AFR stays high forever when hot started

[Supernova_6969]

*I wrote all this, then realized it was text heavy. i went back and highlighted (bold) the most relevant bits of text.. for the time or attention challenged...

so I have an issue similar to that of many others: when my car starts cold, it's fine. (the AFRs are super shitty, but constant, and consistent from one time to the other. I'm tuning now, so it's getting better).

My car is a datson 280zx with a turbo 2.8l, MS2 on v3.00. at the time of editing this, i'm running b&g code (2.89 or 2.90) because i havent got the confidence to install msextra from scratch ( i didnt set this car up originally). I am researching how to set up a car fron scratch on MsExtra so i can do it. I have a post on this but i dont feel like it helped me understand any better, or gain. Confidence in the msq i whipped up.

When I turn her off, if I restart her anytime in the next 4-6 hours (meaning if the engine is still hot), the AFR are around 16 (anywhere between 14 and 18, with dips to 9 (that's partly my tuning problem) and as high as 20).

HOWEVER, and this is were it's different from all the other posts i read, my AFRs STAY HIGH as long as I drive the car, and will be high if I stop and re-start, and that's how she is until she's had at least 5-6 hours to cool (on a hot day. couldn't tell in winter, she's a summer car)

My IAT sensor (open style GM sensor) was originally on the aluminum intake manifold and would get crazy heatsoaking even while I was driving (it would be between 98 degrees on boost (within a second of boost coming on) and 200 degrees as soon as I'd let the throttle go (within a second or two), with cruising temperature around 140 degrees, all that at, say, 80 ambient).
I moved the IAT sensor to the piping just before the throttle body last week, and I now have super stable temperature, at pretty much ambient temp except for when I sit in traffic, at which point it'll climb SLOWLY to 100-105 (which really is only 20 degrees above ambient).

I figured that would solve my issue, but no. I still get insanely high AFR when the engine is started hot, and it just won't go away.

I'm including two Msl that show this. (**note that the AFR goes to like, 22 sometimes. that's an O2 sensor grounding issue that i'll fix soon, but the numbers are very close to the numbers i get from the dial in the car. disregard that weirdness and pretend the AFRs are face value)

The first one is a log of the engine starting after about 4 hours of being parked, on a hot day (today). You can see an average AFR of about 13ish... there is a lot of lean spikes, which I usually don't get on a cold engine (I get a few, but not like this. i think either the engine was not completely cold and starting to show the symptoms of my problem, OR that it's the result of very mild tuning I did in the hour before that specific ride). the dips are normal-ish; I'm working on that with the tuning - having the sensor constantly heatsoaked cause the MS2 to pull some gas (MAT table) and now that it gets the real temperature, which is significantly lower, i have to adjust my very-bad-to-begin-with Ve table

The second one shows a hot restart, after about one hour. I get constant very lean AFR (I'd say the average is 16), which causes the engine to stumble and miss, and even though you can't see it, when the afr is high, the boost needle starts to shaken, almost vibrate, probably since the constant misfiring (or detonating??) creates issues with the the consistency of the vacuum; this goes away when the AFR dips below 14-15, whenever that happens.

When you compare that IAT from one to the other, they are very close together, so I don't think that's the issue. I thought maybe some mechanical component was failing, and creating some misfires, but then the 1:1 correlation between the rough running and the high AFR point to something gas related. I changed my fuel pump a month ago for a walbro 255, so that shouldn't be an issue (problem was there before, and is there after...) I tested and cleaned the injectors this past winter (they were fine and I still cleaned them) so that shouldn't be it (problem was there before, and is there after...).

I'm wondering if the MS is getting some reading from some sensor other than the IAT that causes it to pull back gas when it's hot....

could anyone take a look at the MSL and see if they can spot something weird? I did, but couldn't see anything; i'm a novice at best when it comes to MS...

thanks!

Seb

[slow_hemi6]

Older code so you can't play with the MAT air density so probably my best suggestion is to allow ego control to operate at your idle speed. You currently have it off at 1000rpm.

[Supernova_6969]

Hii!
thanks for the answer.
well the MAT corrections do work in the crappy version I'm using (I'd played with them with some sucess before I swiched the sensor)... That said, the IAT are the same in both MSL, so the MAT corrections from ideal gas law shouldn't really change anything, (although I still want to change to MsExtra current version (I posted a thread on that, but I don't feel confident in the new tune I made in MsExtra to switch over just yet).

I did put on the EGO yesterday all the way down to 800 (after I recorded the second MSL) cause I didn't want to blow up my engine, and it helped, but only at at least 15% correction, and then I'd end up with too rich, too lean, too rich, too lean.. it lessened the peaks, but I still l had issues....

Another thing is that I feel the MAT and the EGO corrections is a bit of a patch, and i would have like to have a car that runs fine, and then add a little extra in the for of ego/map to extra fine tune it....

[turbo conversion]

What MS version and firmware version are you using?

