HELP: Audi TT Cup understeer

RasmusP · Jun 30, 2017

Hi,
as you all can see we had some good discussions going on in the last days and the consensus was something like "Let's have a place to talk about the driving".

So here I am, creating the first of that threads:
I need some advice on fighting the understeer in the Audi TT cup

When I watch the replays I see that "the fast guys" are having a better line and when I compare it to me in detail I see they are faster while cornering, turning in point + braking is almost the same and I see no clear difference.
It's like some "magic" that pushes them tighter around.

But then sometimes I get that too... Just don't know how to do it consistently

My thoughts: when you brake you will gain grip on the front wheels. So touching the brakes slightly while cornering will give you a tighter turning radius but will also slow you down. If you brake too hard you will transfer too much weight to the front and you will get understeer.
If you accelerate you will get understeer because the front wheels lose weight and therefore grip.
If you accelerate too much you will get HEAVY understeer, because FWD...

So somewhere in between there seems to be a "trick" that lets you do the "tighter turning radius" you get when touching the brakes, but without braking and without slowing down.

Can you help out? Do I have to turn the wheel more aggressively or trail brake deeper into the corner? I don't know and before I'm spending a few hours on doing a splitscreen slow motion video, studying the details etc. (maybe I will do that either way) I just thought about asking you all and start a nice discussion about the handling of our beloved (and sometimes hated) friendly event car

Kek700 · Jul 6, 2017

This really baffled me so I had a look at my own replay (the race was 28/06/17 btw, not 30th

).
I get very different numbers when I compare your final lap with that of the leader.
For SF I get 155 for both of you (vs. your 162/154)
For MCS on the first turn, I get 61/59 (vs. your 71/58)
For MCS on the second turn, I get 99/104 (vs. your 105/104).
Not sure what could cause this. Doesn't really seem like lag could do it. I'm taking the speeds from the built-in apps (both the Essentials app and the Gears app show the same speeds). Maybe you used an app which doesn't play nicely with replays?
I stopped comparing after the first few corners, as the figures were so different.

Quite concerned about Neil's statment above. Blows my ship out of the water.

RasmusP · Jul 6, 2017

You can put it down in the water!
Got 2 replays send to me and I will check them both to see if data differs!

@Neilski can you tell me how you get these data? Would like to play with them my self!

Neilski · Jul 6, 2017

RasmusP said:
@Neilski can you tell me how you get these data? Would like to play with them my self!

Just using the built-in apps to actually measure the speeds - quoting from my post: "I'm taking the speeds from the built-in apps (both the Essentials app and the Gears app show the same speeds)."

I used a good bit of slow motion to find the minimum speeds (using Ctrl + space, s, d, f for pause/play, slow, backwards, fast-forwards - so happy I finally found the damn keyboard shortcuts for controlling the replay!).

I still have no idea how Ernie was seeing different speeds.
The car's speedo might give a slightly different reading to the apps (if, like IRL, it's connected to the wheel speed rather than the road speed) due to tyre slip, but that should be a really tiny effect (2% or less at decent speed?) so I don't think that can be the explanation.
I must say btw that I'm rather surprised that the apps do show the true speed rather than the drive-train speed, but that's just another weirdness of AC...

RasmusP · Jul 6, 2017

Neilski said:
Just using the built-in apps to actually measure the speeds - quoting from my post: "I'm taking the speeds from the built-in apps (both the Essentials app and the Gears app show the same speeds)."

I used a good bit of slow motion to find the minimum speeds (using Ctrl + space, s, d, f for pause/play, slow, backwards, fast-forwards - so happy I finally found the damn keyboard shortcuts for controlling the replay!).

I still have no idea how Ernie was seeing different speeds.
The car's speedo might give a slightly different reading to the apps (if, like IRL, it's connected to the wheel speed rather than the road speed) due to tyre slip, but that should be a really tiny effect (2% or less at decent speed?) so I don't think that can be the explanation.
I must say btw that I'm rather surprised that the apps do show the true speed rather than the drive-train speed, but that's just another weirdness of AC...

Ah okay, didn't see that.
Wow, finally we got the shortcuts!

