Pedal vs. Throttle Mapping - How Does This Work?

[Vlad Soare]

Hello,

Vehicle info:
2020 Mustang GT, Euro-spec, manual gearbox.
Log_veh_info.png

In the Engine/Torque Management/Driver Demand section you can configure the relationship between the accelerator pedal and the throttle opening in the various drive modes. Here's how the default configuration for the Sport+ and Track modes looks:

Pedal_settings_Sport.png

Until 20% it's linear (i.e. the throttle opens exactly as much as you push the pedal), after which the throttle starts opening a bit more. In Normal mode the relationship is completely linear from start to end. This makes perfect sense.

However, I scanned the accelerator position and the throttle position yesterday, while driving in Sport+, and what I'm seeing in the log looks very different to what I was expecting based on the above table.

Throttle_vs_pedal_graph_2.png
Throttle_vs_pedal_graph_2.png

The throttle position is red, the accelerator position is white.

As you can see, the throttle position always lags behind. The only situation where the throttle opens more than the pedal is when the pedal is at (or close to) 0%, and the throttle remains a bit open to keep the engine running. As long as there's some input from the accelerator, any input at all, the throttle will open less than that the accelerator.

What gives? Maybe I'm scanning the wrong channels? I'm using Accelerator Pedal Position and Throttle Position. Are these not the right ones?

Thank you.

[GT stang]

it works like the actual vs commanded pedal
for example the actual pedal position is 80% but the commanded output is 100%
maybe you can only see changes in the scanner if you log the pedal position voltage vs pedal position ratio

the throttle is controlled by the driver demand which refer to your final pedal position output

[Vlad Soare]

I see. So what I plotted above was the actual position of the pedal, not the commanded position (which is in fact what the ECU uses in order to determine how far to open the throttle).

There's another channel, called Accelerator Pedal Position (Filtered). I was wondering what that meant. I think that's the commanded position. If I plot that along with the non-filtered one, things start making sense.

Plot_1.pngPlot_2.pngPlot_3.png

Green is "filtered", white is not. The horizontal reference line is at 20%.

So, we can see that the "filtered" pedal position follows the base position almost exactly as long as it is below 20%, but climbs faster than that once it gets above 20%. Just like it is defined in the configuration table.
It all makes sense now. Thank you!

Now, I guess that the reason why the throttle position doesn't agree with the pedal position (either base or filtered) has nothing to do with this particular table, but most likely with the torque management, which depends on a plethora of other criteria. Right?

[engineermike]

The pedal translation simply scales the commanded by that amount before going to the torque demand table. 15% pedal could translate to 20% which means it might demand 300 ftlb rather than 250. Then the 300 ftlb is converted to a load in the torque-to-load table, load, converted to air flow, manifold pressure calculated (SD model), then the throttle body model is used to convert the desired airflow, throttle inlet pressure, and manifold pressure into a throttle angle to be commanded. Then theres the nested PID feedback loops used to correct the actual vs commanded angle and actual vs desired airflow.

The days of translating pedal directly to throttle angle are long gone.

[Vlad Soare]

Thanks, Mike. That's what I thought.

In the meantime I made a new scan of those two parameters, and everything looks as expected.

Scan_2.pngScan_1.pngScan_3.png

In Normal mode the actual and commanded are identical every time. In Sport+ and Track they're identical up to 20%, after which the commanded position rises faster than the actual. And in Rain&Snow it's the other way around, the commanded is lower than the actual unless you push the pedal really hard.
Everything makes sense now. Thank you!

[Aus_spec_ranger]

The pedal translation simply scales the commanded by that amount before going to the torque demand table. 15% pedal could translate to 20% which means it might demand 300 ftlb rather than 250. Then the 300 ftlb is converted to a load in the torque-to-load table, load, converted to air flow, manifold pressure calculated (SD model), then the throttle body model is used to convert the desired airflow, throttle inlet pressure, and manifold pressure into a throttle angle to be commanded. Then theres the nested PID feedback loops used to correct the actual vs commanded angle and actual vs desired airflow.

The days of translating pedal directly to throttle angle are long gone.

So how would u change the throttle body plate to close or open at certain pedal throttle % in this table if other tourqe management tables are jumping in the background trying to compensate for all the above u stated ? Sounds complicated and seems we don't have control over what torque maps do

[engineermike]

The whole logic system from the ground up isn't designed to make the throttle open and close as a function of strictly pedal position. It's much smarter than that, and for good reason. It's better to start with what it's doing that you don't like, figure out why, and fix that instead. If one were hell-bent on forcing the blade to follow the pedal, I think there is a pedal-follower mode you can enable and set up the curve, or play with WOT start and end values. But like I said, I would not recommend that. People's (myself included) first instinct is to try to make these modern systems work more like something we're familiar with (e.g. cable throttle) but you have to let that mentality go because the modern control really is better and it's not meant to work as a direct relationship.

[Aus_spec_ranger]

The whole logic system from the ground up isn't designed to make the throttle open and close as a function of strictly pedal position. It's much smarter than that, and for good reason. It's better to start with what it's doing that you don't like, figure out why, and fix that instead. If one were hell-bent on forcing the blade to follow the pedal, I think there is a pedal-follower mode you can enable and set up the curve, or play with WOT start and end values. But like I said, I would not recommend that. People's (myself included) first instinct is to try to make these modern systems work more like something we're familiar with (e.g. cable throttle) but you have to let that mentality go because the modern control really is better and it's not meant to work as a direct relationship.

