FWD cars steering angle too small
(22 posts, started , go to first unread)
FWD cars steering angle too small
I tried replicating real-life driving school excercises in LFS, and there are couple of issues:

* in real life you can gently leave the clutch over 5 seconds, and get the car rolling and drive without touching the throttle. I replicated the gear ratios (taking into account the top power revs and wheel size), and still could did not make it in LFS.

* XFG steering angle is lower that in comparable cars. They feel much less manueverable than IRL or RWD cars. My calculations show it should be 35-36 degrees (more below).

I failed to reproduce the backpedal-into-garage excercise. Here's the area where we excercised and how the car should go (backwards):



My measurements in QGIS.


I suppose, the dimensions are round, so the area is 5 by 8 m, and the box is 5.5 by 3.

I reproduced it in LFS layout (see in the replays below), and tested it with a 1080-deg wheel and 3-pedal set. The car won't fit in the "garage" if you drive normally. It only fits if I put it forward as much as possible, and turn the wheel to the maximum while still standing. Then XFG does get into the box. In real life you get it going and then turn the wheel to the limit, which takes a couple of seconds.

Replays: UF1, XFG, XFG with mouse for max steering, XRG.

It seems to me the angle of 30° is too low. My calculations show it must have 35..36°.

In this Stackoverflow thread, a simple formula is proposed:

steering angle = atan(wheelbase / (turning circle radius - car width)) (I suppose it's track width.)

Renault Logan 2005 that I drive:
* turning circle diameter 10500 (radius 5250)
* track width: 1480
* body width: 1740
* wheelbase: 2630

Simple Python code:

In [1]: from math import atan, degrees
In [2]: degrees(atan(2630 / (5250 - 1480)))
Out[2]: 34.900118296332984
In [3]: degrees(atan(2630 / (5250 - 1740)))
Out[3]: 36.843770981210255

So the angle is between 35 and 37 degrees.

Why I bother telling this? Driving forward and changing gears is simple. Backpedaling + steering is hard, because you need to watch many things at once and have your feet act independently from hands. So any extra practice in LFS is very valuable. I can dispense with current cars and just expand the area in LFS a bit and make it work, but it would be nice to have things a bit more realistic.
Thanks for the real world comparison.

For anyone reading, the front wheel drive cars in LFS have more limited steering angle than the rear wheel drive cars. This is because in real life there is a drive shaft and the CV joint can only turn a limited amount.

I'm sure the limits in LFS can be improved. I would like to have some more examples of front wheel car steering angles, with reference links. So that in a future version we could make some changes.

Either:

1) Actual turn angle of the front wheels.

Or:

2) Wheelbase of the car and turning circle (and car width if available)
Are you using a comparable setup to a road going car? In LFS I used:

open differential
parallel steer 0%
+.1 rear toe in
no rear anti-roll bar

and could reverse in normally, as long as I kept the speed low enough.
Quote from Scawen :Either:
1) Actual turn angle of the front wheels.
Or:
2) Wheelbase of the car and turning circle (and car width if available)

Tomorrow, I'll take a tape measure, plus angle ruler, and check both turn radius/diameter, and wheels angle.

As far as the area is concerned, today I took a tape ruler and measured the area dimensions. Apparently, I took the fresh asphalt for the entire area, and made it much narrower.

Here are the real dimensions and them in QGIS: (edit: probably, the box size should be ~520, not 500)




And it was no problem to backpedal into the box in LFS as well. My apologies. But I'll check the actual steering angle on the Logan.
Quote from bobloblaw :Are you using a comparable setup to a road going car? In LFS I used:

open differential
parallel steer 0%
+.1 rear toe in
no rear anti-roll bar

and could reverse in normally, as long as I kept the speed low enough.

I had open differential, parallel steer 0%, 0 toe-in both on front and rear wheels. Maybe the antirollbar was the cause, I'll check as well.
I took an angle ruler to the autodrome today, and measured the angles. I checked carefully that I matched the plane of the wheel and of the side wall vertically. The angles seem to be around 40 degrees.

The first car is Skoda Octavia 2, it's angle is about 40 degrees (steered to the lock, I checked).

The Logan seems to have narrower angle, on the photo is below. Seems it's 40 degrees too.

If I did make a mistake in measurements (misaligned), the angles should be at least 36 deg.


Surprisingly, the Octavia that looks like a seriuos and large car is actually just 20 cm longer than the simple-n-sturdy Logan, and has a shorter wheelbase.
Checked Kia Soul (2013 model).

The protractor is misaligned with the camera, but I did check it in InkScape: it's 40°.

Measured another car, Lada Granta 2018. It's similar to Logan. The angle seems to be 38° both by eye sight with protractor, and with InkScape line rotation.

detail : is measuring the inner wheel giving the same result as the outside one ? If I understand correctly, with the ackerman effect, the inside wheel is having a greater angle than the one I would suppose to be to consider when speaking about the "steering angle".
I did a test with the Peugeot 308 (2.0 hdi) of my wife and was measuring ~25° (not very accurate I have to admit).
I did some computations for the cars I have under my hands with a basic formula on the Radius and the wheelbase : arcsin( R / wheelbase ) to give approximatively the outside wheel angle :

I found the turning circle over internet, not sure they are right, but driving speaking : indeed the 308 has the poorer turning circle compare to my two other cars
There is a slightly greater value for the rwd car as expected (I would have suppose even more ...), but the result for the 308 seems to correspond to what I am actually measuring.
Attached images
test_steering_angle.PNG
Flotch: I did not measure the outer wheel angle, but I can check that. Indeed the outer wheel should turn less than the inner one. Then the question is what the setting in LFS relates to.
#12 - HX7
Measured the turning circle diameter of XFG and FXO.
XFG:11.38m
FXO:12.88m
Seems a little large in FWD cars...Uhmm
Attached images
turning circle diameter.jpg
I completely agree with you guys.

