PDA

View Full Version : Temporary solution for FFB BUG: Logitech DFGT (possibly G25/G27 too)



bmanic
02-07-2015, 00:24
UPDATE!: PC patch 2.5 seems to have fixed the FFB issues with Tire Force and such. I have not yet had the chance to verify this myself but the initial reports from people are very good. I suggest starting out trying the "Classic" FFB preset found in the Force Feedback Calibration tab in the wheel preferences. Then tweak from there.

I'll leave my original post here still for reference but keep in mind that it is not perhaps relevant any more as the bugs are fixed!

UPDATE 2: Indeed. FFB is now fixed so this thread is irrelevant from this moment on. FFB is back baby!! And it is awesome! :-)



--------------------------------------------------------------------------------------




Bug symptoms: With a completely fresh install and properly calibrated wheel the default FFB settings for the Logitech DFGT (and most likely all other wheels) causes massive FFB clipping. This will feel as a strong "centering spring" and a complete loss of any fine details while cornering.

Bug description: Something is weird with the way the game scales the FFB forces to the wheel driver. It is impossible to completely prevent clipping, no matter how hard you apply Soft Clipping or Relative Adjust. Even when setting Soft Clipping half-input to 5.0 and Soft Clipping full output to 5.0 at the same time you'll have clipping at default Tire Force values and default car setup spindle master scale values. This was NOT the case before. With Soft Clipping half-input at 0.5 and full output at 2.50 I had ZERO clipping before, no matter how hard I hit the system. Now it clips instantly with these same values. Relative Adjust is similarly affected. It is now impossible to prevent clipping with Relative Adjust, no matter what settings you set.

I also noticed that the FFB clipping metering does NOT respond to overal FFB strength setting the way it used to. If you set overall FFB strength to 5 (it defaults to 100 on the DFGT, as it should) you'll still have a meter that goes all the way from top to bottom. This is different behavior than it was before. It used to properly show the actual FFB output levels. If you had FFB strength at a low value the metering would scale accordingly.

Special note: This temporary solution will not work for Xbox One or PS4 users because console versions of pCars have the values for controller settings at a way too coarse setting. It is for instance impossible to set a value of 0.2 for the Soft Clipping (half-input) parameter. Hopefully the developers realize what a huge problem this is and will fix it accordingly. Heck even for PC users the 0.1 "step size" for the Soft Clip (half-input) is way too large. It should be 0.01 resolution.


Temporary solution (follow my instructions exactly to the letter or you will have a very different experience):

1) Step one, check your logitech control panel settings

Here is my Logitech DFGT control panel. I highly recommend turning off the "Allow game to adjust settings" as it has historically always proven to be a problem with Project CARS. I have NOT verified if it still causes issues but I prefer to rather be safe than sorry. Basically what it does is that it forces the game to use my settings and can thus not apply any spring nor damper effects no matter how much it wants to.

210298

You can experiment with the Overall Effects Strength parameter. I've found 98% to be the most linear for my DFGT (measured with that FFB linearity app found at the iRacing forums). A classic setting for G25/G27 is to set it at 107%.

Set the rest of the values to zero! Make sure you have 900 degrees of rotation.

2) Step two, RESET and re-calibrate your wheel and pedals within the game (hit the big RESET WHEEL button in the lower part of the screen in the controller preferences)

3) Step three, make sure your FFB Strength is set to 100 (DFGT wheel preset defaults to 100) and FFB damper saturation set to zero (it should default to zero)

4) Step four, time to change the actual FFB settings that mater (from the FFB calibration tab in the lower right corner of the screen)

Here are the values that I've found as a solution. They do indeed look "crazy" but trust me.. they work as a temporary solution.

210299

210300

Yes, you are seeing it correctly. Set your Tire Force to a value of 4 (it defaults to 100). It needs to be specifically 4 so that you can drive all the cars with their default setups without clipping, while enjoying the benefits of Soft Clip (half-input) which basically can be used to boost low level FFB information. NOTE!! Make sure you TURN OFF Relative Adjust. Set all the Relative Adjust values to zero.

I've tried about 10 different cars with various amounts of down force and none will clip at a value of 4. If you set it to 5 you will have some clipping with some cars. However, at a value of 4 some cars will feel a bit too light in their FFB but this can easily be remedied within the actual car setup. Bump up the Spindle Master Scale of any offending car. For instance with the Ford Capri group 5 car I put the Spindle Master Scale all the way up to 60 until it started clipping.

You should also be able to use Jack Spade's FFB tweaker files. In this case you may want to experiment with a Tire Force value of 5 or even 6, depending on how hard his FFB settings are hitting the soft clipper. Do keep an eye on the FFB clipping meter though. You should NOT have full clipping while cornering hard, this is the main thing to avoid.


Known issues (and why it is only a temporary solution): The overall FFB strength will feel quite low. There is nothing that can be done about this until the devs fix the soft clipping so that we can hammer it hard with low level forces without causing clipping. The good thing is that you'll have to learn to drive without a clipping FFB and thus you will learn to "listen" to the wheel better which will ultimately help you drive faster laptimes and more consistent laps. Furthermore, I am not 100% sure these settings work for the G25/G27 as the FFB dead zone in the middle of those wheels is a lot bigger than on a DFGT. However, you can try to remedy this by setting a more aggressive Deadzone Removal value. You can probably safely set a value of 0.15 or even more. If the wheel starts rattling too much you can set the Deadzone Removal Falloff to 0.03 or more.

On a positive note: If we wouldn't have these awesomely flexible FFB settings this bug could have ruined the whole game for a very long time. Due to the extreme flexibility of the settings we can now keep on enjoying the game. On the other hand, this kind of bug should simply not be able to get through testing. Oh well.. c'est la vie. :)

Cheers!
bManic

beezley
02-07-2015, 01:01
Nice work bmanic :joyous:

Will test this out tmrw on DFGT....

I have been twiddling today and found some good settings that I think are as good as pre 1.4 patch....

Logitech Profiler exactly the same as yours apart from overall effects strength 100%, allow game to adjust settings ticked.

In game Everything is Default apart from;

Tireforce 18 (You may have to adjust this a click or two either way depending on how strong you like ffb)
Per wheel movement -0.01
Deadzone removal 0.06

then using Jack Spades files

will post tmrw and let you know how I get on with your settings.

the other improvements in the game are coming on tremendously, probably is not said it enough, but is still appreciated all the same, well done all at SMS :)

thx beezley :cool:

poirqc
02-07-2015, 03:12
Furthermore, I am not 100% sure these settings work for the G25/G27 as the FFB dead zone in the middle of those wheels is a lot bigger than on a DFGT. However, you can try to remedy this by setting a more aggressive Deadzone Removal value. You can probably safely set a value of 0.15 or even more. If the wheel starts rattling too much you can set the Deadzone Removal Falloff to 0.03 or more.


From what i understand, Deadzone removal range and falloff works in a ratio. 2:1 range/falloff to be precise. If i put something else, i got massive clipping.

On my G27, i have a range of .16 and a falloff or .08(Keep it mind i'm not talking about the other settings, i'm only talking about Deadzone Removal by itself)

You have to increment the setting with the arrow one by one. If you move the slider with your mouse, you can be between, lets say, 0.081 and 0.089 and it will only show 0.08. What i did was that i moved the slider in the 0.07 range and then incremented with the arrows till it ticked to 0.08.

Hope it'll help.

resmania
02-07-2015, 05:16
DZR set to 0.8 and it works perfect for me. thx for the great thread.

rangerman65
02-07-2015, 07:22
DFGT user here. Thanks bmanic used your suggested settings exactly and Jack Spade's 66%soplateral. Perfect.

Kobus Maree
02-07-2015, 07:46
It feels similar to default settings with
DZ removal 0.18
softclip - both at 5

I tried yours and it feels good thanks
DZ at 0.17
falloff at 0.02
rest as recommended

cheers

Kobus

kurtie
02-07-2015, 09:18
I've also seen that turning off the "Allow game to adjust settings" seem to improve FFB on my G27. What exactly does that checkbox? I mean... you configure some values in logitech profiler, and then you have a checkbox that says that those values will be modified by the game? I find that part of Logitech profiler confusing... I mean: you have a global settings with a "allow game to adjust settings" checkbox, and specific game settings with "allow game to adjust settings" checkbox... if you check any of those checkboxes does it mean that logitech profiler is bypassed and is the game who decides the values to be used?

Regards,
Kurt.-

beezley
02-07-2015, 13:25
I tried your settings bmanic and now I can't decide what I like best haha :D It is nice to have options lol

I feel your settings have more dynamic, but less of the slip angle effect in the settings I have posted (maybe because I have less heavy settings?) with mine you can definitely increase tyre force to 20 maybe even more. I do not like a heavy wheel so my settings on the tyre force maybe are a bit conservative.

Here are few pics of the ffb monitor at the Parabolica with my settings :-) Thanks a lot for your settings btw :cool:

210335210336210337210338

bmanic
02-07-2015, 13:49
I've also seen that turning off the "Allow game to adjust settings" seem to improve FFB on my G27. What exactly does that checkbox? I mean... you configure some values in logitech profiler, and then you have a checkbox that says that those values will be modified by the game? I find that part of Logitech profiler confusing... I mean: you have a global settings with a "allow game to adjust settings" checkbox, and specific game settings with "allow game to adjust settings" checkbox... if you check any of those checkboxes does it mean that logitech profiler is bypassed and is the game who decides the values to be used?

