Is AFR Oscillation normal?
#1
Is AFR Oscillation normal?
Don't really have much to base it on except my car and more of a general question. At cruise with feedback turned off, my afrs can oscillate nearly 2 whole points from rich to lean and bounce around. Im pretty sure something is off. What i'd like to know if its supposed to be smooth and follow my target than jump all around it and average the target?
The following users liked this post:
Vicoor (06-02-20)
#3
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
could be misfire, the engine can do it a little and you won't notice. why is the target AFR so far off from the actual AFR? (misfires would read lean)
more tuning required i think
more tuning required i think
#4
On the noise. What would you recommend? A shield wire? Im running coil on plugs now and they are on the other side of the engine bay, so I feel like it's not the issue. The o2 sensor is pretty close to the turbine and I think that has some contributing factors but not all. I have new bungs on the way and will be relocating the signal further downstream.
I was thinking maybe a possible vac leak, bad fpr, shot wideband sensor or poor map sensor placement at the moment.
Car was professionally tuned prior to the rebuild were it suffered from the jumpy afr on a road trip from GA to CO. Tuner never mentioned a problem to me but he might of used his setup in the tailpipe most likely.
If you all want a log file to view let me know. Its running on Aem's latest infinity tuner.
#5
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
it actually could be the port, when it braps, its a misfire, and it will read lean, since it does fire sometimes its going to be all over the place.
#6
Rocket Appliances
iTrader: (11)
I agree on tuning. I did enlarge my ports a bit more this rebuild. (I'm half bridged and always was, just opened up the regular ports a bit more as well as clean them.) This is the old tune putting down the road to see what's going on. Usually the feedback is turned on and it will pull it more into target but it oscillates just as much.
On the noise. What would you recommend? A shield wire? Im running coil on plugs now and they are on the other side of the engine bay, so I feel like it's not the issue. The o2 sensor is pretty close to the turbine and I think that has some contributing factors but not all. I have new bungs on the way and will be relocating the signal further downstream.
I was thinking maybe a possible vac leak, bad fpr, shot wideband sensor or poor map sensor placement at the moment.
Car was professionally tuned prior to the rebuild were it suffered from the jumpy afr on a road trip from GA to CO. Tuner never mentioned a problem to me but he might of used his setup in the tailpipe most likely.
If you all want a log file to view let me know. Its running on Aem's latest infinity tuner.
On the noise. What would you recommend? A shield wire? Im running coil on plugs now and they are on the other side of the engine bay, so I feel like it's not the issue. The o2 sensor is pretty close to the turbine and I think that has some contributing factors but not all. I have new bungs on the way and will be relocating the signal further downstream.
I was thinking maybe a possible vac leak, bad fpr, shot wideband sensor or poor map sensor placement at the moment.
Car was professionally tuned prior to the rebuild were it suffered from the jumpy afr on a road trip from GA to CO. Tuner never mentioned a problem to me but he might of used his setup in the tailpipe most likely.
If you all want a log file to view let me know. Its running on Aem's latest infinity tuner.
If you can post the map and some logs where this was happening, it should be pretty easy to track down the source of the issue if it is related to the tune's impact on the new motor.
Skeese
#7
Changed the map sensor smoothing from 30 to 80 percent. This helped minimize most of the issue. Found some other mechanical things too like my manifold bolts started to come loose from a powder coat job. The finish pancaked leaving the bolts almost a half turn loose. Still have some oscillating issues though. Replaced the sensor and relocated bung further downstream as it was a wee bit close to the turbine. The one log i have is my drive home from work ~ almost 20 min worth. Ill take another log as soon as i finish up some other odds and ends.
On another note for lambda sensors, I have come to realize that Bosch created several lsu 4.2 sensors. My MTX-L runs the 17014 sensor and plug. (rounder one) and Aem uses the 17018/17205 ones with the square plug. Based on pics the tips are different. I couldn't find squat researching the key differences besides the plug and tip.
I started to edit the ve table to enrich some areas and change my target afr's
As you can see very jittery still. This was in the beginning of my drive home. The area im really concerned with is the highlighted area in time (purple vertical line) The map oscillates with the o2 still. Even after 80% smoothed. This was the worse area in the entire log but everything else looks the rest for the most part.
On another note for lambda sensors, I have come to realize that Bosch created several lsu 4.2 sensors. My MTX-L runs the 17014 sensor and plug. (rounder one) and Aem uses the 17018/17205 ones with the square plug. Based on pics the tips are different. I couldn't find squat researching the key differences besides the plug and tip.
