AFRs Suddenly Lean on Restart
#1
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I have a flyin miata FMII turbo 96 with a fully rebuilt 99 motor. EMS is MS DIYPNP. AFRs come in via wideband LC-1. Upon restart (once motor is warm), the AFRs are incredibly lean.
I just recalibrated the sensor - today. I've seen this happen before and haven't been able to trace a cause. Once the car's been started for a couple mins and I drive just a bit on it, the AFRs return to normal. The car doesn't buck or make any strange noises when accelerating lightly. I'm not going for a hard run when the AFRs are showing 17-19 to find out how real the reading is. Wondering if anyone else has had this issue. Only thing that makes sense to me is that the LC-1 is giving false readings. Perhaps the restart attempts to warm an already warm sensor, which makes it too hot an gives false readings until it's had a chance to cool?
I don't think it's on the fuel side. I've got a brand new 190LPH Walbro pump and RC550s doing the work. I replaced the fuel filter about 10,000mi ago. All other parts less than 5k mi.
Let me know what your thoughts are. Particularly interested if anyone else has seen this before and was able to tell what it was. I'm leaning towards the LC-1 being the issue.
![vash](https://www.miataturbo.net/images/smilies/vash.gif)
I don't think it's on the fuel side. I've got a brand new 190LPH Walbro pump and RC550s doing the work. I replaced the fuel filter about 10,000mi ago. All other parts less than 5k mi.
Let me know what your thoughts are. Particularly interested if anyone else has seen this before and was able to tell what it was. I'm leaning towards the LC-1 being the issue.
#2
Boost Pope
![](/mt2011/images_td/ranks/supporter1.gif)
![](/mt2011/images_td/ranks/smod.png)
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,339
Total Cats: 6,793
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Since the MS bases fuel computation on manifold pressure, it has to compensate for air density based on temperature. So, as air temperature increases, the MS slightly trims the fuel pulsewidth down. The particular variable used in this computation is called Gair, and you can look at it both on a gauge and in the datalogs.
A value of 100 in Gair means "no correction", while values smaller than 100 represent a reduction in fuel. Normally, we expect this value to be somewhere in the mid to upper 90s while running, as intake air temp is usually warmed up somewhat by under-hood temps.
What I suspect is happening in your case is that when you stop the car while the engine is warm, all that under-hood causes what's called "heat soak" of the temperature sensor. Eg: it raises the temperature of the body of the sensor itself. When you re-start the engine, the air flowing past the sensor starts to cool it back off, but for a little while it reads artificially high.
Check your Gair variable after a warm restart. I'll bet you see it start out at a fairly low number, and then start to work its way back up into the mid 90s. (The value of Gair goes up as the temperature reported by the sensor goes down.)
Where is your IAT sensor physically located?
A value of 100 in Gair means "no correction", while values smaller than 100 represent a reduction in fuel. Normally, we expect this value to be somewhere in the mid to upper 90s while running, as intake air temp is usually warmed up somewhat by under-hood temps.
What I suspect is happening in your case is that when you stop the car while the engine is warm, all that under-hood causes what's called "heat soak" of the temperature sensor. Eg: it raises the temperature of the body of the sensor itself. When you re-start the engine, the air flowing past the sensor starts to cool it back off, but for a little while it reads artificially high.
Check your Gair variable after a warm restart. I'll bet you see it start out at a fairly low number, and then start to work its way back up into the mid 90s. (The value of Gair goes up as the temperature reported by the sensor goes down.)
Where is your IAT sensor physically located?
#3
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
upload datalogs of this issue.
