Thrustmaster t300 Recommended FFB MPs each car

I could be wrong but I thought this HDV entry would alter the level of simulated steering arm torque that corresponds to full FFB output:

Code:

[CONTROLS]
NominalMaxSteeringTorque=9.5   // Maximum steering arm torque to effect force feedback strength

which, by the sounds of it, might allow you to test the effect of different STS settings at/near 'full force' in a more reliable manner.

 

Sorry for missing your earlier post Lazza.

I don't think this will help because it's not the/a problem in the first place.

I'm not certain but i think changing NominalMaxSteeringTorque has the same effect as changing the car-specific ffb multiplier. So if the default NominalMaxSteeringTorque=9.5, then the maximum ffb torque output from your ffb will equate to 9.5Nm of virtual steering wheel torque. And when in non-dev mode, increasing or decrease the ffb multiplier from the default ("1.0") will increase or decrease the NominalMaxSteeringTorque per car.

Edit: Ah, i think i remember Terence or Turtle talking about this. It could just be the setting that applies to ffb wheels than can produce more physical torque than (for example) 9.5Nm. That way if you ffb motor is very powerful and you set the car-specific ffb multiplier too high, your ffb wheel won't output more physical torque than is actually occurring in the virtual world (i.e. so it doesn't ever go higher than 1:1).

I need a response from Paul (when he has the chance) to my last 2 questions because although he say's he's done the experiment, i can't help but get the feeling that something still doesn't quite add up or has been accidentally misspoken (or something).

 

Build 906 introduced a setting pertaining to wheel torque:

The NominalMaxSteeringTorque parameter itself has been there for much longer.

 

Oh ok, fair enough. May i ask what this parameter does then? Does it cap the maximum virtual steering wheel torque to maximum ffb torque output?

 

Sorry, had to wait to make sure you weren't going to added any edits.

The game has to know how much virtual steering arm torque corresponds to full FFB output. That's what this parameter does.

With it set to 9.5Nm, a calculated steering arm torque of 4.75Nm gives you 50% FFB output. Your FFB multiplier then goes on top of that, so 1.0 will actually give you 50% FFB force, 0.5 will give you 25%, and 2.0 will give you 100%. >2.0 obviously won't change anything because you're already at 100% output.

Once you reach a calculated torque of 9.5Nm you get 100% raw FFB output. Again this then gets multiplied by the FFB mult. But if you exceed the NominalMaxSteeringTorque value the value just gets capped, regardless of any further multiplier.

For example, reducing the NominalMaxSteeringTorque by a factor of 10 gives you 10x stronger feedback in very simple terms, but then applying a FFB mult of 0.1 doesn't put it back to normal; you just end up with FFB that maxes out very easily and at the wheel is 10x weaker than normal at peak.

What all this means is that instead of having to be driving around at speed and throwing the car into a turn in order to approach 'full force', you can lower the NominalMaxSteeringTorque and approach full force while sitting still and simply turning the wheel. Then any effects of the STS parameter can be more reliably judged.

 

lmao! fair point.

Hmm, this doesn't seem right. It's been my experience that when lowering the car specific ffb multiplier, the maximum virtual steering wheel torque (that corresponds to the maximum ffb torque output from my ffb wheel) increases (to the point where you can make really strong virtual jolts to the virtual steering wheel still not clip the ffb and i wouldn't be surprised if these jolts were easily way over 9.5Nm on the virtual steering wheel). Conversely, increasing the car specific ffb multiplier decreases the maximum virtual steering wheel torque (that corresponds to the maximum ffb torque output from my ffb wheel). And regardless of whether you increase or decrease the ffb multiplier, provided you can get the ffb to reach 100% (i.e. clip), the strength of the ffb torque through your ffb wheel is the same.

And if that is true, then increasing/decreasing the car specific multiplier should have the effect of increasing/decreasing the car specific NominalMaxSteeringTorque value. But your confirming it does not?

 

Did exactly the test I gave as an example before I posted:

- Devmode, skip barber, entered realtime and moved the steering wheel left and right: small-moderate amount of FFB.
- Changed the nominal max force from 9.5 to 0.95: full FFB with any movement. On my G27 with a soft lock of 407° I couldn't tell where I hit full lock... the force didn't change.
- Set the car specific FFB Mult to 0.10: FFB nearly non-existent.

