What is more important for roll in a car: bump settings or rollbar?

I've recently started to doubt everything I thought I knew about setting up F1 cars in realistic sims like Gp4 or rFactor. 

At the moment, I am retrying the 1991 historic F1 mod... and it got me baffled. I thoroughly enjoy it, I just can't get it to change behaviour. It basically understeers when you go off the brakes. I thought I understood rollbar vs bump settings, but apparently I don't...

Yes you are missing a couple of things. A systematic way of measuring what you are doing, and a systematic approach to testing.

Other than that you doing just fine.

When I tune a car's suspension I have a series of measures of responses, such as on-centre yaw gain at three speeds, yaw delay time, and off centre yaw gain at typical cornering speeds. I also invariably look at the understeer gradient. Altogether I probably have 20 responses for a given car that characterise its steering and handling in the linear range. These are investigated systematically in a Taguchi type design of experiments, which typically tests 10-30 different factors, resulting in several hundred runs. Then some rather trivial stats gives a model that runs in excel, so we can diddle around and find the optimum. Bear in mind I usually have other constraints on the tune we select, for instance ride frequencies and % roll stiffness from the sta bar are measures that are critical to the overall ride and handling, even if they make meeting the steering objectives impossible.

Am I missing something?

 

Well, you are assuming the sim physics are correct.

Do you have a 'stability control' function enabled?

Do you have a 'stability control' function enabled?

No, I turn all aids off, except clutch.

Well, you are assuming the sim physics are correct.

That is ofcourse always dangerous.

Yes you are missing a couple of things. A systematic way of measuring what you are doing, and a systematic approach to testing.

 

I am not as systematic as you are, but I am not totally rushing in blindly. I've got a more than working knowledge about studying telemetry and bump settings and so forth. 

When I tune a car's suspension I have a series of measures of responses, such as on-centre yaw gain at three speeds, yaw delay time, and off centre yaw gain at typical cornering speeds. I also invariably look at the understeer gradient. Altogether I probably have 20 responses for a given car that characterise its steering and handling in the linear range. These are investigated systematically in a Taguchi type design of experiments, which typically tests 10-30 different factors, resulting in several hundred runs. Then some rather trivial stats gives a model that runs in excel, so we can diddle around and find the optimum. Bear in mind I usually have other constraints on the tune we select, for instance ride frequencies and % roll stiffness from the sta bar are measures that are critical to the overall ride and handling, even if they make meeting the steering objectives impossible.

But I have to admit that I don't understand yaw very well. In the telemetry I use for gp3 and gp4, (F1perfview) slip-angle is quite easy to understand and working with works well. I study the graph, determine if the understeer/oversteer is on brakes, after brakes, during release, during throttle and then I act accordingly.

With rFactor I use F1 Telemetry by Simon Phillips (designed for At the Limit), and there is no straight slip-angle, but about seven menu's for yaw. The author himself states that manually comparing slip-angles is 'unusable' (It worked fine for me, but okay...)

Then he states I have to use his way, and 'correlate slip-angle vs slipangle'. That I don't understand at all!

I would put up a screenshotm of his explanation, if I only knew how to do it on this board...

Well, I don't know squat about a sim, but I can tell you in real life those cars ran some massive front springs to keep the front low and working aerodynamically. In real life, if you try to make it turn by just softening the front end, then you end up having to stick the front way in the air and lose the thing aerodynamically. No amount of roll-couple shifting is going to help a car that's chronically low on front downforce.

Well, I don't know squat about a sim, but I can tell you in real life those cars ran some massive front springs to keep the front low and working aerodynamically. In real life, if you try to make it turn by just softening the front end, then you end up having to stick the front way in the air and lose the thing aerodynamically. No amount of roll-couple shifting is going to help a car that's chronically low on front downforce.

Yeah, in that aspect I think the sim-physics of the 1991 mod can't be right. Basically in gp2, gp3 and gp4 you set the front springs 400 lbs heavier than the rear, and the front rollbar three times as heavy as the rear. I have  base setup I user since ever, which is roughly 1200 lbs -front suspension, 800 rear, rollbar 1500 front, 300 rear, and from there I work with the slow bump settings. When I need a quicker response, I put more weight on the front rollbars, and the cars respond accordingly.

In the 1991 mod, the default settings are much, much softer, but if I try Gp4-like settings on the cars, the tires heat up  to twice the ideal heat. 

But your posts have some good hints, I will try them.

I think I should change the topic title:

Do rolllbars work as rebound dampers too?

