PDA

View Full Version : [Physics] Brake pressure regulator



LesiU
03-09-12, 04:07 PM
Not something I'd see on top of the "to do" list, but having an implementation of a brake pressure regulator for rear brakes (which limits how much pressure can be generated in brake lines, limiting braking force at rear brakes) would be a nice addition, as almost all factory street legal cars have such device.

lordpantsington
03-09-12, 04:12 PM
:D +1 !

LesiU
03-09-12, 04:35 PM
Hah! I knew I could count on you :D

lordpantsington
03-09-12, 10:41 PM
Do you have any further information to add here on how a proportioning valve works? (Asking because if some of the legwork is done for ISI, perhaps it stands a better chance of getting implemented?)
Example values: 55 bar valve, .46 reduction factor, (5/55) :D

The one thing about rf/2 is that its physics seem to go straight for the heart. I'm not sure how small details of individual components would play into that. Yes, there is code to use a pressure of some sort (0-100%), but the assumption is that it's a multiplier acting directly on Braking Torque.

Along the same subject, how is ABS working currently? As an aid it appears to just increase grip (which isn't that far off). If the braking system is expanded to deal with pressures, it seems we could apply ABS (and a simple version of the triggers to engage it) to the cars that have it, while leaving aids off on a server. Add to it different mapping (Alternate triggers/reactions) and I'm not sure what else would be needed to flush out the braking aspects of the physics engine.

LesiU
03-10-12, 03:05 PM
Proportioning valve is not a simple reduction. It works like a cutoff, sort like a thermostat, allowing to take off additional pressure generated in brake lines. It won't allow you to go above defined pressure.
in rF realms, that might be represented as additional parameter for rear braking torque.
If I have BrakeTorque=3500.0 for rear brakes and then there is BrakeTorqueThreshold (let's call it like that) set to 2500, then brakes would work like right now but rear are cut off at 2500Nm (but overall balance until that point is calculated for 3500).

That's how it works in real cars. If a car has "designed" pressure of 80/80 bar (front/rear) and you use proportioning valve for 55 bar, then the system works like for 80/80 until you hit rear threshold not allowing you to have more rear pressure than 55 bar (there is an additional valve with separate loop letting go excess of pressure above 55 bar point).

ABS and TC can increase tyre grip but don't have to.
In theory, if you set:
TractionControlGrip=(1.0, 0.0)
ABSGrip=(1.0, 0.0)
that should not change grip levels.
Then, you have TractionControlLevel and ABSLevel which affect throttle and brakes. I will do some more tests at some point in the near future, as I will need that set correctly in my road cars.

As for alternative to how currect brake system works in rF, we should probably wait for changes (if any?) in rF 2 that might come at some point. Definatelly that's something to discuss.

jtbo
03-13-12, 02:57 PM
My car has valve that senses load too, it gives more braking power to rear wheels when I put heavy load on car, when there is less load it gives less, but I have never really looked into how it does that.

Car is small station wagon / estate with abs brakes and not much electronics, no central ecu and valve is mechanical, no electricity there either.

You can of course emulate setting front/rear brake bias, but it is only one value, so you would always get smaller amount going to rear and for some situations it would be wrong of course.

Then modern crap as I like to call them, has things like emergency brake assistance, which increases pressure when you push brakes fast and strong.

I made spark cut TC for one of my cars at one time, what a fun on ice BABBANGBABGABANG and people where staring 3 meter flames from straight exhaust, oh the joy of young ;D
Such is of course not possible with rF as from my understanding TC just controls amount of throttle, maybe whole engine does not know about sparks at all.

TC and ABS increasing grip is something I just can't understand, sounds to me that it is just plain wrong, in reality there is grip what there is and it is only how efficiently one can use all that grip, but never more grip by such system, maybe there is again something I don't understand?

lordpantsington
03-13-12, 06:46 PM
Your understanding about ABS is the conclusion I've also reached, in actuality it is more about limiting losses then an actual gain in grip. Technically all ABS does is keep the wheel from locking. The increase in stopping power is caused from keeping the tire/road surface reaction from changing over to sliding friction. It seems to have more to do with the tire/road reaction then the actual car. But how do you simulate that? Does increasing the grip create the same result? The one thing I've wondered about is when does the grip increase? Surely it would only seem fitting that answer is under braking. A more appropriate application would be under braking when the wheel is sliding/starting to slide. I doubt anyone but ISI knows how the ABS system is implemented.

