Same tune - back to back tests - different AFRs...

[2005vetteTT]

Hello all -

Just curious if this is normal behavior. Below I have two 3rd gear pulls, they were about 30 seconds apart.

Pull 1: pull_1.JPG

Pull 2: pull_2.JPG

As you can see, in pull 1 my AFR is 11.9 VS pull 2 where my AFR is 11.4. The IAT differences are only 5 degress (first pull 59, second pull 64). Is this something normal with HPtuners? I know the boost levels are marginally different but the second pull has a little more boost and a lower AFR. They are both within the same point in the tune (5200 rpm, 152kpa).

[mowton]

Assume you are looking at the point on the cursor line. As a sampling system operates there are peaks and valleys based on the sample rate so looking at one time tick and comparing isn't really a fair way to evaluate. In addition your WB has a sample rate as well and they are asynchronous (moving at different rates and different cycles) so there will always be some sampling error which averages out over time. Variance is only about 4% and probably the next pull would yield different readings as well. I wouldn't worry about this minor variance.

Ed M

[2005vetteTT]

Assume you are looking at the point on the cursor line. As a sampling system operates there are peaks and valleys based on the sample rate so looking at one time tick and comparing isn't really a fair way to evaluate. In addition your WB has a sample rate as well and they are asynchronous (moving at different rates and different cycles) so there will always be some sampling error which averages out over time. Variance is only about 4% and probably the next pull would yield different readings as well. I wouldn't worry about this minor variance.

Ed M

I see what you mean but my worry is the whole pull is like this. From 3500 RPM to 6000, i'm around 11.4-11.5 on the second pull and 11.9-12.0 on the first pull. If there variance is that high everytime, how do I know what AFR i'm really at... 12 is too lean for my setup.

[mowton]

I see what you mean but my worry is the whole pull is like this. From 3500 RPM to 6000, i'm around 11.4-11.5 on the second pull and 11.9-12.0 on the first pull. If there variance is that high everytime, how do I know what AFR i'm really at... 12 is too lean for my setup.

An other contributing influence on WOT AFR is going to be what were the LTFT's just as you entered PE. The PCM will add the trim value to your commanded PE EQ if the trims are positivein to help "protect the engine" because it assumes that the leanness will carry forward. If you tuned in either closed loop, or had a hanging fuel trim in open loop, I guess you could have gotten a false rich condition, pulled fuel and when the trims were ok, the resultant fueling with PE only would be leaner.

Hard to say with just a single screen shot of your chart values only..... Need to provide the logs so we can look in detail.

Ed M

[2005vetteTT]

An other contributing influence on WOT AFR is going to be what were the LTFT's just as you entered PE. The PCM will add the trim value to your commanded PE EQ if the trims are positivein to help "protect the engine" because it assumes that the leanness will carry forward. If you tuned in either closed loop, or had a hanging fuel trim in open loop, I guess you could have gotten a false rich condition, pulled fuel and when the trims were ok, the resultant fueling with PE only would be leaner.

Hard to say with just a single screen shot of your chart values only..... Need to provide the logs so we can look in detail.

Ed M

I will post the log + tune once I am home. I have LTFT disabled. The car is running in closed loop for part throttle but immediately goes OL once I go wide open. Thank you for your help!

[mowton]

I will post the log + tune once I am home. I have LTFT disabled. The car is running in closed loop for part throttle but immediately goes OL once I go wide open. Thank you for your help!

The logs and tune should provide more insight for sure :-)

Ed M

[2005vetteTT]

AFRchanges.hpl
ECS_tune_20160218.hpt
layout.Layout.xml

Attached is the log/tune/layout (using HPTuners 3.0). All g/cyl in the tune has been halved because E40 ECU cannot do > 60lb injectors. The car runs great minus this question of changing AFRs i have. Most pulls are perfect and some pulls are too lean or too rich. Not too worried about the too rich but the 12 AFR scares me.

The part of the log I've referenced starts at 8:02.

Thanks again for your help !

[schpenxel]

I don't see a wideband reading in the log. AC pressure instead?

What's the formula you're using to get to AFR?

[2005vetteTT]

Yeah - AC pressure.

