How ECU determines air mass flowrate

[kenamond]

Well, after looking at your other 5 page thread, I'm going to say that it's looking more and more like a knock sensor problem. I also noticed that your logs show excessive timing, expecially for a 2g, so I'm inclined to agree with 2gGSX on that one.

We really need to find you a way to get ahold of some race gas. Oh, and I'm pretty sure that the highest 2g Airflow map occurs at 2.1g/rev, not 2.2g/rev.

You must've posted this while I was replying, so I didn't notice it.

So you're agreement with 2gGSX is what in particular? Timing is off, and that's leading to too much advance and, as a result, knock? Or that the knock sensor is bad? Or ???

I guess I'm unclear on some things with the timing. Here's what I *think* could be happening.

First off, I'm assuming the 2gb cars have a cam angle sensor and no crank angle sensor, but I got all confused about this when I had to change my CAS. The 2ga cars (or maybe just the 95's) have two sensors. Anyway, that's what I'm assuming below.

The timing can be off if I jump a tooth. If I do this so that the intake cam is off one tooth, the CAS will be off one tooth. So the ECU thinks it's 1 tooth worth of angle off. I vaguely recall 48 teeth on the cam pulleys? If so, that's 7.5° off in cam angle which is 15° in crank angle, right? (cam rotates half as much in 1 tooth as the crank)? So the ECU is getting bogus data about crank angle. It's firing WAAAY off. I'm not sure if there's a preferred direction to jump a tooth, but if it jumps so that the intake cam is one tooth late, then I get an artificial 15° retard. The other way gets me 15° advance. Also, I payed to have my t-belt changed (before I started wrenching on the car), so maybe they didn't time it correctly, and this error could be in either direction.

If it was retarded 15°, I think I'd notice, right? If it was 15° advanced, I'd knock all over the place, right? Is this a possibility? 15° retarded seems rediculously off to run right, but that's pure suspicion. And is 15° advanced possible and still have the car run reasonably well (and hold together for 10k miles)?

Back to the ECU... The ECU still only knows what the CAS tells it, right? If so, then the ECU is still *wanting* excessively advanced timing. If it's looking it up on the table, this tells me that the airflow is wrong. Which brings me back to the MAS.

So being off a tooth doesn't seem like it would have any impact on the excessive advance I'm logging, because the ECU just looks up the timing advance off of a map based on whateverthehell the MAS and CAS feed it for mass airflow and rpm.

Ah hell, I just need to go check the timing marks and shut up.

 

[spyderturbo007]

Right on spyderturbo007 - the highest 2g airflow map occurs at 2.1g/rev, not 2.2g/rev. The PDF I uploaded previously in this thread confirms this. I must be getting senile in my old age.

No problem. I didn't even look at the PDF, so I could be getting senile too. God knows I'm getting old.

You would have to purchase or borrow what is probably a very expensive scan tool (I think its called "MUT-II") from a dealership in order to force the ECU into the proper mode to then be able to verify base timing.

Yep, you can do it with a MUT-II, but last I heard they are over 3k, provided you could even get your hands on one.

So you're agreement with 2gGSX is what in particular? Timing is off, and that's leading to too much advance and, as a result, knock? Or that the knock sensor is bad? Or ???

Actually, both. Based on the excessive timing I saw in your logs, it looks like there is a problem there. I would also look into a knock sensor problem because that would also be high on my list of "things to address". I'm hesistant to say that because I'm going to be the last person to advise you to start throwing money at the car. Maybe you could swap it out with a friends?????

As for the MAS, I though I read in your other post that you swapped it out with a working one and didn't see a difference. Or am I with Al and going senile?

 

[kenamond]

Actually, both. Based on the excessive timing I saw in your logs, it looks like there is a problem there. I would also look into a knock sensor problem because that would also be high on my list of "things to address". I'm hesistant to say that because I'm going to be the last person to advise you to start throwing money at the car.

