Bain Racing Podcast brought up some interesting questions

First of the podcast was one of the best. Jake is a master at his craft and everytime I hear him speak, I realize how little I know about induction design.

Andre had made a comment that he would rather tune an NA engine with Speed density rather than an ITB engine (assuming with a throttle based load axis). I am not sure as why this would be the case. I keep hearing and seeing online that an ITB engine should be TPS and everything else is speed density. Every time I ask why, I always get, "its just the way it is".

I have not been able to find any reasonable explination as to what or why this is the case. Let me ask this question, Take a piston engine, No VVT, No boost, No variable runners or cam switching, If at WOT, 6000rpm, 99kpa MAP, my VE is X(pick any number you like), If everything stays the same besides MAP, say down to 80kpa, 100% TPS, 6000rpm, what is my new VE, X+?, X*?

While the VE is the same, the amount of fuel injected should be proportional to the change in the mass of air. With 19.2% less air, you would expect the fuel delivery to be 19.2 % less. if tuning alpha-N (TPS as the load axis), then a baro (or airbox pressure) sensor would typically be used to compensate for the mass of air, with a trim of -20% for 80 kpa (0% for 100 kpa).

So in Speed density terms, the fuel model/fuel trim/baro correction should account for all of the changes happening for MAP at a given RPM range, correct?

A buddy of mine in the US who is very accomplished with Motec platforms tunes NA ITBs using TPS as the main load axis with MAP as a fuel table comp. Throttle position represents load while MAP represents the density of air entering the cylinders.

Yes, the MAP changes represent density (along with air temperature), so the Speed Density tuning uses that as the load axis, and the fuel table could be the VE at that load / RPM (air temp is included in the air-mass calculation).

The reason that ITBs are tuned using the throttle position, is because:

1 - It's difficult to get a good manifold pressure signal since there are multiple separate manifolds - and just connecting them together and logged the pressure at high sample rates will show the pulses of each one. Some ECUs are capable of using only a single runner, and sampling at a specific crank position to get a steady MAP signal.

2- At low throttle positions, they are very sensitive to changing the air-mass. So the hard=to measure MAP is easy to measure with TPS.

On the ignition side you are even more challenged. What you want to use as one axis is cylinder fill. In the case of TPS for load this is not always a good choice since baro changes affect the fill.

Nigel, that technique is useful for Mx00 (hundred series) ECUs, but with the M1 we use the Multi-Throttle Tuning method, where the VE table is a single entry of 100%, and the Throttle Position is used to estimate a "Manifold Pressure Signal", relative to airbox pressure. This allows most other systems (like ignition Timing) to with a more accurate estimate of the actual Manifold Pressure.

I just finished tuning a 1969 Datsun 510 with ITBs using this method, and it can work very well.

David,

If the VE at WOT for a given RPM is mechanically fixed and the compensation for the air density is handeled by the ideal gas law, Baro and Air temp compensations, why are the MAP based fuel tables not just flat in the vertical axis?

Robert, How about some context for your question. Are you talking about ITB's, Single Throttle Manifolds, or something else?

Or is your generic question -- why is there a different volumetric efficiency at different manifold pressures and engine speeds? The answer to that is because engines with cam driven poppet valves and moving pistons have different efficiencies at different pressures and engine speeds due to air flow (velocity) and pressure gradients (pulsations changing the local pressure). How much air and fuel mass can be trapped by the closing valve at the intake stroke depends on the velocity and pressure over the intake stroke.

The VE does flatten out when pressure is above ambient in forced induction engines, (i.e. boost solves everything...)

David, Your comment that for a given RPM the VE is the same and for the change in MAP a change in required fueling is handled by the fuel model. If the target lambda table is also taken into consideration in the fuel model, Why is the VE not still the same? If the VE at 6000rpm at 99kpa is 102, why is it still not 102 at 25kpa?

This doesn't matter; from ITB to single throttle, common plenum, NA, turbo, supercharged, the engine fundamentals at a base level are all the same.

I'm sorry if I have confused you.

I am not confused. You keep talking about using TPS/AlphaN as the load axis when you have ITBs due to their lack of resolution in regards to MAP signal.

Why would you not still use the TPS as load in any other application?

In a single throttle NA engine how do you replicate lower than current atmosphere MAP? You reduce the throttle opening? If this is the case, why is the throttle not the load? MAP is scaled in the fuel model, what is the compensation for throttle position? Nothing? How does this make sense?

The throttle position isn't used as the lookup axis for efficiency because it doesn't define how well the cylinder can pump air.

The pressure before and after the cylinder, or the ratio of pressure across the cylinder defines how efficiently the cylinder can pump.

If you had a very large volume inlet manifold the throttle position can change but the pressure in the manifold wont respond as quickly.

If you had 100 kPa in the inlet manifold it doesn't matter (as much, pumping losses might change) what position the throttle is at.

Scott, Thank you for the reply. This is an answer to the question when it is looking for either TPS or MAP. When you can use both together, why is using both not a preferred method?

An engine’s VE doesn’t change much with only a change in MAP. The fueling requirements from the change in MAP are handled by the fuel trim.

Is there a reason you would not or could not use TPS as a main fuel table axis when MAP is already being scaled in a car with a single throttle/common plenum, turbo charged or supercharged?

I already explained, the cylinders ability to pump air is defined by the pressure ratio across the cylinder.

