Braking: Understand the Subject BEFORE Critiquing

OT: Can you fill in some info, how do we best learn modulating around the point of traction loss? And/or how do we learn optimizing braking to have best traction?

Digressing, what was originally raised in the topic is the amount of sensitivity to be set for LC (50% or not).... and thinking through this makes me believe that there is no single right or wrong answer, because it will depend on the type of LC used, and the driver using it. Most LC brake pedals have some form of foam/PU/rubber damper before the actual LC, and these foams can vary in length and strength too (plus some can even have multiple combined). The sensitivity setting simply allows one to set up the game according to the preference of the driver, based on his hardware. His feel of 50% may be different from someone else's, so there shouldn't be any right or wrong. Now I understand what you're saying that the driver's 'perceived' 50% should match game's 50%, they should match. But the sensitivity setting for this to happen can vary, again, depending on the hardware and the driver's perception. There's no single/hard value for all.

Great tip. Turned out my current sensitivity of 40 still resulted to somewhere 55-60 brake application per widget, so now I'm back to default of 35.

 

While this is true, it's no different than a real car. The amount of force we apply with our foot is not the exact same force that is applied to the caliper pistons. It's "theoretically" a ratio but is not as there are too many variables to call it a constant ratio (i.e. the how the piston contacts the pad, how the pad contacts the rotor, the springs used in the cylinders, etc). If you were to measure the amount of pressure at the caliper verse the distance on the pedal, you'll see that it ramps up quite a bit as you get past the mid point of travel.

Humans however adapt fairly quickly and build muscle memory based on the behavior of the car. Whether it's "linear" or not, as long as it's "consistently the same" the driver will build muscle memory and adapt. Even in real cars, braking is very different. Some have a lot of travel, some don't. Some have firm pedals, some don't.

 

The peak grip generated by the tyre is at the point it begins to slip (similar to maximising cornering speed using slip angle). The best way to learn how to threshold brake is practice progressively braking until the tyres just lock up then backing off slightly until regaining grip and trying to reapply to the point just before locking up. You should get different feedback through the FFB and sounds from the tyres as they cross the threshold from peak grip to slip. There will be tyre squeal before locking up, so the idea is to keep it consistently in that area without locking the tyres. The problem with exaggerated cadence braking is that you will constantly pass through the peak grip to either side so your average will be lower.

True, but...

This makes no difference. The load cell is measuring force applied, not distance travelled. The springs and PU bumpers etc only change the physical feel of the pedal and how much it travels for a given pedal force. The load cell will still measure 50kg of force using soft springs or hard springs when 50kg is applied to the pedal. (This is simplified, the pedal is a lever arrangement so the force applied to the pedal has a constant factor applied for what is experienced by the load cell). You may also have a soft spring so the pedal has some give before engaging the harder PU bumpers, but all that force is still applied to the load cell. The usual thing to do is calibrate the pedal so the soft spring part is covered by a dead zone. That way force won't be registered till you hit the harder bumpers giving a more realistic feel.

It certainly may suit some to set the pedal up using sensitivity in the sim, but this is to make up for other issues with the load cell, like if it is not adjustable. It's better to adjust the load cell using hardware if available as it will not affect the linearity. You can emulate it using brake gamma adjustment in game, but you loose the linearity, which is bad for the reasons mentioned before.

Keep in mind that a real car's braking system is entirely linear. There are components that can change the force applied along the path such as the master cylinder. But the pressure applied by the brake pad is directly proportional to the brake force applied to the pedal by the driver i.e. linear.

 

Which is why we talk about pressure at the pedal and not distance travelled with load cells. As you know the distance travelled on a load cell is very dependant on the setup of the springs. The pedal feel of a real car can also be affected by the factors you mention and elastomeric materials, seals, etc. The feel may be non-linear but the force being applied is still directly proportional.

 

Have a look in that driver61 channel I linked to on page one.

 

From a valve perspective yes. It's linear from pressure, but not necessarily application. There are several variables which come into play which has to do with the caliper/piston design and so forth. For example, not making the correct contact on the pad with the pistons will not apply the same force to the rotor. Therefore, even though the "pressure" is the same, how the car behaves is different and thus will require a different braking technique. If you are speaking purely from math, yes, it's proportional. If you are speaking from a driver, it's is not and depending on many things beyond just the calculations of the fluid.

 

@Mahjik The temperature of the brake pads and discs will also affect the perceived braking response. However all of this is dealt with by the simulation part of the game. The load cell and in game brake sensitivity setting is equivalent to the brake pedal up to the master cylinder in a real car. Which definitely is a linear system. If you could somehow create a nonlinear hinge arrangement for a real car to give a nonlinear brake response then Cliff may have a point, but none exist that I know of and good luck getting it registered for road use if you could make one.

 

Technically, the pedals on 90% of the cars are not directly interacting with the MC. There is linkage between the two. Most sims don't fully replicate a complete braking system (I don't actually know any sim that does today). They translate the pedal input into direct braking force on the calipers which removes a lot of variance that happens in a real life car. This one of the areas where most real life racers have issues getting used to braking in sims because it's too binary today.

 

It is certainly a direct linkage on a Caterham, isn't it Cliff.

I never said a complete braking system was simulated. There will be algorithms that affect how the brakes function and can be tuned for each car which is why people are complaining about some cars and not others in AMS2. And yes there are variances per car in real life. None of which can be captured by using a single gamma curve. Real life racers have a lot of issues switching to braking in sims, the lack of g force / SOP feel being the main one, even if you had the perfect brake simulation they would still have issues without that.

 

Yep, that's part of that 10%.

