Mirrors special placement problem Car has two internal mirrors and two external mirrors: four mirrors total, but LFS support only three: two external and one internal, which have to be centered.

Using two separate "internal mirror" subobjects, mapped with different cutouts (doesn't matter cutout flipped or not) — mappings rotated along Z together with last subobject towards "eyes".


After removing left "internal mirror" subobject, mapping rotated, same as before, but in the opposite direction.


Using united "internal mirror" subobject (doesn't matter build mirrored or not, or using flipped opposite side cutout or not).


Changing internal mirrors subobject to "attachment" (two external mirrors already mapped) leads to all mirrors turn black (doesn't matter build mirrored or not, or using flipped opposite side cutout or not).

After removing mappings from external mirrors, internal works like this:


Сonsidering central seating position, default POV=90deg, vehicle width, looking straight you can't see external mirrors, it makes sense to stick with last scheme for a while.

Last edited by Egor K, Wed 16 Nov 2022, 19:40. Reason : couple minor fixes

I use to have agang exe 24 (2011) street MTB and want recreate it inside LFS.
Who knows maybe LFS someday will support bunnyhop or at least hop (like GTA:SA does).

Engine

I figured that way:
Crank arm length = 0.175m
Rider mass = 70kg
Hand power (pulling handlebar) = 30kg
therefore
Rider force on pedal = (70+30)kg*9.8=980N
Engine moment = 980N*0.175m=171.5N*m
Since model mass (with rider) exceeds actual mass (due to: motor 26 kg, battery 7kg, wheels 19kg) I decided to set 152% torque.
IRL 12.3kg bike plus 70kg rider = 82.3kg
LFS 55.0kg bike plus 70kg rider = 125.0kg (152%)

Final Drive

Bike has 25 tooth front and 11 tooth rear sprocket, so 0.44 gear ratio. Since LFS allows 0.50 minimum, compensate with final drive ratio of 0.88.

Dimensions

According to dimensions from agang 2011 catalog (46.2 MB) (page 63/80):

Mesh

Current progress:


At first I thought to do it quickly, but got a little bogged down in the details

Quote from Flame CZE :

Ah yes, my bad. The way is to select all triangles and then click “select connected points”, then hide those.

Thank you!

Quote from Benzinass :

How to load LOD2 from file? when I try to load with LOD2 selected the the main object just gets replaced.

1 Load *.obj with lod2 (it will replace everything and became main object)
2 Save that main object from Modeller (mesh editor)

3 Load whole Vehicle again
4 Place triangles from main object to any other than first layer (must be empty)
5 Add "new subobject"
6 Load saved subobject from (2) by subobject > load

7 "merge into main"
8 Go to main object and select all triangles from first layer (that's why in (4) you need clean first layer - former subobject moved to it and now it's easy to select only lod2 triangles)
9 "add to : lod2"
10 remove lod2 tris from lod1

No matter which lod I choose, Modeller always shows all the points. Thus, regardless of the active lod "A: all" selects literally everything

A couple of workflow tricks.

While working on external shape to see clearly how adjust shape put marks, such as windows, doors, headlights, etc. by projecting NURBS curves:

Later, these contours were used to build the final mesh:
1 draw nurbs curve
2 add Shrinkwrap modifier (wrap method: Project, select Axis, allow Negative direction as curve is outside)
3 choose curve resolution (object data properties > Resolution Preview)
4 convert curve to mesh (check keep original)
5 join (Ctlr+J) converted curve to the final mesh

Since back of air duct meant to fit smoothly into external shape, I applied Lattice modifier to its part.

Image source: https://cars.mclaren.com/us-en/legacy/mclaren-f1

Triangle mode shows tris only from selected lod. But point mode shows vertices from all three lods.

Spoiler - click to reveal

This not only confuses work, also changing one lod leads to changing another, for example: duplicate triangles from lod2 to lod3 > using "fuse to green" to reduce lod3 tris = points from lod3 and lod2 fused synchronously. How to make visible and selectable points only from current lod?

SOLVED by Flame CZE

Last edited by Egor K, Wed 19 Oct 2022, 17:05. Reason : Problem solved

Thank you guys! Some progress in the following sequence:
1 external shape
2 external shape mesh + SubD
3 semifinal mesh + Subd
4 check
5 after shrinkwrap (3) to mesh (1)
6 final mesh
7 final

Last edited by Egor K, Thu 12 Jan 2023, 13:20. Reason : rearranged images

Quote from Gutholz :

If I sneak a camera into a cinema and upload the recording then that is still illegal. Even if I make small edits like cropping out areas of the movie or mirroring it.

