AC Round 3 of 23 Formula Agile 2023 Thu, March 30th 2023

[RasmusP]

In Z1Analyzer these things are called scatterplots, you just plot Gforces for braking/accelerating/steering.

The accelerating part is not "well rounded" with the Mazdas.

Haha yeah the MX-5 is basically a flat line back to the center, lol.

I checked the channels and it's a raw channel from AC via ACTI:

And like Carsten says, it's a scatter plot:

 

[eN1xes]

TF is going on here ;'D

 

[stigs2cousin]

Carefull Fulvio, they´re still after your speed

(only with technology instead of talent )

 

[stigs2cousin]

Haha yeah the MX-5 is basically a flat line back to the center, lol.

Come on Rasmus,

it´s not that bad:

 

[RasmusP]

Come on Rasmus,

it´s not that bad:


View attachment 653523

That looks a bit stretched
Here's Donington in the Mazda:

 

[stigs2cousin]

I promise I didn't doctor the Scatter Plot

(I may or may not have mixed it up with a Formula Agile plot )

 

[KFMaster]

Carefull Fulvio, they´re still after your speed

(only with technology instead of talent )

No worries Fulvio, technology is no match for talent until I can implement a few i9 processors in my brain

 

[KFMaster]

Haha yeah the MX-5 is basically a flat line back to the center, lol.

I checked the channels and it's a raw channel from AC via ACTI:
View attachment 653435

And like Carsten says, it's a scatter plot:
View attachment 653436

Thanks Ramus, got it!

And here is a plot of my two best laps at Salzburgring in the Ginetta G55 GT4, one from yesterday @1:26:027 (green) and today @1:25:027 (red, and yes, I knocked off another second of my lap time Progress!). But I can't tell from this thing why the red one is faster. On the other hand, looking at the speed/throttle/brake traces I know exactly why the red is better than green.

I guess the scatter plot does tell me some thing about my over all drive technique -- still lacking in fully exploiting the traction circle. It is pretty much a triangle instead of half circle. So coordination between letting out brake (too fast -> could have braked later) and applying steering (too slow) still needs work. Is that right conclusion?

 

[stigs2cousin]

For checking the brake release I'd look at the brake trace with a steep rise (brake stomp) and a gradual release aka trail braking.

As Rasmus explained during brake release with more grip in the front you can get away with more/faster steering than completely of the brakes.

But, depending on your stage of driver development there may be "lower hanging fruits"

For example be certain that your driving lines allow full throttle around the apex depending on the corner type.

For corner types look at:

Yousuckatracing.com

 

[KFMaster]

For checking the brake release I'd look at the brake trace with a steep rise (brake stomp) and a gradual release aka trail braking.

As Rasmus explained during brake release with more grip in the front you can get away with more/faster steering than completely of the brakes.

But, depending on your stage of driver development there may be "lower hanging fruits"

For example be certain that your driving lines allow full throttle around the apex depending on the corner type.

For corner types look at:

Yousuckatracing.com

Thanks Carstan for your insights. The Yousuckatracing blogs are nice reads, very educational.

I did some more work on braking/steering coordination. Together with driving the better racing line as from my experience these are closely related. I was having trouble holding the racing line often due to entry understeer. Better trail braking is help this allot. In the new traction circle plot, I compare my best time today (again red, now 1:24:138) and last practice best (green, 1:25:027). I think my trail brake is improving as the red scatter plot being more rounded out. The benefit in corner speed is clearly visible in the speed plot. Now entry and apex speeds are improved, but exit speed needs to be worked on(corners 2, 3, and 4). I actually didn't notice it until looking at the log data. What I think happened was that I was not quite used to the higher corner speed, had to be more conservative with throttle at exit, ended up losing some of the speed gained earlier in the corner. Now I need to work on get the throttle down to induced some oversteer on exit. So more time to gain. From the variance plot, looks like I can easily get under 1:24 if I work that out.

This Ginetta GT4 car is much easier to drive than the F Agile. Glad I tried it as it has sufficient performance for my learnings to translate to F Agile while avoiding crashing left and right. After tomorrow's GT4 rookie race, I can go back to work on the F Agile and hope to apply the learnings from driving the Ginetta. That will be a much greater challenge.

Would love to continue hearing insights from you more experience drivers...