Well, could you then tell us how you got these graphs? Looks like a telemetry tool

Neilski · Jul 6, 2017

RasmusP said:
Well, could you then tell us how you got these graphs? Looks like a telemetry tool

Yeah, that's a pet project on which I'm making really slow progress. I use a small Python app in AC to dump lots of data to a CSV file during test runs and then another standalone Python app to suck in the data and analyse it. I'm still feeling my way around the oddities of AC telemetry data, but some info is already quite accessible - easy stuff like speed, acceleration, distance seem to mostly work. (If you're familiar with Python, I'm mainly using numpy, pandas and matplotlib.)
Right now I'm struggling to try and quantify the distance advantage you have at the end of a straight if you shift at the optimal revs in an MX5 (next up: the TT), but some weirdness/misbehaviour in the telemetry data is screwing it up. (Headline answer seems to be a few metres.)

RasmusP · Jul 6, 2017

Only programmed in c and c++. But nothing more than getting a festo Beltdrive with light barriers etc to work (universaty, basic machinery programming)

What you are creating seems awesome!

Neilski · Jul 6, 2017

Well, I'm amusing myself anyway

There are lots of properly cool apps out there - some of those really are awesome.
(PS: I'm using it as an excuse to learn how to use the various Python tools and libraries!)

RasmusP · Jul 7, 2017

Finally recorded Ernie's lap and one really fast lap of the fastest guy.
Now comes editing and rendering.

RasmusP · Jul 11, 2017

Video is rendering! Should be online tonight!

It really shows a world... I'm stunned

Neilski · Jul 11, 2017

Sounds good

(Apologies in advance but I'll also wanna hear about the video editing tools. Just a geek at heart!)

RasmusP · Jul 11, 2017

Here it is. And yes, @Kek700 Ernie is right. The slower lap times has higher speeds! But the faster driver is braking later and harder and keeps more entry speed. Combined with a shorter line here and there the almost 2 seconds are gained.
First half is from rear view with halts at crucial points like braking points, speed differences, line differences etc.
The cars are more or less "synchronized" so you will see the lap times gets asynchronous over the lap.
The second half is the pure video with a halt at start/finish and then just the full lap. Timer is synchronous so you see how the faster car crawls slowly far away.
In the end I summarize the differences to:
- braking later and harder
- shorter line
- slightly better apexes

Have fun! After we did a bit of discussion about that video I will summarize again and put our two videos with the two summarizes in the first post

And for @Neilski :
I use fraps for recording (biiiiiiiiiig files, 40GB per lap)
then I send them through Adobe Media Encoder because Sony Vegas has some bugs with the original files....
And after that I put it all in an old Sony Vegas version, cropping them and moving them into position. The halts are done with cuts and "save screenshot".
Rendered to a 16 mbits MP4.

If you want to know more just ask

Neilski · Jul 11, 2017

RasmusP said:
And yes, @Kek700 Ernie is right. The slower lap times has higher speeds!

Well, this has got to be the freakiest thing I've seen for a long time, maybe the entire week

I was baffled by the massive speed differences between the faster and slower car, and I was coming to the (barmy) conclusion that the replay was simply knackered (but how? and indeed why?), when I saw a clear smoking gun which proved it for me.
See pic below: the upper car is supposedly going slower but yet it has higher RPM in the same gear. This happens many times but I first noticed it when I watched the in-car footage. In fact it's visible everywhere once you look.

Was this Ernie's replay?
If so, I'm having to conclude that either Ernie's PC behaves weirdly when saving replays or that AC replays are just plain unreliable for judging speeds

If it isn't his replay, then that's two independent replays of the same race which show the same nonsense speed data.
NB: my own (partial) replay seemed to show both cars with sane speeds - see my reported measurements above. ~~However, I didn't review my own speeds in that replay...~~

Edit: I just re-ran my replay. My own speeds are indeed massively higher than everybody else's. So either "medium" settings (like mine) mean useless speed info in replays for any car other than your own, or AC replays are just generally broken for speed info. (This is probably a known thing on the AC forum??)
Also, I forgot to mention that the conclusions about different lines and braking points are perfectly good still!

Cool video btw! (Am curious to learn how it was generated

)(Edit: ah thanks, I see you added the info, yay

)

Neilski · Jul 11, 2017

One more remark, having watched more of the split-screen footage (omfg, 40GB per lap!