Quote from Scawen :

I agree with everything you said.

If you agree with everything that weisenheimer said while the shape of a car is obviously copyrighted, how can precise recreation of a real model be legal at all without permission from the copyright holder even if car named different, but you can instantly recognize what car it really is?

Quote from Scawen :

... Please visit the 20th Anniversary page to see the screenshots ...

Second screenshot from Fairfield Test Centre pack shows a straight of regular artificial bumps, it would be cool to have also random ones, like Belgian Pavé

Quote from Evolution_R :

LFS needs that!

Yes!

I could not even dream of such a program in the foreseeable future. It is a pity that mainly big V8 engine is demonstrated and the 747cc motor is a bit more bass than you'd expect.

Quote from Flame CZE :

something similar is in review

It is similar ONLY way it looks! It contains default frame (center of gravity and moment of inertia not relevant, at all), engine and fuel tank coordinates = 0,0,0 and so on we have Ambulance Truck lighter than Miata with kart-grade handling. I do not wish Mr. Dygear judge the accuracy of LFS mods by hususu9633 made mods.

As I see workflow:
1 make shape (no mouldings and gaps)
2 make sure general shape is ok
3 apply subdivision
4 make gaps and mouldings by offset edge slide
Gaps and mouldings in low-poly mesh are:
5 hindrance to shape correction
6 enemy of smooth shape transition (not visible on your screenshots because there is no matcap)
Am I wrong?

Mass

According to book "Driving Ambition" kerb weight (incl half tank): 1138kg. Distribution front/rear: 41.2/58.8%.

Engine

According to audiosignal.co.uk engine+all ancillaries, the exhaust and silencer is 266kg. Engine mass 303kg automatically determined by Vehicle Editor a bit heavy though.

Quote from Jason Barlow :

it hasn’t got a flywheel.

Plate claims idle 900 rpm and another engine settings, most of all I like ignition timing string. Whole video by James Cottingham (DK Engineering TV) worth watching.

Steering

• Maximum Lock — I figured maximum lock is 33 degrees while testing on Layout Square — according to Autocar magazine turning circle is 13 m.
  
• Steering Wheel Turn (ONE WAY) 504 degrees — according to Autocar magazine lock to lock 2.8 turns.
  
• Parallel steer — quote from C466/007/93 (Concept and design of the McLaren F1 suspension systems), see attachments to post by user flemke on 26-09-2019 (hereinafter the "C466/007/93"): "...around 40% of full ackerman was adopted for F1."
  

Suspension

• Camber — planned: 45' F&R, but user tortoise claimed: 1.75 F; 0.75 R.
  
• Inclination. Front — according to audiosignal.co.uk — 8 degrees. Rear I tentatively figured by freeze frame tracking. Whole video by Tim McNair (Grand Prix Concours) worth watching.
  
• Caster — according to audiosignal.co.uk — 4.6 degrees (46 is written — I think point just lost).
  
• Scrub radius — according to audiosignal.co.uk is 25 mm.
  
• No rear anti-roll bar. Here's what Gordon Murray said about the anti-roll bar at the front only (interview with Henry Catchpole on Carfection YouTube channel):
  Quote from Gordon Murray :

  When you look at the basic physics of roll compensation in a car you are trying to stop the roll in a corner. And you can, if you get frequency balance and percentage, you don't get too crazy on the percentage of roll compensation that's coming from the roll bar, you can do it all at one end. The reason why I never do it at the rear is you'll get all this trouble for traction. Particulary in a racing car. To get best bossible traction, best possible independent suspension ... and then you join left and right axles with the bar. It's nonsense. So on one wheel bump or coming out of a corner on roll you're unloading one side of the car. Artificially. Which is nonsense. If you want the best traction — you don't join the axles.
  
  
  Quote from Driving Ambition page 159 :

  Jonathan Palmer decided the car rolled too much and argued against Gordon's long aversion to using rear anti-roll bars. Ultimately a very thin and flexible one — which should have had minimal effect — was made up and fitted: "To Palmer's credit he came in after a lap, admitting "I see what you mean! It was virtually undriveable."
  
  
• Roll centre — according to C466/007/93: 3.1 Roll centre "...the centres be placed at 30mm and 60mm above the ground, front and rear respectively."
  