As for the MAS, I though I read in your other post that you swapped it out with a working one and didn't see a difference. Or am I with Al and going senile?

I can afford a knock sensor. I vaguely recall they're ~$80, but I'm old, too. And I can afford to pop the back seat out, unscrew the rewired FP ground and dremel the paint off the metal, too. Hell, I can even afford to search on how to check what voltage my FP is getting! I'm ROLLING IN IT!! But I'm pretty short on motivation at the moment.

I didn't swap out the MAS, and I don't recall lying about it either Must be senility.

Any ideas why I'd get excessive timing? If my LTFTs are good, I imagine that means that the ECU is delivering the correct fuel most of the time (when in closed-loop mode). If that's the case, then it must be getting the right g/rev, which means that the MAS should be working. But I have no idea why I'd get too much timing advance unless the MAS was bad.WTF

Wish I knew someone with a 2gb so I could swap out the MAS.

I'm gonna go skim the classifieds...

 

[zippyshoe]

The only causes of "excessive" timing advance that I can think of are:

1. Using a piggyback fuel controller in conjunction with large injectors. The large injectors would require hiding a lot of airflow from the ECU. This could move you onto a lower airflow timing map, which would result in higher peak timing advance.
2. Using a modified ECU (such as a custom "chipped" ECU) which uses a set of timing maps which differ from the stock maps.
3. Using a standalone fuel controller (or DSMLink) which allows modifications to the factory timing.

Unfotunately, it doesn't appear that any of these would apply to your situation based on your vehicle profile. Would you be able to post a log or a link to a log of a WOT pull that displays RPMs, mass airflow (lbs/min), and timing advance?

Also, have you checked your throttle position sensor? This is a long shot, but I was wondering if a defective TPS could cause your vehicle to get "confused" as to whether its supposed to be in open-loop or closed-loop mode.

 

[kenamond]

The only causes of "excessive" timing advance that I can think of are:

1. Using a piggyback fuel controller in conjunction with large injectors. The large injectors would require hiding a lot of airflow from the ECU. This could move you onto a lower airflow timing map, which would result in higher peak timing advance.
2. Using a modified ECU (such as a custom "chipped" ECU) which uses a set of timing maps which differ from the stock maps.
3. Using a standalone fuel controller (or DSMLink) which allows modifications to the factory timing.

Unfotunately, it doesn't appear that any of these would apply to your situation based on your vehicle profile. Would you be able to post a log or a link to a log of a WOT pull that displays RPMs, mass airflow (lbs/min), and timing advance?

Also, have you checked your throttle position sensor? This is a long shot, but I was wondering if a defective TPS could cause your vehicle to get "confused" as to whether its supposed to be in open-loop or closed-loop mode.

My profile is current, so none of those suggestions would apply. That is, unless someone snuck into my car and put a chipped ECU in it when I wasn't looking.

My TPS readings look fine. They were like 97+% as I recall. Also, the O2 sensor output for all of my logs look like I'm definitely in open loop mode for the pulls. At idle, the O2v fluctuate, and I've noticed that when I'm on my way to do a pull with the logger on and when I'm at part throttle, so it seems that the open/closed loop behavior is okay.

I'll go dig up my original post on this subject (has several logs) and add it to this post...May or may not have any with airflow in them.

 

[2gGSX]

The only causes of "excessive" timing advance that I can think of are:

1. Using a piggyback fuel controller in conjunction with large injectors. The large injectors would require hiding a lot of airflow from the ECU. This could move you onto a lower airflow timing map, which would result in higher peak timing advance.
2. Using a modified ECU (such as a custom "chipped" ECU) which uses a set of timing maps which differ from the stock maps.
3. Using a standalone fuel controller (or DSMLink) which allows modifications to the factory timing.

Unfotunately, it doesn't appear that any of these would apply to your situation based on your vehicle profile. Would you be able to post a log or a link to a log of a WOT pull that displays RPMs, mass airflow (lbs/min), and timing advance?

