iRacing Physics (Brake Bias in Dallara P217)

[BT7 Driver]

Some GG plots from real life:

1) 1977 Formula 1, three world class drivers driving the same car, Paul Ricard Circuit. Note the degree of variance, with two variables (car and circuit) constrained.

2) 1987 McLaren Formula 1 car, Ayrton Senna, Adelaide. Showing the inverted tear-drop resemblance that Jackson noted.

3) 1970s road car. Road conditions, car capability and driver ability all constraining full exploitation of the traction circle.

Finally, in the fourth image, a visualisation of what is being claimed about iRacing's tyre model in the quoted post from the Rennsport thread (re-quoted in post #14 here). But imagine a more extreme version of the diagram, where the red lines bow even further inward toward the X and Y axes. If the claim were true, the GG traces for top iRacers would sit within that red outline. Clearly they do not.

Interestingly, the densest part of the road car driver's plot does resemble the less extreme visualisation presented.

 

[mclarenf1papa]

Is this the correct document?

https://www.fia.com/sites/default/files/2022_lmp2-2017_wmsc_2021.12.15_published_2021.12.16.pdf

I can't find anything related to steering racks.

Or is that specified/constrained elsewhere?

Those are just the technical regulations. Each car is homologated with the FIA and has its own technical document.

You'd be looking for the page with this:

(I remembered it as 73 but it's actually 74). Then you'd have to model the suspension geometry to figure out what that means in terms of the overall steering ratio.


I don't have the Dallara's or I'd be able to answer your questions directly. Not something you'll likely be able to find anywhere online.

 

[BT7 Driver]

I can't find any IBT (iRacing Binary Telemetry) files for any of the LMP2 cars, otherwise I would post some GG plots for them also.

Getting top-level iRacer IBT data is hard and this is a feature I would like to see VRS add to their service, i.e. to have their coaches upload the IBT data for their best laps so that nutcases like me can do deeper analysis with specialised tools such as ATLAS.

So if any of the VRS Team are watching this, how about it?

I really don't think it would threaten your service in any significant way and it might attract some people that do not currently use it. The vast majority of your users would IMO continue to use the VRS telemetry tools because they are far easier to use to find lap time improvements, the major point of your service. Most people have little or no interest in struggling with external telemetry tools such as ATLAS and their massive complexity.

VRS does have a basic traction circle implementation in its advanced telemetry mode, but it lacks the analytic power and flexibility I'm looking for.

 

[BT7 Driver]

Lacking top-level iRacer P217 IBT files, here's the next best available GG telemetry, from VRS using their traction circle tool (available in Advanced Telemetry Mode).

Driver is Martin Krönke, one of the VRS coaches and has competed at World Championship level in iRacing since 2012 (iGPWCS World Champion in 2016). These are from his fastest lap (3:29.135) in the 2022 S2 IMSA Series datapack.

VRS inverts the more common plotting convention and shows deceleration plotted upward. The service is optimised for driver laptime analysis and coaching, so it can sometimes be difficult to get cleanly plotted isolated corners, as there are no zooming options available for the traction circle. Tertre Rouge, Mulsanne Corner and Arnage plot well, so attached below.

 

[BT7 Driver]

Seven seasons of the evolution and refinement of the iRacing tyre model tracked through GG plots.

These are all for the Dallara P217, same driver throughout (Martin Krönke), IMSA series, Watkins Glen (Boot), T7.

There's one exception - 2021 S4 - when the series did not race at the Glen. For that season, I have substitued Okayama (Full) T1, the most similar GG plot I could find from a quick search of Martin's driving for that season.

These demonstrate IMO that the iRacing tyre model as presented in the Dallara P217 has good grip/force combining behaviour, particularly from 2021 S4 onwards, when it seems to me that a significant improvement was delivered by iRacing.

