Tweaking VE Tables on GPEC2A

[Stoopalini]

For dialing in VE tables, is this the right approach?

Take a log and in VCM Scanner, create a chart which compares "LTFT+STFT %" math against "PRatio" and apply this result to the VE table itself?

So using this log as an example ... Custom-Throttle_3_cold-start.hpl

I copied the column and row axis from the VE table of my tune file out of VCM Editor, and pasted those axis into the chart in VCM Scanner to result in this:

VE_Error.JPG

For the LTFT+STFT chart, I have it set for 1 cell hit. Should I increase this to get better data, or is one cell hit good to use so long I have average selected? Or maybe I need to choose something like 5 cell hits to not use anomalous data to make adjustments?

This chart should give me a % of error for non-WOT fueling, right?


Do I then just copy this out and apply the % adjustments to the VE table directly?

VE_Table.JPG

When applying to the VE table itself, how do you go about doing this? Do I simply select the entire chart of the log result in VCM Scanner, right click and copy .... then go to Editor, select the entire VE table, and right click choosing "Paste Special" -> "Multiply by %"?

I assume if the above is correct, I would then need to smooth it out, as applying these %'s to the VE table would result in some jagged edges if not.

I'd like to start refining my own VE table, but just want to be sure I am doing it correctly.

[Stoopalini]

So reading different threads on the forum, and looking into my current tune and logs ... I'm a bit confused on how commanded AFR is interpreted.

In my tune file, I see Stoich set as 0.0720 in table 34290. This should be an AFR of 13.89, right? (1 / 0.0720 = 13.89)

Then I see FA PE set across the board to 0.0209 in table 34230. This is an adder, so 1 / (0.0720 + 0.0209) = 10.76

Stoich_Setting.JPG

So my AFR target at part throttle, with no FA Enrichment, is 13.89 and WOT AFR target with full PE enabled should be 10.76 .... correct?

So now I go to my log, and look at what is happening there. With just cruise, and units set to Stoich Gasoline (14.7) AFR, I am seeing 14.7

Log_AFR_Stoich.JPG

Of course, if I change the units to Lamda, then I see appropriate % numbers ... in the case of closed loop cruise it's 1

Log_AFR_Lambda.JPG

At 1st I thought this makes sense as the scanner is assuming Stoich is 14.7 when selecting Gasoline as the units, and this is why folks prefer to log and tune using lambda ... so when a stoich value different than one of the standards is set in 34290, it is less confusing when looking at the log data.

But then I consider the fact I have an AEM X-Series OBDII Wideband AFR Gauge, and the gauge reading matches what I see in the scanner while driving ... so this makes me think the 14.7 AFR number is accurate. This would mean for some reason, the ECU is not honoring the Stoich value of .0720.

Here's is what AFR looks like at WOT:

Log_AFR_WOT_Stoich.JPG

And here it is with Lambda selected for units:

Log_AFR_WOT_Lambda.JPG

So if stoich of 13.89 is being honored, then this would be the 10.7 as expected (13.89 * 0.77 = 10.70). But I'm pretty sure my gauge is matching up to the 11.3 AFR number.

One of the tuners who tried to get my vehicle figured out was tuning with ANN enabled, but related to misfires and stoich ... he said:

The issue we seem to have with your tune is that sensed MAP value does not seem to be attached to anything, and it is determining the point at which WOT is at barometric and switching to MAP, which holds off the power enrichment needed to account for positive cylinder displacement. Even with additional fueling after the misfire, it cannot be recovered. The fueling has to be there before positive cylinder pressure occurs to keep misfire from happening.

That brings me to the next issue. My tune runs at 14:3-14:00 for stoich, this OS is not honoring that and AFR is moving around during non power enrichment functions. So even if MAP is off by a couple percent, we can not use a slightly rich cruise to cover up lean fueling tip in, which is usually a safer way to run boost. It seems to be stuck using calculated mass airflow with a defined injector constant to accomplish 14.7.

Now I know with NN disabled and tuning by VE, some of that is different ... but the concept of the ECU sticking with 14.7 stoich seems to be something both tuners observed.

[LilSick]

why cant that variance be from the small amount of alcohol in the gasoline?

i believe it is

that is why lambda is used

[Stoopalini]

In a previous tune, Stoich was set to 0.0757 which is 13.2 AFR, and my AEM gauge would still read ~14.7 during cruise. I understand commanded EQ will always read 14.7 so long as Gasoline is selected for units .... but I would expect the AEM gauge to show the real AFR. Or is the AEM gauge reading the value from the ECU and using it to determine stoich and lambda?