Also, have you checked your throttle position sensor? This is a long shot, but I was wondering if a defective TPS could cause your vehicle to get "confused" as to whether its supposed to be in open-loop or closed-loop mode.

His original assumption of a hugely miscalibrated MAS would also give rise to the possibility of the ECU seeing such a low g/rev count that he's on a lower load timing map, which would also give him excessive timing.

Open/closed loop operation on a 2g is only defined by baro, intake temps, and coolant temps: coolant < 179F, intake >= 133F, baro < 22.9 inHg, baro > 31.6 inHg. Under the assumption that he's stuck in open loop, there should be other signs such as the O2 not cycling during cruise.

A malfunctioning TPS sensor would however cause the ECU to give lower airflow readouts, which as stated above could give him excessive timing.

 

[kenamond]

Here is the 5-page thread where I was first trying to figure all of this out. Post #44 has a log with airflow numbers. I need to convert that to g/rev and see what the 2g stock timing map shows. Then I can *in theory* see how much timing is being pulled by comparing the logged timing and the map timing advances.

 

[2gGSX]

You must be logged in to view this image or video.

Voila! Data from: http://www.dsmtuners.com/forums/showpost.php?p=50443158&postcount=44 "First log"

Any deviation in the second graph between your observed and expected timing looks consistent enough to not be due to knock, and more likely coolant temps. I've NEVER seen anyone hold a constant 2 degrees of timing retard due to knock throughout a full pull. It will always get worse or go away.

 

[zippyshoe]

His original assumption of a hugely miscalibrated MAS would also give rise to the possibility of the ECU seeing such a low g/rev count that he's on a lower load timing map, which would also give him excessive timing.

Open/closed loop operation on a 2g is only defined by baro, intake temps, and coolant temps: coolant < 179F, intake >= 133F, baro < 22.9 inHg, baro > 31.6 inHg. Under the assumption that he's stuck in open loop, there should be other signs such as the O2 not cycling during cruise.

A malfunctioning TPS sensor would however cause the ECU to give lower airflow readouts, which as stated above could give him excessive timing.

Yup, I forgot to list the miscalibrated MAS as an additional cause of excessive timing advance.

I was actually thinking that kenamond may have been unexpectedly operating in closed-loop during a WOT pull. I was thinking that if his TPS was reporting a value much less than 100% at WOT, then this might cause the ECU to remain in closed-loop operation(??). However, this seems pretty unlikely, especially based on the logs contained in his other post. Also, his TPS seems to be fine.

Here is the 5-page thread where I was first trying to figure all of this out. Post #44 has a log with airflow numbers. I need to convert that to g/rev and see what the 2g stock timing map shows. Then I can *in theory* see how much timing is being pulled by comparing the logged timing and the map timing advances.

I was looking over this post and the one item that was alarming to me was that you appear to lean-out pretty badly at the upper RPM's based on the O2 voltage. That could certainly cause some significant knock. Looking at Post #44, your logged airflow does seem pretty low compared to what I would expect at 15 psi boost. For example, at 17 psi of boost, I log about 31 lbs/min of airflow (FP Big28). At 15 psi of boost, I would expect you to see roughly 29 lbs/min of airflow instead of only 21 lbs/min as shown in the Post #44 log. Of course if your MAS is reporting lower-than-actual airflow numbers, this would also explain the leaning-out condition since insufficient fuel would be delivered.

I forgot to ask - Is your MAS hacked in any way?

 

[kenamond]

Wow, 2gGSX, you shouldn't have! But thanks! I was actually looking forward to that little exercise, but this cuts to the chase for me.

My MAS is 100% stock and untouched on the inside (have to touch the outside ).

So I'm wondering why I am getting such low airflow values. If the MAS was miscalibrated, wouldn't my LTFT values be way worse than 4% based on the long-term closed-loop data? I'm intelligent-guessing about LTFTs...they record over a long period of time how much extra/less fuel is needed than what the fuel map says based on long-term O2v readings during closed loop mode?? Something like that? If so, that seems like a way to compensate for MAS miscalibration. So 4% LTFT tells me the MAS calibration is close enough.