Until you've sure you tuned your car I would disable IAT correction. I have the same issue and found it to be not just caused by their stupid IAT correction, but something else as well, causing me to run lean after a hot restart of sitting around for 5-10 minutes or more.
also upload the tune file
Until you've sure you tuned your car I would disable IAT correction. I have the same issue and found it to be not just caused by their stupid IAT correction, but something else as well, causing me to run lean after a hot restart of sitting around for 5-10 minutes or more.
also upload the tune file
#4
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
yeah, this happens to all of us. I did 50 correction value for awhile, but then it would run lean when it got really cold (not enough correction), so I just left it at 100.....tuned idle a little richer, and deal with the rough idle with hot starts. some people would recommend against this but I've had no issues, I have EGO set to run at idle as well, so 30 seconds after start, EGO kicks in and bring it back down to stoich. Then in normal idle scenarios driving around, EGO kicks in right away and leans it out slightly to stoich.
#5
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Thanks for the responses guys. After reading another forum on an issue with a temp sensor, I was thinking it could be that. I like the idea of EGO correction, but I'll probably limit it to 20 to 80kPa to avoid changes at WOT and high load. Should I keep the change as "SIMPLE" or should I go for PID correction? If PID, what settings are reasonable. I've searched before and not had good luck finding settings. Seems that it would be wideband dependent. I'm running the LC-1 reading in as a lambda sensor with whatever standard time averaging exists on the MS and LC-1 side.
I also made some non-linear MAT corrections inline with what FM uses on their hydra units. I'm using the FMII kit and I think they also use the GM IAT, so I figured their values would be good. I might need to take another look at the differences in fueling eqn with respect to MAT between the Hydra and the MS. Is the recommendation to disable non-linear MAT corrections altogether?
And a really important question: After a really hard pull, will the MAT sensor become heat soaked and make my engine run **** lean and potentially detonate under load, or would the elevated temps read in be more likely real heat soak in the charge pipes, intercooler, turbo, etc? I would think this would be an issue regardless of EMS.
Not on my tuning computer now. I'll try to upload my tune this afternoon. By the way - how do I go about uploading anything to this forum? WTF happened to the good old attach button.
Thanks for not being a dick like Hustler. Hustler gives this forum a really bad rep.
I also made some non-linear MAT corrections inline with what FM uses on their hydra units. I'm using the FMII kit and I think they also use the GM IAT, so I figured their values would be good. I might need to take another look at the differences in fueling eqn with respect to MAT between the Hydra and the MS. Is the recommendation to disable non-linear MAT corrections altogether?
And a really important question: After a really hard pull, will the MAT sensor become heat soaked and make my engine run **** lean and potentially detonate under load, or would the elevated temps read in be more likely real heat soak in the charge pipes, intercooler, turbo, etc? I would think this would be an issue regardless of EMS.
Not on my tuning computer now. I'll try to upload my tune this afternoon. By the way - how do I go about uploading anything to this forum? WTF happened to the good old attach button.
Thanks for not being a dick like Hustler. Hustler gives this forum a really bad rep.
#6
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
Alright, before Hustler jumps on to ream me about attachments, I figured out that its in the advanced screen. MSQ before enabling EGO correction in vacuum is attached.
I had a thought. Could we counteract the effects of heat soak on startup with ASE. I thought it was logical at first for ASE taper and % values to decrease with temp increase. Now, I'm thinking they should start to go back up a bit with coolant temp increase. It took about 2mins for the AFRs to settle in my car after restart at about idle or 1200rpm, making for about 2400revs (cycles?). I'm not sure of the exact coolant temp, but I'm guessing somewhere in the 200F range. I'll need to take a closer look at the MATs on restart next time, but lets say they were about 160F, where true inlet temp is going to be more like 90F on average startup. In metric absolute temp, the difference between the temp reported and actual requires 220% the fuel that our false MATs are giving us. That seems a bit high, but might not be too rich on startup. Also remember to factor in your non-linear MAT correction values at the supposed MAT temp vs. actual MAT temp. For me, that's about another 4%.
Now - the problem with my ASE solution. Not every restart at elevated coolant temp will involve heat soak. I know we're all professional drivers and nobody ever stalls and that everyone's tune is perfected to the point it never stalls at idle, and nobody's car stalls when rewriting to the controller. Any of these situations will cause a nasty rich restart with the fix outlined above. However, perhaps there is a happy medium involving only partial enrichment and then EGO correction after 600 cycles, or approximately 30secs?