David

EDIT: text

[Supernova_6969]

Sorry, should have said.

MS2 with v3.00 board.
I am running by default the b&g code that was in it (ish) when i bought it, version 2.89.
Ill edit the original poat to include this.

I know this site is for msExtra, but i figured the data in the msl would be universal, and if anything was weird or out of place it would just stand out.

Like i said, i do plan on making the jump. And with this happening it'll probably be sooner rather than later....

[turbo conversion]

I am running the same basic set up as you with a few variations I'm sure.

Non cross flow heads as we have give more hot start issues than their counter parts.

I would strongly recommend making the switch to 3.2.4 firmware.

I have struggled with hot start but have a very good tune now that works.

Not sure how it would translate to B&G code, here is a data log of a hot start.

David

[Supernova_6969]

David.. i'll check your log when i get home, thanks!

Did your issue last, or did it go away after a few minutes like most people?.

[turbo conversion]

The lean restart lasts for a minute and a half or so once dialed in.

I have ASE set to time out just after the o2 sensor becomes active.

David

[slow_hemi6]

Hii!
thanks for the answer.
well the MAT corrections do work in the crappy version I'm using (I'd played with them with some sucess before I swiched the sensor)... That said, the IAT are the same in both MSL, so the MAT corrections from ideal gas law shouldn't really change anything, (although I still want to change to MsExtra current version (I posted a thread on that, but I don't feel confident in the new tune I made in MsExtra to switch over just yet).

I did put on the EGO yesterday all the way down to 800 (after I recorded the second MSL) cause I didn't want to blow up my engine, and it helped, but only at at least 15% correction, and then I'd end up with too rich, too lean, too rich, too lean.. it lessened the peaks, but I still l had issues....

Another thing is that I feel the MAT and the EGO corrections is a bit of a patch, and i would have like to have a car that runs fine, and then add a little extra in the for of ego/map to extra fine tune it....

Ok when I looked at your logs for idle areas that had the same rpm and map but different IAT, I could see that the Gair was changing from 98 to 96%. But what was not changing was the injector pulsewidth! That means one thing, it is not doing any real correction at all. 2% correction from 85 to 95 is doing nothing. This is why I said to turn on ego correction. By doing that you have seen that you had to throw 15% at it to make the AFR change significantly. That is probably too much correction due to the over/under shoot nature of the ego correct but it is definitely showing MAT correction as lacking impact as it is now.
I don't know much about B&G so I went looking and did find in "Other Tuning" the "Non Linear IAT fuel correction" table. Seems to me you could make the last 3 temperature points 85, 90, 95 and add the extra fuel it wants to stabilise the AFR there.

[Supernova_6969]

slow hemi.

this is why I wanted to post here for; I wanted people who know what they are talking about to look at the data and find issues. hopefully what you found is a good lead.

ok. so without your comment, I'd be 100% convinced that the IAT correction table had nothing to do with it. let me ask you guys to validate a few bits of information for me, cause i'm not 100% certain i fully understand everything.

1- Gair is, from what I understand, a percentage of fuel that is either added or removed depending on air temp.
2- the Gair is regulated partly by the ideal gas law, which is in the code, and also by the "Non Linear IAT fuel correction" that you mentioned, which i can correct at will.
3- the injector pulsewidth is how long the injector stays open so that it can send the right amount of fuel. the pulsewidth should be longer when the air is colder/denser and shorter when the air is warmer/less dense.
3-the ego correction changes the amount of gas by also changing the pulsewidth based on AFR readings. the percentage of correction needed to get anything from the ego is a good idea of how much the Gair should push, if that's what needs to be fixed...

from what you read in my data, the temperature goes up, but the pulse width stays the same (or almost). the way I understand things, if the pulsewidth is the same when the air is hot as it is when the air is cool and the AFR is right, I should get TOO MUCH gas, not too little.. right? (now if i'm right here, that points out to the MS2 being right and something else being wrong. maybe fuel pressure, or as I read somewhere, maybe he resistors driving my low impedance injectors are heating up, creating issues with the actual real pulsewidth that the injectors get).

Is this making any sense so far? I'm not leading an argument, I'm genuinely asking if my reasoning is right....

One way or the other, I could correct that by changing the "Non Linear IAT fuel correction" so that it makes the pulsewidth longer at higher temps which would compensate for this (I did that last week, and I think it did help. but when I moved the sensor, I figured the issue would get corrected.. it didn't, even though the temps when from 100-200 degrees to a mere max of 100). Doing that could counteract the symptoms (potentially).. but the cause of the problem would still be there, i guess, whatever it is (I'f my previous assumption of the cause is right)

does all of this make any sense or am I going the wrong way?

now I re-looked at the data, and the average AFR does seem to be lower (with lots of spikes though) as tha IAT goes down toards the end of the log... that seems to support your solution, but it does not explain why things appear to work backwards, if i'm right....

