Track Map Looks Wrong

[dbm55]

I cannot get a correct track map for Daytona (using infield) to be generated - using an M48 in a Porsche 911 race car w/ADL Dash Manager & I2. I'm fairly new to I2 but have researched everywhere I know to resolve the problem.

In Dash Manager: Wheel Speed is wired to Dig In 1 Speed (unitless); Beacon wired to Dig1 (pin 52); Lat G wired to AV1 (pin 45).

In I2: Dig In 1 Speed channel is mapped to "Vehicle Speed" using units km/hr; Corrected Speed uses "Vehicle Speed" w/units=mph and looks to be exactly accurate for the logged data; both Lat G and Long G sensors appear to be oriented properly (Lat G is showing positive for left turns, right hand force); Speed curve is smooth - no excessive noise apparent; Long G has some noise, but Lat G noise is minimal and appears near zero on straights with realistic values in the turns (up to 1.5G). Venue is set to a closed circuit and Track Distance is set properly.

Generated Track map is completely bizarre - almost like it's "inside out" - and doubles back across itself several times, see below:

motec.gif (15.22 KiB) Viewed 24335 times

Am at a loss - what are we doing wrong?! Any help would would be greatly appreciated!!

[JamieA]

what is the track distance shown on the generate track dialogue? does it match the actual track length?

Are you setting the track up as crossover, open or closed, I assume it should be closed, as there are no bridges on the track?

Jamie

[dbm55]

Hi Jamie, thanks for responding! It is set up as a closed circuit - but you are correct, the Length shown in the Generate Track dialogue is 6010m vs. the 5729m that I set up in the Details dialogue. By the way, I set it as meters rather than miles because the Venue Length textbox in the Details dialogue only allows whole numbers (no decimal places), so the 3.56 miles that I wanted to set results in 4 instead; so I set it as 5729 meters to get the granularity I wanted.

Why is the Generate Track Length different from the Details Venue Length? Is it being calculated from Corrected Distance? 6010m = 3.734 miles, yet my Corrected Lap Distance on my fast lap is listed as 3.27 miles - ??? Since there are a number of conversions going on (Digital In 1 Speed [unitless] mapped to "Vehicle Speed" [kph] which calculates Corrected Speed [mph] which generates Corrected Distance [miles]), I wonder if there is some loss of accuracy?

I've attached the log files in case this would help - I've been trying to map Lap #2. I will try and attach a .zip of the project files in a subsequent post but not sure if that's allowed on this board. Thanks again!

[dbm55]

Here are the project files. Also, here is what the map should come close to looking like (it is run counter-clockwise):

DaytonaRoadCourse.gif (3.21 KiB) Viewed 24233 times

[Sean]

Just curious, where was the beacon? At the start line?

Also, has the track map worked okay before elsewhere?

[JamieA]

Ok, Thanks for the info. After looking at your data in detail, I believe that you are not going to be able to get a good track map using G force sensors at a track like this.

G force sensors measure lateral, longitudinal and vertical G forces. We use the G force lateral sensor along with the ground speed to determine the corner radius for drawing of the track map.

Unfortunately on a banked track, things dont go so well for this method. Some of the forces that would usually be G force lateral, end up becoming G force Vertical (see attached picture)

g force.jpg (308.9 KiB) Viewed 24178 times

So in the picture, you can see a place where you should be cornering, but there is no lateral G, but there is .5 of a G vertical. So, when using using only G lat to determine corner radius, to us, you are going straight, but in reality, you are cornering, but the G force is going into Vertical instead.

You can generate a basic channel that combines G lat and G vert like this:

'G Force Lat' [G] + 'G Force Vert' [G]

that will give you a combined G lat and vert channel. Creating a track map based on that channel ends up looking like this (keep in mind, you are seeing the driven line, not the actual track map)

track map.jpg (65.99 KiB) Viewed 24174 times

[dbm55]

That's It!!!!!! Thanks a million Jamie, I never would have figured that out.

Sean, no the beacon was originally at pit-out (see round red cursor at start of infield in image below), but I offset it -.42 miles to place it accurately (see red section marker above direction arrow at start-finish at bottom of image). And no, we had never tried to generate the map before - just coincidence that the first time was Daytona on the banking!

Jamie, why is the Track Length in the Generate Track dialogue so different from the Venue Length? The docs say it should be no more than 100 meters. I realize it represents the driven line, but a 281 meter discrepancy seems to be too much to me. I assume it's calculated - ideas? Also, do you think there's too much noise showing in the Longitudinal G Force sensor?

Thanks again for solving this!!!

map.jpg (184.04 KiB) Viewed 24137 times

[JamieA]

Driven distance (on the racing line) can often be vastly different to the advertised track length.

Where are you getting speed from? Wheel speed sensors I assume.

Check how you have calculated you tyre diameter.

[dbm55]

Yes, wheel speed sensor, which is set to use "Dig In 1 Speed". But here's the part that's confusing me: Within ADL Dash Manager (v 3.20U2), my pin settings show the Beacon set up for Dig1; I also show settings for Spd1 & Spd3 (Wheel Speed Front and Rear respectively) - these are both set to pulses/rev=16 and circ = 2100 mm, but I'm pretty certain these are not in use. "Dig In 1 Speed" appears to be input from the ECU instead. Is the calibration for that made directly in the DOS ECU Manager (if so, where?) or can it be set within ADL?

Also, I noticed in the ADL logging setup that the Anti-Aliasing filter is checked; however it is shown as "applied" on all channels except G Force Long and G Force Vert - could that account for the excess noise? How do I get it to apply to those as well?

Thanks again!

[Mazi]

Here is datalog of street use 996TT. Speed signals are collected from stock ABS via CAN. You can see the sleep at light-moderate load. If I remember correctly there is more then 15km/h slip at 300km/h full load highway. I believe that in race condition you can see much higher slip.