Cars Physics reworked for the two Maserati 250F 2019-09-13

[alekabul]

@alekabul Correction the tires are now Dunlop L15 550 (204 compound) on the front and L15 650 on the back. See here: https://www.longstonetyres.co.uk/vintage-tyres/dunlop-racing.html

I will make a first draft of changes to the V6 referencing your remarks and will give you a link in PM. When we agree on the final changes then I will release an update here.
Thanks for the care!

 

[Fat-Alfie]

So, no new data for the tyres, then?

 

[Kyuubeey]

You're not gonna find any data for tires from this era.

 

[binijh]

You're not gonna find any data for tires from this era.

@Fat-Alfie @Kyuubeey Yeah not much to do there do. The folks over at Blockley Tires reckon they have the closest to the period ones as they tried to recreate the molds. Have no idea of course, but something to look into: https://www.blockleytyre.com/

 

[alekabul]

Reference lap times:
- 2.05 in the dry at Nurburging GP (without the Schumacher curves)
- ~2.30 at Silverstone GP w/ the motorcycle chicane
- ~2.55 at Spa, no configuration change

I intend to use Spa for the tests, but tell me first - do you find the traction on the Assetto's dry track accurate itself.

I mean not only from driving the 250F's but comparing other cars from real and AC too. If needed I can make myself a proper Spa for the tests. That's why I asked for a cornering speeds breakdown. Or a peak readings from a G-meter in a clean lap.

 

[GTV]

Can someone please send or upload the original files?
I have stupidly overwritten the originals.

Thanks in advance!

 

[alekabul]

250F orig - Google Drive

drive.google.com

The data for the two original cars. Better option - use Steam to check and re-hash the game install. That would revert any changes of original files and place new copy of the missing ones.

 

[GTV]

Thank you! Steam check would overwrite many files which have changed overtime .
To prevent that, I should have secured everything. So this is the easier way.

 

[alekabul]

As always I will advise anyone who likes to modify the original cars - at least rename the "data.acd" to "data.acd1" before doing alterations. That way it's easy to have the car ready for online racing just by reverting to the original name of the small file. No use of wasting storage space for complete backups.

 

[AadHofman]

Fun fact: With the original Kunos cars, minimize positive camber and maximise toe-in in the setup screen, and your wobble is gone.

 

[alekabul]

Fun fact:

A man who drives one disagrees with you. And even my efforts to make the two cars more stable were not to the point with him. I know your fun fact and yes, it can help the 6 cyl to get rid of the pendulum but not on the 12 cyl.

Anyhow, if someone does feel the added cars are unnecessary - simply remove them and forget about it.

 

[AadHofman]

Nah, I have nothing against your effort. Just an experiment I did coming back to AC. I remembered these cars being wobbly, and was mildly surprised how much it can be corrected with basic setup tuning, even the 12 cyl (adjusting the diff and ARB helps).

 

[AadHofman]

I should add that with the 12 cyl, don't maximize toe in for the right front, as it goes beyond the left front value. Keep it the same, otherwise you are back in the wobbly land.

 

[AadHofman]

Am I right in assuming that because the suspension.ini has Link Count of 4, it disrecards the values for the 5th Link (Parnard)? Then the rear roll center is all over the place, causing the front end to wobble.

At least when I tested it with the 5th link, front end didn't wobble with the default setup, and I was significantly faster in a couple of tracks I tested.

@Kyuubeey?

 

[Kyuubeey]

Am I right in assuming that because the suspension.ini has Link Count of 4, it disrecards the values for the 5th Link (Parnard)? Then the rear roll center is all over the place, causing the front end to wobble.

At least when I tested it with the 5th link, front end didn't wobble with the default setup, and I was significantly faster in a couple of tracks I tested.

@Kyuubeey?

Yeah, probably.

Although be warned, 5link will bind up pretty badly due to the too stiff links in AC. They are effectively not that stiff IRL, even solid links.

Build a 4-link model where the more straight of the arms is combined into a single arm. It will preserve the kinematics if the arm is completely straight when viewed from above, but not bind up in deflection.

If both UCA and LCAs are angled, combine the less angled ones into a single arm and perhaps angle the opposite arms just a tiny bit more. Proper way would be to graph the kinematics of course to maintain the RC.

Watch out because with angled LCAs and a panhard bar, there are two geometric centers and two kinematic centers. Even if the arms are not angled, it will still bind up with 5link in all configs I tested. So use 4link.

 

[AadHofman]

Thanks for the info! Do you mean a 4-link model turned into a 3-link?

 

[Kyuubeey]

Thanks for the info! Do you mean a 4-link model turned into a 3-link?

4 links + a panhard is 5 links. A 4link model, with 3links + a panhard.

 

[AadHofman]

Gotcha!

 

[alekabul]

Am I right in assuming that because the suspension.ini has Link Count of 4, it disrecards the values for the 5th Link (Parnard)? Then the rear roll center is all over the place, causing the front end to wobble.

At least when I tested it with the 5th link, front end didn't wobble with the default setup, and I was significantly faster in a couple of tracks I tested.

@Kyuubeey?

I too was wondering if that "LINK_COUNT=4" was a typo but since the 250F has DeDion rear there was no Panhard and the stability gain with it enabled as experiment comes at the cost of ruined FFB.

More interestingly, the front DWB is way different than what the exposed suspension links suggest. Kunos made a choice to lie about the pivot points in order to get a nice FFB without judder at very low speed.

If the pivots are correct the scrub becomes big enough to deal with the stability issues

 

[AadHofman]

I too was wondering if that "LINK_COUNT=4" was a typo but since the 250F has DeDion rear there was no Panhard and the stability gain with it enabled as experiment comes at the cost of ruined FFB.

Are you saying that De Dion tube (and the transverse leaf spring) is better modelled without a transverse link? Doesn't make sense to me. Also the FFB was not affected at all with the 5th link -- or if it was, it was for the better with a better handling car.