If they worked directly with/against each other this shouldn't happen; you should be able to make them cancel each other out across the whole range of forces.

 

Can you tell me if the following is correct then with the NominalMaxSteeringTorque at 9.5Nm?

So with the ffb multiplier at 1.0, a calculated steering arm torque of 4.75Nm gives you 50% FFB output. A calculated steering arm torque of 9.5Nm gives you 100% FFB output = max torque output possible from your ffb wheel.

With ffb multiplier at 0.5, a calculated steering arm torque of 4.75Nm gives you a 25% FFB output. A calculated steering arm torque of 9.5Nm gives you 50% FFB output = half the max torque output possible from your ffb wheel. And since NominalMaxSteeringTorque is limited to 9.5, any calculated steering arm torque greater than 9.5Nm will give you the same constant 50% of max FFB output? Meaning your ffb wheel's max torque output is limited to half of what the motor is actually capable of?

With ffb multiplier at 2.0, a calculated steering arm torque of 4.75Nm gives you 100% max FFB output. And any calculated steering arm torque greater than 4.75Nm will give you the same constant 100% of max FFB output? Meaning your ffb wheel's max torque output now occurs when the virtual steering arm torque is at or over 4.75Nm?

finished editing

 

Correct on both counts.

FFB Mult will reduce your maximum FFB force if <1.0, and will start clipping higher input forces if >1.0. NominalMaxSteeringTorque sets the input range for your FFB, so if the calculated torque only ends up being half its total you'll only get half full force, and the opposite is true as well - if the nominal value is lower than the calculated, you'll get periods of full unwavering FFB force and lose fidelity.

It's a two stage process: the game calculates steering arm torque (presumably in this phase it also takes "Steering torque minimum" into account), then scales that between 0.0 and 1.0 based on NominalMaxSteeringTorque. With that done, it then multiplies by your FFB mult. The result has to be between 0.0 and 1.0, so whether you end up never reaching 100% force or spend too much time on 100% depends on how well the NominalMaxSteeringTorque matches your vehicle. The skip barber has 9.5Nm, a typical F1 mod has 22-25Nm. Switching the two wouldn't produce great results but you could compensate for the overly high nominal value on the skip barber by pushing up the FFB mult. Lowering the FFB mult on the F1 wouldn't stop the FFB being clipped, it would just make all FFB weaker (so you're clipping at half strength or less, instead of at 100%... and you've made the lower forces too low as well... )

 

Sorry Lazza, i'm just really shocked if all this is true. I mean, it even goes against everything i wrote in part 1 of the optimal ffb guide thread. I haven't played rf2 in a few months nor have i experimented with the ffb multiplier for yonks but the whenever the last time i did, i could have sworn that decreasing the car specific ffb multiplier did not start to cap the max ffb torque output from my ffb wheel. So for example, if "1.0" made my ffb clip on the straight, that level of physical torque output from my ffb wheel would feel equally as strong as if i set the ffb multiplier to "0.5" and only just clipped on the high speed corners instead.

If your observations are correct, how are you able to tell that the ffb motor torque output range has been limited to half the max when setting the ffb multi from 1.0 to 0.5? Just by feel or something else?

Also, did you make these observation in dev mode only? What about in non dev mode. Any difference?

If everything you say is indeed correct, i will need to rewrite part 1 of my guide and advise changing the car-specific NominalMaxSteeringTorque instead of the ffb multiplier.

Has NominalMaxSteeringTorque been around since before i wrote that guide?

I will confess to being in a little bit of denial atm. It's just really quite different from what i thought i knew and remembered observing in the past.

Is there any chance you could be mistaken?

 

What does this mean for us normal folk.

 

I already gave you the normal folk answer two pages back.

 

I'm sorry if I gave the impression I've got 100% confirmation of everything I've said, but if I sounded confident in my last post it's because the results of my 5 minute test agreed with what I'd assumed.

If I set my FFB Mult to 0.5 to lower the force of my wheel, for whatever reason, I wouldn't expect to get a jolt at 100% force because of a specific event in the sim or because a car had its NominalMaxSteeringTorque set wrong. I'd expect a maximum force output of 50%, so I'd also expect all the calculated forces to be scaled to fit within that 50%.