I assume the AEM gauge does not do that, and something is commanding 14.7 as Stoich regardless of what is entered into table 34290. I just looked at my stock calibrations, and it's 0.0720, so 13.89.

For comparison, I looked at a stock 2015 Charger R/T, and it has 0.0688 entered for Stoich. So this would be 14.53 AFR. Maybe this is because E10 was not so common in 2015?

Maybe turning ANN off in my OS is not completely disabling it somehow, and thus 14.7 continues to be targeted?

In any case, is the approach I outlined in my 1st post valid for adjusting VE tables? If so, it seems my fueling is really out of whack in some areas (ie: 18 and 20% in some cases).

[Spray-Cam Hell-Ram]

I could be wrong in my thinking here but as long as your O2 sensors are making fuel corrections, you'll likely see 14.7 being targeted. Turn ANN off and put it in open loop and hopefully you'll see your requested AFR being targeted. You would need to input 376 degrees for temperature for both the front and rear O2 sensors under Fueling>Oxygen sensors to override the O2 sensors and run in open loop to be able to do what you are trying to do.

[B00STJUNKY]

In a previous tune, Stoich was set to 0.0757 which is 13.2 AFR, and my AEM gauge would still read ~14.7 during cruise. I understand commanded EQ will always read 14.7 so long as Gasoline is selected for units .... but I would expect the AEM gauge to show the real AFR. Or is the AEM gauge reading the value from the ECU and using it to determine stoich and lambda?

I assume the AEM gauge does not do that, and something is commanding 14.7 as Stoich regardless of what is entered into table 34290. I just looked at my stock calibrations, and it's 0.0720, so 13.89.

For comparison, I looked at a stock 2015 Charger R/T, and it has 0.0688 entered for Stoich. So this would be 14.53 AFR. Maybe this is because E10 was not so common in 2015?

Maybe turning ANN off in my OS is not completely disabling it somehow, and thus 14.7 continues to be targeted?

In any case, is the approach I outlined in my 1st post valid for adjusting VE tables? If so, it seems my fueling is really out of whack in some areas (ie: 18 and 20% in some cases).

Your gauge isn't pulling any data from the PCM to determine stoich. Widebands are pretty much all setup to display AFR for gasoline without any ethanol influence calculated in the displayed reading.

Furthermore, O2 sensors don't actually measure the amount of fuel that was burned in the combustion process. They measure the oxygen content in the exhaust stream and derive a Lambda value based on that content. From the Lambda value, you can reverse calculate what that translates to in AFR through a simple algebraic equation (Lambda value multiplied by whatever the stoich value is of the specific fuel in use).

Once you actually understand that, and realize that the "AFR" number a wideband displays is literally just a calculated value derived from Lambda and the assumption that the fuel in use has a stoich value of 14.7:1 (which obviously pump gas with ethanol mixed into it does NOT have), it starts to make way more sense why folks recommend forgetting AFR and they tell you to focus on Lambda.

The scanner is showing a target of14.7:1 at cruise because it assumes gasoline with 0% ethanol. The wideband displays 14.7:1 when running at stoich because the wideband AFR also assumes (wrongly, in nearly all pump gas scenarios) gasoline with 0% ethanol is in use, and, again, it simply doesn't measure the ratio of fuel that was burned in combustion. Again, it measures oxygen content, and based on that oxygen content, it correlates that to Lambda. And from Lambda, it calculates against the stoich value it was programmed with.

I understand this post was a bit redundant, but I figured I'd phrase it a couple different ways to drive home the way these things work in hopes that it helps folks understand why there's so much confusion around this subject.

[Stoopalini]

I definitely don't want to do an open loop only tune. My end goal is to have ANN enabled to maximize VVT. This is the wife's daily driver, so smooth, compliant, safe power is the real goal. Not looking for maximizing output or chasing dyno/time slip numbers or anything. My questions about VE, stoich, misfires, etc ... are all in efforts to learn and gain a better understanding of how it all works.

That said, it's my understanding with ANN turned off and operating in closed loop, the ECU should use the stoich value defined in table 34290 for lambda, even though VCM scanner will report 14.7. Actually, the ECU doesn't know anything about AFR at all as it all appears to be in FA ratio. So I believe whatever is set as stoich in table 34290 should be considered lambda by the ECU. Then any deviation from that (rich or lean) would be a modification to lamda. VCM scanner will convert that to an AFR number for you, but it uses preset stoich values based on particular fuels, and for gasoline, it's 14.7.