• Roll compensation — planned 75% roll compensation (4.4 degrees for 80 mm travel). Later in C466/007/93: "...55-65% roll compensation...". So let's set it to about 3.7 degrees for 80mm travel.
  
• Motion Range — according to audiosignal.co.uk: 170mm (90 bump + 80 rebound).
  
• Ground clearance — 120mm.
  

No anti-dive confirmed.

Quote from Jason Barlow :

more body roll than was fashionable in supercars

Quote from Tiff Needell :

rolled too much on the track

I like road car behavior with noticeable suspension travel, that real feel no other sim but LFS simulates so well.

Final drive

Lock is 40% — according to netcarshow.com.
Maximum speed with claimed gearbox ratios from Autocar magazine allegedly doesn't match well-known record 386.7 kph as redline was raised for record run, detailed information on Racing Line magazine.

Tyres

Tyre pressure information.

Aerodynamics

Maybe the coolest thing about this car, described by Gordon Murray (on Speedvision channel):

Quote from Gordon Murray :

did a little bit of styling and a lot of aerodynamic ... Styling was part of the process rather than dictating what the car should be

Production cars don't have front fog lights as early prototypes.

Quote from Driving Ambition page 164 :

At least five different nose undertrays were tried, some with and some without the distinctive spot lights. ... These changes were found to have dramatic effect — actually reducing drag from a Cd of 0.34 to 0.32.

Mesh

Mainly using this main reference sketch, which has accurate object positions.
Some progress (GIF)




Sizes correspond to manual.

Rims modelled by converting NURBS spoke profile to mesh with shrinkwrap modifier.

Rear lights: red COBO 02.279, amber COBO 02.282.
Article in Racecar Engineering (June 1996) has very detailed monocoque description, including center beams dimensions: height 250mm (match main reference sketch) and width 150mm on floor, as well as distance between them — 22 inches.

Vehicle Editor does not allow place two kinda-rear passengers due well-known unique three-seat layout.
Steering wheel was modelled using pic from book "Driving Ambition" which I redrawn to better mesh align.

Clocks
Interesting statement about about the first implemented feature:

Quote from Gordon Murray :

Ron Dennis suggested the needles swept the dials on start-up. This is now mimicked by many others.

Thanks to Jay Leno we can see McLaren F1 fuel cell and know it must be replaced proximly every five years (same tells whichcar.com.au).

People working on IRL

Barry Lett had arrived from Lotus to oversee body engineering.
Paul Martin was in charge of composites.
Steve Randle did stress analysis and ended up working on the car’s suspension.
(source)
Peter Stevens — stylist

Last edited by Egor K, Wed 16 Oct 2024, 16:30. Reason : add some info (engine and suspension, also quotes), add ARB info, deleted Bellm's complaint — it concerns GTR, added quote about rear ARB, 241016:removed questions, will ask in the Assistance topic

Quote from pärtan :

ability to move the bearing objects in 2D (x and z) directly from the suspension diagram view

A good idea.
It would also be very useful to see in real time: coordinates of roll center and its offset from COG.

Trying to understand units of measure in dashboard texture.
It's 512*512px centered, and X and Y coordinates of speedo and tacho are unitless fractions of half (256px) of texture (see attachment).
One unit of dimension of multi function display is 1/64 of whole texture width 512/64=8px

But radius of speedo and tacho does not correspond to fractions nor pixels.

Quote from terga :

...increase the upper caster limit ... only 6 degrees. It's also less than some real cars are capable of.

I vote for it, even regular everyday normal e34 has caster almost of 8 degrees

Snoop serious as always!

I like the fact that stock version comes first. Despite stock wheels sinking in arches (which is real).

Quote from Amynue :

I have an idea how to make different rims for the front and the back, but not sure if it will work.

Can we know-how?

Trying to understand why there is custom access to mods since there is no way to monetize?

I don't complain about custom access at all. Of course, the right of the author to decide who to give access to the mod. Sometimes I ask author online for access, but quite often they ignore it — can't blame them, cause what's the point of custom access when you're ready to give access to random dude online?

But why do I even see here on the forum when it's private? It makes it some kind half-private: see, but don't touch?
Also, why is custom access mods moderation is public?
Isn't this one of the reasons for delay of "public mods" reviews?
How about at least making viewing public mods a priority over custom access mods? Or even make it true private.

Is it possible to set distance of spokes from edge of rim (or from center) separately for front and rear?
When editing either rim or spokes, I did not find buttons to switch axis.