The fact that this ISI comment says this:

Code:

NominalMaxSteeringTorque=11         // Maximum steering arm torque to effect force feedback strength

would, to me, indicate that any higher steering arm torque will not produce a higher force feedback force; it says it's the maximum torque that will effect(sic) the FFB strength. So I wouldn't expect that if my FFB mult is set to 0.2 but I manage to generate 55Nm of steering arm torque that I'll get 100% FFB strength at my wheel, I'd expect the effective torque input to still be 11Nm.

To answer your middle question, the line I've just quoted comes from a Devmode rtrainer.hdv I found in my old rF2 files. The file date is 12th Oct 2011.

Regarding testing, I prefer to reduce the chance of wishful thinking or placebo having an effect, so I avoid trying to feel 'small' changes. I know there are people willing to try one thing, make a change, and say it feels 'about x% different' (20? 10? less?) which makes me suspicious, even if I can't be sure they're mistaken. Hence, when trying to check whether NominalMaxSteeringTorque set a limit on the resulting force or was just a scale, I reduced it by a factor of 10 and offset the resulting increase in FFB by also reducing the FFB mult by 10. Big enough figures that if I couldn't feel any change I could be fairly confident I was wrong about it. As it happened the FFB did just about turn off completely, which appears to confirm what I'd been thinking.

I only did this test in devmode, and to be honest I'm not invested in this enough to bother taking different figures over to single player to test it there. My mention of NominalMaxSteeringTorque was in regards to Paul trying to check the effects of STS on FFB (gotta love acronyms, eh?!) by turning the wheel at speed to generate forces approaching that nominal torque figure. Why make it so hard to approach the limit and potentially introduce inconsistent testing methods when you can just move the limit closer?

As for changing the NominalMaxSteeringTorque as part of an FFB guide, that doesn't seem a great option for the end user as it's part of the mod, not a player.JSON or controller.JSON setting. It's up to the modder to set it to a reasonable value, which is a whole other can of worms entirely.


To sum up: I'm only relating the results of the test I outlined which appear to confirm what I'd suggested. I could well be wrong, but I prefer to test things rather than make assumptions where possible.

That leads me to a question for you: if you wrote a guide partly under the premise that setting the FFB mult below 1.0 would still produce full force at the wheel in the right circumstances, what testing had you done to confirm it?

As Marc has implied, if you want something specific to the thread topic you'd have to go back several pages...

 

That is a very fair question.

I based it on my own observations (at the time) of the effects on ffb wheel torque output with changes to the car specific ffb multiplier through my hands (so not exactly a perfect/scientific measurement device by any standards ofc). I distinctly remember the max strength of the physical torque output from my ffb wheel remaining the same when the rf2 ffb was clipping (i.e. maxing out based on TechAde's plugin bar) regardless of the car-specific ffb multiplier. Could i have observed this wrong and the max torques not be exactly the same? Sure, that is quite plausible. It's just that I'm finding it hard to accept that i could have gotten it so wrong that i did not notice the max torque output from my ffb wheel becoming halved when the ffb multiplier was also halved. It's the sort of observations that you'd think no one could make a mistake on.

To be off by 10%, even 20% in judging the change (if any) in the max torque output from my wheel when lowering the ffb multiplier from 1.0 to around 0.5, i can imagine being completely plausible. But to be 50% out....it's just such a large margin of error.

So from ffb multi 1.0 to 0.5 (whilst keeping NominalMaxSteeringTorque constant), when in the game and you get constant ffb clipping occurring, the strength of the torque output from your ffb wheel literally felt like it's been halved to you?

finish edits.

 

I just realised as well, if my interpretation of how ffb multi affects the ffb output was incorrect, then my guide should not work in the first place yet i have my own personal experience that it works like so and many other people who have followed the guide have also confirmed so.

For example, if the default ffb multiplier of 1.0 is causing ffb clipping issues within the racing envelope (i.e. not driving over kerbs, gravel or crashing), then by lowering the ffb multipler, at some value i am able to completely avoid encountering ffb clipping within the racing envelope. But that doesn't seem compatible with your own explanation of how ffb multi works that you concluded based on your observation.

