Given the limited setup options, hardcoding the curve looked like a viable solution to me. Certainly far from being the cleanest, most elegant one - but it would still make the VWS closer to the real thing without the need for more changes.

Quote from Bob Smith :

how do you decide what shape of curve to hard-code?

So if you go with a constant stiffness spring, how do you choose its value?

BTW if the VAG guy didn't hand out spring information (shame on them), I'd take a look at the Golf GTI springs, the car has been out for some time, shares a lot of components with the Scirocco, and thus is reasonably similar. If you get ahold of a couple springs, it shouldn't be too hard to extract a realistic looking curve.

For choosing a value, you can pick a value in the range of spring deflection that is most often used given the driving conditions you expect the vehicle to be put through. Seeing as spring rates haven't been given, it's also easier to pick a value in the range and still be correct, then estimating the entire range. Not that modelling the progressive springs wouldn't be the more accurate way to go about things, but as usual, the more complex your modelling, the more reliant you become on accurate data.

Finding out the real spring rates is only so useful if you know the motion ratios that go with them. I've managed to find stock spring rate values before, but never motion ratios. For McPherson struts you know the motion ratio is going to be very close to 1.0 due to the nature of the design, but if that's not the case, the numbers aren't very helpful.

I think I understand where you coming from - i.e. you say what matters the most is the amount of the error. If this is small enough you won't notice it because of flaws in LFS or limitations of the controllers masking it, right?

I still have some questions if you don't mind - I find all of this very interesting.

Quote from Bob Smith :

For choosing a value, you can pick a value in the range of spring deflection that is most often used given the driving conditions you expect the vehicle to be put through.

So you would use an iterative process, using what you think is a good guess or the softest value and then progressively increasing it through testing on various LFS track configs?

Quote from Bob Smith :

Finding out the real spring rates is only so useful if you know the motion ratios that go with them.

But even if you don't have the ratios you could still extract the springs curve and use it adimensionally, am I wrong?

And motion ratios could be estimated through a geometric study of the suspensions?

That would be easier if they had CAD drawings of the suspensions and not just the body - but there's a lot of stuff available online and offline the thing being basically a refactored GTI.

Assuming the devs can't come to a good estimation - if the motion ratios are 'made up' then it makes no difference whether they choose to go model progressive or stick to constant stiffness, the springs would still be 'made up' as well

If you knew the spring rate curve, you could find the normal spring deflection during hard cornering, and pick a value close to that point, if you are to model as a constant rate spring. That way body roll druing cornering is as accurate as possible, with the downside of too stiff springs over light bumps. For a racing simulator, it's more important to get things like body roll correct. Of course, anti-roll bars are an ever more of an unknown.

I do agree that, if you have the real curve for the springs, it can just be scaled. Good luck finding such a curve though. I also agree that, with good diagrams or access to a real car with a tape measure, a very reasonable estimate of the motion ratio could be obtained.

Quote from Bob Smith :

Good luck finding such a curve though.

I read there are instruments to find out the spring stiffness, for a curve you'd need something quite sophisticated I suppose.

I wonder how the guys at iRacing handle this they have stated sth along the lines of testing the cars in their labs IIRC.

Quote from Bob Smith :

I also agree that, with good diagrams or access to a real car with a tape measure, a very reasonable estimate of the motion ratio could be obtained.

So VAG just gave em a CAD drawing of the body? They weren't interested at all in the handling then.

Was it the same with the other RL cars in LFS?

The main issue with spring stiffnesses is not measuring the curve when you've got the spring in front of you, it's anyone else (including manufacturers) publishing the data. Any manufacturer who knows what they're doing must know what this curve is going to look like before the build the spring in the first place, as surely that's all part of the design.

I don't know exactly what data the devs have been given for each real car; I'd imagine it varies quite a bit between cars. I do know that for the RaceAbout, real spring stiffnesses were supplied but that for the Scirocco, they were not.

Quote from Bob Smith :

Any manufacturer who knows what they're doing must know what this curve is going to look like before the build the spring in the first place

That's understood.

I was suggesting it's not impossible, even for an individual with some money to burn, to find out the hard data.