Let's start a movement: "Give 36 degrees to FWD cars!"
If you take a modern civic Type R it is given for 12.56m ... slightly less than the FXO ...
With my "computation" (is it correct ?) I would find it having a steering angle of 25,45° (outside wheel).
I do not know the wheelbase of the FXO, but with HX7 measures, this would correspond to Wb = (Turning circle / 2) * sin(angle). In our case FXO is having 30°, so sin(30°) = 0.5 => Wb = (Turning circle / 4)
So, 3.22m ... seems too huge ...
If I take the computation proposed by detail in :
Quote from detail :...
In this Stackoverflow thread, a simple formula is proposed:

steering angle = atan(wheelbase / (turning circle radius - car width)) (I suppose it's track width.)
...

Quote :For the outer wheel don't subtract the width.

This means : steering angle of outer wheel = atan(wheelbase / (turning circle radius)
=> this results in lower values for my previous computations with sinus (but the measure for the 308 seems more in line with what I have seen...), and then the wheelbase of the FXO would be even greater ???
Houston we have a problem ...
Attached images
Capture.PNG
Here are some figures that should help and explain how to calculate the turning radius, for the correct "Ackerman type" of the steering rack. Note the abbreviations: I-inner, O-outer, F-front, R-rear. We need a little bit of clarification from Scawen for what exactly the parallel steer option means in car setup. I'm guessing that if you set 0% that would correspond to a full Ackerman steering and going towards a higher % makes it less Ackerman until wheels are completely parallel.
Attached images
Fig 2 - equations.jpg
Fig 1 - car_scheme.jpg
Stackoverflow implies that turning cirlce is measured at the rear outer wheel, Flotch's "asin" implies front outer wheel, HX7's implies "wall to wall", meaning it includes the front bumper overhang.

Those are 3 different values, and rear outer wheel makes no sense tbh, so i'd through overflow's function out.
The "curb to curb" front outer wheel circle + asin function will yield correct values for outer wheel. Please recheck me on this statement though Smile

As far as measuring inner wheel, I think this is the formula:
atan(WB/(sqrt(r^2 - WB^2)-Track))

Measuring FXO at track = 1.6, r=6.2 (took HX7's value and subtracked a bit), WB = 2.65 I've got:
outer: 25.3
inner: 33.5

I assume that "maximum lock" in car setups is for average value between both wheels.

Once again, please recheck my calculations and conclusion.
Checked Renault Logan inside and outside angles.
The protractor is again slightly misaligned, opposite of the initial measurement, now it reduces the angle, and I got around ~36° inside (I saw 40° the previous time) and ~30° on the outside.



Inside (left front)


Outside (right front)
Quote from detail :Checked Renault Logan inside and outside angles.
The protractor is again slightly misaligned, opposite of the initial measurement, now it reduces the angle, and I got around ~36° inside (I saw 40° the previous time) and ~30° on the outside.



Inside (left front)


Outside (right front)

This is not accurate at all. The best way to try and use this method would be car jacked up and wheel off TBH. Doing this is an at best guess and hardly accurate.
Quote from booplesnoot :This is not accurate at all. The best way to try and use this method would be car jacked up and wheel off TBH. Doing this is an at best guess and hardly accurate.

I know, but that's all I have, and this is driving school car. Anyone having a jack is welcome to do that.
Another observation on realism: I have to do U-turn maneuvers in limited space on road. Either I do it in 3 steps (left turn upto left kerb, backpedal right to kerb, forward left and leave the place), or I do it using the junctions with service driveways:


I checked if Blackwood is a good place to try these maneuvers, but the roads there are just very large.



Meanwhile, most European roads/streets have lanes up to 3.5 metres.

European Roads Safety Observatory 2018 document:
Quote :Lane widthshould be examined in relation to the expected operational speedas well as the percentage of heavy vehicles. Very narrow lanes cause problemsby not providing adequate lateral clearance, especially as far as heavy vehicles are concerned. In general, increasing lane width leadsto improved road safety, but very wide lanes may result toexcessivespeeds and encourage unsafe overtaking manoeuvres. A lane width of 3,50m is considered typical in most countries and research indicates that accident rates seem to reduce for lane widths of 3,50m to 3,70m (ERSF,1996;Proctor et al., 2015;Pilot4Safety 2012).

The narrowest one I found in Blackwood is 4 m. In South City, lanes are just at the maximum allowed, 3.75 m.
Quote from detail :le quote

I guess they are larger because its an Industrial Estate, and its meant for trucks to drive on.
Quote from Ibtasim6781 :I guess they are larger because its an Industrial Estate, and its meant for trucks to drive on.

Hmm, I thought of that, but had not bothered to check. Now I looked at Southhampton port in UK, and indeed, lanes are up to 5.7 m to make life easier for trucks.

FWD cars steering angle too small
(22 posts, started )
FGED GREDG RDFGDR GSFDG