So if stoich is 0.0720 (ie: 13.89 AFR) in table 34290, and in the scanner you select Gasoline for the display unit ... then when "Commanded EQ" is showing 14.7 in VCM Scanner, this is just representing lambda or stoich, which is really 13.89. So I'm assuming my gauge (which is using a wide band to measure) should read 13.89 when commanded EQ is showing 14.7.

So dealing in lambda means you just watch for deviation from 1 to see how rich or lean you are, when 1 = stoich/lambda (or 13.89 when stoich is set to 0.0720).

But what I'm observing is that the ECU is continuing to use 0.068 for stoih even though something else has been defined in 34290.

[Stoopalini]

Thanks BoostJunky, that helps to confirm what I'm starting to understand.

So what is table 34290 actually used for then? I thought this would be where you can define the stoich of the fuel in use, but it seems the ECU is ignoring it.

[Jim P 2.0]

Lambda is far better to use, doesn’t matter the fuel type and its stoic value. No matter the fuel type a value of 1 is stoic and anything above is lean and anything below is rich.

[Stoopalini]

Makes sense Jim ... how can I command a cruise target of 14.1 AFR and then track it with lambda?

I thought this would be accomplished by changing the value in table 34290 to 0.0709, so the ECU thinks the stoich of the fuel you are using is 0.0709 ... but that doesn't seem to be working. Or ... I am still not fully understanding how the scanner and my wide band are reporting the numbers.

If I enter 0.0709 in this table, does it mean the target mixture is 14.1 AFR in reality when Commanded EQ is reading lambda of 1 in VCM scanner? If so, wouldn't my wide band then be reporting 14.1 AFR since I can't change stoich in the gauge (hard set to 0.0680 I believe)? Or if I change the gauge to read in lambda, would it report 0.93?

Or do you accomplish a slightly rich cruise mixture by somehow telling the ECU to target 0.93 lambda? I just haven't figured out how you can do that, if it's even possible.

[B00STJUNKY]

Thanks BoostJunky, that helps to confirm what I'm starting to understand.

So what is table 34290 actually used for then? I thought this would be where you can define the stoich of the fuel in use, but it seems the ECU is ignoring it.

You're obviously still missing something.

.0720 is the manufacturers assumption of the stoich value for gasoline being used from the majority of station pumps. Is it going to be accurate for all stations at all times? Nope. But that doesn't matter much, since the vehicle is equipped with closed loop feedback from O2 sensors.

If the value were changed to .0680 instead of .0720, you'd see a constant closed loop feedback that was around 5.9%, assuming trims were at or near zero before changing said valve to .0680.

If you were to shut off the O2 sensors, the PCM would use this stoich value to calculate fuel requirements in open loop, with the addition of any modifiers that were applicable, including PE Enrichment whenever that was appropriate.

So, let's say the airflow tables are perfectly calibrated for the engine, all modifier tables were 100% calibrated correctly, injector characteristics are 100% calibrated with no errors, and you shut off O2 sensors. The fuel requirements would be calculated at the ratio of .0720 grams of fuel for every gram of air entering the engine. The PCM "knows" the exact amount of air being ingested into the motor at all times based on all these calibrations.

Now, since we are assuming everything was 100% calibrated with no errors, and the fuel we're using does indeed have a stoich value of 13.89:1, you would see a Lambda of 1.0 in the exhaust. The injectors will have injected fuel at a rate of .0720 grams of fuel per gram of air in the cylinder air charge for each cylinder, which equates to 13.89 parts of air per part of fuel. So, you would be running at 13.89:1 AFR, yes? Yes. Would your wideband then read 13.89:1? No. It would not. Why? Because it was programmed to assume e0 gasoline from the manufacturer. It's reading 1.0 Lambda from the exhaust stream, and calculating AFR to be 14.7:1, because that's what it was programmed to do.

Your 34290 table is doing exactly what it's supposed to. It's giving the PCM the base rate of fuel to air ratio for open loop operation, and a base from which to start before O2 sensors make minor corrections for any errors in either calibration, fuel inconsistencies, factory tolerances (as no 2 engines are going to flow exactly identical if measuring on a micro-scale), and even making corrections for wear and tear in the engine.

[Stoopalini]

Ahh, ok ... so the 34290 table is in fact a value used to define stoich, but it only comes into the equation when the narrow band O2 sensors are unavailable for the closed loop feedback (like Spray-Cam suggested).