Regards,
Kurt.-

The check box forces the game to use the settings you provide. If you tick that check box the game can freely alter the Strength, Spring and Damper FFB settings of the wheel driver. I always have this setting unticked (except in Assetto Corsa as it uses these additional effects for it's FFB so make sure you have it ticked for that game).

bmanic
02-07-2015, 13:54
I tried your settings bmanic and now I can't decide what I like best haha :D It is nice to have options lol

I feel your settings have more dynamic, but less of the slip angle effect in the settings I have posted (maybe because I have less heavy settings?) with mine you can definitely increase tyre force to 20 maybe even more. I do not like a heavy wheel so my settings on the tyre force maybe are a bit conservative.

Here are few pics of the ffb monitor at the Parabolica with my settings :-) Thanks a lot for your settings btw :cool:

210335210336210337210338

Those look good! No clipping at all. It seems that by using Jack Spade's tweaker files + a very low TF you have no clipping and that is very good. It's the only way of getting rid of the problem, to make sure it is never clipping while cornering. My settings are using (or abusing, depending on how you look at it! :) ) the soft clip half-input to force lower level forces into the wheel. This way you get to feel a lot more low level forces at the expense of overall dynamic range (which explains your slip angle observation).

Keep in mind that I usually tweak Fy in each individual car to be a lot lower than any of the other forces because it is dominating the FFB. I know Jack Spade does this too in his tweaker files but he then also adds additional SoP so that you can feel the rear wheels and I never use that. It's possible that you have better settings for using Jack Spade files without clipping. The main thing is to not have clipping so use any means necessary to achieve that.. then it's down to personal taste how you set the rest. :)

konnos
02-07-2015, 14:21
Profiler at 100% FFB, the rest at zero, 900degrees and tick on Allow game. In-game all settings to default (Jack 66% tweak). As suggested in the other thread about g27, i reduced Relative Adjust Clamp to 0.32 which seemed decent enough for me. No clipping and decent force on the wheel. If you want a lighter wheel reduce RAG by a tiny bit every time. I tried Z4, porsche GT3 on Oulton they seemed to be pretty good. Again, default settings otherwise, so TF100 and FFB100 etc

bmanic
02-07-2015, 14:27
Can you post a picture of the FFB meter while cornering? It's possible that Jack Spade files are enough to reduce the tire force to below clipping but I remain skeptical until I see some proof. Your settings seem awfully high. I'll try it myself once I get a chance to play again.

konnos
02-07-2015, 14:42
210353210354210355210356210357

It can probably use a little less of RAC, but it feels decent on Oulton, i feel i might be missing some dynamic forces in there. Not sure on the deadzone, i could use a little less in the centre.

Sir_Stig
02-07-2015, 17:51
I'm getting similar results with RAC of 25 and RAG at 1.2. I need to play a bit more to see if I can do deadzone removal at all, but I get no clipping that I can see with these settings, while having a much more active FFB than I get from lowering TF.

David Semperger
02-07-2015, 18:38
Just here to report that the FFB in the 2.0 patch made all cars completely undriveable with my 911 Carrera wheel as well. Really heavy centering in-game, none in menus.

Sampo
02-07-2015, 19:42
David, have you tried deleting your profile after 2.0? And is the FFB still heavy after that?

David Semperger
02-07-2015, 19:48
I haven't and it would be pretty bad if we had to resort to that, since our career progress is stored in the profile as well. However, Bmanic's settings help me quite a lot as well, although it's very obvious that they only mask the underlying issue, because no matter what I do I can't get FFB detail back to where it was before 2.0.

Sampo
02-07-2015, 19:52
Yes, it would be bad. Could you still make a backup copy of the profile and try with a fresh one and report how it feels?

Edit: Oh and if you try it, disable cloud saves in steam before doing it.

David Semperger
02-07-2015, 20:12
Doing a full profile reset doesn't help in my case.

FeFernandes88
02-07-2015, 20:15
These settings work well in the G27 ? Someone already tested ?

Sampo
02-07-2015, 20:16
Just tried it again with a clean 2.0 profile and changed TF between 10 and 20. It almost feels like that at some points in the FFB calculation TF is considered as 100 and at some points as 10 or something like that. The tyre force gets heavier but the fidelity disappears.

Sir_Stig
02-07-2015, 20:35
Try out the settings I'm using, let me know what you think?

FeFernandes88
02-07-2015, 20:48
You can tell me yours?

FeFernandes88
02-07-2015, 20:49
You can tell me yours Sir Stig??

Sir_Stig
02-07-2015, 21:20
I can confirm that 0.25 relative adjust clamp and 1.20 relative adjust gain with 0 deadzone and 66% jack spade makes the ffb so so much better on my g27
210179210180210181210182210183210184

I started on pCARS with a DFGT and upgraded to a G27 a week ago, and like many people have been searching for the holy grail of reducing the heaviness of the wheel whilst still retaining road feel.

After an enormous amount of thread lurking, guide reading, trial and error and rage quitting, and very little posting, I think I have found the setting which helps most for my setup:


Reducing Tire Force lightens the wheel but diminishes most of the other FFB.
Reducing Relative Adjust Clamp lightens the wheel yet retains most of the other FFB feel.


Lowering Relative Adjust Clamp to 0.20 - 0.30 on a freshly reset/recalibrated and otherwise default FFB setup immediately made a massive improvement across the board. The wheel feels lighter yet more informative and responsive. With HUD telemetry you can see that the FFB trace has significantly more "wiggle" than that produced by lower tire force. More importantly, it comes nowhere near clipping, thus preventing the horrid Logitech rattle! I adjust Relative Adjust Gain to 1.20 which brings the FFB back up, but still within limits.

Adding Jack Spade's 66% SOP lateral tweaks to the mix further improved the liveliness of the FFB for me.

If someone can tell me why I might be throwing the FFB baby out with the bathwater with these settings, I'd be all ears. I greatly appreciate SMS including so many ways to adjust settings, but FFB Wrangling seem to be an arcane art not helped by cryptic/non-existent parameter descriptions and some apparently contradictory opinions on their interactions. Many many thanks to those on this forum who have helped contribute to improving the largely unenjoyable defaults.

PS If someone knows of a way to reduce the "clunk" as the wheel moves past 12 o'clock when up to speed, that would be appreciated too!

Robski
02-07-2015, 22:09
Tried all the fixes mentioned here, DFGT still feels lifeless and missing effects in certain situations. FFB has been a massive issue for me for years in pCARS, I have always said it's pretty underwhelming, got shot down countlessly... The only way to fix it is... well, re-do it from scratch, but that won't happen now, of course. Think I'm done with this game.

poirqc
02-07-2015, 22:35
I did some tests about Deadzone removal. I still think the 2:1 ratio is needed. However, i like some put a little bit less Range then wheelcheck.exe tells me.

Right now, with my G27, i get 0.16 with wheelcheck.exe. However, it seems this will flaten some tarmac bumps feeling. If lower it to Range 0.14 (Falloff 0.07) I feel them back without getting too much deadzone.

Hope this helps,

Sampo
02-07-2015, 23:17
Just tried it again with a clean 2.0 profile and changed TF between 10 and 20. It almost feels like that at some points in the FFB calculation TF is considered as 100 and at some points as 10 or something like that. The tyre force gets heavier but the fidelity disappears.

Quoting myself, but after another clean profile, I think I found settings that I like. TF 30-35, DRR 0.08 and DRF 0.01. I just had too much DRR before I think. Felt pretty good for me. It's still strange that I used TF 100 before and now I have to use a third of that to get the same result.

Edit: Really nice feeling for F Rookie and RUF GTR-8 on WG Short so far. I take back everything I said in previous posts :)

nicardy
03-07-2015, 00:05
Quoting myself, but after another clean profile, I think I found settings that I like. TF 30-35, DRR 0.08 and DRF 0.01. I just had too much DRR before I think. Felt pretty good for me. It's still strange that I used TF 100 before and now I have to use a third of that to get the same result.

Edit: Really nice feeling for F Rookie and RUF GTR-8 on WG Short so far. I take back everything I said in previous posts :)

Sorry for the silly question, but just quickly whats the best way to delete the profile ?.

Sampo
03-07-2015, 00:36
I think you'll be fine with just resetting the contreller settings, but if you really want to delete your profile (which includes all your settings and your career), delete this file:
\Steam\userdata\[your userid]\234630\local\project cars\profiles\default.sav

You can also take a backup of that file before deleting it. Remember though that the steam cloud functionality can restore the previous file for you. Better uncheck the steam cloud in the steam game settings before doing anything with the file. I never use the steam cloud myself. I take backups when I deem it necessary.

sherpa25
03-07-2015, 06:36
I can confirm that 0.25 relative adjust clamp and 1.20 relative adjust gain with 0 deadzone and 66% jack spade makes the ffb so so much better on my g27

W/ my G27 (Profiler, TF, FFB = 100, Deadzone = 0), using your settings and @konnos, I find both still a little too heavy (at some spots) . How would I lower it if ever?
And what deadzones are you both using, BTW?