([7101.10]/.5)+9.55 is the formula

[schpenxel]

At first glance the VE/boost VE tables look really rough and not smooth at all. With such big swings up and down between cells in the boost VE table I suspect even slightly different MAP/RPM combinations would yield different AFR's. This is what happens when people do the "multiply by %" method over and over without looking at the graph of what they've created.

I would spend some time smoothing the boost VE table out by hand (and using the smooth feature in HPT) to see if that makes things more consistent for you

If you later see variances in AFR on cold days vs. hot days, I've had success in using the Flow rate vs. IAT multiplier table to cancel those out

I see the tune is scaled.. if I were starting from scratch I would probably use the "double stoich / half injector flow rate" method instead so you can get more resolution in the timing table since you aren't using the MAF.

Add manifold air temp to the parameters you're logging

Overall if this was tuned by who I think it was, I'm not impressed (no offense)

What MAP sensor are you running? I've never seen those settings before.. almost looks like someone tried to BS the values to make a sensor work that won't work correctly on a 2005 / E40

[GHuggins]

Something else, the 105kpa rows should be the same between VE tables. Same concept as the last maf cell to first maf cell in the low to high maf tables... Then I would recommend the same as previously mentioned with smoothing out the VE tables....

Just some thoughts....

[2005vetteTT]

At first glance the VE/boost VE tables look really rough and not smooth at all. With such big swings up and down between cells in the boost VE table I suspect even slightly different MAP/RPM combinations would yield different AFR's. This is what happens when people do the "multiply by %" method over and over without looking at the graph of what they've created.

I would spend some time smoothing the boost VE table out by hand (and using the smooth feature in HPT) to see if that makes things more consistent for you

If you later see variances in AFR on cold days vs. hot days, I've had success in using the Flow rate vs. IAT multiplier table to cancel those out

I see the tune is scaled.. if I were starting from scratch I would probably use the "double stoich / half injector flow rate" method instead so you can get more resolution in the timing table since you aren't using the MAF.

Add manifold air temp to the parameters you're logging

Overall if this was tuned by who I think it was, I'm not impressed (no offense)

What MAP sensor are you running? I've never seen those settings before.. almost looks like someone tried to BS the values to make a sensor work that won't work correctly on a 2005 / E40

I am running a cobalt 2 bar sensor. The values are correct as far as I know.

As for the double stoich / half injector flow rate - I am running half injector flow rate right now, and I was told the E40 ecu will not acknowledge the stoich value in the calculation thus this method does not work. I will give it a try though.

The VE tables definitely need some smoothing - even though the histogram says it is in a certain cell, does the ECU blend the cells around it? Since it is hitting the exact same cells for each pull, that would be my only guess as to why smoothing would help?

Thanks again for everyones help on diagnosing this!

[schpenxel]

The double stoich method works just fine on e40's. I'm driving one right now done like that with ID1000's. Stoich is factored in on all ECM's so I'm not totally sure why someone thought it wouldn't work.

Yeah it's pretty much constantly interpolating between neighboring cells when calculating airflow, that's why having things fairly smooth is important

I'll have to look into that sensor. Sounds like there may be an option I didn't know about. I hope so because I've been using map sensors that I've made myself by cutting open stock ones and replacing the sensor with a 2.5 sensor from digikey that works within the confines of an e40 OS

[2005vetteTT]

The double stoich method works just fine on e40's. I'm driving one right now done like that with ID1000's. Stoich is factored in on all ECM's so I'm not totally sure why someone thought it wouldn't work.

Yeah it's pretty much constantly interpolating between neighboring cells when calculating airflow, that's why having things fairly smooth is important

I'll have to look into that sensor. Sounds like there may be an option I didn't know about. I hope so because I've been using map sensors that I've made myself by cutting open stock ones and replacing the sensor with a 2.5 sensor from digikey that works within the confines of an e40 OS

Yeah unfortunately I can't use the whole boost table since it is only 2 bar but it works great for what I need. Part number is 12580698.

Thanks again for your help - I think the double stoich method will really clean the tune up and hopefully smoothing will help the variance I am seeing! Will update once I have a chance to mess with it.

[schpenxel]

Wish I had known about that one before I went through the trouble of making one!

[5_Liter_Eater]

I've run that sensor in the past but it's only 2 bar. The digikey route allows you to use the whole 2.5 bar range of the E40.