Reveal all secret hardcoded friction and braking coefficient tables

According to the SDK: Several empirical/experimental tables are set in hard in
the code to provide friction and braking coefficients Cside friction, Crolling
resistance and Cbraking resistance depending on:

  • the type of contact (eg: wheels, skids, scrape points, etc…)
  • the type of surface (eg: concrete, snow, grass, etc…)
  • the surface condition (eg: normal, wet, icy, etc…)

Please reveal the values contained in these tables and allow developers to
change them. Also, provide information on the source of the values currently
being used.

yes please.

I’m not sure if I should start a new topic or just add to this one, but here
goes… It seems that more information has been added to the SDK regarding
ground friction modeling, including the addition of friction coefficients used
for “static” and “dynamic” friction at low and high speeds on dry and wet
runways. However, this has raised some additional questions needed for me to
better understand this model. It is stated that the coefficients are expressed
as in the coulomb model. That’s all well and good, but the coefficients
provided are for 2 surfaces sliding against each other, not that of a rubber
tire rolling over a surface. The equation provided does have a “less than or
equal to” sign for the friction force determined from the friction
coefficients, but there is no explanation given for how the “less than” force
is computed. There is also a statement that using this model, “weight does not
play a significant role on tarmac.” I may not be understanding what you are
trying to say here, but weight (or more accurately, the normal force) plays a
very significant role on the friction force generated between the tires
and the tarmac (or any other surface). You even have it correctly represented
in your friction force equation. Why do you think jet transports have very
effective lift dumping systems (ground spoilers) in order to get more normal
force applied as quickly as possible? It is not the drag produced by these
devices that is most important in achieving shorter stopping distances. It is
the reduction in lift, and hence, increased normal force that results in
higher braking friction force being generated. For modeling of an airplane
moving over a surface on the ground, one is not normally concerned about
“static” and “dynamic” friction, but rather “rolling” and “braking” friction.
Static and dynamic friction may come into play when the vehicle is stopped
with the brakes on, preventing the wheels from turning. In that case, you have
to consider a surface (rubber) resisting slippage against another surface
(tarmac). There is an equation for rolling friction given, but it is a
summation of the rolling and braking resistance coefficients multiplied by the
strut force. Perhaps there is more (hopefully a lot more) to the ground
friction force model than has been provided in the SDK? For a rubber tire
rolling on a surface, there is very little resistance. This is what is
referred to in the aviation industry as “rolling friction.” For example, for a
transport category airplane, the rolling coefficient of friction is on the
order of 0.0165. For non-skidding braking on a dry surface (e.g., an airplane
with an effective anti-skid system), the maximum braking friction does not
vary much with speed and is on the order of 0.40. For a wet or snow covered
surface, the maximum braking friction does vary greatly with speed and is (
very roughly ) about 50% and 25% of the dry braking friction,
respectively. On ice, the maximum braking friction is about 0.05 to 0.08,
depending on speed. Thus, a non-skidding braked wheel on a dry runway cannot
achieve the friction coefficients in your table even though skidding
drastically reduces the effective braking coefficient. Using your current
model makes it difficult to accurately model both the taxi movement situation
(rolling friction) and the landing or rejected takeoff braked distances.

Hello Asobo, Could you please take into account the above post from @donstim?
This is actually a big problem with the 747 as it impacts the take-off ground
roll and thus makes it impossible to have realistic performance.

Actually, it impacts the 310R, Beaver, Otter, Skyraider and all of the others
we’re working on…

Hi Asobo team. Can you please provide any update on this? Not expecting an
immediate fix but would be good to know if this is being worked on. Thanks
Shervin

Hello. Here’s some additional info, partly answering
@donstim interrogations. The meaning intended by
the sentence “weight does not play a significant role on tarmac” is that a 10
tons aircraft doesn’t have 10 times the stopping distance of a 1 ton aircraft.
Braking distance is not directly proportional to aircraft mass. An heavier
aircraft has more normal force but also more kinetic energy to dissipate so
there’s a kind of balance between the two. Of course we have rolling and
braking friction in the simulation. Rolling friction has a base value of 0.02
and is modulated based on speed, wheel size and wetness Braking friction is
obtained from the input and toe_brakes_scale parameter and is also
modulated based on wetness. @boufogre What
precisely do you think is the problem with the 747 for example? The rolling
friction would be too high? Regards, Sylvain