N0body Of The Goat
03-07-2015, 06:47
I'm getting similar results with RAC of 25 and RAG at 1.2. I need to play a bit more to see if I can do deadzone removal at all, but I get no clipping that I can see with these settings, while having a much more active FFB than I get from lowering TF.

I tried those setting with Tyre Force back to 100 and an earthquake hit the south coast when using the Formula A! ;)

Back to 10 TF for now for this car, ~15-20 for others.

N0body Of The Goat
03-07-2015, 07:50
Bmaniac, tried your settings, to maximise FFB without clipping in the Formula A @ Spa, I just had to double Tyre Force to 8 on my Fanatec. :)

Thanks!

konnos
03-07-2015, 10:48
woops double post.

konnos
03-07-2015, 10:49
W/ my G27 (Profiler, TF, FFB = 100, Deadzone = 0), using your settings and @konnos, I find both still a little too heavy (at some spots) . How would I lower it if ever?
And what deadzones are you both using, BTW?

I havent tweeked my deadzone removals yet, i will play again soon. If you want it to become lighter, reduced Relative Adjust Clamp more, an increment at a time, don't go too low at once, until you like it. I think i will lower it too.

bmanic
03-07-2015, 18:29
I did some tests about Deadzone removal. I still think the 2:1 ratio is needed. However, i like some put a little bit less Range then wheelcheck.exe tells me.

Right now, with my G27, i get 0.16 with wheelcheck.exe. However, it seems this will flaten some tarmac bumps feeling. If lower it to Range 0.14 (Falloff 0.07) I feel them back without getting too much deadzone.

Hope this helps,

I'm not entirely sure if the dead zone removal range is actually measured in percentage.. so it's perhaps not possible to use values from wheelcheck directly. Would be nice if AJ could confirm or deny this.

bmanic
03-07-2015, 18:31
Tried all the fixes mentioned here, DFGT still feels lifeless and missing effects in certain situations. FFB has been a massive issue for me for years in pCARS, I have always said it's pretty underwhelming, got shot down countlessly... The only way to fix it is... well, re-do it from scratch, but that won't happen now, of course. Think I'm done with this game.

I call BS on this.. unless you have a dead DFGT or just a very different idea of "lifeless". If you followed my directions to the letter and then set Fx and Fz for any car to maximum you'll have a seriously bumpy and lively ride, even when compared to iRacing at it's most bumpy track (something like sebring).

.. and judging by your comment in general, I assume you've already made up your mind and never even bothered trying the settings. Your loss.

poirqc
03-07-2015, 18:57
I'm not entirely sure if the dead zone removal range is actually measured in percentage...

For what it's worth, i tried 0.16 DRR and 0.04 DRF and it wasn't good.

I did try 0.15/0.075, 0.14/0.07, 0.12/0.06 and 0.10/0.05 and they all behaved linearly. They resulted in greater deadzone but they never induced clipping.

Hope this helps,

nicardy
03-07-2015, 20:31
I think you'll be fine with just resetting the contreller settings, but if you really want to delete your profile (which includes all your settings and your career), delete this file:
\Steam\userdata\[your userid]\234630\local\project cars\profiles\default.sav

You can also take a backup of that file before deleting it. Remember though that the steam cloud functionality can restore the previous file for you. Better uncheck the steam cloud in the steam game settings before doing anything with the file. I never use the steam cloud myself. I take backups when I deem it necessary.

Thanks mate, I think it would be worth as I've still got the in helmet view flying above the track !

poirqc
03-07-2015, 21:39
I forgot to mention, and some forum members pointed it out too,

The G27 deadzone is directly tied to the FFB overall strenght of the profiler.

100% ~ .16 DRR / 0.08 DRF
107% ~ .06 DDR / 0.03 DRF ( that one is vague, i get value between 5.5% and 7%)
110% ~ .04 DRR / 0.02 DRF

But anyway, you get the point, if you move the Logitech overall force value, you have to get adjust the DRR accordinly(and falloff).

Also, thanks to dw123 for the graph (http://isiforums.net/f/attachment.php?attachmentid=13556).

sherpa25
04-07-2015, 00:21
I havent tweeked my deadzone removals yet, i will play again soon. If you want it to become lighter, reduced Relative Adjust Clamp more, an increment at a time, don't go too low at once, until you like it. I think i will lower it too.

Yeah, lowering it helps, thanks. Will experiment with DRR/Falloff. So far better, except my only issue are the 'gear rattles' that now seem more pronounced, even with DDR=0. Of course only on bumps/kerbs, but these didn't seem to be as pronounced as before. Do you have this as well? Perhaps there's a setting that could greatly reduce this on kerbs w/o sacrificing the other FFB? Or could it be because I'm using Jack's 66%? Will test others.

poirqc
04-07-2015, 02:21
First of all, i want to tank bmanic, dw123 (http://forum.projectcarsgame.com/showthread.php?32999-FFB-Ready-to-give-up&p=1021203&viewfull=1#post1021203) and Photonenbert, the guy who wrote the FFBguide (http://forum.projectcarsgame.com/showthread.php?29907-Force-Feedback-Guide-for-Project-CARS&p=969970&viewfull=1#post969970) and everyone who's helping understand this beautifull FFB system!

This is only my understandings of how it apply to the G27, i still have to learn and understand much. At some point, if we find enough, this would apply to other wheels too, the values would just be different. If we understand the maths, getting the data will be easy. On a side note. It seems that for a lots of settings, when it's at 0, the seting is simply disabled.

I'll try to explain my findings and the thinking behind the settings. In my view, there's a linear way to calibrate the wheel. This part is not about setting the feeling for this or that, but more about letting the wheel output the best range it can. At that point it should be linear. I didn't mess with Jack's files. I figured the default car settings should be ok. It will probably be handy at some point. All the tests i did were at Doninghton National with a Ginetta Junior. I know this track by heart now and i figured this small car can be easily pushed hard while looking at the FFB graph. :D

In steps (https://youtu.be/ay6GjmiJTPM?t=164), the setup should be:


Configuring the logitech profiler
Dealing with Deadzone
Removing wheel drag(Per wheel movement)
Dealing with signal shaping(Clipping removal)



Configuring the logitech profiler

This part will directly affect the deadzone one. Big credit to whoever made that graph! dw123 posted the original link (http://isiforums.net/f/attachment.php?attachmentid=13556),
210590

The profiler config is in the upper left part. As you can see, 100% force is clean and linear but it starts late to respond. 110% on the other hand, start faster and is linear enough to be used. Witch leads us to the other point.

Here's come Wheel Checker. Well, if you read the graph closely, you won't really need Wheel Checker.

Dealing with Deadzone

As you can see in the graph, the deadzone of the wheel is directly tied to the overall effects strenght. 110% gives us only a deadzone of 4%. This gives us a Deadzone Removal Range of 0.04. With the help of the FFB and dw123, as i posted elsewhere, it seems to me the DRR and DRF works in tandem with a ratio of 2:1. I did a couple tests and it seemed to be linear. This gives us a DRR 0.04 and a DRF 0.02**

**You can't really miss the DRR value but the DRF value needs to be put precisely with the arrows. DRF 0.02 to 0.029 will always show 0.02 only.

Removing wheel drag(Per wheel movement)

This part is handled with Per Wheel Movement, Per Wheel Movement squared and wheel smoothing. Like the deadzone values, i extrapolated some from the one in the FFB guide. This lead me to:

Per Wheel Movement: -0.04(-0.02)
Per Wheel Movement squared:0.02(0.01)
wheel smoothing:0.03

Even with small values, there's less drag right away. The first reason to mess with this is the remove the drag(input lag) from the wheel so it respond accordingly. A side effect of it is that small signal are easier to feel.

Dealing with signal shaping(Clipping removal)

Tire Force
The first part is to lower the value of Tire Force, as a starting point, depending on the usage or Soft Clipping or Relative adjusts. The theory here is that we shouln't use both at the same time.

Relative Adjusts and Soft Clipping
I mostly copied values from others as i didn't have enough time to mess with those. Howerver, bmanic soft clipping values or pCars default Relative adjusts are both good starting points.

Personnaly, with the settings i had access to, i prefered Relative Adjusts. I find it a little bit more dynamic. A carefully tune Soft Clipping could surely be as effective. Who knows.

What i can say is that Relative Adjust Bleed could help with rattling. When you put more of that, the FFB signal takes more time to decay. Too much of it and the signal will stick. There's more answers in the FFB guide.

Scoop Knee and Reduction
Normally this should be used with a high Tire Force, to lower the feeling of low FFB signal is too present. Right now, for the G27, it seems Scoop Knee and Scoop Reduction should be set at 0 for both. If we check back the graph you can see that naturally, the G27 doesn't respond to small forces. There's no need to lower those forces. Personnally, i like to feel the road surface and turning this off seems to help.

With all this, i like to down the sensibility a bit to 48 overall and the deadzones to 2(will save as 1) in the control section.

P.S.: It's late, i'll try find mistakes tomorow morning! :D English isn't my first tongue.

Hope this helps!


TL;DR - Since 2.5, those value are a bit different.

Profiler strengh 110%

Tire Force 26
PWM -0.04
PWMS 0.02
WPS 0.03
DRR 0.04
DRF 0.02
LS 0
LS 1.0
LD 1.0
RAG 0.98
RAB 0.21
RAC 0.96
SK 0
SR 0
SCHI 0
SCFO 0

VictoriSV
04-07-2015, 11:30
I'm sorry, but I think you are wrong in some statements of your last post.