One of the previous tuners who was trying to get this vehicle sorted had said something about trying to target an AFR of 14.1 during closed loop, so when tip in occurred, it would compensate for the lean spike which was occurring. This is what made me start to dig on the stoich table ... but obviously I misunderstood what this was used for.

It seems when in closed loop, there really is no way to deviate from targeting stoich for whatever fuel you are burning. Which is fine by me, as I imagine that is the best way to run the engine and there has to be some other root cause for the lean spike at tip in.

[B00STJUNKY]

Makes sense Jim ... how can I command a cruise target of 14.1 AFR and then track it with lambda?

I thought this would be accomplished by changing the value in table 34290 to 0.0709, so the ECU thinks the stoich of the fuel you are using is 0.0709 ... but that doesn't seem to be working. Or ... I am still not fully understanding how the scanner and my wide band are reporting the numbers.

If I enter 0.0709 in this table, does it mean the target mixture is 14.1 AFR in reality when Commanded EQ is reading lambda of 1 in VCM scanner? If so, wouldn't my wide band then be reporting 14.1 AFR since I can't change stoich in the gauge (hard set to 0.0680 I believe)? Or if I change the gauge to read in lambda, would it report 0.93?

Or do you accomplish a slightly rich cruise mixture by somehow telling the ECU to target 0.93 lambda? I just haven't figured out how you can do that, if it's even possible.

If you were to change the stoich value in that table to .0709, you would be telling the PCM that the fuel you're running has a stoich value of 14.1:1 rather than 13.89:1, and your fuel trims would increase in the positive direction from wherever they sit now by about 1.6% (meaning that if they were previously averaging -5.0%, they would become -3.4% now - and if they were previously averaging +5.0%, they would now average +6.6%).

You aren't changing the value that the O2 sensors target when changing this parameter. You are only telling the PCM what the fuel stoich value is assumed to be.

You cannot change what the O2 sensors target with the existing parameters in HP Tuners. And it becomes even more complicated and nuanced when differentiating between factory equipped narrowband o2 sensor feedback vs factory equipped wideband O2 sensor feedback. Narrowband sensors are really only accurate at Lambda 1.0, whereas wideband O2 sensors are accurate over a much wider range. So, assuming you COULD change the target of the O2 feedback system, it would only be reliable to run something other than Lambda 1.0 if it were factory equipped with wideband O2 sensors.

[Spray-Cam Hell-Ram]

Makes sense Jim ... how can I command a cruise target of 14.1 AFR and then track it with lambda?

I thought this would be accomplished by changing the value in table 34290 to 0.0709, so the ECU thinks the stoich of the fuel you are using is 0.0709 ... but that doesn't seem to be working. Or ... I am still not fully understanding how the scanner and my wide band are reporting the numbers.

If I enter 0.0709 in this table, does it mean the target mixture is 14.1 AFR in reality when Commanded EQ is reading lambda of 1 in VCM scanner? If so, wouldn't my wide band then be reporting 14.1 AFR since I can't change stoich in the gauge (hard set to 0.0680 I believe)? Or if I change the gauge to read in lambda, would it report 0.93?

Or do you accomplish a slightly rich cruise mixture by somehow telling the ECU to target 0.93 lambda? I just haven't figured out how you can do that, if it's even possible.

You can accomplish this at least 2 different ways that I know of... the way I mentioned in my previous post, open loop, or, since you don't want to operate in open loop, then simply add the difference between 14.7 AFR (or in our case, .0688) and your target cruising AFR into the cells in the particular area of the power enrichment tables that represent cruising conditions. So in your case .0709 - .0688 = .0021. And if your cruising air charge is under 350mg, you can either adjust the air charge range of your PE tables to accommodate your cruising air charge before inputting values. Or if you have a later model hemi (not sure which year starts adding additional AFR tables under Fuel>Oxygen sensors, guessing around 2014/15) there are a couple of tables under Fuel>Oxygen sensors>upstream called "Bank 1 Control Goal FAR" and "Bank 2 Control Goal FAR" where you can also adjust FAR values based off of air charge and rpm. And then under "Downstream" there should be offset control goal tables you can create voltage offsets within to adjust commanded FAR. Ive had success altering commanded FAR using these tables

[Stoopalini]

Or if you have a later model hemi (not sure which year starts adding additional AFR tables under Fuel>Oxygen sensors, guessing around 2014/15) there are a couple of tables under Fuel>Oxygen sensors>upstream called "Bank 1 Control Goal FAR" and "Bank 2 Control Goal FAR" where you can also adjust FAR values based off of air charge and rpm. And then under "Downstream" there should be offset control goal tables you can create voltage offsets within to adjust commanded FAR. Ive had success altering commanded FAR using these tables

This is great, thanks! I have a 2023, and the upstream tables are there but I don't see the downstream offset control goal tables. I do see upstream FF offsets in table 11464, but nothing for downstream non-FF goals.