Also, keep in mind that air up here is at .75bar, not 1.0bar. 15psi boost puts me at 26psi whereas this would be 29.7psi at sea level. I'd have to calculate what that translates into in density difference, but that may explain the lower g/rev numbers. If it's a linear relationship, then 21lb/min vs. 29lb/min is still too large of a discrepancy (linear would mean I should get 25.4lb/min instead of 21).

Are things just too hot under the hood? My IATs in some of the pulls were between 80 and 84°F with the normal knock problems. On hotter days it was as high as 95°F.

One of my posts in that other thread mentioned that my coolant temps were 190°F throughout the pull, but I think my t-stat was stuck open at the time. I recall that the ECU will pull timing for low coolant temps, so maybe that explains the fixed 2° shift.

So I need to do another pull with the new O2 sensor. I know I was leaning out in that log (I mentioned it in the other thread), so I've been shifting before 6k and holding boost to 12-13psi most of the time.

I'd like to put the stock airbox back on to get the cold-air intake effect, but the aftermarket UICP is in the way. And if I revert back to the stock IC pipes, I have to unbolt the TB and will probably introduce boost leaks. Ugh.

Anyway, I have more info now. Thanks again guys! I need to swallow my miser attitude and just buy the 255 FP, AFPR, and fuel pressure gauge.

 

[zippyshoe]

Based on the Ideal Gas law, it looks like air density would increase linearly with absolute pressure (as opposed to gauge pressure):

PV = nRT,

density = n/V = P/(RT)

Therefore, if temperature is held constant, then density increases in proportion to absolute pressure.

With the LTFT's in-line, I think its still possible that there could be an air metering issue. If the MAS measures accurately at lower airflow values, but becomes inaccurate at higher airflow values, perhaps this might not be apparent from the fuel trims. Since the fuel trims are only updated during closed-loop conditions (when airflow is relatively low), they may not reveal an inaccuracy that only occurs at high airflow levels.

What concerned me the most was the leaning-out condition during the WOT pull - I think this may be what is producing the knock. This could either be caused by air metering inaccuracy or a fuel delivery issue, or a combination of both.

 

[kenamond]

Based on the Ideal Gas law, it looks like air density would increase linearly with absolute pressure (as opposed to gauge pressure):

PV = nRT,

density = n/V = P/(RT)

Therefore, if temperature is held constant, then density increases in proportion to absolute pressure.

With the LTFT's in-line, I think its still possible that there could be an air metering issue. If the MAS measures accurately at lower airflow values, but becomes inaccurate at higher airflow values, perhaps this might not be apparent from the fuel trims. Since the fuel trims are only updated during closed-loop conditions (when airflow is relatively low), they may not reveal an inaccuracy that only occurs at high airflow levels.

What concerned me the most was the leaning-out condition during the WOT pull - I think this may be what is producing the knock. This could either be caused by air metering inaccuracy or a fuel delivery issue, or a combination of both.

Okay. I just went out and did two pulls, both at about 15psi. The first one I logged rpm, airflow, O2v, and timing and I let up at 6300rpm in case I was leaning out. Second pull I logged rpm, O2v, timing, coolant temp, and IAT and I took it to 7k rpm. I'm not gonna post them until I do what 2gGSX did, but the major points are as follows:

Coolant temps held at 203°F the whole pull.
O2v were 0.98v to 5k and then 0.96v to redline (I checked this after the 1st pull to make sure I wasn't leaning out).
IAT was at 87°F throughout the pull
Timing advance dipped to 10 at first boost and rose to 21 at redline EDIT: Timing dropped in two places much like it did before.
Airflow was 20.5lb/min at 6300rpm (didn't log this in 2nd pull to redline)
EDIT: Boost dropped to 11psi by redline

So I now don't lean out like I did. Could just be the new O2 sensor. So for the first time in a year, turning up the boost has actually crossed my mind...but not until I'm convinced I know what's going on now.