[slow_hemi6]

Yes it could possibly be something on the physical side but one thing for sure, the correction table exists because there has been a need for it. I guess a judgement call has to be made on when correction is excessive and maybe covering up a real issue that could be rectified. Pretty much the same debate rages on about barometric correction, why people need to correct, and why the ideal gas law is not behaving as expected.

[Hahns5.2]

This is a very common issue, I doubt the cause of yours is any different than anyone else's. One thing that helps quite a bit is making sure your injector deadtime is correct, have you done that?

[ol boy]

I've been fighting this also. I recently installed a water injection kit to pull down iat during boost. I began to notice that the hot starts were improving the more I used the water injection. I'm thinking that oil deposits and such build up in the intake valve and cause some weirdness on hot starts. The valve will be cooled even if the engine is hot and any deposits would absorb the fuel as it goes past until it comes up to a steady state temp upon hot start. This is all a SWAG. I think the hot start problem most are seeing is a mechanical problem than MS hot calculating the right PW.

Sent from my SM-G920V using Tapatalk

[turbo conversion]

Can you post your current msq?

David

[ol boy]

Anyone figure out how to add a msq using tapatalk? It won't see it in any of my folders.

Sent from my SM-G920V using Tapatalk

[Supernova_6969]

Hi!

a few things....

1) Hans:
yeah, I saw that a lot of people had brief hot start lean issues, but never found someone who had hot start lean issues that lasted more than a few minutes.
I haven't' checked my Inj dead time, just left it as is from when I bought the car. it's at 1.2ms, which, now that I checked, might be a little high (low impedance with resistors injectors). I'll be trying the values that are suggested by the MsExtra (0.8-.0.9ms) manual when I install MsExtra, soon. I'm reviewing the tune I created to put in the car a while ago so it should be very soon.

2) David :
I didn't post my msq because of two reasons: 1) it's B&G, and I didn't want to bother people here with that non MsExtra crap (I'll post today's latest one here though. feel free not to look at it, for the reasons i mentonned), and 2) it changes Daily. which brings me to point 3....

3) I played with the "Non Linear IAT fuel correction" table. first cold start, no issue, no touchy. second start today, kindof hot, got the lean issue. I went ahead and generously added some values where the temp was at (98 degrees): first 2, then 4, then 10. only the 10 made enough of a difference to compensate. I drove away happy for about 15 minutes, and then realized that the AFR had become rich. I had to set the values back to 0 to get it back to normal. I had to do two more hot starts later on, and had to do the same.
Conclusion:
a) the Non Linear IAT fuel correction works for this issue, until the car cools down. then it seems to overcorrect. I have to fine tune it so that it only works when the car is hot, and not during normal operating temps. which is a problem because in traffic, it'll be hot-ish, but usually without the high AFR if it was not a hot start. to be continued.
b) somehow, my car seems to all of a sudden have the "normal" hot start issue that goes away after a few minutes.
c) i'm very happy but somewhat confused. I'll have to drive her some more to see if this last. somehow, my car seems to do that sort of stuff; be ok one week, be weird another, and then back to normal, USUALLY without my doing anything.

4) ol boy. sorry. I don't use tapatalk.

[Hahns5.2]

The extra firmware should allow you to easily test your injectors, I'd really recommend doing that, just need a 100cc graduated cylinder. I'd also recommend using PID EGO control.

I will say since OEM vehicles don't suffer from this problem, even with IATs located right in the intake manifold, something about the implementation of the MAT correction with Megasquirt code must be wrong, or some sort of feature that ignores the MAT after a hot start for some time is needed (separate from the ignore MAT during ASE, since ASE after a hot start is quite short), or whatever it might be, since this is a constant issue for most users. I know they've implemented the MAT/CLT correction curve as well, I've tried messing with that and it seems to make no difference with my AFRs no matter how I adjust it.

But the biggest help for me so far was getting the injector deadtime right. Unfortunately my fuel pressure is a returnless system with no pressure reference, so it will never be truly right until I invest in a proper fuel system.

[ol boy]

Heard OEM uses a long time delay for ASE into WUE. What's the max ign events allowed in the ASE counter? I'm at 500 for most of the curve, so I'd assume it's an uint16_t in size. 0 to 65535..

Sent from my SM-G920V using Tapatalk

[turbo conversion]

Have a look at my msq, it may answer a lot of your questions.

David

[Supernova_6969]

hi!

so a quick, possibly final update.

I migrated to MSextra code 2 weeks ago, with David's help (thanks a million). that made it so I speak the same tunning language as most people here. I did end up changing the injectors settings like david suggested, which might or might not have helped. I know my VE table needs a lot of work now, so it did do something.

I also ended up bypassing a fuel damper, which leaked and possibly created fuel pressure issues (although I don't see how it would change anything about hot starts).

but overall, the issue, or the non-issue, has been stable for two weeks. the changes to GAIR table, where I give it a bit more gas when it's got some hot air, seems to have pretty much fixed the issue.

I wanted to thanks everyone who posted, and I hope some people find this useful...

Seb