 

[KFMaster]

For checking the brake release I'd look at the brake trace with a steep rise (brake stomp) and a gradual release aka trail braking.

Working on this one too. I got the Moza SR-P peddle which I though is a great deal at $175. With nice construction and load cell design, I can't see why is not as good as the $500 ones besides fancier construction and better configurability. When I got it, everything was as I expected -- the construction and build quality were just amazing for $175 -- except the design. They did this hybrid position sensor (first half of the travel) + load cell (second half of the travel -- actually it doesn't really move much after the first half, as the second half is a rubber bushing) design. This kind of defeated the load cell design if I use the linear mapping by the default Moza setup -- the first half of the braking force comes from the encoder position which makes it hard to modulate. So what I did was to remap first half of the travel to just 25% of the braking. But this makes not possible to "stomp", or fast ramp, on the brake. So last night instead of practicing, I end up making a mod to the brake so that the position based braking is now only ~10% of the travel. Fortunately the Moza software is very flexible and I am able to set the position based braking effect to zero. Now the SR-P peddle is just perfect and I can get the very steep ramp on brake application

 

[RasmusP]

Short post about the Track Map that Dai sent me. Looks like this:

It's absolutely great and if you ever want to REALLY work on your driving, doing this, then getting a fast lap from the front runners and then comparing it should really help.

Here's how it looks against a worse lap. Red was 0.25s quicker:

 

[RasmusP]

I was having trouble holding the racing line often due to entry understeer. Better trail braking is help this allot.

Without seeing your setup:
If the car is neutral while rolling or at 0-50% throttle, it's probably the brake bias too far at the front.
Brake bias makes a huuuuuge difference!
While practicing in ACC at the Nurburgring with the BMW M4 GT3, I was stuck at 2s slower than another BMW driver. I knew that he also used the Coach Dave setups so it was either the driving or something else.
Luckily you can hop onboard another car and, other than in AC (afaik?), see the TC, ABS & Brake Bias settings!
The CD setup had TC at 5 and BB at 54.1. The fast driver had TC at 8, BB at 50.1 !!

So I just switched to 8 / 50.1 and holy cow, the entry rotation was massive, while the exits got a lot calmer.
10 laps later I was 1.5s quicker!

I messaged Coach Dave Academy and he more or less acknowledged, that TC/ABS and BB settings can be a bit different to what they did the fastest laps from MoTeC & YT with.
MoTeC doesn't show these values, but the hotlap videos on YT do.

I think my trail brake is improving as the red scatter plot being more rounded out.

It's important to look at the scatter plots ONLY for a single turn. Even in Chicanes you need to select just the first or second half.
Otherwise You might have some spots from the counter steer or bumps from another turn showing up in the spots of the actual turn. And since our brain likes to draw imaginary "trend lines" through all the spots, you can get very wrong impressions.

It makes more sense when using the "combine dots" setting in MoTeC.

Looks like this then:

And for the full Albert Park lap:

So.. Would you have connected the correct dots to each other?

What I think happened was that I was not quite used to the higher corner speed, had to be more conservative with throttle at exit, ended up losing some of the speed gained earlier in the corner.

Yep! When working on one part of your driving, you'll always become slower at first.
So you first work on your braking, then entry, then apex, then exit. And from time to time you combine the improved bits.

Depending on the pattern of your "issues", you need to work on one thing for all corners or all things for one corner for best efficiency of your practice time.

I can highly recommend to create something like a OneNote notebook where you write down every thought and analysis you do.

For me, the most efficient way is to watch a video from a fast driver, note down some bits that seem important to me and then go on track until I reach a "plateau".
Then watch your replay from the chase cam and you'll see that your brain is off by 1-3 metres.
Write down every spot that needs adjustments.

Drive again until your stop improving.

Then either compare your driving via replay + youtube or in MoTeC (or both).

Write everything down, print it out, put the sheets next to your monitor or whatever.
I write them on my iPad and put it in front of my monitor.

Repeat until no practice time left or reaching the final plateau, which is your skill level of controlling the car and feeling/seeing/noticing things.

This last point is where better equipment comes into place, depending on your strength. If you can spot everything with your eyes, it depends on your eyes what grade of monitor and fps you need.
Some only need 45 fps on a cheap TN panel, others need 100 fps on an OLED.

