Bill Lehman

Looking forward to getting back on the track, I looked at my last event at Lightning where I couldn't get near my PB with similar tire condition and compared it to my old PB. . I have an oversteer/understeer math channel (Thanks Frank) and used it as well as steering angle to see what was going on. I never payed much attention to steering angle but was surprised that its range was +/- 1000 degrees, Are these .1 degrees? I saw one of my values @ 963. Can one of you Pros explain this?

ProCoach

I would think you would HAVE to have steering angle (steering wheel angle) or better yet steered (at the outside front wheel, measured on turn plates) angle to do any meaningful US/OS measures.

You have a 987.1? Please clarify.

Matt Romanowski

Not positive on your car, but the easiest way is to connect to the dash and look at the live measures for the steering. You'll see the straight value and left/right changes.

I understood the post to be that he has steering angle, but it has a range of 1,000 units. He is asking if the units are .1% or something else.

ProCoach

To answer that question, all Bill has to to check the Solo config and steering angle channel measure divisor.

By default, it is .1 degree. I think 1 is good enough for steering wheel angle.

boxer-11

[let me know if this would be better as a separate thread but since the comment was made here and it's something I've been thinking about anyway...]

Steered wheel angle is uncommon in most template/decode schemes in the various loggers I've played with but the value in calculating US/OS is clear. I know my '18 GT3 and '16 GT4 both claim to have "variable steering ratio" racks. Is measuring using turn plates going to work for that kind of car?? If so, what would the process be like to gather the measurements?? I'm inferring that you have to measure on the plates and then correlate to something that you do have data recorded as you drive (steering wheel angle??) otherwise I can't see how you'd make use of turn plate info. In other words, are you suggesting that it's possible to build a mapping for steering ratio (cockpit wheel to road wheel)?? Forgive my ignorance but, when Porsche say it, does "variable ratio" mean non-linear fixed in a fixed way or does it mean that the ratios are varied with speed?? (If the latter then it seems like measuring on turn plates won't help much...)

Squeaky

I'm interested in the answers to the steering ratio questions as well.

924RACR

Well... when we talk about steering ratio, it's typically thought of as a single, fixed ratio - 13:1, 15.5:1, 18:1, etc. The "ratio" being discussed refers to the motion ratio of hand wheel angle (the one connected to the loose nut) vs. road wheel angle. Road wheel angle is what really matters, since that's what actually affects the dynamics of the vehicle.

So it's actually a major simplification to refer to a fixed ratio; while these ratios are approximately correct on center, even a little applied knowledge of steering geometry - Ackerman geometry, to use a likely more familiar term - reveals that in fact the ratio changes over the range of travel of the rack.

In the world of Chassis Controls (aka Stability Control, ESC, PSM etc) we correct for this with a transfer function in the software to relate the hand wheel angle to our measurements of road wheel angle throughout the range of motion. This gives us more accurate representation of what's happening at the road wheels lock-to-lock - since we're only actually measuring angle at the steering wheel/rack. Of course, suspension compliance can come into play, no pun intended. This could be measured statically, using instrumented turn plates as on an alignment rack, or dynamically out on the skid pad.

There's also the point that, again due to that lovely Ackerman geometry, the inner and outer wheels turn different amounts. Usually, for most discussion purposes, these are simplified to what we usually call the bicycle model, just one front and one rear wheel. Saves a LOT of trouble in the calculations, and provides adequate accuracy.

However... usually, on the racetrack (and really in any road driving at any speed of consequence), there's nothing good to be discussed if you've turned the steering wheel past 90 degrees or so. Certianly, we're not talking about refinement of control at that point, be it driver or electronics! This being the case, the linear approximation can be quite sufficient, using the on-center ratio.

So if you're looking at a variable-ratio rack, it's not necessarily substantially different than what you'd consider a "fixed ratio" rack - there is variation anyway, and it can be handled in the same manner. I believe the ones I've heard about are speed-dependant, which could get in the way of applying the turntable style of measurement - though again, you'll likely still be close enough to use the data.

Of course, if you have rear-wheel steer... that's a whole 'nuther level of fun!

(Side notes - 1000deg +/- sounds right to me for steering signal range, and I agree whole degrees of resolution would be sufficient.)

ProCoach

Thanks, Vaughan!

Squeaky

So you're saying using a ratio 16.5 is close enough for a 981 cayman, for most on-track steering angles and regardless of speed?

924RACR

Assuming that's the 0kph on-center measured ratio - yes, I should expect so. If you want to check, and can drive the car on public roads - take a chunk of data at speed at around 0.2-0.4 lat (which will be well within the linear range of tire and chassis response) and see if the model lines up with the sensor measurement...

For that matter, if you can't drive the car on public roads, you could give that a shot on a racetrack during an out- or in-lap...

Squeaky

Yes that's the stated center position ratio. Any thoughts on how to measure Ackerman?

924RACR

Well, as mentioned, look at how the model and actual deviate at low lat, with speed. But again, likely will be lost in the noise for the steering angle range you're interested in.

boxer-11

FWIW, I did get a chance to put my car up on an alignment rack, with the motor running, and I took measurements by looking at the live trace of the steering wheel position data from the MoTeC Dash software, adjusting the wheel by 5 degrees each time and then reading the measured degrees of horizontal turn from the alignment machine (shout out to the Dundon guys for allowing me to use their rig!). I captured the inside and outside road wheel angles for steering wheel angles from 0-200 degrees of turn both to the left and to the right. The resulting ratios varied but not very much at all in the grand scheme of things; barely more than I'd put down to measurement error really. I did make a 2D table in i2pro based on this data to translate steering wheel angle into steered wheel angle for use in a over/under steer math channel. Results for that channel look plausible, although I think you need to think about it as a relative measure of car balance as opposed to anything absolute to get any use from it. There were some interesting wobbles in the signal in places that I didn't expect though that, for example, made me decide to change braking strategy a bit to good effect in one place on the local track.

Squeaky

Would you be open to sharing such table?

ProCoach

The nice thing about Motec 2D and 3D tables is that when loaded into the config, it processes onboard and outputs the O/S-U/S as a value.

I suppose you could make a math (or even a custom sensor) in AIM that could incorporate that.

Great work, boxer-11. I think the goal is relative balance rather than a set number, and this is perhaps the best way to do this.