When you adjust the scoop parameter or any other you don't have to think about the dead zone of your force feedback any more if you set it correctly, the game will take it into account. Now, as you can see there are two linear parts in the g27 force, one slope is bigger than the other. At 100% force in the profiler the "division" or knee is around 40% of the force, and at 110% force it is situated around the 20%.

So, what does this mean? If you want to have a, more or less, linear response with this wheel you need to lower the first slope and make them just one.

Now the important part, to do this simply set in the first parameter the position of the knee related to your force in the profiler, if it's 100% in profiler, put 0.4 here as the knee is at the 40% output force. The second parameter, scoop knee reduction, reduces the forces below the knee by that proportion. You have to trace the ideal lineal progression of the force from the end of the dead zone to the maximum output force and apply this factor to the value of the knee to set it to the ideal value at that point. In my case the value of the knee was 1800 and the ideal value to make a linear progression was around 1250, so dividing 1250 between 1800 you get about 0.7, a 0.3 reduction rate, that's the value you should put to this parameter.

Another thing is the part of the drag reduction, I tested you values and they don't work well in my g27. In the FFB guide tells you that to test this you have to set master scale to 0 in the car's specific FFB and go to the track. You should notice a light wheel and no drag but keeping it without accelerating when turning, and that's exactly what happened to me, when I turn the wheel it continues turning even faster.

Once all these things have been said, I ended setting the parameter like this:

Scoop knee: 0.4
Scoop reduction: 0.3
Per Wheel Movement: -0.01
Per Wheel Movement squared:0
wheel smoothing:0.01 or 0.02

Keep in mind that I use 100% force in profiler and please, let me know if you disagree with anything I've said and forgive my writing faults, English is not my first language :p

sherpa25
04-07-2015, 12:17
@VictoriSV, what are your values for the other parameters, other than those mentioned in your post? Also, how'd you get the value of the knee of 1800?

I'd like to test the values in my G27. Thanks.

poirqc
04-07-2015, 12:20
Thanks for your input VictoriSV! This is exactly the kind of discussion i was after.

Scoop knee

Your explanation makes sense.

Scoop Reduction

I would like to know how you got the values to calculate the Scoop Reduction. I don't understand where 1250 and 1800 come from. I think i understand with the tracing part. What do you think of this?
210634


Wheel Drag

I ended up with those value because i wasn't sure if i needed to test this while moving or standstill. The way i did it was using the outside cam, turning the car and looking at the wheel as i turned it.

How do you test it?

I agree with you that you can remove more smoothing if you remove less wheel drag.


What do you guys think?
Thanks

Edit - I tried your settings VictoriSV

Per Wheel Movement: -0.01
Per Wheel Movement squared:0
wheel smoothing:0.01 or 0.02

They were better on the cars i tested(Ginetta junior, McLaren F1, Audi V8, Caterham classic). The best case was for the Audi V8 road. With -0.04/0.02, i could feel the acceleration you where talking about. With your settings, i could see the wheel of the car was following mine. It seemed good drag wise. The Audi was also more predictable to drive.

I put the Scoop knee as you suggested. It was 0.22 to my 110%. Since i don't understand how to calculate it, i tried 0.17 and 0.13 Scoop reduction. I'll get back to that later. Smaller outputs were still there, they were just less bold than before. It felt natural.

We are getting close guys! I can smell it! Keep it coming!

VictoriSV
04-07-2015, 13:56
Thanks for your input VictoriSV! This is exactly the kind of discussion i was after.

Scoop knee

Your explanation makes sense.

Scoop Reduction

I would like to know how you got the values to calculate the Scoop Reduction. I don't understand where 1250 and 1800 come from. I think i understand with the tracing part. What do you think of this?
Click image for larger version.

Name: scoop.jpg
Views: 0
Size: 126.5 KB
ID: 210634


Wheel Drag

I ended up with those value because i wasn't sure if i needed to test this while moving or standstill. The way i did it was using the outside cam, turning the car and looking at the wheel as i turned it.

How do you test it?

I agree with you that you can remove more smoothing if you remove less wheel drag.


What do you guys think?
Thanks

Those values of 1800 and so come from the value "deltaX", not from "deltaXdeg", there is no difference between the graphs that these forms, is just because I choose "deltaX". You can use any of the them, both are result of the force created by the wheel.
Relating to the wheel drag, the process is explained here: http://forum.projectcarsgame.com/showthread.php?29907-Force-Feedback-Guide-for-Project-CARS&p=969970&viewfull=1#post969970 (http://forum.projectcarsgame.com/showthread.php?29907-Force-Feedback-Guide-for-Project-CARS&p=969970&viewfull=1#post969970)
But answering your questions, you have to be moving and turn the wheel at different speeds.


@VictoriSV, what are your values for the other parameters, other than those mentioned in your post? Also, how'd you get the value of the knee of 1800?

I'd like to test the values in my G27. Thanks.

Tire Force: 20-25 (depending of you how close you want to be to clipping)
Per Wheel Movement: -0.01
Per Wheel Movement squared:0
Wheel smoothing:0.01 or 0.02
DRR: 0.18
DRF: 0.05
Scoop knee: 0.4
Scoop reduction: 0.3

Other values like Relative Adjusts and Soft Clipping comes to your tastes, but now I'm with the default ones. These are the values that Jack Spade recommends, you can test them if you want:

From left to right (A - D) - heavy compression ----> moderate compression - based on FFB = 100, Tire Force = 100

Relative Adjust Gain-----(A) 1.30 - (B) 1.20 - (C) 1.15 - (D) 1.10
Relative Adjust Bleed--------0.10 ------ 0.10 ------0.10 ------0.10
Relative Adjust Clamp ------1.15 ------ 1.10 ------1.03 ------0.95

Soft Clip (Half Input)---------0.50 ----- 0.80 ------1.10 ------1.29
Soft Clip (Full Output)--------1.39 ----- 1.29 ------1.20 ------1.10

I don't know how useful they could be because they are based on 100 Tire Force and since 1.4 you have to lower a lot the Tire Force to avoid "killer wheels".

I played a lot with these values and I never found a combination that makes me think "this is the one!" but I will tell you what I know.

The Relative Adjusts and Soft Clipping can help you to reduce/avoid clipping.
If you play with the Relative Adjusts you can get a signal with a lot of spikes and "spicy" or you can get a more smooth signal without to much "noise" or spikes in it, but maybe less responsive. It produces an effect of compression as well.
In the other hand, Soft Clipping something opposite to the scoop parameters, their objective is to compress the upper signal to give more room to play to the weaker forces and boost them. The cons are the non-linearity that results from the compression, specially on the stronger forces.
The full explanation is in the document attached to the other thread that I linked earlier.

I hope this can help to solve your doubts ;)

EDIT - Referring to how I calculate the Scoop Reduction:

In you graph you get a "deltaXDeg" value at the knee around 90 and you "ideal" graphs shows that the value at that point should be around 30, so you do 30/90 and you get 0.33, that it's to say a reduction of 0.66=(1 - 0.33) to archive the "ideal" value, and that's what you have to set in the Scoop Reduction.

Keep into mind that i have no means to check that the conclusions that serves of support to these calculations are true without no doubt, but this is what I've come after reading and testing since the game was released.

poirqc
04-07-2015, 15:51
I think you picked the 120% line to give my example. I'm currently at 110%

If i understand you correctly here it goes:

Edit - Forget about the part in italic. I just tried 0.22/0.65 and it was awefull. I obviously didn't understood the scoop part.
The knee is at 85 deltaXDEG. With the line i draw, i get a deltaXDEG of 30. This leads us to 30/85= 0.35 . By your calculation, it means a value of 1-0.35 = 0.65 Scoop Reduction. I'll write again to be sure.

Profiler at 110%:
Knee 0.22
Reduction 0.65

With a 100% profiler i get:

Knee 0.4
Reduction 1-(42/88)=0.53

With a 90% profiler i get:
Knee 0.42
Reduction 1-(41/82)=0.5

I had a look at my spreadsheet, from wheelcheck, i understand picking 1800 as a knee because steps get linear after that. What i don't get is picking 1250 to calculate the Reduction.

P.S.: I'm really enjoying this thinking process but i gotta go, i'll have another look tomorow

VictoriSV
05-07-2015, 09:44
Your calculation are OK. I was saying the numbers approximately because I couldn't see exactly witch were.
Don't bother about my calculations, I was checking them and I had some error with the numbers, but the principle is the same.

Usually are recommended 0.7 in Scoop Knee and 0.15 in Scoop Reduction, but is clearly that the Knee isn't at 0.7 in the G27 so... I'll continue testing my config ;)

bmanic
05-07-2015, 12:09
The games default value of 0.7 scoop knee and 0.15 scoop reduction are the values I used in my FFB settings that then became the default for the game (not sure why they were chosen because my settings only worked if you also tweaked each car specific FFB setting as I had them). My settings were entirely based on "feel" and not any kind of research nor wheeltester.