If you're not spotting things via visuals, you might improve with good headphones, where you can hear the nuances of grip.

For me, going from a G27 to my CSW 2.5 was the biggest improvement. Thanks to the bigger range of ffb, I could FEEL my braking and throttle input due to the front tyres getting loaded.

The better pedals (Clubsport V3) only made things more consistent and a bit more intuitive. But it was the wheelbase, that let me feel what my feet are doing!

If you're not a sensitive guy regarding your arms+hands, you might need something stronger than the 8 Nm of my wheelbase. And if you're a sensitive kid, a G27 with the gain at 70% might be enough.

If you're not sensitive with your legs+feet, you might need a loadcell that can handle 120 kg or a way stiffer throttle with a smoothing hydraulic damper attached.

This Ginetta GT4 car is much easier to drive than the F Agile. Glad I tried it as it has sufficient performance for my learnings to translate to F Agile while avoiding crashing left and right.

Yes, it's simply a great car! I've learnt a lot from it in an easy way. Lots of grip, some sliding possible but overall you can concentrate on your line and inputs without the fear of spinning if you try something.
The F Agile simply wants to kill you all the time. Everything happens way too quickly.
But the speed rush I got from hotlapping around Albert Park for 2 hours non-stop was something truly amazing
Whenever I alt+tab out to analyze stuff, everything was warping, lol.

I can't see why is not as good as the $500 ones besides fancier construction and better configurability.

As an ongoing engineer, I can tell you that configurability is very time consuming and costly.
Making a fixed design is a lot easier.

Also times have changed. Other brands had to "invent" the designs, while newer brands can basically copy the best parts from different available products.

They did this hybrid position sensor (first half of the travel) + load cell (second half of the travel -- actually it doesn't really move much after the first half, as the second half is a rubber bushing) design. This kind of defeated the load cell design if I use the linear mapping by the default Moza setup -- the first half of the braking force comes from the encoder position which makes it hard to modulate.

Sounds awesome and full of issues, depending on the configuration haha.
But you clearly understood how things should work in theory.

But this makes not possible to "stomp", or fast ramp, on the brake. So last night instead of practicing, I end up making a mod to the brake so that the position based braking is now only ~10% of the travel. Fortunately the Moza software is very flexible and I am able to set the position based braking effect to zero. Now the SR-P peddle is just perfect and I can get the very steep ramp on brake application

Yeah I can completely follow your thoughts and doing! Sounds great now

 

[stigs2cousin]

Thanks Carstan for your insights.

My pleasure, just don´t use them to embarass me (any more than necessary)

What I think happened was that I was not quite used to the higher corner speed, had to be more conservative with throttle at exit, ended up losing some of the speed gained earlier in the corner. Now I need to work on get the throttle down to induced some oversteer on exit. So more time to gain.

Actually Prof Korf teaches it the other way around:

use ( a slight) of brake oversteer to straighten the car up earlier to ( in a nagging Driver61 voice) "get on the throttle as early as possible."

If you have corner exit oversteer you have to lift and lose time (and a lot of it as it adds up until the next braking zone)

 

[stigs2cousin]

Yep! When working on one part of your driving, you'll always become slower at first.
So you first work on your braking, then entry, then apex, then exit. And from time to time you combine the improved bits.

Depending on the pattern of your "issues", you need to work on one thing for all corners or all things for one corner for best efficiency of your practice time.

That´s sound advice, Rasmus.
So maybe some hobby coaches (plural) here should deliver their advice in manageable portions and give Dai some time to work at his own pace?

I can highly recommend to create something like a OneNote notebook where you write down every thought and analysis you do.

For me, the most efficient way is to watch a video from a fast driver, note down some bits that seem important to me and then go on track until I reach a "plateau".
Then watch your replay from the chase cam and you'll see that your brain is off by 1-3 metres.
Write down every spot that needs adjustments.

Drive again until your stop improving.

Then either compare your driving via replay + youtube or in MoTeC (or both).

Write everything down, print it out, put the sheets next to your monitor or whatever.
I write them on my iPad and put it in front of my monitor.

Repeat until no practice time left or reaching the final plateau, which is your skill level of controlling the car and feeling/seeing/noticing things.