I started to edit the ve table to enrich some areas and change my target afr's
As you can see very jittery still. This was in the beginning of my drive home. The area im really concerned with is the highlighted area in time (purple vertical line) The map oscillates with the o2 still. Even after 80% smoothed. This was the worse area in the entire log but everything else looks the rest for the most part.
Last edited by wankel88; 02-13-20 at 03:20 AM.
Trending Topics
#8
Rocket Appliances
iTrader: (11)
I'm downloading the infinity software now, I've never used it before and want to see how the O2 feedback gain settings are handled. Had something similar happen with a car I was tuning on a haltech recently and it ended up being a combination there being too large of an offset between the AFR and the target AFR there for the PID gain that was being used which was causing a massive overshoot there and then ongoing oscillation.
#9
Senior Member
Luckily the Infinity software is easy to play with once you organize everything and have your setting set up right. I know on my Integra it did some weird things, but I would almost bet that if close loop is active, it's probably being over worked like Skeese is saying. PIDs have to be set up just right and the Infinity is very very touchy with it's VE map set up in the way that if it isn't expectionally smoothed out, it will do weird stuff like this. What does your fuel pressure look like and do you have a fuel pulse damper?
Double check your wall wetting table and set up while you're looking through as if you don't have any spare columns next to the zero throttle per second column, it would trigger transient events an unstable AFRs like you're seeing and you'll be chasing your tail for quite a while. My Integra had random variations from zero to ten throttle per second changes because feet aren't precise in holding throttles in one place and caused a similar event.
Double check your wall wetting table and set up while you're looking through as if you don't have any spare columns next to the zero throttle per second column, it would trigger transient events an unstable AFRs like you're seeing and you'll be chasing your tail for quite a while. My Integra had random variations from zero to ten throttle per second changes because feet aren't precise in holding throttles in one place and caused a similar event.
Last edited by newtgomez; 02-13-20 at 09:56 AM.
#10
No fuel pulse damper. Just replaced my old fpr with a newer compact one. Its static pressure is set to 43. Lambda feedback was disabled for these logs to show the raw values. The tables for the lambda control are in the wizard under advanced setting tab and I have not messed w them. Those tables I didnt want to play with. They have a "p" and "i" table based on rpm if my memory serves me right. Never messed with a pid controller except my greddy boost controller for the mac valve. I can email the log and tune together to whoever's intrested in it. The compressed file size exceeds posting it. PM me your email and I'll send it over. Earlier I mentioned the different lsu 4.2 sensors. When I built my harness I used the round style body as I had a spare connector. It's the same one innovate uses. Aem however uses the square type one. So I've been using the Bosch 17014 compared to the 17018/17205. Think this is the issue? I ordered a new connector to pin into my harness. Really isnt much info comparing the 2 sensors.
Edit. To add more I do log fuel pressure as a data plot. It's pretty smooth from the logs based on manifold pressure so not sure if a dampner is neccessary.
Edit. To add more I do log fuel pressure as a data plot. It's pretty smooth from the logs based on manifold pressure so not sure if a dampner is neccessary.
Last edited by wankel88; 02-13-20 at 03:28 PM.
#11
I'm downloading the infinity software now, I've never used it before and want to see how the O2 feedback gain settings are handled. Had something similar happen with a car I was tuning on a haltech recently and it ended up being a combination there being too large of an offset between the AFR and the target AFR there for the PID gain that was being used which was causing a massive overshoot there and then ongoing oscillation.
#12
Rocket Appliances
iTrader: (11)
But if everything is in open loop, none of that should have any effect.
Skeese
#13
The closed loop "feedback" is handled by a PID controller, which is algorithm that is used to determine the output required to correct an offset from a target value. P, I and D are the coefficients that affect how large the correction can be and how fast it will try and correct. If the values used for these are to aggressive for your system and the size of the error you are trying to correct for, the most common issue is oscillation just like you are seeing there.
But if everything is in open loop, none of that should have any effect.
Skeese
But if everything is in open loop, none of that should have any effect.
Skeese
#14
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
That's what I'm saying. This log was in open loop. Does the afr bounce seem right or should it be smoother? I'm wondering if its cause of the different lsu 4.2 sensor im using. I posted a new thread about what's the differences and I haven't had any answers and a internet search left me blank.
in summary, at part throttle the o2 sensor is being fooled, it is somewhat consistent, but you really need to tune it so it runs the best, and set the target AFR accordingly.
Example: my o2 sensor rig says the P port idles at a 14.4 AFR, and if that were true, we would not bother with side ports...