Now as far as MAS calibration, one idea I had was to do a 3rd gear pull at 3/4 throttle and see if I can get the airflow up there while in closed loop. I'm not sure what to look for, so I'll think some more about this. I don't know how "short term" the STFTs are...maybe I should look into that.

As far as the low airflow, I'm wondering if it's just that I have stock exhaust restricting flow through the motor. I don't know enough to think through how exactly the increased backpressure in the exhaust mani would affect CC pressures, scavenging, intake flow, etc. Perhaps one of you guys know that stuff
EDIT: The fact that I dropped to 11psi at redline is something we didn't consider. That's surely gonna drop the redline airflow compared to 15psi at redline. If you have airflow data at 4200 rpm or so, that would probably be a more apples-apples comparison.

Anyway, I'm gonna go geek out on the log data and 2g timing map for a while.

Before the logs, I'd decided to go ahead and buy the FP, AFPR, and FP gauge. Now I'm not sure if that's what I need next.

 

[2gGSX]

I would suggest the next mod to be......

You must be logged in to view this image or video.

From that one screen I can tell that my 50 hz airflow metering is off, I need to work on my cooling system, I should adjust my global fuel down 1%, and that's IT!

I get the feeling you'd like it

 

[kenamond]

Yeah, yeah, yeah...I'd love to have DSMlink. But that's almost 3x the FP/AFPR/FPG.

Argh. Once I got going on this, I realized that I need a program to interpolate from the 2g map. I never installed my F90 compiler on this machine...but I'd have fun writing the program. I guess I could learn more Python; it's free and easy to install

When I manually pick off data that is close to a 2g map point, the data is off at high rpm by 4° (retarded). The data at mid-range (4k-5k) looks to be 1° retarded.

I see something strange between the two logs. It looks like the timing and rpm values are out of synch in one of the two logs. When I plot both rpm vs. advance together, there is an offset left-right (in rpm) between the curves as boost is hitting (a very smooth part of the curve).

Anyway, here are the logs. I included g/rev airflow in the 1st log (no airflow in 2nd log).

RPM Timing Airflow(lb/min) O2v Airflow(g/rev)
2520 26 4.71 0.94 0.848
2596 26 5.11 0.94 0.893
2676 24 5.56 0.94 0.943
2768 23 6.03 0.94 0.989
2868 22 6.57 0.94 1.040
2980 21 7.25 0.96 1.104
3108 19 8.23 0.96 1.202
3260 16 9.27 0.98 1.290
3412 13 10.56 0.96 1.405
3620 11 12.18 0.96 1.527
3824 10 13.36 0.96 1.586
4020 11 14.58 0.98 1.646
4232 12 15.52 0.98 1.664
4420 10 16.96 0.98 1.742
4624 10 17.90 0.98 1.757
4812 10 17.78 0.96 1.677
4992 12 18.02 0.98 1.638
5132 14 18.40 0.98 1.627
5300 13 18.89 0.96 1.618
5508 14 19.26 0.96 1.587
5620 14 19.68 0.96 1.589
5780 15 19.95 0.96 1.567
5924 17 20.05 0.96 1.536
6044 17 20.09 0.96 1.509
6160 17 20.21 0.96 1.489
6240 16 20.43 0.96 1.486
6344 17 20.52 0.96 1.468


RPM Timing Air Temp O2v
2472 26 91 0.94
2564 24 91 0.94
2692 22 90 0.96
2808 21 90 0.96
2944 20 90 0.96
3096 16 90 0.98
3252 13 88 0.98
3448 11 88 0.98
3656 10 88 0.98
3876 10 88 0.98
4112 11 88 0.98
4352 11 88 0.98
4568 11 88 0.98
4816 11 90 0.98
5012 13 88 0.98
5236 14 88 0.96
5408 15 88 0.96
5572 14 88 0.96
5736 17 88 0.96
5904 15 88 0.96
6092 17 88 0.96
6232 17 88 0.96
6400 18 88 0.96
6552 20 88 0.96
6676 21 88 0.96
6784 21 88 0.96
6904 21 88 0.96

 

[zippyshoe]

Your O2 voltages now appear to be fine. You do not appear to be leaning-out anymore - at least as best you can tell with a narrowband O2 sensor. It looks like you may still be seeing some knock based on the dips in the timing advance at the higher RPM's.