In fact, all of the current default FFB settings are based on the settings I mentioned in one of my early FFB threads. But like I said, they only made sense when combined with my per car settings and of course these default to 100 for fx, fy, fz and mz (my car settings were all heavily Mz biased which meant I had very different settings).

poirqc
05-07-2015, 14:45
Well, it seems 0.22 Knee / 0.66 Reduction isn't that bad after all. The time i tried it was with 26 Tire force. I wasn't sure how the reduction was applied and it made the well so much lighter going from 0.15 to 0.66. Afterwards, i incremented Tire Force to 43. This gave me an overall similar feeling to 0.22 / 0.15.

The problem i had with that is that some cars(Ford MK40, RWD P30) clipped a bit too much for my taste. From there i had to lower the Tire Force again to avoir that clipping but the wheel got a little too light.

For now i'll get back to a lower Scoop Reduction, and a lower Tire Force at the same time. I find it easier to manage 1 slider than a couple.

So in essence, a higher Scoop Reduction means a higher Tire Force? Now we need a way to find the balance.

Thanks

CCS86
06-07-2015, 18:37
Hmmm, not sure why my DFGT would be different, but with the 2.0 patch, swing my tire force that low nets almost no perceivable force feedback.

flyoffacliff
06-07-2015, 18:57
I just use the default settings, and reduce overall strength in game. All those settings look so complicated. Is default really that bad with G27?

I also never calibrated the wheel in-game. I thought the G27 does automatically when you turn it on.

OppaErich
07-07-2015, 16:00
I've reset my wheel again and applied the settings exactly like in your pictures. There's shaking of the wheel, even with GT3 cars. There's a steering resistance but I can't feel the car. I'm understeering, oversteering, anything without noticing. Briliant, DiRT Rally doesn't even start anymore and this one is for engineers only. No racing game to enjoy for me and I have a week off...

poirqc
07-07-2015, 16:39
The first thing you can do to chase the problem is start a practice run, turn on the telemetry and check if the FFB signal clip(yellow line going straight to top or middle of the graph.)

From there, you can start to get solutions.

Toxic
07-07-2015, 17:21
I just leave everything on default and reduce the master scale in each cars setup to between 10 - 14.

Helps me.

poirqc
07-07-2015, 17:26
Dialing the deadzone removal with the wheelcheck program is somewhat important. You can do without, but the wheel will be more precise with the correct setting. Keep in mind you have to tune the falloff depending of the range setting. a ration from 4:1 to 2:1 should give you nice results.

OppaErich
08-07-2015, 09:56
I have TF 0.1, DR 0.02 and overall FFB 82 now, there's still a tendency to oscillate but it's weak enough to live with it. The yellow FFB line never reaches the top or bottom.

bmanic
08-07-2015, 21:49
Hmmm, not sure why my DFGT would be different, but with the 2.0 patch, swing my tire force that low nets almost no perceivable force feedback.

Even with all the other settings exactly as I wrote them?

It is possible that the FFB bug happens only for a new profile. I reset my profile before starting to play patch 2.0 and that caused the bug. Apparently people who have profiles dating back to pre-1.4 patch are not affected. Soft clip and Relative Adjust are still bugging though, even with the old profile. TF values are the ones working correctly if you have an old profile.

In short: It's a weird bug related to the profile somehow.

konnos
09-07-2015, 08:30
bmanic, I wasn't entirely satisfied with your proposed settings for the DFGT, I would like a more lively wheel. Given your significant experience with the FFB system, can you help me out with my settings? I started off with only putting the ReltvAdjustClamp at 0.20 or so, like I described in a previous post. Everything at default values with 66% JackSpade tweak. The FFB graph is lively and doesn't clip, however, when forces appear, they re a bit too abrupt, like too forceful maybe.

If i didn't want to reduce TyreForce to eliminate those, what is the proper solution to adress this, without making the steering heavy?

I also tried a few values with Scoop Knee Reduction up to 0.30 (from the default 0.15), it was a little better but somehow the FFB died a little? Not sure if that was the case and not sure what the slider actually does... I also tried a bit of Soft Clipping Half. If i understand this right, it amplifies small values when it is set below 1.00 and it lowers them above 1.00? Not quite sure where to go with it and tampering with that didn't produce the results i was expecting, although the forces were restrained.

What do you propose I fiddle with more?

poirqc
09-07-2015, 11:00
You have to dial the deadzone first, you can use wheelcheck.exe to do a minimum Minforce test.


Scoop is used to change the shape of the signal a wheel can produce. It's a software help for the hardware limitation of wheel. Idealy, the forces a wheel should produce should be linear. Often, it is not.

Scoop knee is where the slope will bend(knee).

Scoop Reduction will lower the forces to try to get them more linear. When this is dialed, you can start to move Tire Force up to it's max level before clipping.

VictoriSV explained it better (http://forum.projectcarsgame.com/showthread.php?33699-Temporary-solution-for-FFB-BUG-Logitech-DFGT-(possibly-G25-G27-too)&p=1030178&viewfull=1#post1030178).

There's a bunch of posts there talking about that.

CCS86
09-07-2015, 13:51
Even with all the other settings exactly as I wrote them?

It is possible that the FFB bug happens only for a new profile. I reset my profile before starting to play patch 2.0 and that caused the bug. Apparently people who have profiles dating back to pre-1.4 patch are not affected. Soft clip and Relative Adjust are still bugging though, even with the old profile. TF values are the ones working correctly if you have an old profile.

In short: It's a weird bug related to the profile somehow.




I did. But I have also noticed what you have, where for the same given settings the wheel can behave differently. Maybe relating to resetting the wheel.

I went ahead and graphed my DFGT wheel force. 100% is a little bit digressive at the top, but overall pretty damn linear and has nearly no "knee" at the bottom. 60% is the most linear of the bunch, but introduces a knee.

What do you guys think?



211262

konnos
09-07-2015, 14:13
@poirqc: deadzone is fine, it s 0 for DFGT, so defaults are ok. I don't really understand that graph. xdeg etc i don't know what values those are, and they re for the G27 anyway. I ll dial the g27 settings in anyway, apart from deadzones and see how it is.

@CCS86: how do you do that graph? hope we can sort our DFGT out in this thread ;)

poirqc
09-07-2015, 15:02
Nice graph CCS86! the DFGT seems like a nice piece of a wheel!

I think you're better off setting the force in the profiler to 60%. From there you can disable the scoop knee feature because the force is almost linear. From the graph, you need a Deadzone Removal Range of 0.16. After that, you set the Falloff between 0.04 and 0.08, Lower value means a stiffer wheel with less understeer feeling. Higher means more understeer feeling but the wheel will get lighter.

If you remove Scoop settings, you'll need to adjust the tire force a bit again. Just put it to the max before if start to clip

Hope this help.

@Konnos: As you can see, the DFGT has a deadzone between 2% and 16% depending on the logitech profiler force.

CCS86
09-07-2015, 15:33
@CCS86: how do you do that graph? hope we can sort our DFGT out in this thread ;)


I used "wheelcheck" and the log2 function. Then pulled the data into excel for graphing and analysis.



Nice graph CCS86! the DFGT seems like a nice piece of a wheel!

I think you're better off setting the force in the profiler to 60%. From there you can disable the scoop knee feature because the force is almost linear. From the graph, you need a Deadzone Removal Range of 0.16. After that, you set the Falloff between 0.04 and 0.08, Lower value means a stiffer wheel with less understeer feeling. Higher means more understeer feeling but the wheel will get lighter.

If you remove Scoop settings, you'll need to adjust the tire force a bit again. Just put it to the max before if start to clip

Hope this help.


Thanks! It seems like a really nice force curve compared to some more expensive wheels. I was thinking about upgrading, but maybe I'll hang on to it a little longer.

I'll have to read up a bit more about the effect of each adjustment slider. Does the scoop knee function help to straighten the force curve in the lower sections where falloff occurs?

I noticed from the Asseto Corsa thread where I got some direction on Wheelcheck, they incorrectly described which data field went on which axis. They said "force" was plotted on Y and "deltaXdeg" (rotation) should be on X. That gives a graph that does not match any other wheel force chart I've seen. I should have labeled mine, but "force" is on the X axis, and "rotation" is on Y.

Does that mean that in pCARS the FFB "deadzone removal" remaps the force output to align "0 required force" with "force required to affect wheel"? (in my case 16% force @ 60% profiler strength, becomes the new "0 force")

konnos
09-07-2015, 16:08
As you can see, the DFGT has a deadzone between 2% and 16% depending on the logitech profiler force.

Where can you see that value? Is that the x axis? At the Min Force test in Wheelcheck it says 0% for me. Also what is the Rē value in the graph?

poirqc
09-07-2015, 16:36
Where can you see that value? Is that the x axis? At the Min Force test in Wheelcheck it says 0% for me. Also what is the Rē value in the graph?

The linear force test generate forces from 0% to 100%. The wheel sensors report the position after a movement occured. The value will be 0 if the wheel didn't move. This is the deadzone. It'll be >0 if it was able to move.

So, on the X axis, you can see, depending on the force of the profiler, the deadzone gets bigger.

Now, the question is: Do we trust the min force test or the linearity test?

Edit - CCS86, i'll ask you something later about the linear values.

CCS86
09-07-2015, 22:11
Where can you see that value? Is that the x axis? At the Min Force test in Wheelcheck it says 0% for me. Also what is the Rē value in the graph?