Whether the max torque output possible from the ffb wheel remains the same for different ffb multipliers or not is a separate issue though and like i've already said, it's certainly possible I'm wrong about that (despite my doubts as explained in my previous post). But i cannot accept that the other part of my observation on how ffb multi works being incorrect. It's the bed-rock of the guide and no one has commented that changing the ffb multiplier does not work as how i described in the guide.

 

Sorry to pull you into this Lazza. I know you are/were only trying to help.

I'll try to fetch my wheel tomorrow out of the storage center and try the tests myself.

 

The never ending story. Did i already mentioned that a dev could help, and save a lot of time. I think i did.

Because it doesn't happen, here another one speculating.

I see the STS just as an feature to adjust the response curve, in other words it just shifts the sensibility.

Than what i already said in another context is , that the software can't know the limit of your hardware without any sort of checkback signal, the would report the charecteristic to the software.

So because of this i don't understand how it means the STC would do nothing. Of course it can't increase the capability of the hardware, but it sets a range, in other words, it tells the software that the steering wheel is capable of any torque you set, no matter the real physical power of the steering wheel plugged in.

The result should be a differnet torque charecteristic, which also would change the torque increasing rate at all, while the car multi and the real physical capability of the steering wheel sets the max torque and caps it at the limit of the absolut max, the is set trough the car multi or the steering wheels max limit, but the torque charecteristic should be changed still, because the software still don't knows the limit, so it scales the charecteristic into the range the is set by the STC.

This is how i understand it and it works well for me. I never felt more real before, as it is with my personal ffb balance, and even if i'm wrong, it works, and that matters.

 

Here's some data comparing SteeringShaftTorque and FFB output at Steering Torque Sensitivity of 0, 1 & 2. I've included the raw csv files in case anybody wants to do their own analysis.

Note that I edited the HDV and set NominalMaxSteeringTorque=6.0 as during an initial test run I wasn't seeing steering torque go much above 6.0Nm. It was set at 9.5Nm but that meant we weren't really using all the available range with a FFB multiplier of 1.0. This can be seen in the data as clipping in the FFB output when Steering Shaft Torque > 6.0Nm. FFB multi was set at 1.0.

If we look at the data when we're in the middle of the steering torque range we see that for a steering shaft torque of ~3Nm STS of 0 produces ~25% FFB output, an STS of 1 produces ~50% FFB and an STS of 2 produces ~75% FFB output.

When we look nearer the upper limit of 5.5Nm we see an STS of 0 produces ~85% FFB, an STS of 1 produces ~92% FFB and an STS of 2.0 produces ~99% FFB. Again this is entirely consistent with DrR1pper's graphs i.e. as steering torque approaches NominalMaxSteeringTorque the difference between STS 0, 1 & 2 becomes less and less, until both are producing an output of 1.0 for an input of 6.0Nm.

This suggests that the steering torque sensitivity curves are applied as per DrR1pper's graphs i.e. in the same way as the driver input sensitivity curves. From this data I can't tell for sure whether the curve ends at NominalMaxSteeringTorque (as Lazza suggests), however this would make sense.

STS=0:
View attachment 16189

STS=1:
View attachment 16188

STS=2:
View attachment 16187


Edit: It wasn't obvious which attachment was which, so I've added them inline.
Edit2: And the forum has resized the images down to 1920 width from the original 3840, making them illegible even when zoomed in.
Edit3: I've put the full size images on dropbox:

STS=0 https://www.dropbox.com/s/oelgjwpx7636ez0/STS_0.0.png
STS=1 https://www.dropbox.com/s/4nrkd57sviw6cdf/STS_1.0.png
STS=2 https://www.dropbox.com/s/gdeyymcjuudsfm9/STS_2.0.png

 

From what I can see NominalMaxTorque is used to scale the steering shaft torque into the range -1 to +1.

Paul, the FFB output is filtered by the default 9 samples, I think that's why the FFB appears to lag behind during fast gradients. The figures I took were from areas where the filter was having little effect.

Sent from my SM-G900F using Tapatalk

 

Quick note Paul, the steering torque and FFB traces aren't scaled the same in my plots. That was a bit silly of me, will post properly scaled plots tomorrow.

Sent from my SM-G900F using Tapatalk