So if we're defining VE ideally we would look this up against this pressure ratio, however exhaust pressure sensors aren't particularly common and in NA applications we can assume this is constant for a given operating condition.

We can infer the operating conditions based on the throttle position but it's not what defines the VE.

Imagine you have a 1000 litre plenum with a 80mm throttle body, the throttle can change position without influencing the operating point until the engine consuming the air in the plenum causes the change (or the pressure rises due to the flow through the throttle). Obviously, this is an exaggeration but the principal still applies

I am not disputing any of that info. I asked two questions and didn’t get an answer to either.

When you can use both together, why is using both not a preferred method?

Is there a reason you would not or could not use TPS as a main fuel table axis when MAP is already being scaled in a car with a single throttle/common plenum, turbo charged or supercharged?

Anybody?

The answer depends on the application.

OEMs airflow models take lots of factors into account to try model a cylinder airmass.

In aftermarket tuning it's simplified because the tuner doesn't have a simulink model of the engine and 12 months to spend on an engine dyno.

If manifold pressure and throttle position can both influence the calculated air mass and that's necessary for the particular engine you're tuning then maybe that's the correct answer.

Throttle as the load axis on the efficiency table, as I already described, isn't necessarily going to represent the conditions that influence how well the cylinder can pump air, but in some cases it might.

It depends on what is available in the software and the airflow/mechanical characteristics of the engine.

If the ECU doesn't model airflow and it's just injection time based then you might go about it a different way.

"The answer depends on the application."

The applications that dont benefit from blending TPS and MAP are the ones withouth a throttle. Or where the throttle is barely used.

"OEMs airflow models take lots of factors into account to try model a cylinder airmass.

In aftermarket tuning it's simplified because the tuner doesn't have a simulink model of the engine and 12 months to spend on an engine dyno."

This is just the canned answer explaining why your tune will always end up "good enough". With a Speed density fuel table, part throttle tuning is barely accounted for. WOT is great, idle is good. Cruise is stoich, dynamically drive it and see the fuel trims at part throttle accel. Make some 15-35% throttle pulls and show me that the lambda control isnt doing all of the work.

"If manifold pressure and throttle position can both influence the calculated air mass and that's necessary for the particular engine you're tuning then maybe that's the correct answer."

It is always the correct answer. Because you choose to ignore it, doenst make it right only part of the time.

"Throttle as the load axis on the efficiency table, as I already described, isn't necessarily going to represent the conditions that influence how well the cylinder can pump air, but in some cases it might."

When you have already scaled for MAP, you are not just using TPS, you are using both... Saying the throttle has no influence on air mass is the crazy part here.

"It depends on what is available in the software and the airflow/mechanical characteristics of the engine.

If the ECU doesn't model airflow and it's just injection time based then you might go about it a different way."

Again, this is just a limit you have placed on yourself. This works on DOS based M48 MoTeC up to nearly all of the current standalones. Some of the better ECUs actually calculate throttle mass flow. It borderline easier to do than tuning a speed density map. Efficiently tuning the model requires a decent understanding of it and it takes time to hone that skill.

I can bet that you have never tried blending TPS and MAP. Which is ok, most people havent. This just leads to the same, "ITB is Alpha N and everything else is Speed density...".

Case study, engine is a 2.3l EcoBoost, Storker crank, Rods/Pistons, Cams/springs, Stock ports and valves. EFR6258. 100oct fuel, Retains GDI, DBW, Dual VVT, Dual knock. OEM 2.0l intake manifold and stock 60mm throttle body. Running on a LifeRacing F88 GDI4. Road Race sprint, 25-40min races. Power limited due to class limitations, right around 375ft/lbs + 285whp.

ECU is injection pulse width (injPW) based, with no VE calculation or background fuel model. Main fuel table load axis is TPS; the main ignition load axis is MAP. Only a simple MAP compensation table is set for 100% per 100 kPa. Engine was calibrated per throttle row and engine torque output was observed. Between 35-40% TPS, torque was 99% of peak at 100% throttle. This is visible in the fuel table's curve and confirmed with delta throttle pressure data.

Boost target is TPS based. Actual TPS is clamped at 75%. You can tune this engine Speed density with a linear throttle/pedal curve. Driveability will be terrible, especially since anything over 35-40% throttle is nonlinear. Utilizing basic tables to maximize driveability, especially in a road racing application allows the pedal to torque output to be very linear giving the driver a lot of usable control.

Admitted EMAP at peak stays between .9 and 1.1:1. This makes it so no additonal fueling changes are required on MAP alone.

How does this work? Single throttle, common plenum, turbocharged, running a TPS fuel map? impossible... No extra tables, no fuel model, no long drawn out tuning time, additional data gathered from these methods that allow me to easily maximize the setup for increased driveability.

sorry if this is the wrong thread but when I searched this came up as close.

4.2ltr stroked to 4.8

triple weber's acting as throttle bodies , fuel injection through m130 , running e85 , tps driven

having an issue when first application of throttle from say a roll around a corner normal aim of .96 drops to .76 and then climbs to 1.20 before levelling off

just assume you are not our first port of call and not our first rodeo so to speak but I am stuck

Graeme, If you move through the VE table sites with constant throttle, do you still have issues? or is it only in transient throttle application?

only off throttle then back on , so transient

i have enclose a pic of the log file where I tried to make a series of throttle applications to make it happen

You need to adjust the accel enrichment. How it feels and the lambda trace need to go hand in hand. In some cases making the lambda perfect, might not feel that great.