That is very strange that your boost is dropping to 11 psi at redline. The Big28 is rated at 36 lbs/min max airflow and you are nowhere near the airflow limit of this turbo, even at redline. Normally, I would suspect a boost leak in this situation. However, boost leaks tend to inflate your airflow readings, which doesn't seem likely based on your logs. The Big28 is generally able to sustain at least 20psi to redline, so this is puzzling. Perhaps elevation is a factor here.

 

[2gGSX]

The boost dropping is a relatively common issue with T28s from what I remember. It has to deal with the actuator being way too soft. FP sells a 17 psi actuator which seems to help whenever people upgrade. Based on the DSMLink forums, people see an average of 3 lbs/min increase in peak airflow, up to 5 lbs/min. Most people generally end up at ~35 lbs/min afterwards.

 

[kenamond]

You don't think this has anything to do with my stock exhaust?

And an FP Big28 is not a T28. Don't know if that makes a difference.

And 2gGSX, did you manually interpolate from the map to get the "expected" timing advances or do you have a program?

 

[2gGSX]

Neither the stock exhaust nor the wastegate actuator is helping, and I believe that they are compounding problems. I'm only reporting what I saw from other threads. That and, the wastegate is much cheaper than a full exhaust setup!

The FP T28 should still use the T28 compressor wheel (it's not explicitly stated, but this its predecessor was the FP T28/28 after all, and I'm not sure anything bigger can get shoved in there). The wastegate actuator should have a native spring pressure of ~10 psi from what I remember, which is what was reported by other users as well (FP T28's, not other variants).

When I did the timing advance table I chose RPM values close to the values of the table itself (increments of 500), and then rounded down to the nearest RPM value and chose that as the expected timing advance. An example might be easier: for RPM 2524, I rounded it to 2500 and looked up the assumed timing for 2500 RPM and the given g/rev. Hope that helps!

 

[kenamond]

The FP T28 should still use the T28 compressor wheel (it's not explicitly stated, but this its predecessor was the FP T28/28 after all, and I'm not sure anything bigger can get shoved in there). The wastegate actuator should have a native spring pressure of ~10 psi from what I remember, which is what was reported by other users as well (FP T28's, not other variants).

The T28 and FPB28 compressor maps are completely different, but I don't know if it's just the housing design or the wheel itself.

When I did the timing advance table I chose RPM values close to the values of the table itself (increments of 500), and then rounded down to the nearest RPM value and chose that as the expected timing advance. An example might be easier: for RPM 2524, I rounded it to 2500 and looked up the assumed timing for 2500 RPM and the given g/rev. Hope that helps!

Gotcha. That's what I was doing, but there weren't many values that had both the RPM and g/rev close to the map values. I think I'll write a python script to read the map and log in and dump out the timing retard. I'm not sure if anything other than coolant temp and IAT will cause a timing adjustment (other than knock), but if that's it and I log those as well, I may also be able to compute knock count. My logger used to sample at 11S/s, but now it's down to 6 for some reason.

And yes, you've been extremely helpful. Much thanks!

 

[zippyshoe]

I have also heard of this issue with the regular FP Big28 wastegate actuator (rather than the optional 17psi actuator). Under high boost pressure conditions, the wastegate flapper can blow open due to insufficient tension at the wastegate actuator. However, in your case, at only 15psi boost pressure, I would be skeptical that this would be an issue unless your wasgate arm was bent or defective in some way, or your wastegate flapper wasn't sealing for some reason.