Googling a bit around I was able to find interesting stuff regarding the GTI which is why I was mentioning it. The same holds true for many popular models, and I think big repair shops also have huge DBs with technical data about an almost endless list of models.

Quote from Bob Smith :

I'd imagine it varies quite a bit between cars. I do know that for the RaceAbout, real spring stiffnesses were supplied but that for the Scirocco, they were not.

Spring stiffnesses are just a tiny bit of the puzzle, I would have expected the average 'manufacturer' to provide enough information to Scavier to model the car with a certain degree of accuracy - certainly not because they are happy to share it, but because I put faith in Scavier not accepting too many compromises about this.

Quote from NightShift :

I was suggesting it's not impossible, even for an individual with some money to burn, to find out the hard data.

Absolutely. Unfortunately, I don't see the devs being in a position to collect all this data first hand, hence their reliance on other sources, both for hard data, and good estimations to fill in the blanks.

Quote from NightShift :

Spring stiffnesses are just a tiny bit of the puzzle, I would have expected the average 'manufacturer' to provide enough information to Scavier to model the car with a certain degree of accuracy - certainly not because they are happy to share it, but because I put faith in Scavier not accepting too many compromises about this.

From what Scawen has told me, and from what I've learnt at work, manufacturers are generally not willing to part with data, and it can take quite a bit of persuasion and reassurances to get anything from them. Rest assured the Scirocco will be based on real data, and that the devs do want to represent the car in the most accurate way possible in LFS, just that some numbers have had to have been estimated.

TBFH I am a bit concerned since I tried a VWS mock-up set for the FXO. The thing was next to undriveable, despite the cars not being all that different. Now there are of course differences, so there's no point in attaching too much value to a test like that.

But I am under the impression that LFS being a racing simulator, it's optimized to give more faithful results when dealing with the ranges found in race prepped cars, and as such there are some troubles when settings approach the values used in RL road cars. E.g. that springs and and dampers will have to be a goo deal stiffer than their RL counterparts.

If anything, it's the stiffest springs and especially dampers that will cause issues in simulations. One issue we have with soft springs in LFS, is that we don't have your typical road tyres to go with them. Even road normals are pretty good. So you get a little more roll than you should. Plus I'm not sure if LFS models anti-pitch geometry, which only requires a mild setting in road cars to keep the car a lot more planted with the spring stiffnesses they tend to run.

Quote from Bob Smith :

Plus I'm not sure if LFS models anti-pitch geometry, which only requires a mild setting in road cars to keep the car a lot more planted with the spring stiffnesses they tend to run.

And wouldn't you use stiffer springs (and dampers) to cope with that? - OTOH race prepped would be more stiffly anyway so the problem is less felt. While 'optimized' might not have been the best choice of a word, it does not sound to me like we're saying entirely different things.

BTW the FXO was indeed rocking back and forth like there was no tomorrow, and that was with 38 N/mm springs at the front, vs the 26-28 N/mm rating I've found for the GTI.

These are a couple other cars I've found, the ratings are similar.

http://www.automobilesreview.c ... -news/mazda3-mps-2/10593/
http://forums.streetfire.net/showthread.php?t=24138

What are the respective ride frequencies of the cars in the links? 1.7 Hz without any sort of anti-dive/squat geometries and progressive springs and dampers would be a bit washy I guess.

Using this data:

VW Scirocco 2.0 TSI - mass 1298 kg - distrib ?/? (should be close to the Golf though)
VW Golf GTI Mk 5 - mass 1303 kg - distrib 62/38 - source
Mazda 3MPS - mass 1485 kg - distrib 64.1/35.9 - source
Chevrolet Cobalt SS/TC - mass 1352 kg - distrib 60/40 - source

and assuming motion ratio=1, the actual front frequencies should be:

1.19 Hz - for the Mazda 3MPS.
1.30 Hz - Golf GTI Mk V, Chevy Cobalt SS/TC,
1.31 Hz - RL VWS assuming the same wdist as the Golf (incorrect)
1.55 Hz - LFS VWS assuming 38 N/mm and the same wdist as the Golf.