As I've said before, any remaining issues regarding behaviour with realistic and unrealistic brake bias are more likely IMO to be a result of a dynamic balance issue (weight shifting front-rear). GPL had exactly that characteristic years ago, heavily exploited by the aliens of the time with their extreme throttle-braking techniques. There's probably more to GPL's behaviour than that, likely also involving its differential and its baked-in 'torque converter' effect, but the throttle-braking was a major part of how it behaved. Again IMO, from many years of driving it and experimenting with it.

iRacing is a now distant descendant of GPL with many of its simulation quirks long since eliminated, but maybe there is still something problematic remaining?

 

[BT7 Driver]

A selection of nine road cars, bringing the total road car sample size to ten with the P217 included:

• Global MX-5
• Dallara F3
• Formula Vee
• Porsche 911 GT3 Cup (992)
• Hyundai Veloster N TC
• iRacing iR-04
• Skip Barber RT2000
• Spec Racer Ford
• Holden Commodore

GG plots are all taken from VRS's 2022 S2 datapacks (the season prior to current). It's a selection of drivers this time, but they are all VRS coaches.

Wherever available in the series, I have plotted Watkins Glen (Boot) T7. Where that track wasn't driven in the series, I have selected a suitable alternative track and a turn that demonstrates clear trail braking, thus grip/force combining. These are all top-level drivers, so it's not hard to find such examples.

Be aware that VRS and iRacing often do not agree on turn numbering. The turn numbers shown are those used by VRS.

These samples demonstrate IMO that the iRacing tyre model as presented in multiple road cars has good grip/force combining behaviour as of 2022 S2. The samples probably all include the significant tyre model improvement which appears to have been deployed by iRacing in the 2021 S4 release; however, it's not possible to be certain of this because iRacing do not include an exhaustive account of tyre model fine-tuning in their release notes.

 

[David Rolleston]

Just to clarify your last paragraph. There was an update to the tyre model S21s4 when the Mercedes F1 was released. Mercedes shared a load of data and provided feedback to iRacing on what they though iRacing could do to improve it.

 

[BT7 Driver]

Just to clarify your last paragraph. There was an update to the tyre model S21s4 when the Mercedes F1 was released. Mercedes shared a load of data and provided feedback to iRacing on what they though iRacing could do to improve it.

Yes, agreed, it seems clear that the Mercedes-AMG F1 Team have been assisting iRacing, but the W12 was released in 2022 S1, not 2021 S4.

My point about the release notes generally is that their descriptions are usually quite bland and generic statements such as "Tire construction parameters have been updated".

Sometimes there's a little more detail, such as this for the GT3 cars in 2021 S4: "Tires have been updated for the GT3 class. You can expect different limit feel and degradation behavior. Grip will be slightly lower on cold tires, peak after a few laps, and then drop noticeably before holding steady until the tires are heavily worn."

But there's nothing specific enough to correlate with the force combining improvement that appears to me to be visible over time and especially from the 2021 S4 release onwards, based on the P217 history shown above and a selection of road cars for 2022 S2.

 

[David Rolleston]

Either way it’s good news that there is continual development. Interesting the new Radical SR10 ‘suffers’ from hot rear tyres that have to be managed. Just about manageable in the shorter sprint races, but looking forward to how they perform in longer 90 minute leagues races I’m in. Wondering if it’s by design or if they have over done it a bit

 

[BT7 Driver]

Interesting. I've resisted buying the SR10 so far, but I do own the SR8.

 

[BT7 Driver]

Here's a comparison of IMSA series P217 telemetry from the Virtual Racing School (VRS) and Pure Driving School (PDS) services. PDS supplies telemetry as downloadable files for use with Motec, whereas PDS provides its own custom web-hosted telemetry analysis tool. I have inverted the VRS GG plot to align with Motec's plotting convention (braking downwards)..

Watkins Glen Boot T7 (Toe).

VRS coach is Martin Krönke - 1:32.368 lap.

PDS coach is Dominik Färber - 1:32.040 lap.