#15
Rocket Appliances
iTrader: (11)
with a big ported engine, the AFR is going to bounce, the sensor is picking up the misfires. it also will read leaner at part throttle because of this. i also feel that the actual AFR and target are too far apart, when you turn closed loop on, its going to have to have a lot of control to bring the AFR inline, and while this is possible, it is better to have the basic map closer to the target. the more authority the closed loop has, the higher the chances are that it will over-correct or get into oscillations (especially with a big port)
in summary, at part throttle the o2 sensor is being fooled, it is somewhat consistent, but you really need to tune it so it runs the best, and set the target AFR accordingly.
Example: my o2 sensor rig says the P port idles at a 14.4 AFR, and if that were true, we would not bother with side ports...
in summary, at part throttle the o2 sensor is being fooled, it is somewhat consistent, but you really need to tune it so it runs the best, and set the target AFR accordingly.
Example: my o2 sensor rig says the P port idles at a 14.4 AFR, and if that were true, we would not bother with side ports...
#16
True, some bounce is to be expected with a large port engine due to the overlap and reingestion, however it shouldnt be 2 AFR points. My semi peripheral will idle at 1800 and steady chop withing +/-0.2 AFR on the AFR500V2 with the NTK sensor. Even if its not reading the true AFR, it shouldnt be oscillating that much.
#17
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
True, some bounce is to be expected with a large port engine due to the overlap and reingestion, however it shouldnt be 2 AFR points. My semi peripheral will idle at 1800 and steady chop withing +/-0.2 AFR on the AFR500V2 with the NTK sensor. Even if its not reading the true AFR, it shouldnt be oscillating that much.
#18
Rocket Appliances
iTrader: (11)
Skeese
#20
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
What size are your peripherals? Mine is a semi peripheral with huge street ports on the irons and the LIM has both the primary and secondary runners bridged together to the peripherals. I can get it to idle at about 1400 rpms but its beyond picky that low due to the extreme low vacuum and it doesn't brap so much as just sounds like struggling ***. You'd be the only person I've ever met in the rotary world that has suggested a peripheral of any sort can/should idle that low.
Skeese
Skeese
my engine has HUGE ports, its probably so big its down on power.
#21
Rocket Appliances
iTrader: (11)
Skeese
#22
Moderator
iTrader: (3)
Join Date: Mar 2001
Location: https://www2.mazda.com/en/100th/
Posts: 30,841
Received 2,605 Likes
on
1,848 Posts
back in the old days, Mazda had a book that told you how to make your car into a race car. there might have been one for the FD, but i've never seen it in English, just the parts catalog
Rx7 http://foxed.ca/rx7manual/manuals/comp.manual.pdf
Rx3 there is an early, with a full bridge, and a late with a PP Foxed.ca - Mazda RX-7 Manuals
Rx7 http://foxed.ca/rx7manual/manuals/comp.manual.pdf
Rx3 there is an early, with a full bridge, and a late with a PP Foxed.ca - Mazda RX-7 Manuals
#24
#25
Ok so i made some changes and things are starting to look even better. First off my injector data was indeed inaccurate. I run the fuel injector connection 850/2000cc setup. Pretty much a stock ev14. I was running the ID flow rates which were 100cc's more at the same fuel pressure. The Infinity had preset data for the ones I run and that made a difference in control. I started tuning from scratch again due to this as my ve table was now inaccurate. I wish the guy i paid to tune my car saw this or knew something was off before he delivered it to me. Also switched from speed density to Alpha N for my load axis in my VE table. I came to a conclusion that due to the way the REW throttle body works this was more efficient. The primary plate opens alone the first 10 to 15 percent then the secondary plate opens. I think this is where things were starting to fluctuate as i tend to cruise around in this throttle angle. The REW engine doesn't share a common plenum to the ports, similar to an ITB setup i'd say. If this were so i think the speed density tuning would be preferred. The speed density i know works most of the time for us, but this seems to not favor me. I also read an article that the factory ecu also runs alpha n at lower rates. Must be for a solid reason. Attached is a cruise i just did with the o2 feedback enabled. I'm working the table now so dont knock on my ve table graph lol. You can download the "Infinity Tuner" software to view the log along with all my settings. The AEMData is a separate software designed to playback and and view data only. I use the tuner software as its an all in one. You will need a layout before you can load the tune. There is a "basic Mazda rotary" layout in the 506 ver 96.2 folder.
https://drive.google.com/file/d/1Smb...ew?usp=sharing
https://drive.google.com/file/d/1Smb...ew?usp=sharing