Like Poirqc said, the deadzone is the point on the X axis where the curve intersects.

The R^2 value indicates the fit between the trendline (linear dashed lines) and each curve. an R^2 of 1 represents a perfect fit. It's just a numerical method to analyze fit.




Now, the question is: Do we trust the min force test or the linearity test?

Edit - CCS86, i'll ask you something later about the linear values.



I would be much more inclined to trust this method for deadzone checking. I found that my logged data showed a number of very small, non-zero results in each curve. I manually threw them out because in the scope of this test, they are effectively zero. I bet they are enough to trigger "movement" in the minimum force test though and show you a falsely small deadzone.

Also worth mentioning, I think you should manually move the wheel to 0 position before each force test. It will try to self center, but usually miss by a decent amount (remember, there is a deadzone after all!), which will likely throw off results.

CCS86
10-07-2015, 13:46
Even with all the other settings exactly as I wrote them?

It is possible that the FFB bug happens only for a new profile. I reset my profile before starting to play patch 2.0 and that caused the bug. Apparently people who have profiles dating back to pre-1.4 patch are not affected. Soft clip and Relative Adjust are still bugging though, even with the old profile. TF values are the ones working correctly if you have an old profile.

In short: It's a weird bug related to the profile somehow.




I tried it again last night.

I removed all FFB tweaks, set my profiler to 60%, reset and recalibrated my wheel, and applied your settings.

Now I do in fact get decent FFB, but am actually still clipping pretty heavily. This was in a Formula A car with high downforce, for reference.

I just wish we could change FFB settings from within a session!!! It would really help accelerate testing each slider independently.

Mahjik
10-07-2015, 14:47
I just wish we could change FFB settings from within a session!!! It would really help accelerate testing each slider independently.

Use the Master Spindle force setting within your session to fine tune the strength of FFB for each vehicle.

konnos
10-07-2015, 22:55
How can you plot the graph in excel? The wheelcheck i used is v1.72 and i didnt see any buttons to copy the log or any other button for that matter.

poirqc
10-07-2015, 23:17
When you do a Log 2 linear test, it will create a .csv file into your documents folder, the same one that has Project Cars where you put jacks file. You just have to import that file in excel or google sheet. From there you create a graph. There's plenty of guide on the web.

You should redo the test a couple times with different strenght values in the config panel of your wheel.

CCS86
11-07-2015, 15:32
Use the Master Spindle force setting within your session to fine tune the strength of FFB for each vehicle.



True. I'm still trying to find baseline FFB settings though. Don't feel like I'm anywhere near fine tuning per vehicle yet.

BuToNz
11-07-2015, 20:03
Spent sometime reading through this thread today and testing settings. Basically nothing has even got me close to what we had before the 1.4 patch.

It's actually unplayable for me with the oscillation/centre spring, anything that gets rid of that seems to completely break the rest of the feeling (I may as well be playing an arcade cabinet game with linear FFB). Tried combinations of everything mentioned in the first post, mixed with various other suggestions here. I have lost all force feedback when standing still/1-3mph then when accelerating it progressively gets heavier and heavier, it's constantly clipping no matter what settings I use.

Quite the disappointment this hasn't been fixed already as over the past few years it's been hands down great. Now I get heavy middle deadzone 'notch', loss of road feel and a rattle everytime I even try to turn a corner.

If anyone has some suggestions or even some acknowledgement that it is not something to do with my settings but is the issue everyone else seems to be having I would be grateful. From what I've seen I'm really surprised more people aren't swarming the forums complaining?

bmanic
11-07-2015, 20:33
bmanic, I wasn't entirely satisfied with your proposed settings for the DFGT, I would like a more lively wheel. Given your significant experience with the FFB system, can you help me out with my settings? I started off with only putting the ReltvAdjustClamp at 0.20 or so, like I described in a previous post. Everything at default values with 66% JackSpade tweak. The FFB graph is lively and doesn't clip, however, when forces appear, they re a bit too abrupt, like too forceful maybe.

If i didn't want to reduce TyreForce to eliminate those, what is the proper solution to adress this, without making the steering heavy?

I also tried a few values with Scoop Knee Reduction up to 0.30 (from the default 0.15), it was a little better but somehow the FFB died a little? Not sure if that was the case and not sure what the slider actually does... I also tried a bit of Soft Clipping Half. If i understand this right, it amplifies small values when it is set below 1.00 and it lowers them above 1.00? Not quite sure where to go with it and tampering with that didn't produce the results i was expecting, although the forces were restrained.

What do you propose I fiddle with more?

For a lively wheel I suggest looking into the car specific FFB settings. Try first by removing any FFB tweaker .xml files then set the wheel up using my original post, then do this for your chosen car:


Master Scale = set so that it doesn't clip too heavily but so that you get the amount of force you desire. Will probably end up being a quite high number.
Fx = 200
Fy = 40
Fz = 200
Mz = 70

Fy smoothing filtering = 0.3 (important!)
Fz smoothing filtering = 0.1 (this can also be left at 0.0 but try 0.1)

What should happen is you get a really lively wheel with tons of "road noise" and such. It may be a bit too much though.. if it's too much then try lowering the Fx and Fz values until you are happy.

bmanic
11-07-2015, 20:36
Spent sometime reading through this thread today and testing settings. Basically nothing has even got me close to what we had before the 1.4 patch.

It's actually unplayable for me with the oscillation/centre spring, anything that gets rid of that seems to completely break the rest of the feeling (I may as well be playing an arcade cabinet game with linear FFB). Tried combinations of everything mentioned in the first post, mixed with various other suggestions here. I have lost all force feedback when standing still/1-3mph then when accelerating it progressively gets heavier and heavier, it's constantly clipping no matter what settings I use.

Quite the disappointment this hasn't been fixed already as over the past few years it's been hands down great. Now I get heavy middle deadzone 'notch', loss of road feel and a rattle everytime I even try to turn a corner.

If anyone has some suggestions or even some acknowledgement that it is not something to do with my settings but is the issue everyone else seems to be having I would be grateful. From what I've seen I'm really surprised more people aren't swarming the forums complaining?

Which car are you driving? I've noticed the Formula A was really weird but that was due to the odd default car setup and was simply fixed by setting the Caster a bit higher (3.9 or 4.0).

As for the FFB being weird. I agree. The soft clip and relative adjust have gone bonkers and it's really hard to fine tune the values so that you get what you want. This is due those two parameters being bugged (and apparently only seriously bugged after resetting the profile.. people with old profiles pre-1.4 patch don't seem to suffer which is very odd indeed).

bmanic
11-07-2015, 20:39
True. I'm still trying to find baseline FFB settings though. Don't feel like I'm anywhere near fine tuning per vehicle yet.

You'll have to do both to fully understand how the system works. :(

A question that can and should be asked: What kind of wheel response do I want? "Road feel immersive" or "fast lap times/informative"? Then tweak from there basically. I find it difficult to get both at the same time. When I want to truly understand a car and how it handles right at the limit, I usually end up with a Mz biased car specific setup (except for the Formula A which doesn't like too much Mz) but if I want a more "road feel" and lively setup I then tweak Fx and Fz to dominate. After this I may go back and tweak the wheel parameters to enhance the experience.

BuToNz
11-07-2015, 20:53
Which car are you driving? I've noticed the Formula A was really weird but that was due to the odd default car setup and was simply fixed by setting the Caster a bit higher (3.9 or 4.0).

As for the FFB being weird. I agree. The soft clip and relative adjust have gone bonkers and it's really hard to fine tune the values so that you get what you want. This is due those two parameters being bugged (and apparently only seriously bugged after resetting the profile.. people with old profiles pre-1.4 patch don't seem to suffer which is very odd indeed).

Tried most of the trackday cars (Radical, Caterham R500 etc), GT3 cars and a few road cars. Stayed clear of the open wheelers as I've put less time into them. I tend to hop around cars fairly often so spending time just on one car FFB tweaking really isn't worth it, the 100% default G27 settings were perfectly fine pre-1.4 patch.

The main problem is the centre feeling, I can be cornering and I feel it pulling a little so I let it pull to centre but it's so strong it 'overshoots'. So the wheel goes from say 2 o'clock to 11 o'clock then back to 2 o'clock until it gets to a point where I no longer want to hold the wheel and it spins 180. It's constantly clipping too which obviously is why the detail is now missing, but no matter how I change settings I can't get them back.

Seeing 'FFB improvements' in the 1.4 patch notes I thought I'd give it a fresh profile, since then it's been unplayable. I will give some more settings a try tomorrow (Thanks for the PM Sampo) and see how it goes but I've pretty much played with all the settings today (Bar the car specific tuning).

poirqc
12-07-2015, 04:40
I'm running with those setting (http://forum.projectcarsgame.com/showthread.php?34461-Yet-another-G27-FFB-template) currently. Maybe they could help?

Good luck

bmanic
12-07-2015, 15:33
Tried most of the trackday cars (Radical, Caterham R500 etc), GT3 cars and a few road cars. Stayed clear of the open wheelers as I've put less time into them. I tend to hop around cars fairly often so spending time just on one car FFB tweaking really isn't worth it, the 100% default G27 settings were perfectly fine pre-1.4 patch.

