismissing the mis-fire possiblity, but as well as the post-combustion mentioned, you may be looking at an ignition timing irregularity. By that I mean the timing may be getting pulled (retarded) because of 'knock' - is this enabled in the ECU and is it being logged?

With the light throttle AFRs being unstable, and assuming the fuel curves are smooth, it's a common result of an ignition misfire - if the charge isn't ignited and the fuel oxidised, the unused oxygen in the exhaust gas is read by the lambda as a lean condition. It's not uncommon for people to add fuel, thinking the engine is running lean when this happens, which can make the problem even worse as a rich mis-fire may start.

My personal suggestion would be to review the ignition side, consider it basic maintenance, with fresh spark plugs gapped at, for example 25 thou', HT leads' resistances checked and replaced if required, coils wiped down to reduce the chance of flashover, etc.

Then have a careful check over the map you've made and the map that doesn't seem to have a problem, for the changes you made and anything that you may have overlooked or accidentally changed. Maybe give the car a try to see what, if any, difference it made first?

Was this run done on a dyno, or on the road? And does the engine have a torque peak around 3600-3900 RPM?

I looked at the WOT file (Log-1-v2.csv)

Look at the ignition timing trace and you'll see that it's starting to advance (from 0 - 4 deg) just as the rough running starts. What happens if you reduce the timing in this area?

Notice that the ignition timing trace is back to 1.5 degs of timing when the RPM trace settles down.

Do you have a log of the "good tune" in a similar situation?

BTW -- your AFR is not "on point" when the issue is occurring, it shows exactly signs of misfiring with lean/rich swings corresponding to the RPM changes.

Here is a picture of what I was looking at:

Attached Files

Hello both thank you for the full bodied replies.

Gord, I do have Knock being recorded as well as cylinder roughness. I figured I would see something in those if/ when the car misfired but guess I was not correct in thinking that.

"With the light throttle AFRs being unstable, and assuming the fuel curves are smooth, it's a common result of an ignition misfire"

I didn't know this would be a thing, figured it wouldn't have been misfiring with such a low load. The AFR target stays consistent just the measured AFR fluctuates.

David, These logs were taken on the road. I'm not exactly sure where peak torque is currently, I think it is closer to 3000-3300 range from the good tune however that tune may have different ignition timings.

I do see the ignition timing changing and AFR wavering during the stuttering events, thank you for taking a look and confirming these are signs of misfiring I wasnt really sure what that meant from the logs.

I do not currently have a log of going WOT from the good tune, I only have cruise logs. Ill generate a WOT log and post back here.

I'm going to go through my spark side of things to make sure those are all Kosher. I'll do some better comparisons with the ignition timing on the current map and then the good one that I have to see if its possibly too advanced or similar.

Do you know how/ if AVCS timings would play a role in this? I have read online that it could also cause misfiring and possibly the WOT and cruise issues I have. I'm not too sure if that could be playing a role here as well since I didnt even realize this was a misfiring issue. ACVS has been changed to match a TGV delete setup I had found online and also matched these to the values from the good tune so I figured those were ok. I can provide the tune I have been working on if it would help.

Here's the logs in a better format as well

Log1 https://datazap.me/u/walljohnson/log-1641262134?log=0&data=16-20

Log0 https://datazap.me/u/walljohnson/log-1641103245?log=0&data=17

Hello again!

I have cleaned the MAF and MAP sensors, did a leak down test on the car(pressurized all the piping), and replaced spark plugs.

Below are logs from the known good tune on the car, I did not do a WOT pull this log, this is cruising and getting into 5-10lbs of boost as well as other light throttle moments.

Had to split them up as they were too large to upload, its the same drive



I am having a hard time telling if the logs are OK or not? My AF learning seems way off but I'm not sure if thats alright or if there is something else I should look into that is wrong?

If you are using Datazap, I can see how you might have a hard time deciphering what is happening. I just wasted 15 minutes and have zero insights. I won't be looking at any more Datazap files, unless there is a webinar showing me how to make it useful. Megalog Viewer HD is much better for looking at data in my opinion.

To debug a problem, you need to present a log where it's happening, and if possible log where it's not happening.

What problems did you notice when driving and capturing those logs?