In the PDS (black) plots, the deceleration spike at -0.7 lateral G and highlighted by the data cursor correlates with the downshift to 4th gear. There are similar spikes for the downshifts to 3rd, 2nd and 1st visible at -1.7, -2.5 and -2.2 lateral G respectively. Also two upshift spikes are visible on the upper part of the plot. I've not been able to correlate the flat-topped 'spike' at -1.3-1.4 lateral G with any other telemetry feature so far.

From these images, it seems fairly clear to me that the VRS tool is applying some sort of smoothing when plotting, given the absence of downchange gearchange spikes - one of the upshift spikes does seem to be getting past the smoothing. In the PDS/Motec GG plot all six gearchange spikes are clearly visible - four downshifts on the lower braking quadrant and two upshifts on the upper acceleration quadrant. I think Motec and Atlas are better suited to this sort of analysis; the VRS telemetry tool is optimised toward coaching.

Based on the Motec plot and discounting the downchange spikes, the P217 does seem to be presenting more of an inverted teardrop shape, as Jackson suggested would be the case with substantial downforce present.

 

[BT7 Driver]

Another low downforce car example (MX-5), the most circular trail braking GG plot I've found so far.

PDS coach Marcello Pagnan - Production Car Challenge - MX-5 - Watkins Glen Boot (Week 7) - T8 (Heel) - 2:08.138 lap.

Driver inputs followed by GG plot. Data cursor placed on the downshift to 2nd deceleration spike.

 

[mclarenf1papa]

Here's a comparison of IMSA series P217 telemetry from the Virtual Racing School (VRS) and Pure Driving School (PDS) services. PDS supplies telemetry as downloadable files for use with Motec, whereas PDS provides its own custom web-hosted telemetry analysis tool. I have inverted the VRS GG plot to align with Motec's plotting convention (braking downwards)..

Watkins Glen Boot T7 (Toe).

VRS coach is Martin Krönke - 1:32.368 lap.

PDS coach is Dominik Färber - 1:32.040 lap.

View attachment 592787View attachment 592789
View attachment 592790View attachment 592791

In the PDS (black) plots, the deceleration spike at -0.7 lateral G and highlighted by the data cursor correlates with the downshift to 4th gear. There are similar spikes for the downshifts to 3rd, 2nd and 1st visible at -1.7, -2.5 and -2.2 lateral G respectively. Also two upshift spikes are visible on the upper part of the plot. I've not been able to correlate the flat-topped 'spike' at -1.3-1.4 lateral G with any other telemetry feature so far.

From these images, it seems fairly clear to me that the VRS tool is applying some sort of smoothing when plotting, given the absence of downchange gearchange spikes - one of the upshift spikes does seem to be getting past the smoothing. In the PDS/Motec GG plot all six gearchange spikes are clearly visible - four downshifts on the lower braking quadrant and two upshifts on the upper acceleration quadrant. I think Motec and Atlas are better suited to this sort of analysis; the VRS telemetry tool is optimised toward coaching.

Based on the Motec plot and discounting the downchange spikes, the P217 does seem to be presenting more of an inverted teardrop shape, as Jackson suggested would be the case with substantial downforce present.

T7 has a bunch of elevation, pretty bad corner for analyzing combined grip. T8 would be better but WGI in general is not the best track for this. Sebring or a similarly flat track would be better.

 

[BT7 Driver]

The main purpose of the first post (#31) was to compare the two telemetry tools, Motec (for PDS telemetry) versus VRS (their web-hosted telemetry tool), but agreed some care is needed with track and corner selection.

What's your view on the MX-5 telemetry for Okayama T1 back on page 1 (post #17)? Is that a suitable track and turn selection? It seems suitable to me as the track is fairly flat at that turn and the turn follows a long straight.

Here's Motec telemetry from three more laps at Okayama T1 (from Discord user crexLive's session driven January 29th 2022):

Also, following your suggestion, two Pure Driving School samples for MX-5 Advanced Mazda Cup at Sebring Club (coach Marcello Pagnan, 2022 S3, 1:12.938 lap, T6 and T8 as labelled by iRacing):