I also have an FP Big28 with the regular wastegate actuator and I have no trouble holding 17psi boost pressure right up to redline. It may hold at even higher boost pressures which I am unable to run due to my stock intercooler.

 

[kenamond]

I have also heard of this issue with the regular FP Big28 wastegate actuator (rather than the optional 17psi actuator). Under high boost pressure conditions, the wastegate flapper can blow open due to insufficient tension at the wastegate actuator. However, in your case, at only 15psi boost pressure, I would be skeptical that this would be an issue unless your wasgate arm was bent or defective in some way, or your wastegate flapper wasn't sealing for some reason.

I also have an FP Big28 with the regular wastegate actuator and I have no trouble holding 17psi boost pressure right up to redline. It may hold at even higher boost pressures which I am unable to run due to my stock intercooler.

Yours is a good datapoint to know. I was thinking the same thing, but I have no idea what backpressure is knocking at the door of the flapper. Maybe it's not a problem. Maybe the stock exhaust bumps up the backpressure so that it is a problem.

But that leads me to believe that I need to start thinking about an exhaust upgrade. I still have to get a couple more logs and get airflow to redline at 9psi and 15psi. Then I can look at the expected and actual timing numbers and find out more about knock. Might be I wasn't knocking in my 9psi pulls as much as I thought. Damned complications!

Have you considered water/meth injection instead of FMIC? That's probably the route I'll go. Cheaper and better intercooling. And if you use the windshield washer bottle, the idiot light even tells you when you need to refill it!

 

[2gGSX]

Gotcha. That's what I was doing, but there weren't many values that had both the RPM and g/rev close to the map values. I think I'll write a python script to read the map and log in and dump out the timing retard. I'm not sure if anything other than coolant temp and IAT will cause a timing adjustment (other than knock), but if that's it and I log those as well, I may also be able to compute knock count. My logger used to sample at 11S/s, but now it's down to 6 for some reason.

And yes, you've been extremely helpful. Much thanks!

Ah! You meant the load values. I'm not sure if the ECU interpolates it or not, and in fact I hadn't considered that. I was working under the assumption that your spot on the timing map gets placed based on the highest exceeded value, similar to the RPM function (e.g. 1.5 g/rev will stick you into the 1.4 g/rev slot on the timing table).

Your idea of interpolation is more likely correct, however, since DSMLink operates by interpolating between values--0 timing change at 3000 RPM, +2 timing at 3500 RPM --> +1 timing at 3250 RPM.

Your best bet if you *really* want to know is to do a quick linear regression between two data points for both load and timing for every point you want to know. Though at this point, we're getting into an exercise of math rather than tuning.

As far as I know, IAT and coolant temp are the only things on a stock ECU that will pull or advance timing (other than knock obviously). From the DSMLink wiki--

2G intake air temp (IAT) ignition advance adjustment:
185F = -3°
132F = -2°
100F = -1°
73F = 0
48F = 0
20F = -1°
-23F = -2°

1G intake air temp ignition advance adjustment:
185F = -3°
132F = -1°
100F = 0
73F = 0
48F = 0
20F = -1°
-23F = -2°

Finally, the drop in sample rate is most likely caused by an increase in the number of logged variables.

 

[kenamond]

Finally, the drop in sample rate is most likely caused by an increase in the number of logged variables.

I thought that the S/sec reported by the logger is total samples per second regardless of number of variables. So at 6S/sec logging 6 variables, I'd get 1 "row" of data per second. I'll check the wallclock times for my last log and see which looks more correct.

Thanks for the other data. I think the document zippyshoe posted (or was it you?) may have described how the ECU computes timing advance including coolant temp and IAT considerations. I'll have to peek at that again. It would be a good test to see if I can predict knock count from DSMLink logs and compare to the DSMLink knock values. Maybe us 2g/pocketlogger guys can read knock count after all, though indirectly and via additional log data. Need the pocketlogger guys to incorporate the 2g ECU behavior into their logger.