The main problem is the centre feeling, I can be cornering and I feel it pulling a little so I let it pull to centre but it's so strong it 'overshoots'. So the wheel goes from say 2 o'clock to 11 o'clock then back to 2 o'clock until it gets to a point where I no longer want to hold the wheel and it spins 180. It's constantly clipping too which obviously is why the detail is now missing, but no matter how I change settings I can't get them back.

Seeing 'FFB improvements' in the 1.4 patch notes I thought I'd give it a fresh profile, since then it's been unplayable. I will give some more settings a try tomorrow (Thanks for the PM Sampo) and see how it goes but I've pretty much played with all the settings today (Bar the car specific tuning).

So even if you follow my original post 100% to the letter you are unable to get rid of the clipping??

konnos
17-07-2015, 13:48
Ok i ve been trying to get my wheel together all this time, i havent even been online or raced at all, that's how much it's bothering me, the only racing i do is a couple laps in TimeTrial to change my settings. I have gotten close to something I like, so please try and give me some pointers as to what I should do next.

I m using 60% force in the Logitech Profiler since i mapped my wheel with wheelcheck and saw that, like another poster, it is the most linear response, so I don't have to use Knee at all. When i get more competent with the FFB settings and truly understand how they interact with each other, I will use 100% force and readjust soft clipping etc, for now i want to take the Knee out of the equation.

So Deadzone Removal Range at 0.03 and Falloff at 0.02 feels ok, i can tweak it more, but it's not my main issue at the moment.
Relative Adjust Gain 0.25, RABleed 0.50, RAClamp 0.20
Scoop Knee and Reduction at 0.00 as described above
Soft Clipping Half 0.70, Full 2.59

My problem is that I try to adjust Soft Clipping Half to stop the clipping while maintaining some liveliness, I get to a point where i feel it's decent (it still clips) and then i start to fiddle with the SCFull, and it doesn't produce the result i thought it would. In the beginning it clips more and after i raise it some of the clipping goes away (but not all of it) but instead, now the clipping happens at a lower point in the FFB graph. I thought SCFull would reduce the range of the FFB graph to within values that i wanted but i don't understand what it's doing now. If i raise it more, the clipping plane remains but the wheel gets softer and that's it, I don't see any range compression.

How do i move from here without reducing TireForce? Is there a good reason to do so?

212184212185

konnos
17-07-2015, 14:34
And as i was writing this the new patch came out... huh! Lets see what has changed :)

konnos
17-07-2015, 15:01
HELL YEAH! Problem fixed!!! <3 <3 <3 <3 <3 <3

mattzildjian
17-07-2015, 16:36
HELL YEAH! Problem fixed!!! <3 <3 <3 <3 <3 <3

So the new FFB doesnt need a low Tire Force setting anymore?

Or do you mean its fixed because you went back to Classic (Pre 1.4) FFB?

konnos
17-07-2015, 17:32
Yes classic. No clipping, the wheel is soft enough to my liking, depends on the car of course, I only adjusted my deadzone removal and falloff and it s golden.

M0dMaSt3r
17-07-2015, 19:00
Yes classic. No clipping, the wheel is soft enough to my liking, depends on the car of course, I only adjusted my deadzone removal and falloff and it s golden.

Change to default and set the Steering Gain to 1.60 ! Works perfect for me ;)

kontza
17-07-2015, 19:29
Yes classic. No clipping, the wheel is soft enough to my liking, depends on the car of course, I only adjusted my deadzone removal and falloff and it s golden.
How about Jack Spade files? Are you using those as well?

I have a Logitech Momo, and the 2.5 patch did nothing for me; massive clipping still occurs. What I did after patch update: I removed Jack Spade files, let game adjustb ffb in Logitech Profiler and reset wheel in game settings.

With 2.5, is the correct order of operations as follows:

Logitech Profiler: prevent game from changing ffb settings
L-tech Profiler: set ffb strength to your liking
In game: reset wheel
All done, no need to tweak tire force, or install Jack Spade files

CCS86
18-07-2015, 03:43
Make sure you calibrate the wheel and pedals after resetting in game

OppaErich
18-07-2015, 15:57
Spent sometime reading through this thread today and testing settings. Basically nothing has even got me close to what we had before the 1.4 patch.

It's actually unplayable for me with the oscillation/centre spring, anything that gets rid of that seems to completely break the rest of the feeling (I may as well be playing an arcade cabinet game with linear FFB).

Same here, wasted some more hours of my life fiddling with the f*****g wheel. And my last hopes are gone now. It's not unplayable for me but it's also anything but fun. The steerings gets heavier when I turn, I do feel curbs but that's it. No sense of weight transfer or which end is going to slide soon. DiRT Rally gets more and more playtime and I hardly touch pCars anymore. And I'm a track racer, Rally should be just 'something different'. Really sad this, it's been so good...

Roger Prynne
18-07-2015, 16:01
Please refrain from swearing OppaErich, next time you get an infraction.

What about if you use the Classic setting, and you don't say what wheel you have.

poirqc
18-07-2015, 16:02
Same here, wasted some more hours of my life fiddling with the f*****g wheel. And my last hopes are gone now. It's not unplayable for me but it's also anything but fun. The steerings gets heavier when I turn, I do feel curbs but that's it. No sense of weight transfer or which end is going to slide soon. DiRT Rally gets more and more playtime and I hardly touch pCars anymore. And I'm a track racer, Rally should be just 'something different'. Really sad this, it's been so good...

Does this (http://forum.projectcarsgame.com/showthread.php?34461-Yet-another-G27-FFB-template&p=1052849&viewfull=1#post1052849) give you better results?

OppaErich
18-07-2015, 17:31
Please refrain from swearing OppaErich, next time you get an infraction.
What about if you use the Classic setting, and you don't say what wheel you have.

I thought there's some filter in action and took care of spelling :) Sorry.

Erm, DFGT like in the thread title.

Edit: Now I see your edit remark. If you've cleaned a lot of posts this way then I know why I was thinking there's a filter... ;)

OppaErich
18-07-2015, 17:57
Does this (http://forum.projectcarsgame.com/showthread.php?34461-Yet-another-G27-FFB-template&p=1052849&viewfull=1#post1052849) give you better results?

TF > 46 and the line is gone through the roof. I'm at 37 now, at higher values the steering gets too heavy for my taste. I feel the tyres when steering and the curbs, sawtooth like in the graph. There are some sines too but I can't feel them. Lower than 37 and I can't feel the tyres enough anymore and get brake lock and go too fast into corners. And YES -
I did reset and calibrate in game. I did this approximately 3 billion times over the last weeks.

Roger Prynne
18-07-2015, 18:15
I thought there's some filter in action and took care of spelling :) Sorry.

Erm, DFGT like in the thread title.

Edit: Now I see your edit remark. If you've cleaned a lot of posts this way then I know why I was thinking there's a filter... ;)

All good mate just think next time.

konnos
18-07-2015, 22:50
Erm, did you reset the wheel? I turned Classic on, adjusted DRR to 0.06 and DRF to 0.03 (or a bit more if you like), everything else as it is in Classic and Profiler at 100%. No clipping, the wheel is a lot better than anything i could come up with all this time of testing again and again. I can't believe you are using Classic and you re having trouble, it should be the same for you as it is for me.

Gaaazzzaaa
19-07-2015, 10:57
Silly question , but I have the IT skills of a baked potatoe, how do you reset the wheel?

Roger Prynne
19-07-2015, 11:16
Not the red line obviously but the one to the left.

212438

Gaaazzzaaa
19-07-2015, 11:29
Thank you. Not sure how I missed that.

OppaErich
19-07-2015, 14:09
Erm, did you reset the wheel? I turned Classic on, adjusted DRR to 0.06 and DRF to 0.03 (or a bit more if you like), everything else as it is in Classic and Profiler at 100%. No clipping, the wheel is a lot better than anything i could come up with all this time of testing again and again. I can't believe you are using Classic and you re having trouble, it should be the same for you as it is for me.
Was this for me ?
I left DRR and DRF at default 0.0 and 0.01. WheelCheck does report 0.0% (-1). If I raise this, I get wheel oscillation. I have this now too but it's so weak I can't feel it, starts only if I take my hands off the wheel.

poirqc
19-07-2015, 14:50
If you have a G27, with profiler stengh at 100%, Practical DRR is 0.16. You then put a DRF of 0.08.

From there, you can lower DRF to your liking. People uses ration from 2:1 to 4:1

FuriousDemon
19-07-2015, 14:50
This FFB clipping bug is there still. At least for me with TX wheel.
The way to fix it, the same way it has been since 1.4. Is to delete the profile and recalibrate the wheel. None of the new FFB settings fix that bug. After I delete the profile and start anew, even with "default" profile all is well. If you keep using the old profile, you get clipping. I really hope Ian and the team look into this. I personally don't mind deleting profile since I don't have much time for a career anyway these days, but many others would.

Please, SMS, when you work on the new patch, don't delete your profile. Test on a PC that has a 2.5 installed, then install 3.0(or w/e the next patch is) and see if you get massive FFB clipping. If you have a TX wheel at least, you probably will.

Lanius1984
20-07-2015, 12:36
Guys, I'm having trouble setting up my DFGT and I think its likely my problem has already been solved on this forum, but there is so much info, including a lot of out of date info as well.

