Discussion and questions related to the course Understand AFR
Hi, I am new to tuning and trying to am trying to correlate my learnings from the EFI Fundamentals course to the functionality contained in the HP Tuners editor software (v3.6.26). I understand that the E38 doesn't have a VE table and instead uses a "virtual torque" function. This is where I get lost. Is anyone able to enlighten me on how you can map fuel in this environment? Am I only to use PE to set the desired AFR?
Please provide more context. What engine is this and in what vehicle with what trans? Post screenshots of the exact tables you're looking at. Your username says LS3 - is this a naturally aspirated GM 6.2L LS3 engine? What mods are done to it?
GM's control system is a sort of hybrid between speed density and mass airflow based. So the gas filling model and the mass airflow work together and you have to look at the fuel trims to adjust them. Power enrichment is one of the ways that the AFR is controlled.
thank you. Normally aspirated LS3 with T56 gearbox. It has a MAFless tune with a mid-sized cam. The tuning fuel delivery module showed fuel being set in a 3D table. I am not seeing this in HP Tuners and am wondering if instead is achieved simply by the EQ ratio vs RPM table. If so, how does this work with the virtual torque tables?
TL;DR The power enrichment table is just a target, it doesn't tell the engine how much air is entering it like a standalone ECU's VE table does. The virtual torque is part of a bigger much more complicated speed density style system that takes into account spark timing, accelerator pedal position, etc.
Andre can certainly chime in on this one but this should help. This is a complicated answer because modern stock ECUs are very complicated.
Since this is more of a general "how it works" question, let me explain the basic concept of torque-based control. The details are from documentation from Bosch, who were early adopters of torque based control. Bosch is one of the main suppliers for stock ECUs on German cars (VW especially) and have been used by American OEMS. GM for example had Bosch develop their 3.6 liter direct injected engine, and then later wrote their own software controls.
See the attached images.
Here's a simplified version of the idea. You put your foot down on the gas pedal and the ECU looks at the pedal mapping and says "the pedal is at 20%, he must want 300 Nm of torque." Then it asks itself: "Ok, what do I need to give the driver 300 Nm?" It knows that torque depends on charge air, air fuel ratio, and spark timing. So then it asks itself "Ok, how much air would I need if I weren't knocking and my air fuel ratio is Lambda=1 for best emissions?"
The LS3 is a pretty simple engine by modern standards. It doesn't have variable camshafts or a turbo or EGR or any of that. So now that it gas calculated how much air it needs, it's going to adjust the throttle to get that airflow. It's got a bunch of tables that tell it what throttle position corresponds to what airflow. So now it wants to know how much torque it's making with that throttle position. The tables you posted are part of calculating the modeled actual torque. If I have this amount of airmass per charge, at this specific rpm, with this specific spark timing, what torque do I get?
The Power enrichment table you posted is one part of the target AFR calculation GM uses. The other main part is a catalyst temperature model, which is usually overriden or otherwise disabled/modified in aftermarket tunes. Best torque is going to come from enriching for power (power enrichment) but that hurts fuel economy and emissions. The ECU doesn't want to give it any more fuel than absolutely necessary to meet regulations. There's an algorithm which calculates whether the requested torque can be achieved without retarding the spark from MBT or enriching the mixture to increase power output or enriching the mixture to cool down the catalytic converter.
So power enrichment is a target AFR calculation, "virtual torque" is part of the modeled engine torque that relates to requested torque and a bunch of other controls (torque model is sent to an automatic transmission for example to help with shifting). Actual airmass calculation on a standalone ECU is basically one VE table and some basic math to calculate how much air is entering the engine. On a GM control system and a lot of other modern Stock ECUs it's a bigger, more complicated calculation that takes into account exhaust pressure and temperature, valve timing, etc. So what GM does is calculate airmass using the model and measure airmass with the mass airflow sensor. The MAF and the modeled airmass are compared to each other. Then it uses o2 sensor (narrowband o2 on an LS3) for dynamic learning of the airmass model. The MAF sensor is there to meet on board diagnostic regulations.