By the way - does the 30sec come from the time it takes for EGO to initiate after startup by default, or is this a setting I need to modify? I had issues with what I thought was EGO attempting to kick in before the sensor had warmed up, so I had set the min kPa to 30 to avoid startup, but this also makes it avoid idle. Thoughts?
I had a thought. Could we counteract the effects of heat soak on startup with ASE. I thought it was logical at first for ASE taper and % values to decrease with temp increase. Now, I'm thinking they should start to go back up a bit with coolant temp increase. It took about 2mins for the AFRs to settle in my car after restart at about idle or 1200rpm, making for about 2400revs (cycles?). I'm not sure of the exact coolant temp, but I'm guessing somewhere in the 200F range. I'll need to take a closer look at the MATs on restart next time, but lets say they were about 160F, where true inlet temp is going to be more like 90F on average startup. In metric absolute temp, the difference between the temp reported and actual requires 220% the fuel that our false MATs are giving us. That seems a bit high, but might not be too rich on startup. Also remember to factor in your non-linear MAT correction values at the supposed MAT temp vs. actual MAT temp. For me, that's about another 4%.
Now - the problem with my ASE solution. Not every restart at elevated coolant temp will involve heat soak. I know we're all professional drivers and nobody ever stalls and that everyone's tune is perfected to the point it never stalls at idle, and nobody's car stalls when rewriting to the controller. Any of these situations will cause a nasty rich restart with the fix outlined above. However, perhaps there is a happy medium involving only partial enrichment and then EGO correction after 600 cycles, or approximately 30secs?
By the way - does the 30sec come from the time it takes for EGO to initiate after startup by default, or is this a setting I need to modify? I had issues with what I thought was EGO attempting to kick in before the sensor had warmed up, so I had set the min kPa to 30 to avoid startup, but this also makes it avoid idle. Thoughts?
#12
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
What firmware are you running. I didn't seem to have this issue with my diypnp. I ran 50% MAT correction value, and used the MAT correction table. It worked pretty well.
FYI. I liked the 50% correction because it allowed (~140 degrees) me to only need 6% correction at max, and -2-3% at the low end. 100% correction required me to have a 13% increase at my highest table setting. My temp sensor is just after the IC
FYI. I liked the 50% correction because it allowed (~140 degrees) me to only need 6% correction at max, and -2-3% at the low end. 100% correction required me to have a 13% increase at my highest table setting. My temp sensor is just after the IC
#14
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
OK, my issues don't appear to be MAT - related, though they might still be heat soak. I was able to fix a bit my upping the ASE % and extending the ASE taper cycles out to 2000 for all coolant temps (what FM uses for their units). That brought me back to about 17:1 on hot startup instead of 19:1. I also noticed that WUE appeared to need correction through 180degF and wasn't fully warmed up until closer to 200F, so I'm still correcting 10% at 180F and taper down to 0% at 200F. That seemed to make my EGO not work very hard at all 102-103% max at idle.
Why no EGO at idle Brain? Unstable? It's actually the most stable idle I've ever had at about 950rpm. I was lucky to dip below 1050rpm before.
Why no EGO at idle Brain? Unstable? It's actually the most stable idle I've ever had at about 950rpm. I was lucky to dip below 1050rpm before.
#15
![Default](https://www.miataturbo.net/images/icons/icon1.gif)
I think the idea is the closed loop PID idle control and EGO will fight each other and make it unstable. Meh, I haven't had issues with it, plus it makes hot starts 10x better. I make my EGO step size around .2 and update rate pretty slow, so it's not causing AFR to jump around all the time.
Thread
Thread Starter
Forum
Replies
Last Post
Greasyman
General Miata Chat
2
09-28-2015 10:44 AM
Motorsport-Electronics
ECUs and Tuning
0
09-05-2015 08:02 AM