Comparing 42994 and 42995 tables, it looks like the 2nd tuner I engaged was trying to leverage these tables to target a different FAR for closed loop. He has 0.0700 populated in those tables where OEM had 0.0688.

Here is how these tables are populated in the stock calibrations:

O2-Sensors_Stock-Tune.JPG


Whipple left these values alone in their canned tune, and the 1st tuner I engaged also didn't adjust these for his VE tune.

The 2nd tuner I engaged did change these for his ANN enabled tune, and here is what he changed them to:

O2-Sensors_2nd-Tuner.JPG

Another interesting point is with the 2nd tuner's ANN enabled file, the vehicle has much better performance (of course there's a lot of other changes as well) but I experience high misfire counts and KR under cruise conditions with his calibrations. Using the Whipple supplied tune, there are no misfires and no KR, but the 0-60 performance isn't much better than it was before the supercharger install.

0-60 time stock was 6.2s
0-60 time with tuner #1's VE tune is 5.5s
0-60 time with tuner #2's ANN tune is 5.0s
0-60 time with Whipple's ANN tune is 6.0s

I don't really have a goal for 0-60, I'm just using it as a comparison point. I think when everything is running efficiently balanced, I should probably be somewhere in the mid to high 4 second range for 0-60.

I really appreciate the engagement guys, it's extremely helpful as I try to get my mind around how this is all working.

[Stoopalini]

Here's an example of the lean spike which is happening at throttle tip in. No KR or misfires with this tune, but the lean spike is still showing in the log.

Anytime I go from 0v pedal to say 0.80v pedal, a lean spike is recorded by the wide band. This happens from a stop or while rolling. In other tunes, this was causing misfires but in this latest revision, it seems to be handling it. This is basically the tune Whipple provided to me with a few changes I made to it.

Here's the tune file: Whipples_2nd_Calibrations_TH2.hpt

And here's the log: Whipples_2nd_Calibrations_TH2.hpl

This was a cold start, 1st of the day.

Any ideas why this is happening? It happened with the VE tune, as well as with ANN tunes.

[LilSick]

you cant compare 0-60 with a street tire on a dirty street and what? a dragy?

[Stoopalini]

It's just a point of reference. The tires have never spun during take off (it's AWD with 295 continental extreme contacts) and I get the time from looking at the VCM Scanner log. The vehicle does have the SRT timers, which contain a 0-60 timer ... But I've read it can be off by 1-2 tenths, so I use the log for consistency. Same stretch of road, no brake torque at launch, in normal drive mode ... I just go from full stop with brake applied direct to WOT.

I realize if I put it in track mode and brake torqued the engine up to 2000 RPM or so, it should be a faster time ... but this is just for comparison purposes though. It is pretty obvious when driving it with the different tunes as well. The 0-60 just helps to put some objective numbers behind my subjective evaluation.

[LilSick]

playing around with the map sensor settings while not logging knock voltage on a blown hemi...

you have bigger balls than i do

how much fuel pressure does it have?

[Stoopalini]

Hmm, I do have both Knock Sensor voltage channels in the template, but it looks like they stopped reporting. I probably need to repoll for channels. This tends to happen from time to time with VCM scanner for some reason. Random channels will just drop off and stop reporting until I do a repoll.

I didn't notice it here because there's really no KR being reported, so I didn't go and look for them.

Not sure on the fuel pressure. I installed the Whipple kit which came with an MSD boost-a-pump. From what I have read, the Durango's OEM pump is pretty stout and Whipple says it can handle this setup with the low amount of boost being generated. Injector duty cycles don't seem to be getting very high and I don't have any issues with the mixture going lean when under WOT. It's only the initial tip in which generates a short lean spike.

Do you think it could be a fuel issue related to the pump or the boost a pump setup? If I do end up having to upgrade the fuel system in the future, I've settled on this KPM drop in replacement. But for now, I would think the OEM pump with the MSD unit would suffice; especially with it being a fresh install on a new car.

https://kpmfuelsystems.com/product/k...dodge-durango/