Damn, your Jim must be a gigantuous nerd

And I thought my keeping a (paper) file about my driving was desperate

I suggest we keep the (Sim) tecnologie out of it for the time being, at least then Dai has an excuse when things don´t go his way.
An excuse I paid a lot of money to remove, with my rig I can´t blame it on substandard equipment.


Se you on track Carsten

BTW, somehow these city tracks Han is "treating us to" remind me of phase of my youth:

 

[KFMaster]

Without seeing your setup:
If the car is neutral while rolling or at 0-50% throttle, it's probably the brake bias too far at the front.
Brake bias makes a huuuuuge difference!
While practicing in ACC at the Nurburgring with the BMW M4 GT3, I was stuck at 2s slower than another BMW driver. I knew that he also used the Coach Dave setups so it was either the driving or something else.
Luckily you can hop onboard another car and, other than in AC (afaik?), see the TC, ABS & Brake Bias settings!
The CD setup had TC at 5 and BB at 54.1. The fast driver had TC at 8, BB at 50.1 !!

So I just switched to 8 / 50.1 and holy cow, the entry rotation was massive, while the exits got a lot calmer.
10 laps later I was 1.5s quicker!

I messaged Coach Dave Academy and he more or less acknowledged, that TC/ABS and BB settings can be a bit different to what they did the fastest laps from MoTeC & YT with.
MoTeC doesn't show these values, but the hotlap videos on YT do.

My plan was not doing anything to the setup for a while, for F Agile I have been using the default setup. For the Ginetta at the Rookie GT4 race, I noticed that the default setup was 10 km/hr slower on the straight so tried the setup Brian shared and stay with it. I haven't done a diff to see what is different from default setup. For now I will concentrating on working my driving.

It's important to look at the scatter plots ONLY for a single turn. Even in Chicanes you need to select just the first or second half.
Otherwise You might have some spots from the counter steer or bumps from another turn showing up in the spots of the actual turn. And since our brain likes to draw imaginary "trend lines" through all the spots, you can get very wrong impressions.

It makes more sense when using the "combine dots" setting in MoTeC.
View attachment 654084


Looks like this then:
View attachment 654083

And for the full Albert Park lap:
View attachment 654085

So.. Would you have connected the correct dots to each other?

Aha, this is what I should do. I will try this out, makes a lot sense.

Yep! When working on one part of your driving, you'll always become slower at first.
So you first work on your braking, then entry, then apex, then exit. And from time to time you combine the improved bits.

Depending on the pattern of your "issues", you need to work on one thing for all corners or all things for one corner for best efficiency of your practice time.

I can highly recommend to create something like a OneNote notebook where you write down every thought and analysis you do.

For me, the most efficient way is to watch a video from a fast driver, note down some bits that seem important to me and then go on track until I reach a "plateau".
Then watch your replay from the chase cam and you'll see that your brain is off by 1-3 metres.
Write down every spot that needs adjustments.

Drive again until your stop improving.
Then either compare your driving via replay + youtube or in MoTeC (or both).

Write everything down, print it out, put the sheets next to your monitor or whatever.
I write them on my iPad and put it in front of my monitor.

Repeat until no practice time left or reaching the final plateau, which is your skill level of controlling the car and feeling/seeing/noticing things.

This a great training strategy, I will definitely doing my improvements long this way.
I have been making notes in word document

This last point is where better equipment comes into place, depending on your strength. If you can spot everything with your eyes, it depends on your eyes what grade of monitor and fps you need.
Some only need 45 fps on a cheap TN panel, others need 100 fps on an OLED.

If you're not spotting things via visuals, you might improve with good headphones, where you can hear the nuances of grip.

For me, going from a G27 to my CSW 2.5 was the biggest improvement. Thanks to the bigger range of ffb, I could FEEL my braking and throttle input due to the front tyres getting loaded.

The better pedals (Clubsport V3) only made things more consistent and a bit more intuitive. But it was the wheelbase, that let me feel what my feet are doing!

If you're not a sensitive guy regarding your arms+hands, you might need something stronger than the 8 Nm of my wheelbase. And if you're a sensitive kid, a G27 with the gain at 70% might be enough.

If you're not sensitive with your legs+feet, you might need a loadcell that can handle 120 kg or a way stiffer throttle with a smoothing hydraulic damper attached.