I've found a solution for my problem, I just dont understand how it works. I've put about 1500 lbs on the front rollbar, 300 on the back... and now the car DOES want to roll into the car after braking. Which is contra-intuitive. The only thing I can imagine is that a soft rear rollbar allows the car to dip forward, even though the front of the car is very tightly sprung in the front roll bar.

Weird, weird, weird

J.

Perhaps some other poor sim-guy can learn from this, so this is (probably) the last post I make in this thread. Not many responses but some were very helpfull.

I figured it out. I always thought that the softer the front rollbar, the more oversteer you have and the less understeer. Basically, with a F1 car, also in sim, you have to put the front rollbar very stiff to work with the high downforce on the front.

So when I play the F1 1991 historic mod, in which the default setups are really soft set up (especially the rollbas), I was rather baffled to have a car that just does not want to steer in after releasing the brakes. Even if I put the slow bumps in front really, really in the low numbers.

However, I forget something. The rollbars are sideways springs. That means that if they are soft, the car can lean more easily over to the side that gravity pulls it to. Meaning: a soft rollbar at the front gives you more oversteer DURING the corner, but gives you UNDERSTEER while braking and steering in. Because the car with a soft rollbar leans over easily and pulls you to the outside...

Anyways, back to the drawing board...

I've recently started to doubt everything I thought I knew about setting up F1 cars in realistic sims like Gp4 or rFactor.

At the moment, I am retrying the 1991 historic F1 mod... and it got me baffled. I thoroughly enjoy it, I just can't get it to change behaviour. It basically understeers when you go off the brakes. I thought I understood rollbar vs bump settings, but apparently I don't...

I have a lot of experience in modding and setting up F1 cars in the virtual world. In Gp4, in historic mods like the 1980 and 1982 mod, I can set up the car exactly like I want. I know how to set fast and slow bumps, diffs, etc, etc, etc...

However, with the 1991 mod, for example on Interlagos, for example, I can't get the Tyrrel-car to steer into the corner after releasing the brakes. Neither can I with the Dallara or the Lotus.

I've tried softer springs,

Ofcourse, if I set the rollbar really soft in the front, the car will respond... but then it will be undriveable in quick sweeps left and right.

Am I missing something?

Edited by Nigel Beresford, 21 February 2014 - 23:09.

However, I forget something. The rollbars are sideways springs. That means that if they are soft, the car can lean more easily over to the side that gravity pulls it to. Meaning: a soft rollbar at the front gives you more oversteer DURING the corner, but gives you UNDERSTEER while braking and steering in. Because the car with a soft rollbar leans over easily and pulls you to the outside...

 

Anyways, back to the drawing board...

Theres no reason a soft anti roll bar should cause understeer on turn entry. Also, body roll does not pull the car to the outside at all.

Brake release understeer can be caused by a few things. Obviously the behaviour of your car depends on the quality of the sim model.

Look at the pitch movements of the car as you come off the brakes. If there is a reasonable amount of lift at the front then 2 things could be happening:

1. Damper rebound forces could be trying to pick the front wheels off the ground and killing their grip - Solution, reduce low speed REBOUND damping

2. If the game models downforce as a function of ride height, the lift on the nose could be causing either a drop of overall downforce, or a shift in the balance to the rear. The solution here is to stiffnen the front springs. This may induce mid corner understeer so you should either soften the front bar or stiffen the rear springs (by the same amount) to compensate.

Static toe angles also have a large effect on turn in.

Softer front roll bars equal oversteer? 

I know we're dealing with all or nothing generalisations, which are always a bad idea. But the first thing I remembered was a McLaren engineer's comment about Hamiton's general setup being super stiff front, soft enough rear to keep the tail in the same country. 

Stiff front, softer rear sounds pretty intuitive for car with a front wing in significant ground effect and a rear wing well clear (ignoring any interaction between the lower element and the diffuser) but how does suspension setup affect undercar df?  Might pitch under braking or acceleration significantly affect either total df or its center? Perhaps even usefully?

Aren't F1 cars, or F1 cars of the early 90s, fairly pitch sensitive because there wasn't much downforce to speak of between the wheels? Compared to previous iterations or Indycars or whatever. 

We're in broad-sweeping-statements territory, but you know what I mean.

Softer front roll bars equal oversteer?

I wouldn't say "equals oversteer" but rather "moves in the direction of less US, more OS".

Softening the bar on either end reduces the lod transfer on that axle which in turn increases its "grip".

Thanks for the replies. Perhaps I should just post a bloody lap on youtube. I am driving the Ferrari 640 in the game now, because it is the only one who I sort of can handle.