):
@Kek700 is shifting too late. This is most important in 4th and 5th gears. Optimum shift speeds in 3/4/5 are about 114/154/195, but you can just download Sidekick which will get it pretty near perfect for you.
If you wait until the redline in most cars you won't feel the hit, but in this car it's a bit of a disaster cos the power curve falls soooo fast after the peak. I'd say this is worth at least 5 tenths over the whole lap, but it could be quite a bit more (hard to guess). See my post about this from Saturday: http://www.racedepartment.com/threa...-gp-wed-12-07-17ed.137953/page-2#post-2521055

BhZ · Jul 11, 2017

Neilski said:
Well, this has got to be the freakiest thing I've seen for a long time, maybe the entire week
I was baffled by the massive speed differences between the faster and slower car, and I was coming to the (barmy) conclusion that the replay was simply knackered (but how? and indeed why?), when I saw a clear smoking gun which proved it for me.
See pic below: the upper car is supposedly going slower but yet it has higher RPM in the same gear. This happens many times but I first noticed it when I watched the in-car footage. In fact it's visible everywhere once you look.

Was this Ernie's replay?
If so, I'm having to conclude that either Ernie's PC behaves weirdly when saving replays or that AC replays are just plain unreliable for judging speeds If it isn't his replay, then that's two independent replays of the same race which show the same nonsense speed data.
NB: my own (partial) replay seemed to show both cars with sane speeds - see my reported measurements above. ~~However, I didn't review my own speeds in that replay...~~

Edit: I just re-ran my replay. My own speeds are indeed massively higher than everybody else's. So either "medium" settings (like mine) mean useless speed info in replays for any car other than your own, or AC replays are just generally broken for speed info. (This is probably a known thing on the AC forum??)
Also, I forgot to mention that the conclusions about different lines and braking points are perfectly good still!

Cool video btw! (Am curious to learn how it was generated )(Edit: ah thanks, I see you added the info, yay )
View attachment 201232

The quality of the replay could be the cause of such imprecision. What anyone can see anyway is that the faster guy is braking later and is using the track better. These 2 are probably the most important parts of the driving style (and the more basic). When i drive i focus entirely on the driving line because THAT define the speed you will carry on the next straight. Braking is also important. If you brake 5 meters later, you are going through those 5 meters at 250kph instead of 60 and you will carry a higher speed for the entire braking zone (just the other day i gained 3-4 tenths just that on AMS).
My advice is to ignore the speed, just consider the driving line and eventually the braking zone (keeping in mind that in a replay they can be slightly wrong).

RainhamIron · Jul 11, 2017

Blimey, looks like you fellas are doing some top work there

I just skimmed quickly through the last few posts in the thread and had a look at the video on my phone screen so might well be missing something fundamental, but the thing that struck me about the car at the top was the almost digital braking. Seems it's pretty much either on or off with very little modulation. Do you know if that's a quirk with the replay, or actually how the controls were being used? Seems odd that this wouldn't induce even more understeer through lock ups, but actually ends up being quicker.

Anyway after trying to literally get to grips a bit this week with the TT Cup, I'm off to go and watch this again on the big screen and take notes...

Eddy · Jul 11, 2017

Andy Dixey said:
Blimey, looks like you fellas are doing some top work there

I just skimmed quickly through the last few posts in the thread and had a look at the video on my phone screen so might well be missing something fundamental, but the thing that struck me about the car at the top was the almost digital braking. Seems it's pretty much either on or off with very little modulation. Do you know if that's a quirk with the replay, or actually how the controls were being used? Seems odd that this wouldn't induce even more understeer through lock ups, but actually ends up being quicker.

Anyway after trying to literally get to grips a bit this week with the TT Cup, I'm off to go and watch this again on the big screen and take notes...

ABS should handle it fairly well which is why you can ram them on full straight away (even without ABS that works down to a certain speed where the brake force overcomes the tyre grip), but as you are hitting into the turn lifting off the brakes slightly will help balance out the car, or apply a little throttle too otherwise it's diving down at the front and will struggle to turn in

RasmusP · Jul 11, 2017

Yeah it seems like both, speed and pedals are a bit off...
But that doesn't matter THAT much in the end because the important one was to see the difference in the line and the start of the braking. I think everyone can learn from it

By all means 2 seconds are sounding like a massive gap but if you pause the onboard at the end on start/finish it's only a tiny bit, not really more than the bottom car has to the car in front...

And I would really like to see how the replays differ! Can we please all upload the replay from tomorrows race? I will do some all vs all Screenshot comparison