Which is why a braking isn't isn't truly linear. Yes, the MC behavior is linear (a ratio) but the entire system isn't not as whole. There is deviation and variance within the system as a whole so stating "braking" is "linear" isn't really true. Ideally, the system as a whole would be but "real world" gets in the way.

 

I feel this is getting rather pedantic. It is closer to linear than a high factor gamma curve.

 

That's fine but don't speak to others as if it's a fact as it's not really a fact (it's an approximation based on math without accounting for variances). If you don't like using the in-game sensitivity adjustment, great. But don't tell others they are wrong for doing it as that's not right either.

 

Wow, I've backed up all my comments with theory and how it is related to real life braking systems and you attack me? I notice you didn't say anything about the OP who was saying matter of fact that people need to "understand the subject" and that a very high sensitivity is closer to reality when that is very far from the truth. You've also agreed (within the realms of pedanticism) that real braking systems are pretty much linear.

If you read my comments I have stated that people can use sensitivity, all I have been trying to do is get the information out about why a linear response is better and closer to reality than using high gamma curves when using load cells.

 

Theory is one thing, but how things actually feel and behave in a real on track situation is surely more relevant. I don't think Mahjik is attacking anyone, he just has a different point of view, and if it comes from experience then it's worth sharing here.

 

I've probably been a bit robust in some of my comments so I apologise for that. We're all here to have fun with this awesome new game and I've had my part in making this thread overly serious. I stand by my belief that linear response is best with load cells, but of course it comes down to preference, what people are most comfortable with and find best for them. With that, I'll bow out.

 

All of those other relevant factors that can be modelled will be modelled in the sim. If they are not, or cannot, they won't. But fiddling with the linearity of our external-to-the-sim controller will only make any weaknesses in the modelling (or realism) of the sim worse. This is a digital realm, not an actual mechanical device attached to other parts in a real car that can interact. I think your analogy is missing that key point.

Reiza is trying to simulate as much of the realistic braking performance as possible--they will implement the curves away from linear so that we will feel something "realistic" when our hardware is set to linear/default. If you want to bend a bent curve (that is not transparent to us users), you are of course free to do so. But as I said way earlier on, sims will only ever be designed assuming people are using default and linear settings. No one can design a sim for a billion permutation and combinations of settings.

The only normal time one wants to move off a linear controller setting is if weak or defective equipment is in play. I am remembering my old G25 pedals with the ridiculously short travel and soft brake pedal. Some help to make it less sensitive was great, but as noted above by @NuScorpii, that was just transferring the problem from one part of the braking process to another. I could bend the curve so that the majority of typical braking was a bit easier to handle, but in exchange I had to give up fidelity elsewhere--because no digital manipulation can solve the inadequacies of the hardware.

Load cell brakes are already designed to help make a linear braking curve feel more realistic. Despite that, braking performance has been all over the map in AMS 2 because Reiza is adjusting the cars themselves and global settings (for the better, so far). If you think that they are making these adjustments with non-linear controller settings in mind... If, after the best possible sim experience is programmed you still don't find the dev's approach to be plausible or realistic, then you get to override it with your controller settings. But for the most part, those adjustments are to account for inadequate hardware (which no load cell brake should be in that category), or, for people with mobility issues that cannot use a standard controller (e.g., because you cannot generate enough force for full braking against a heavy load cell device or have the fine motor skills to adequately or precisely control the amount of pedal application in a lighter force situation). Even here, load cells have their own mechanical force requirement adjustments to prevent the need to resort to a non-linear setting in the sim itself...because that in turn just distorts whatever the sim devs intended.

 

This is pretty much where I stand on this as well. I think it's a good idea to start with 50 and tweak from there. Nothing wrong with using any other value if that works well for you.

 

But that doesn't mean that others aren't. Unlike the game code, which is the same for all of us, the pedal hardware and setup can vary enormously.

So 70% of the pressure gives you 25% of the total braking? That sounds completely unnatural to me. Maybe that's why you don't understand the issue many of us are having.

It depends on the rating of the loadcell. Some pedals have 100kg+ loadcells - that's the pressure they need to produce maximum output. If I want to use the maximum pressure my loadcell can take, I have to press very hard to get maximum braking.

I have LC brakes. When I use the Windows calibration routine, I see a linear increase in brake readout from starting to press the pedal to maximum pressure/braking. I doesn't start slowly and then suddenly increase in response. That certainly appears to show that the loadcell is producing a linear output which is what the game reads.

 

So you're from the Church of the 50% sensitivity, no compromise? I notice that people are now saying that their LCs are being defaulted to 35% by the software. The developers obviously know less about load cells than you do.

I'm not the one who can't stop locking his brakes, or destroying them inside 5 laps.

 

This I will agree with.. However, what the devs cannot account for are the various implementation of hardware. While they may be able to test with certain known widely used hardware, they cannot account for all hardware. This is where settings like sensitivity fill the gap of not all hardware created equal.

Load cells are created to simulate pressure rather than travel, which we all agree. However, not all load cell implementations are the same. Using a Fanatec pedal setup with a load cell is a different than using Heusinkveld pedal. Now, with some pedals, you have a lot of adjustments you can do to them for your own desired feeling. That's no different in most cases to adjust options provided inside of a sim title. It's six in one hand, half a dozen in the other. What I'm suggesting is that it doesn't really matter which way a sim racer uses to get their desired result.

For some of the load cell implementations, travel is fairly short (at least compared to a real pedal). This actually makes modulation of the brake a little harder as there is a more narrow window in the travel. In those situations, using something like sensitivity to adjust the linearity can make the experience better for those sim racers. I will say, the majority of sim racers who used load cell pedals for PC2 did not use 50 for sensitivity.