I have been playing with my setup to adjust this various inputs and learning what I can get from each.

I have a Moza 5Nm wheel base. Seems like plenty strong feedback. I have my CM FFB gain set at 23% for GT cars. Anything more than that I feel like doing a gym work out. F Agile puts back much less feedback amplitude, probably because it is such a light car. I bump up the gain to 30%, still plenty. I have the slip gain setup way up there at 80% if I remembered right. Same thing for the audio, I have tire noise set at 100% with everything else 50% or so.

I use the steering feedback for reactive response in real time, the tire noise as calibration whether to go faster or slower the next time around. I have not been able to use visual cue for reactive control of the car, only for aiming braking points, apex, etc. I think I need allot more seat time get more nuances from each of these feedbacks.

I see some peddles have load cell to handle force north of 100kg. My Moza CR-P can handle up to 70kg. But I map 100% braking to 75% peddle range. More than that I feel like going to gym again.

Yes, it's simply a great car! I've learnt a lot from it in an easy way. Lots of grip, some sliding possible but overall you can concentrate on your line and inputs without the fear of spinning if you try something.
The F Agile simply wants to kill you all the time. Everything happens way too quickly.
But the speed rush I got from hotlapping around Albert Park for 2 hours non-stop was something truly amazing

Absolutely agree! I love to drive the F Agile even though it is hard to master. And just when I feeling like I can start doing really learning at a track, first Jeddah then Albert Park, it's time to go to next track

Whenever I alt+tab out to analyze stuff, everything was warping, lol.

As an ongoing engineer, I can tell you that configurability is very time consuming and costly.
Making a fixed design is a lot easier.

Same here, full y appreciate that.

Also times have changed. Other brands had to "invent" the designs, while newer brands can basically copy the best parts from different available products.

I guess I am the end beneficiary, getting pretty much everything I would need from a high end peddle in a budget price

Sounds awesome and full of issues, depending on the configuration haha.
But you clearly understood how things should work in theory.

Yeah I can completely follow your thoughts and doing! Sounds great now

Yeap!

Thanks Rasmas for your great insights, really helps my learning!

 

[KFMaster]

My pleasure, just don´t use them to embarass me (any more than necessary)

Actually Prof Korf teaches it the other way around:

use ( a slight) of brake oversteer to straighten the car up earlier to ( in a nagging Driver61 voice) "get on the throttle as early as possible."

If you have corner exit oversteer you have to lift and lose time (and a lot of it as it adds up until the next braking zone)

Would you apply brake past apex? My quandary is this: I can feather in throttle way early, i.e., before the apex, but I end up with a speed that will run wide on exist if I am tentative with the throttle. But if I apply full throttle, some time I can get the rear to come out in a nice way to rotate the car and get good exit speed. But this was very hard to do for me and often indeed I end up lifting. Even then I still have good chance of not catching it and spin out.

Hmm, I need to rethink this and try working on getting the cat pointed in the right direct after the apex. That is probably the more reliable way of gaining exit speed.

Thanks Carsten for this great pointer!

 

[KFMaster]

That´s sound advice, Rasmus.
So maybe some hobby coaches (plural) here should deliver their advice in manageable portions and give Dai some time to work at his own pace?

Haha, keep all the good advice coming. It is overwhelming indeed, so much so I have book marked all these good posts so I can come back to revisit periodically. I already read through you guys post 3 times each

 

[stigs2cousin]

Would you apply brake past apex?

Depends on the corner type (YSAR)

If it´s the slow corner before the long WOT zone, absolutely not

use ( a slight) of brake oversteer to straighten the car up earlier to ( in a nagging Driver61 voice) "get on the throttle as early as possible."

If you have corner exit oversteer you have to lift and lose time (and a lot of it as it adds up until the next braking zone)

That´s where the brake bias mentioned by Rasmus comes in.
With my recent setup I use 54 to "rotate" the car into the corner.
But that also means that I have to start braking with a straight wheel or the FA will spin


Let´s say it was a medium left closely followed by a tight right and 5sec WOT after that then I´d try to trail the car in in 3rd, have the Nadir (YSAR) before the apex of the right and WOT early.

Sadly that´s where my inaptitude let´s me down because the FA is just too fast fore me.