So with the MAFless tune, they turn off the MAF and rely on the airmass model. They tweak the airmass model by looking at the fuel trims basically, sort of like an adjustment of a MAF sensor scaling table when you change the MAF sensor tube.
So I pulled up VCM Editor. So the target Air fuel ratio with the engine warm is going to be the power enrichment table as you've seen. If you go to Fuel-->Power Enrich, you'll see (at least on my 2014 Corvette E92 sample calibration that comes with VCM Editor) the table you're familiar with. You should see pedal and calculated torque entry conditions for power enrichment mode. You'll see there's a delay section that in this case is calibrated to be disabled. That's an emissions and fuel economy thing.
Then there are all sorts of other enrichment modes: knock enrichment, over temp enrichment. There's a cat temp enrichment. For transient fuel, it uses a wall film model that's found in a lot of standalones now. That accounts for how much fuel sticks to the walls as the throttle opens quickly or when the engine is cold and the fuel doesn't vaporize well. There's also a bunch of stuff for when the engine is warming up.
So that's the "Target" side of fuel control. But the fueling isn't accurate unless the engine is measuring fuel accurately. As I said previously, it's like a blend of mass airflow control and a very sophisticated speed density system. If you go into the help file and look at Dynamic Airflow (See attached screenshot) you get a better idea of how the speed density side works. The dynamic airflow is sort of like the speed density VE table in a standalone ECU (or an early 1990s GM engine). There are a bunch of different zones where the airflow model operates with baseline VE values, and then based on the short and long term fuel trims the learning changes. So if there are deposits on the valves or some such thing that is restricting airflow at low loads but has only a small effect at high loads, the particular airflow learning zones will account for that.
In your case, you've got an aftermarket cam that changes the overlap and will affect the VE. So that impacts the "Dynamic" and "Speed density" tabs. The help file explains what the tables do. It would be helpful for you to do a comparison between your speed density tune and the stock tune to see all the stuff that's been changed.
Raymond, thanks for your most comprehensive response. I will now need to spend some time trying to understand this aspect as no doubt it is complex. I do agree the best way is to compare a stock tune against my "cam" tune to observe the differences.
Now the torque control is starting to make some sense. I have attached a screen shot of part of my log of my quarter mile run. Note that he actual throttle position maxes out at 83.9% even though I commanded 100%.
I don't think there's much I can add to what Raymond has already posted. To answer your last question though if you log ETC position instead of TPS % you will get a normal 0-100% trace for your TPS so don't worry 83% is normal, it's just the wrong PID.
One thing to understand about throttle is that under some conditions they are highly non linear. As the throttle gets closer to fully open, you get less and less additional airflow. In math terms, it is a logarithmic relationship.
The attached file throttle_2.jpg shows open area of the throttle on x axis vs the throttle % open on the Y axis. You can see that as you get past 60% throttle you don't have much more air entering the engine.
The file throttle_3.jpg shows the air velocity through the throttle valve on the Y axis and the pressure ratio between the throttle inlet (air charge tube) and outlet (intake manifold). You can see that at a large pressure drop across the throttle (a mostly-closed throttle) the velocity is fixed. The air moves at the speed of sound. This physical relationship can be used by an ECU with electronic throttle to calculate airflow as part of an idle control algorithm, or as a failsafe if the mass airflow sensor and/or MAP sensor fails.
The file throttle.jpg shows a 3D map of rpm, torque, and throttle opening % on an older naturally aspirated V6 engine with electronic throttle. Notice how most of the torque that the engine can produce is achieved with 30% or lower throttle angle. Notice the drastic increase in throttle angle near the top "full torque" line. This is due to the non linear relationships described above. You would think going from 30%-->90% throttle would be a big deal, but it actually doesn't make as much difference as you would expect.