Yeah I'm using this for a diesel engine so I would like to know how to close the throttle body plate...what's this pedal follower mode your talking about? I would like to close the plate %50 at 0% of pedal position

[engineermike]

Ecm 44365 and 44366

[Vlad Soare]

I think there is a pedal-follower mode you can enable and set up the curve, or play with WOT start and end values. But like I said, I would not recommend that. People's (myself included) first instinct is to try to make these modern systems work more like something we're familiar with (e.g. cable throttle) but you have to let that mentality go because the modern control really is better and it's not meant to work as a direct relationship.

I'm not sure it's necessarily better. Modern systems try their best to give you the torque you require, but only insofar as they can keep polar bears happy. If the computer thinks you're requesting too much torque for the current gear/speed combination, then it simply won't comply.
A cable-operated throttle, on the other hand, will open instantly no matter what. Once opened, the computer will notice an increase in the incoming air volume and will have to immediately inject more gas in order to keep the engine from stuttering. Polar bears be damned, you'll get the correct amount of fuel the instant you request it.

At least that's my felling after having driven both for many years. If I lugged the engine in an older car I would feel the engine struggling to meet my demand, and then a few seconds later it would succeed. If I lug the engine in a modern car, nothing happens, it feels like it simply ignores my input. I push the pedal to the floor and it behaves as if I had barely pressed it 15% or so.

I could be missing something, though.

Ecm 44365 and 44366

Table 44366 looks a bit strange to me. Why is the maximum throttle angle 45 degrees? Shouldn't it be 90?
Or maybe those aren't degrees?

Untitled.png

[Aus_spec_ranger]

Ecm 44365 and 44366

cheers but i dont think i have these maps in my ford ranger 3.2 Diesel HPT version......i looked thru all my maps one by one and cant find map 44365 or 44366

[engineermike]

@ vlad that table isn’t activated from the factory so the data shouldn’t be regarded as useful.

As far as a cable throttle vs dbw, the driver’s desire is actually torque and not airflow. Modulating airflow with the pedal when you really desire torque is actually an indirect relationship. What makes it worse is that now all other aspects like fueling and spark are playing catchup because the airflow happens first and everything else has to react. With dbw it can command everything to future state at once. As far as lugging the engine goes, if you desire more throttle at low rpm you can tune for that, but there is typically a reason the oem is limiting it, such as lspi.

[Vlad Soare]

Indeed, the airflow is an indirect command. What I really want when I put my foot down is torque. That's correct. But I want the maximum amount of torque that the engine is physically capable of at that particular rpm. With cable-operated systems, like you say, everything else has to react - "has" being the key word here. The ECU can't say, "well, he did open the throttle, but those poor polar bears aren't gonna be happy, so I'd better slow him down a bit". No, it has to react, otherwise the car would be undrivable. Whereas DBW systems can indeed command everything to future state at once, but they can also decide what that future state should be - which is not necessarily what I want it to be.

I'm not talking specifically about the Mustang. The Coyote has loads of torque, which renders this point kind of moot. I'm talking about DBW systems in general. I've seen this in cheaper, low-powered cars. My mother's previous car (a cheap, small, low-powered hatchback) had a cable throttle and drove just fine. The next generation of the same car, which she is driving now, same engine, same power, has DBW and feels like it's broken. You press the pedal and nothing happens. I hate it. On paper it may have the same power at 5500 rpm, but everywhere else it's awful. The cable was perfect. The cable had to work. The DBW doesn't have to unless it wants to.

[engineermike]

Vlad it sounds like you don’t have so much a problem with the dbw as you do the limiters. Just keep in mind even if they had stuck with dbw they could still apply limiters. And besides, releasing the available torque is what “we” do.

[Vlad Soare]

I'm sure they could have applied limiters even if they had stuck with cables, but maybe not to the same extent. If the air flow increases (and you can't do anything about it), then there's only so much you can limit before the engine starts stuttering. If the air keeps coming in, then the fuel and spark must follow, otherwise the car would become undrivable. Whereas DBW allows you to set the limits as low as you (or polar bears) wish.
At least that's how I imagine it. I could be wrong.

I think the ideal tune would be like this. If I press the pedal to the floor, then I want to get the absolute maximum torque that the engine is physically capable of at that particular rpm - emissions and anything else be damned. If I press the pedal to 45% of its travel, then I want to get 45% of the maximum torque the engine is physically capable of. And so on.
Is this possible? Can we tune a Coyote like this?
Come to think of it, even if we could, it may not be a good idea, because it would damage the cats/oxygen sensors/etc. in the long run and would perhaps also affect the engine's longevity. Some limits are there for a reason. Then there's the TC/ESC which also works by limiting the torque. But if we define T as "the maximum amount of torque that the engine can physically produce without causing physical damage to any component", then, provided there's no wheel slip to trigger the TC/ESC, I expect to get exactly T when I put my foot to the floor, or exactly 0.7 * T when I press the pedal to 70% of its total travel, and so on.

But then, assuming we could tune the engine like that, would I really like it? Or would the car be twitchy and unpredictable and a pain to drive?

[engineermike]

As far as I know, the Coyote doesn't have any torque limiters outside of the usual shift torque modulation, traction control, etc. WOT start and end make sure the throttle goes all-in. At part throttle it goes off of the driver-demand tables, which of course can be easily manipulated to get whatever pedal feel you like. I don't see any component protection torque or load limiters in there outside of the transmission. However, the 2024 does appear to have some LSPI load limiting below 2000 rpm. Ecoboost is a different story altogether, but easily modified by tuning.

[K44]

I understand exactly what Vlad is saying. Sometimes the DBW system over rides what I want. This is mostly noticed on a race track.

I would prefer a cable throttle, at least then I know what the car will do every time.