The thing is: it does not respond to bump settings like Gp3 or Gp4. Still not! The thing is: the 1991 Historic Mod is one of the most lauded mods of rFactor. I can't imagine that the sim is that bad. It must be my driving (it usually is).

PS: Yes, the Tyrrel has no front rollbar, it has a third front spring... which does not do what I expect it do either.

Here is a typical turn in (alas) for the first corner at Interlagos:

https://www.youtube....eature=youtu.be

And here the telemetry for the ride height:

http://imageshack.co...3/6047/7bfv.jpg

As far as I can tell, the car does not dip forward at all... It seems it has no pitch!

PPS: I really appreciate the help. I really like this mod, it is excellent, but the driving is so difficult and the setting up so unrewarding that I keep returning to gp4. At least if I change bump settings something bloody happens there!

Edited by Nemo1965, 25 February 2014 - 20:55.

PS: Yes, the Tyrrel has no front rollbar, it has a third front spring... which does not do what I expect it do either.

 

Here is a typical turn in (alas) for the first corner at Interlagos:

 

https://www.youtube....eature=youtu.be

 

And here the telemetry for the ride height:

 

http://imageshack.co...3/6047/7bfv.jpg

 

As far as I can tell, the car does not dip forward at all... It seems it has no pitch!

 

 

PPS: I really appreciate the help. I really like this mod, it is excellent, but the driving is so difficult and the setting up so unrewarding that I keep returning to gp4. At least if I change bump settings something bloody happens there!

If the car is reasonably stable everywhere else, move the brake balance back until it either fixes the problem or the car becomes too unstable. Looks like the fronts are saturated under brakes and have no grip left for turn in.

If you dont want to move the brake balance, try lifting off the brakes earlier.

Put up a GG plot too, just for the first corner.

Edited by NotAPineapple, 27 February 2014 - 21:38.

If the car is reasonably stable everywhere else, move the brake balance back until it either fixes the problem or the car becomes too unstable. Looks like the fronts are saturated under brakes and have no grip left for turn in.

 

If you dont want to move the brake balance, try lifting off the brakes earlier.

 

Put up a GG plot too, just for the first corner.

Thanks for the tip. I am getting good help here and elsewhere. 

Another cause for the understeer, so I have discovered, is that the designers of the mod built in a 'blip' in the gearchange (don't know how to call it) but I have to blip the throttle to get the car to brake on the engine. Yes, very realistic, but you just have to know. 

To get back to the title of the thread, at the limit the tire is sliding and it is pretty much irrelevant exactly which direction it is pointing. So at the limit Fz matters more than delta (the steered angle), or in your terms, rollbar matters more than roll steer /at the limit/ and don't quote it out of context. If you are not at the limit then it is quite easy to have a car that doesn't 'respond' to front sta bar. Everybody knows you should get more u/s with more rollbar, but if that extra rollbar cause the vehicle to roll less it will reduce the roll steer, which is a huge part of your u/s budget. It so happens that for real cars the two effects are of similar power.

To get back to the title of the thread, at the limit the tire is sliding and it is pretty much irrelevant exactly which direction it is pointing. So at the limit Fz matters more than delta (the steered angle), or in your terms, rollbar matters more than roll steer /at the limit/ and don't quote it out of context. If you are not at the limit then it is quite easy to have a car that doesn't 'respond' to front sta bar. Everybody knows you should get more u/s with more rollbar, but if that extra rollbar cause the vehicle to roll less it will reduce the roll steer, which is a huge part of your u/s budget. It so happens that for real cars the two effects are of similar power.

Thanks. I've had an extensive (by now) e-mail contact with someone who used to work as an engineer for the team of which I now try do drive an virtual version. And he confirmed that, indeed this car, with a mono-shock setup, had the annoying tendency to understeer after you release the brakes.

Regarding your comment of not driving at the limit: that makes sense, I am not driving at the limit at all, so the thing in the virtual cockpit has some negative input. But ofcourse you CAN drive at the limit if you have some sort of thrust as a (virtual) driver, how the car is going to respond once you release the brakes. It is my impression that the designers of the mod (FSR 1991 Historic Mod) have done a wonderfull job of trying to implement real-time physics. At the same time, that creates problems for someone who try to drive the darn thing just based on visual imput and not the seat of the pants...

Anyways, you make a very good point. If you are at the limit, the tyre is sliding... It is not something I thought about. My idea in general is to use the slide of the tyre during braking and turning in... and that I stop the sliding of the tyre gradually by releasing the brake and upping the throttle...