 

[zippyshoe]

Yours is a good datapoint to know. I was thinking the same thing, but I have no idea what backpressure is knocking at the door of the flapper. Maybe it's not a problem. Maybe the stock exhaust bumps up the backpressure so that it is a problem.

But that leads me to believe that I need to start thinking about an exhaust upgrade. I still have to get a couple more logs and get airflow to redline at 9psi and 15psi. Then I can look at the expected and actual timing numbers and find out more about knock. Might be I wasn't knocking in my 9psi pulls as much as I thought. Damned complications!

Have you considered water/meth injection instead of FMIC? That's probably the route I'll go. Cheaper and better intercooling. And if you use the windshield washer bottle, the idiot light even tells you when you need to refill it!

I haven't been able to decide between a FMIC or water/meth injection. For all-out performance, its hard to beat a good FMIC. However, for my near-term goals, the water/meth injection option is very appealing. The positives for the FMIC include excellent cooling efficiency and minimal restriction to intake airflow. What I am concerned about is experiencing overheating issues due to reduced airflow to the radiator. Also, once you install a FMIC, it is difficult to return your car back to stock due to all the hacking and cutting involved.

What I like about the water/meth injection option is the slealthiness of this approach and the ease of reverting back to a stock set-up. This approach improves the cooling efficiency of the stock sidemount. The one limitation of this approach is that at high airflow levels, the stock intercooler presents significant restriction to flow and water/meth injection can do nothing for this limitation. But for my goals, I will probably never reach that airflow level.

I thought that the S/sec reported by the logger is total samples per second regardless of number of variables. So at 6S/sec logging 6 variables, I'd get 1 "row" of data per second. I'll check the wallclock times for my last log and see which looks more correct.

Thanks for the other data. I think the document zippyshoe posted (or was it you?) may have described how the ECU computes timing advance including coolant temp and IAT considerations. I'll have to peek at that again. It would be a good test to see if I can predict knock count from DSMLink logs and compare to the DSMLink knock values. Maybe us 2g/pocketlogger guys can read knock count after all, though indirectly and via additional log data. Need the pocketlogger guys to incorporate the 2g ECU behavior into their logger.

I also remember there being a trade-off between sample rate and number of parameters that are logged with Pocketlogger. When I used Pocketlogger, I only logged the parameters I was most interested in at the time in order to maximize my sampling rate.

The timing and fuel maps shown in the PDF I believe are only showing a subset of all the g/rev maps used by the ECU. Using interpolation to determine the intermediate g/rev maps makes sense. I think the one exception is if your airflow is any value greater than the highest map value of 2.1 g/rev. The factory ECU will normally provide 16 degrees max timing advance at 2.1 g/rev airflow. If your airflow exceeds 2.1 g/rev, the ECU will always provide 16 degrees of advance (as long as there is no knock.) - Even if you are flowing twice this value (4.2 g/rev), the ECU will still try to provide 16 degrees of advance.

 

[kenamond]

Actually, with a perfectly efficient FMIC, you can only lower the intake temps to whatever the temperature is outside. However, with water/meth injection, you can actually achieve lower-than-ambient temperatures (impossible with an FMIC). You can run a water/meth injection system without *any* intercooler. You just have to have enough pipe length to evaporate the water/meth spray enough before it is burned in the cc.

What I don't know a whole lot about is how to tune the water/meth system. I forget the "damn good" brand, but it had a fairly sophisticated controller that let you tune the injection based on similar variables as the ECU uses to drive the fuel injectors. As I recall, it was around $300, but that's based on cobweb-covered memories.

I did a bunch of half-assed (okay, maybe three-quarters-assed) research on it many months ago, so I don't recall every detail. I recommend you start looking around before you unload the cash for an FMIC. It seemed very attractive. The biggest downside is that you now have two tanks to keep full (gas and water/meth).