NB: The FXO, even when ballasted to match the VWS, should be well under 62F IIRC.

PS: I've found another source for the 3 MPS which states the front coilovers are rated for 33 N/mm, that would take in line with the others at 1.33 Hz.

What values are you using for the unsprung mass for those calculations? Unsprung masses are very light in LFS, meaning there is more load on the springs - another reason we need stiffer springs in LFS.

i don't know alot about this stuff, but wouldn't the fact the cars bodies in LFS are completely rigid, mean that we would generally run softer springs (thus making some settings copied from real life into the game, unrealistic) being as the suspension mount points don't "flex" like they do in real life... I know putting a strut brace on a real car makes a pretty big difference, i can only assume it works the same in LFS, with the bodies being completely rigid, almost like a race prepped car?

I was sure I had forgotten to say something important :doh: Sorry about that :ashamed: all I could find is the tyre+rim weight, so those numbers come from using the full corner weight.

Assuming the unsprung mass is comparable among those cars, the numbers are still useful for a rough comparison, I.e. the VWS with the 38 N/mm springs is still about 0.25 Hz stiffer than a Golf GTI Mk V.

Subtracting the unsprung mass from the corner weight, each 50 kg of difference would result in an increase of 0.08-0.10 Hz. If 50 kg was to be exactly the unsprung weight for one of the front corners, and assuming I did not make any mistake, the frequencies would be:

1.40 Hz - for the Mazda 3MPS (33 N/mm coilovers)
1.39 Hz - Golf GTI Mk V, Chevy Cobalt SS/TC,
1.40 Hz - RL VWS assuming the same wdist as the Golf (incorrect)
1.66 Hz - LFS VWS assuming 38 N/mm and the same wdist as the Golf.

Hi
Great setups Bob!
Really nice job.
But I am missing something. In one of those files to download, you have written in his name "S2 06A". What does this mean? That the setups inside that package are designed to 06A when it comes out?

Thanks

yeah, but from what i understand, they should still work on z.

I named the pack for 0.6A as, at the time, I was expecting the patch to come out very shortly, and that there were going to be no physics changes to the cars, so it would equally apply to that as for 0.5Z, and didn't want to give the impression of the pack being out of date so soon.

As it happens, there are some physics changes, so they might not apply correctly, 0.6A has been substantially delayed, and there might not even be a need for a third party road going pack in the first place. So in hindsight, I should probably just have called them the 0.5Z road going sets as I should of anyway.

Quote from Bob Smith :

I named the pack for 0.6A as, at the time, I was expecting the patch to come out very shortly, and that there were going to be no physics changes to the cars, so it would equally apply to that as for 0.5Z, and didn't want to give the impression of the pack being out of date so soon.

As it happens, there are some physics changes, so they might not apply correctly, 0.6A has been substantially delayed, and there might not even be a need for a third party road going pack in the first place. So in hindsight, I should probably just have called them the 0.5Z road going sets as I should of anyway.

Thanks for this update

awesome setups. they're fun. and in that VWS.zip, what do you do with the vws.vd3?

Put it in VHPA\vehicles with the rest of the vd3 files. They appear to be for VHPA's internal use, possibly stuff like geometries and stuff.

Hey Bob! So I devoted some time to play around with the RG set of sets and I'm really glad I did that. All of the LSD cars are very pleasant to drive, and respond in predictable ways to modulations of the throttle through corners.

I'm not too fond of e.g. the open diff XFG as IMO it's too easy to spin the inner wheel out of corners, but I suppose that's more of an issue with the current version of LFS rather than the sets.

Thank you for sharing

I'm going to try these sets later on 0.6B, but is anyone using these still (especially on cruise servers)? I'd be interested to hear any comments about these on the latest version.

And oh, Bob, if you're reading this, great stuff - thanks. I think we should have road going sets in official releases of LFS.

Quote from Zero7 :

And oh, Bob, if you're reading this, great stuff - thanks. I think we should have road going sets in official releases of LFS.

You're welcome. Thankfully, you're not the only one who thinks this. Scawen is in agreement also. We'll see what the future brings.