Traction Circle Plot

[gplracerx]

In my opinion the new scatter plot of lateral and lineal acceleration otherwise known as a g-g or traction circle plot of an entire lap is not very useful. If you want to go faster, you need to be able to see the transitions between braking, turning and acceleration for a particular corner. That means the dots need to be connected in chronological order and the dots for anywhere else on the course can't be on the display. For a race track, most of the points are going to be in the middle because acceleration forces are low on straights. It might be slightly more useful for autocross because you're almost always braking, cornering and accelerating hard. But even then, you need to be able to focus on small parts of the course, not the whole run.

[Harry]

You are right it is not the best chart to analyze any particular position on track. But it allows to understand the driver's general capabilities in a second The ideal form will actually look like the sketch below. There will always be some noise in the middle, but the better the concentration is in the "shaded" zone, the better the driver will usually be.

Scan-131228-0001.jpg (88.34 KiB) Viewed 2535 times

The dark shaded area in the top centered is the area you will have at full throttle driving a straight line. For very long straights and low powered cars, it will move towards the center. For a Corvette on standard tracks this will hardly ever happen Any dots except the noise from "Ferraria effects"(? I mean load change - from throttle to brake and vice versa) are not what we want.

I have a "cursor" mode for this chart on my list (the charts except overall acceleration have it already) but will not add it to v18. Btw. I have a cursor planed for the map chart in v18.5... It will allow comparison of the effect of the line driven for the current lap (against the reference).

- Harry

[gplracerx]

That's fine for comparing different drivers and/or cars, but I suspect that the difference between laps on any one day won't be discernible by eye. A single average total acceleration number would probably be more informative. Average percent throttle would be nice too.

[Harry]

An average number is influenced by the driver AND the car, so it does not tell you anything on the distribution.

Harry

[gplracerx]

An average number is influenced by the driver AND the car, so it does not tell you anything on the distribution.

Harry

I don't care about the distribution. I want a chart that I can look at in the few minutes I have between runs and see where I could go faster. I can do that with the old total acceleration plot vs distance. I can't with the traction circle plot, at all. For me, it's completely useless. It drastically reduces the utility of the app.

[gplracerx]

I take it back about the traction circle plot being useless. The ability to correct the yaw angle calibration in the traction circle plot is extremely useful, as it's usually off, at least for me. Would you please give more details on that. I couldn't find anything much on the forum. Maybe I didn't phrase the search correctly.

Does changing the yaw in the TC plot affect the lineal and lateral acceleration charts for that lap including data exported after the change? I assume that the straight ahead button sets the yaw correction to zero, which assumes that the device lineal acceleration axis is exactly parallel to the vehicle lineal acceleration axis. What's the angle increment for the clockwise and counterclockwise rotation around the yaw axis?

[Harry]

The rotation recalculates a potentially missing 3rd calibration step (yaw compensation). This 3rd step requires a long high speed straight with a finishing hard brake. Probably not as common on AX tracks like for permanent circuits...

In the G-G diagram, the plus and minus rotation buttons add / subtract 2 degree each click. The (middle) auto rotation button calculates the aggregated lineal acceleration on the left and on the right side and counter rotates it in a way it is correctly adjusted (in theory). So pressing the middle button should bring you to a near o.k. value and it is *not* zero degree. Fine adjustments can be applied using the minus / plus buttons. Once the dialog is left, the rotated values replace the current values in the database. This means the corrected values will show up in all diagrams afterwards.

- Harry

[gplracerx]

I've been exporting the data to Excel and then calculating a correction by comparing the integrated yaw rate from the product of speed and lateral acceleration with the GPS heading. This is a lot easier and can be done at the event between runs. Thanks.

There are usually still small errors in roll and pitch calibration because the car may not have been dead level initially, but that's only important if you're trying to, say, integrate the lineal acceleration to calculate speed.