Your answer in short:

if you track out too much your apex is too early!! Turn in later, if necessary sacrifice entry speed until you have the line right.

As always it takes practice to get it right (and the confidence that car will turn in when you need it. )

Carry on, driver Carsten

 

[RasmusP]

I have not been able to use visual cue for reactive control of the car, only for aiming braking points, apex, etc

My wheel is not stationary. I have the f-gt lite and fold it away when not driving.
This lead to using "mouse steering" with arrow up/down for throttle/brake and arrow right for upshift, arrow left for downshift.
You can even do throttle blips while braking and downshifting that way

AC had some kind of TC and abs algorithm when using keyboard pedals and it works kinda well for cars like the Tatuus F4 or for cars that actually have TC/ABS.

It doesn't work for F-Agile though... Car is too crazy

Anyway my point is: I drive only via visuals and sounds then and I could use this to dial in my visuals!

I locked all G-Force effects, since you can't really see what the suspension is doing with them activated.
They are immersive, but I put them to "locked" in CM.

If you want to try this without mouse+keyboard controls:
Set the ffb gain to 0% so your wheel turns smoothly. You might need some damping to have some resistance for precision.

If you then also turn off audio, you'll know what your eyes are doing for you

My plan was not doing anything to the setup for a while, for F Agile I have been using the default setup.

That's not a great plan.. Having a bad setup is also not a great plan, so you're screwed anyway if you don't get a good setup from someone fast.

I'll upload my Albert Park Setup. I managed to do a 1:40.9xx with it.
So P3 in qualy, in theory.

Spoiler: softest springs and aero at 14/7!

For now I will concentrating on working my driving.

That's way more important than the setup, of course. But they do go hand in hand.
Not so much with the GT4 as you can't really make a corner flat out or not depending on the setup, but with the Agile the aero has such a massive impact, that you can't get fast with it being wrong.

I have been making notes in word document

OneNote has infinite page size and you can drop pictures without then being locked within page borders or text blocks.
You should try it out, I love it for stuff like this.

I'd never write letters or so in it though.

I have a Moza 5Nm wheel base. Seems like plenty strong feedback. I have my CM FFB gain set at 23% for GT cars. Anything more than that I feel like doing a gym work out.

Hehe I'm using 50% gain with 8Nm of my csw 2.5.
But it's like wine & coffee. You work your way up

And just when I feeling like I can start doing really learning at a track, first Jeddah then Albert Park, it's time to go to next track

Yeah I know that pain.. Only invest much time in tracks that will definitely be used regularly.
Or if a track is a real challenge and mastering it can help you overall.

Would you apply brake past apex?

It depends.
That's really the only thing that can be applied to all corners.
Depending on the car, the grip level, the setup, cambered/off-cambered etc. all change the perfect line.

In short:
- tighter line = less distance
- wider line = higher speed but more distance

And then it also depends on the next corner. Or rather the "combination" the corners are part of.
Being 0.3s faster through T1 but that leading to losing 0.4s through T2+3 is not worth it.

Some people do this by "instinct", which is called "talent".
I personally struggle a lot with this. Using different cars on the same track or different tracks with the same car does help.

Also driving different sims helps me a lot, since every sim gives you different ques.
rF2 gives you massive feedback about how loaded the front tyres are. You touch the brakes and the ffb shoots up.
But it the official GT3's feel like a sponge for me.

AC is a brilliant mix of everything but it's really difficult to keep a slight drift at the exit. It just goes without hope of catching it.
I drive GT3 cars at TC 1.
At TC 2+, I spin.
With TC at 0, I spin.

It also has the most sensitive trail braking oversteer.
Coming back from other sims, I always spin under braking for the first 15 minutes.
I can't see the difference in releasing the brake while turning in in Motec, but it makes a huge difference.
After 15-30 minutes, it's totally fine and I have no idea how I could spin like I did..

Meanwhile I'm drifting through every corner with a mild TC setting in rF2 and ACC.

ACC lacks feedback but it feels really rigit and how I'd imagine a real GT3.

With my recent setup I use 54 to "rotate" the car into the corner.
But that also means that I have to start braking with a straight wheel or the FA will spin

54 with the Agile is pretty far to the rear. I used 55.5 at Albert Park and got that slight tickle from the rear when trail braking.