Data Logging Metrics

[Jersey Tom]

I'm not familiar with CDS. I may have to check it out.

[Belatti]

Relate corner radius (and steering angle) to speed

Radius = (1.467 x mph)2(squared) divided by 32.167 x Lateral G's

The above formula shows that by multiplying steering angle by the square of speed is effective in correction. Though you are conspiring against Ackerman with this formula and tends to indicate too much steering, so the following works better.

Adjusted Steer= steering angle x mph x (square root of MPH).

Excellent handling graph for judging understeer, though it's not an absolute number thing but comparing the "gap" of this signal to the lateral G and knowing what the gap "looks" like with a neutral car.

I think I cant follow the formulas... can you explain the conversion factors and use the "tex" option, please?





Is this right???

[DaveW]

I think I cant follow the formulas... can you explain the conversion factors and use the "tex" option, please?





Is this right???

Belatti, The first equation is simply the normal centripetal acceleration equation, written for "english" units, with a pair of brackets missing after the division symbol. Hence 1.467 is a rough conversion from mph to ft/sec, & 32.167 is go in ft/sec/sec. Implies that the Radius is defined in feet, speed in mph & LatG is in "go" units. Generalizing, it might properly be written,



where go, V & R can be specified in any consistent units (e.g. metres/sec/sec, metres/sec & metres).

[speedsense]

Relate corner radius (and steering angle) to speed

Radius = (1.467 x mph)2(squared) divided by 32.167 x Lateral G's

The above formula shows that by multiplying steering angle by the square of speed is effective in correction. Though you are conspiring against Ackerman with this formula and tends to indicate too much steering, so the following works better.

Adjusted Steer= steering angle x mph x (square root of MPH).

Excellent handling graph for judging understeer, though it's not an absolute number thing but comparing the "gap" of this signal to the lateral G and knowing what the gap "looks" like with a neutral car.

I think I cant follow the formulas... can you explain the conversion factors and use the "tex" option, please?





Is this right???

Yes and if I didn't explain myself well enough, it is best to use the second one "adjusted steer". And the resulting units should be considered a calculated dimensionless index, as the units do not directly relate to physical data.
***As the use of this graph, overplotted on top of lateral G, is subject to appearance as the determination of understeer, having the driver drive the racing line as he is warming up the tires, will serve as a good base line to "see" a neutral gap. Should the gap become smaller (in relation to the Lateral G)..then the car is headed in the direction of understeer...passing through and beyond the lateral G signal indicates very bad understeer.... O/S is obvious and in the opposite direction (the gap gets wider).

[Belatti]

Dave, Speed, thanks both!

Check this data:



Where:
Red id Long G
Orange is Lat G
Black is Combined G

Yep! There is room for better braking there. The combined G (in black) begins an upward slope after the decreasing slope of the Long G, in red.

and then:
Green is a calculated steering angle with this formula:

Black is the steering wheel sensor

I think the is something wrong with this sensor. Maybe I should recalibrate it. Cant it be that the o-ring that acts like a belt between the potentiometer and steering column is slipping a bit?




Here I plotted the adjusted steer formula in blue, on top of the Lateral G, in orange:

[Steering]*[Speed]*pow([Speed],0.5) according to Pi software.
What do you think of it?

[Electronics]

Toolbox has C# based metrics allowing users to create their own libraries.

Also, neither CDS nor Motec are ever used in Formula 1 racing for analysis. Most teams are obviously using Atlas or Toolbox.

[Jersey Tom]

Toolbox has C# based metrics allowing users to create their own libraries.

Also, neither CDS nor Motec are ever used in Formula 1 racing for analysis. Most teams are obviously using Atlas or Toolbox.

In my experience this is because of hardware, not because of software. Pi has a stranglehold on a variety of pro race series because their data logging hardware has been better than Motec's. If you log it with Pi hardware.. you have to analyze it in Toolbox, or export it for use in Matlab or something such.

From purely a software perspective I'd much rather take Motec over Pi... though with some spare time, Matlab is king since you can write whatever you want with it!

[DaveW]

The correlation between measured & calculated steer angles is impressive, Belatti, but I would be grateful if you could explain the derivation of the expression used to calculate steer angle.

My simple mind would suggest that the argument of your inverse trig function might look something like:



Would you agree?

[Jersey Tom]

I'd imagine that's instantaneous Ackermann angle, Dave... or the steering angle expected for a perfectly neutral car of a given wheelbase, at a given corner radius.

Steering trace doesn't look bad to me in terms of the sensor. Looks perfectly normal, typical of a somewhat neutral car... good balance at low speed though a tick free on initial throttle application, and definitely to the free side at high speed.

[DaveW]

I'd imagine that's instantaneous Ackermann angle, Dave... or the steering angle expected for a perfectly neutral car of a given wheelbase, at a given corner radius.

Steering trace doesn't look bad to me in terms of the sensor. Looks perfectly normal, typical of a somewhat neutral car... good balance at low speed though a tick free on initial throttle application, and definitely to the free side at high speed.

Agreed, JT. I can see no reason to be overly suspicious of the steer angle trace, either.

[Belatti]

or the steering angle expected for a perfectly neutral car of a given wheelbase, at a given corner radius.

Thats the one.

My concern is that while the calculated angle is, for example arround 3°, the steering wheel seems to be moving in the 0° to 90° range and thats arround 5° in the wheels... (not taking ackerman into account, that is rather parallel)

[Jersey Tom]

or the steering angle expected for a perfectly neutral car of a given wheelbase, at a given corner radius.

Thats the one.

My concern is that while the calculated angle is, for example arround 3°, the steering wheel seems to be moving in the 0° to 90° range and thats arround 5° in the wheels... (not taking ackerman into account, that is rather parallel)

And? Continue...

[Belatti]

I should install a camera to actually watch driver steering wheel input... but I dont think he moves the wheel 90° in rather plain and constant radio curves.

[Electronics]

Toolbox has C# based metrics allowing users to create their own libraries.

Also, neither CDS nor Motec are ever used in Formula 1 racing for analysis. Most teams are obviously using Atlas or Toolbox.

In my experience this is because of hardware, not because of software. Pi has a stranglehold on a variety of pro race series because their data logging hardware has been better than Motec's. If you log it with Pi hardware.. you have to analyze it in Toolbox, or export it for use in Matlab or something such.

From purely a software perspective I'd much rather take Motec over Pi... though with some spare time, Matlab is king since you can write whatever you want with it!

I find that toolbox is what sells anymore. The hardware and embedded software is great and competition is still catching up to Sigma hardware produced in 2002, however anyone can make a logger that can do the analog inputs, control outputs, CAN, serial and all other car based items.

The ability to use toolbox over other software is what seems to sell the hardware nowadays especially with the club based hardware.

Either way, it is all based upon preference. I find mapping to be completely unrealistic unless you can use government GPS (as opposed to lousy 5hz 20 ft civilian GPS) as anything else is really just based upon logarithms rather than 100% logged data.

[Jersey Tom]

I should install a camera to actually watch driver steering wheel input... but I dont think he moves the wheel 90° in rather plain and constant radio curves.

If you've got a free or loose racecar I sure wouldn't be surprised to see that kind of steering input.

I've seen MUCH larger steering inputs on big sweeping oval tracks.. about as constant radius as you can get.