I've yet to find any data as to how much sliding friction increases stopping distances (or reduces friction). Don't know if the default ISI values are just what feels good, or has roots in test data. Also don't know how rf/2's ABS activates. Each wheel is monitored so it is possible there is a hardcoded wheelspeed difference that activates ABS.

What I'd like to see is: a WheelSpeedDifference parameter made available, rf tests against it, when that test fails ABS is activated. Add to it, frequency;the number of tests per wheel rotation (in my case 48), and different mappings and I'm not sure anything else would be needed.

As it is now ABS is an aid to help drivers. It cannot be turned on for individual cars without disabling it completely for other cars (negating any use of the aid for drivers that need it). I would still like to see it in place as an aid. I would also like to see it implemented as a system, a characteristic of the car itself, so that cars that are allowed to run/have the system in place can use it without effecting others. The ABS aid would stack so drivers that need it can still use that. This allows servers to ban aids, but keep the system.

I imagine TC has the same bit of conflict. Some cars have it, some don't. If you want to run those that do and don't together, you need to negate the ability of those that don't to run it. Either you allow those cars that cannot use it an advantage, or you zero the parameters. If the latter, you also negate the possibility of use for drivers that need it.

I ran across a book with a section about bias valves (was Puhn or Milliken or the like). I should dig that back up an add it here.

jtbo
03-14-12, 05:11 AM
Depends from tires and surfaces, bias tires from 70's and before had shortest stopping distances with wheels locked (there is scientific documents backing this up, also their slip peak is at full lock, longitudal at least) and modern scandinavian winter tires are able to stop car much faster on ice without abs than with abs. Usually ABS on street cars is increasing stopping distances a tiny bit on tarmac too compared to experience good driver doing optimal braking, but with abs average joe manages to avoid accidents.

Tire model is responsible of doing right thing so that whatever braking aid or human is braking it provides correct results compared locked wheel and not locked one, if tire is make right, it should produce correct results, but I have not really tested this too much, that is what it should do, but does it?

I think there was parameter for ABS to use individual 4 wheel abs or simpler abs, that of course affects also to how abs acts.

For certain, help and feature of TC or ABS should not be related to each other, as well as automatic gear changing and automatic gearbox should be complete different things, well we don't have automatic gearboxes yet, maybe we will eventually? :D

ABS efficiency parameter gives me impression that ABS is rather simple model, it is perhaps need to rewrite some of that to be able actually use any realworld data to get similar results, I guess same is with rear brake pressure valve too, it perhaps would need more complex braking system model?

MaXyM
03-14-12, 09:31 AM
Also, as far as I know, there is no brake pressure balance settings in racing cars (at least in some) in form we know from ISI engine based sims. pressure proportion is set just bu setting rear brakes pressure. Front one remains untouched.

lordpantsington
03-14-12, 04:18 PM
OP was about road cars. Reliable race car data is hard to find, it would be nice to see the sim expanded out to help simulate all vehicle types. Not all tracked cars are purebred race cars. To better replicate your club level type racing, this addition would be needed.

LesiU
03-17-12, 04:04 AM
Also, as far as I know, there is no brake pressure balance settings in racing cars (at least in some) in form we know from ISI engine based sims. pressure proportion is set just bu setting rear brakes pressure. Front one remains untouched.
Brake pressure is not something you set with a knob or a pump, like you'd set tyre pressure ;)

We are talking here about road cars, so typical situation is:
- fixed brake bias
- proportioning valve is present in the rear brake loop


In higher class race cars you have no proportioning valve (as you don't want it to be there) but you have adjustable brake bias (usually set by moving brake balance bar left/right).

In mid class race cars you might have every combination in between. Depends on cars and racing series.

Doc_
03-17-12, 04:10 AM
Guys, you have brake bias adjustment in rfactor. No need for brake pressure adjustment valve.

LesiU
03-17-12, 04:22 AM
That is not the same.