I only picked the game up on Friday this week for PC, so my only experience is with patch 2.5 only. Looking at this thread in particular, I chose to set the wheel in 'Classic' mode as this seems to be the common approach for the best setup. Apart from that, I've done a basic profile setup in the Logitech profiler, with the main force at 100%, rotation at 900 degrees, everything else at 0% and letting the game adjust... I haven't changed any other settings (not the calibrate settings or the individual car settings in-game).

The general feel of the wheel is fine, however I cannot get on with high speed straights. It seems very hard to stop the car from moving side to side (this is only gentle, but I feel very silly not being able to keep a smooth line). I didn't have this problem when playing Gran Turismo, so I'm sure its a setting and not my wheel.

My questions are:

1. Is there a single setting, or group of settings that would solve my specific issue?

2. Is there a selection of other changes that will improve the wheel on the whole? I know that Jake Spade's settings were used previously, is this still required/recommended to improve the feel? If so, how do i go about using these?

I'm looking to get the best setup for realism. To be honest, the options in-game are as clear as mud to me, so I just don't know where to start. I don't even know what the Logitech profiler settings do, although I just copied these from a guide I saw on here. I'm worried about changing too much based on posts earlier in the thread or in other posts, as it's not clear what is applicable to the new patch.

Any help would be greatly appreciated.

bmanic
20-07-2015, 15:29
This FFB clipping bug is there still. At least for me with TX wheel.
The way to fix it, the same way it has been since 1.4. Is to delete the profile and recalibrate the wheel. None of the new FFB settings fix that bug. After I delete the profile and start anew, even with "default" profile all is well. If you keep using the old profile, you get clipping. I really hope Ian and the team look into this. I personally don't mind deleting profile since I don't have much time for a career anyway these days, but many others would.

Please, SMS, when you work on the new patch, don't delete your profile. Test on a PC that has a 2.5 installed, then install 3.0(or w/e the next patch is) and see if you get massive FFB clipping. If you have a TX wheel at least, you probably will.

For me it was enough to hit the big RESET button on the controller page. I did NOT have to delete my profile. So it's worth hitting the reset button a few times first.

bmanic
20-07-2015, 15:32
Guys, I'm having trouble setting up my DFGT and I think its likely my problem has already been solved on this forum, but there is so much info, including a lot of out of date info as well.

I only picked the game up on Friday this week for PC, so my only experience is with patch 2.5 only. Looking at this thread in particular, I chose to set the wheel in 'Classic' mode as this seems to be the common approach for the best setup. Apart from that, I've done a basic profile setup in the Logitech profiler, with the main force at 100%, rotation at 900 degrees, everything else at 0% and letting the game adjust... I haven't changed any other settings (not the calibrate settings or the individual car settings in-game).

The general feel of the wheel is fine, however I cannot get on with high speed straights. It seems very hard to stop the car from moving side to side (this is only gentle, but I feel very silly not being able to keep a smooth line). I didn't have this problem when playing Gran Turismo, so I'm sure its a setting and not my wheel.

My questions are:

1. Is there a single setting, or group of settings that would solve my specific issue?

2. Is there a selection of other changes that will improve the wheel on the whole? I know that Jake Spade's settings were used previously, is this still required/recommended to improve the feel? If so, how do i go about using these?

I'm looking to get the best setup for realism. To be honest, the options in-game are as clear as mud to me, so I just don't know where to start. I don't even know what the Logitech profiler settings do, although I just copied these from a guide I saw on here. I'm worried about changing too much based on posts earlier in the thread or in other posts, as it's not clear what is applicable to the new patch.

Any help would be greatly appreciated.

Do you have this zig zag problem with ALL the cars? Usually if a car wants to "wander" a bit on the straights it's actually a problem with the car setup, not necessarily with FFB. You can try to add a little more positive rear toe for the car. If it is an issue with all the cars then it is a FFB problem. Probably the best way to solve it is to set some Deadzone removal range (try 0.08) and then counter that with some Deadzone removal falloff (set it to 0.04).

Lanius1984
20-07-2015, 16:01
Do you have this zig zag problem with ALL the cars? Usually if a car wants to "wander" a bit on the straights it's actually a problem with the car setup, not necessarily with FFB. You can try to add a little more positive rear toe for the car. If it is an issue with all the cars then it is a FFB problem. Probably the best way to solve it is to set some Deadzone removal range (try 0.08) and then counter that with some Deadzone removal falloff (set it to 0.04).

Thanks, I'll try that tonight.

It's all cars to answer your question. Like If I'm not holding the wheel precisely in the centre, it's turning (I know that sounds silly, but it doesn't feel natural), and a small adjustment pushes me the other way. I would just expect a little more play there, but didn't think i should need to be setting an actual deadzone.

What do the deadzone removal range and falloff actually do?
deadzone removal range suggest it removes a deadzone present in the wheel, but my issue seems to be the opposite of that...

OppaErich
20-07-2015, 16:12
What do the deadzone removal range and falloff actually do?
deadzone removal range suggest it removes a deadzone present in the wheel, but my issue seems to be the opposite of that...

I does but we're not talking steering deadzone here. That's the FFB deadzone. In a real car you won't feel much through the wheel when running straight. But we have only the wheel so we can enable feedback there. If I got this correctly...

Sadly, I had to set this to zero and lower Tire Force to get rid of the shaking. There's still a slight tendency left but really only when I take my hands from the wheel. Even then it does not go mad anymore, just slightly sweeping left and right.

GTsimms
20-07-2015, 16:21
I does but we're not talking steering deadzone here. That's the FFB deadzone. In a real car you won't feel much through the wheel when running straight. But we have only the wheel so we can enable feedback there. If I got this correctly...

Sadly, I had to set this to zero and lower Tire Force to get rid of the shaking. There's still a slight tendency left but really only when I take my hands from the wheel. Even then it does not go mad anymore, just slightly sweeping left and right.

The main thing that, I notice in all sims is bumps, the wheel needs to pull more to the left or right as in real life. But, since playing with the lofi wheels, not sure if it is better on a brushless wheel. If you are going ex: 20 mph and hit a spot that has a 3-5 inch change in the surface it should pull to the left or right in the direction of the surface change
.

FuriousDemon
20-07-2015, 21:42
For me it was enough to hit the big RESET button on the controller page. I did NOT have to delete my profile. So it's worth hitting the reset button a few times first.

Have not tried it with patch 2.5, didn't help with patch 2.0.
Still, even if it does solve it, there is a bug that needs to be fixed. Some people will not know to do that.

bmanic
20-07-2015, 23:51
What bug? The FFB is 100% working here. I know 2.0 was still bugged but that's what 2.5 fixed. I'm afraid I do not understand what you are trying to say?

jhonatas
21-07-2015, 00:54
Using the wheelcheck program?
I auditioned and he saved a file log_mforce 07/20/2015 20-54-55.csv My Documents.
How do I import this file and generate the chart?

sherpa25
21-07-2015, 01:42
Using the wheelcheck program?
I auditioned and he saved a file log_mforce 07/20/2015 20-54-55.csv My Documents.
How do I import this file and generate the chart?

For min force, no need for the log/chart. After running Min Force test, it will display the result at the bottom:
212627

Roger Prynne
21-07-2015, 10:46
Have not tried it with patch 2.5, didn't help with patch 2.0.
Still, even if it does solve it, there is a bug that needs to be fixed. Some people will not know to do that.

So how can you comment if you haven't even tried Patch 2.5?

jhonatas
21-07-2015, 12:18
For min force, no need for the log/chart. After running Min Force test, it will display the result at the bottom:
212627

Yes, I did and got the result of 0.05% for the T300, so I should use 0 in DDR and 0.01 in DDF.
But I would like the chart to see the linearity of the forces, but do not know how to do it.
Can you help me?

poirqc
21-07-2015, 12:29
Yes, I did and got the result of 0.05% for the T300, so I should use 0 in DDR and 0.01 in DDF.
But I would like the chart to see the linearity of the forces, but do not know how to do it.
Can you help me?


Have a look at this thread (http://forum.projectcarsgame.com/showthread.php?35147-Linear-FFB-Test-Excel-sheet-Universal&p=1049991&viewfull=1#post1049991).

jhonatas
21-07-2015, 14:27
Have a look at this thread (http://forum.projectcarsgame.com/showthread.php?35147-Linear-FFB-Test-Excel-sheet-Universal&p=1049991&viewfull=1#post1049991).

Thank you, please answer me on another topic.

konnos
21-07-2015, 14:36
The minimum force is probably not something you can trust to input as a % for the deadzone. The best method is to take that *.csv, open it with excel and graph the force and xdegdelta. You will then see at what value of force % your wheel starts to output force that can be felt. For example, my DFGT reports a minforce of 0% (at 100% profile force) but from the actual graph from log2 you can see that the force only begins to appear at around 3-6%. So it's best to use that. try it anyway, and if it feels good then use that. However your main question has to do with road anomalies. If you re braking with all your power at the edge of traction, even small bumps on the road surface will upset the cars' balance and give you that zig zag feeling. Also some cars have different technologies at their axls which lead to instability at very high speeds and can shake when you brake them hard. If that is your problem then it is normal and it is not the FFB problem. Also notice that some tracks are more detailed than others and that some tracks are just smoother to begin with.