1. The statement about the weight affect is in a section discussing friction
force, not stopping distance. It is important to be clear in the SDK. We are
not mind readers. 2. How exactly are rolling and braking friction “modulated”
by the variables you identified? In my experience with a major transport
airplane manufacturer, longitudinal rolling friction was never adjusted for
runway wetness. And we did not add rolling and braking friction to determine a
total braking frictional force (which the SDK calls a rolling friction force).
Rolling friction only applies when the wheel is free rolling. If you are
braking, then the corresponding friction is simply braking friction. But I
guess that is your convention that we would need to take into account if/when
we get the ability to include our own rolling/braking friction values. 3. How
is the braking friction determined from the input and the toe_brakes_scale
parameter? Is it a “base” braking friction value determined in some way
multiplied by the toe_brakes_scale parameter and the percentage of braking
input applied? 4. I won’t answer for @boufogre, but we found the rolling
friction too high on the Asobo A320NEO (at least at the inception of MSFS; I
don’t know if it has been changed since then). Regards, Don

I meant to also add that rolling friction seemed to be substantially increased
when the runway is wet compared to dry, at least the last time I checked. Why
is this? That might be true for standing water, but not for just a wet
surface.

@FlyingRaccoon, Indeed, rolling friction is way too high. Wheel radii are set
to the real correct dimensions. With an empty aircraft, when rolling, the
aircraft speed is decaying when it should be able to taxi on idle thrust. Just
tested: taxi up to 20kts, thrust to idle, the aircraft immediately looses 1 kt
and speed keeps reducing afterwards until aircraft stops. This also greatly
affects the take-off performance. According to the FCOM, it should be able to
take-off from LEMD 36L at MTOW, but I can’t even get to lift off the runway
because the speed is too low. And the runway is 4179m long. According to a
747-400 pilot (less powerful engines), the aircraft accelerates at idle thrust
when empty or near. The aircraft
FCTM states: “Idle thrust is adequate for taxiing under most conditions. A
slightly higher thrust setting is required to begin taxiing. Allow time for
airplane response before increasing thrust further.”
This is not possible at
the moment, and I’m pretty confident that my thrust is correct.

There is also the problem of steering on the ground. As soon as rudder is
applied, the friction and tangent forces shoot up to 70%, acting as a massive
brake and causing the bug where you need +50% N1 to keep going in tight turns.
Sometimes, the value will decrease under 70% after a while, but when releasing
to rudder to neutral, it will shoot back up to 70% before decreasing again, in
sync with the animation of the nose wheel if I’m correct. I think this problem
has been experienced by many users.

On the above screenshot, you can also see something that I don’t know if
related, but I couldn’t find a way to fix, but the wheels seem to be under the
ground surface (-20cm). Changing contact points vertical coordinates will only
move the 3D model up or down, but won’t change the wheels height below the
ground.

1. @Nocturne will check this part of the doc and
make sure we are more clear. 2. Rolling and braking friction are not added in
our model either. Sorry if my explanations were not clear. We take the maximum
of the two. 3. Basically base_coefficient * toe_brakes_scale * braking_input,
with the base coefficient being given in those tables: Flight Model Physics
(flightsimulator.com)

We plan to review this part of the code later that year and exposing some
coefficients in cfg files is considered.

Thanks. Regarding #2, sorry it wasn’t something you explained. Rather, it was
this equation in the SDK that has me confused. Although it’s called “Rolling
friction,” it appears to add both the rolling and the braking friction.

Regarding #3, airplane
manufacturers would love to have such braking coefficients! On an integrated
basis, I’m used to seeing braking coefficients in the range of 0.4 to 0.5 on
dry runways and about half that on wet runways. I guess we use the
"toe_brakes_scale parameter to get realistic stopping distances until such
time as we can edit the rolling and braking coefficients.

Hi Sylvain. Yes that is correct. Rolling friction is too high. Even on our
current project, as soon as I land, without using brakes, spoilers or
reversers, the plane starts decelerating quickly.

“review this part of the code later that year” When do you plan to review this
part of the code? In this entry, write like this:
<https://devsupport.flightsimulator.com/questions/12151/setting-tire-pressure-
to-affect-ground-static-fric.html> “We have more long-term plans to improve
the ground handling.” When might it be in the future? Finally, I have a
question. Maybe I should start a new thread about this but I think this is
somewhat related. What mostly affects the speed to drop sharply when making a
sharp turn, and how can I adjust it? You want there to be a sense of inertia
in the plane, which is a little lacking.

It isn’t just ground friction that’s a problem: all floatplanes and flying
boats behave as if they are on ice. Currently we (partially) get round the
problem by throwing out non-existant spoilers to provide a friction effect,
but unless the on-water friction problem is solved that one is going to come
back to bite us one day.