Poor Man's Ride Height Sensors: some data
 

Some of you may remember that I was trying to find ways to measure the effects of different aero setups on my 911, and was considering different options for ride height sensors. Since the laser sensors start at around $500 each, I went a different route and located some infrared sensors that are used in photocopiers (and cost $12).

http://forums.pelicanparts.com/uploa...1154556007.jpg

They claim to have a .3mm resolution, and output analog voltage for my data logger to read in the following, not-quite-linear way:

http://forums.pelicanparts.com/uploa...1158794641.jpg

I tested them out, and realized the amount of suspension travel I would be trying to measure would probably be too small, given the amount of noise from bumps and the relative stiffness of my suspension. I had an idea for amplifying the suspension movement as a way of artificially increasing my system's resolution, and I threw it together, even drilling a hole in my chassis. :eek: (Hey, it can be patched back up.)

http://forums.pelicanparts.com/uploa...1158796084.jpg

I did a setup in the back as well. When I did a quick test, I wasn't getting a good reading from the rear sensor. And since I ended up not having time to test the wing on a straight road before I tested it on the track, I basically forgot about the sensors.

Well, when going through last weekend's data of the wing and splitter versus a ducktail alone, I realized the ride height sensors were still attached. The rear one is still not working. I think it's wired wrong, but I haven't looked at it. But the front one was giving readings, and -- surprisingly -- the readings weren't just a lot of random noise.

Here are full laps graphed next to each other of a session with the ducktail and a session with the wing and splitter.

http://forums.pelicanparts.com/uploa...1158795047.jpg

As you can see, when I group similar laps together, there are repeatable patterns to the front ride height. You can even see the front coming down sharply for each braking zone.

But here's where the surprise comes. When I overlay a lap of the wing/duck/splitter aero kit against a lap where I'm just running the duck and no splitter, my front ride height is higher with the aero stuff in place than it is with just the ducktail.

http://forums.pelicanparts.com/uploa...1158795509.jpg

If I also had rear readings to compare, I'd be more comfortable saying this, but it looks like the big wing in back might actually be lifting up my front end at high speeds.

It could also be that I'm simply working with bad data. I never really finished the lever system for either the front or rear, and there's still more play in the mechanism than I'd like. But I'm surprised at how readable the ride height changes appear to be, and how clear the differences are between two different aero packages on the same track on the same day.

I'll definitely do some more testing.

Very cool and interesting.

Looks like the big wing is creating enough downforce to cause a "wheelie" effect! It is sitting a couple feet behind the rear axle, so you have a teeter-totter.

You could prove this out by comparing low speed sections with high speed sections. We would expect smaller height differences in the low speed sections.

Have you been able to calibrate the graphed data to actual height change? Unit-less data may be exaggerating the differences.

At this point, I'm only half convinced the data is meaningful. But it was an interesting thing to find.

The next step is to get the rear sensor working, and then do some straight-line tests over the same piece of flat road, repeating and repeating with different aero combinations.

why not try to calibrate the data using a jack to raise the front end a bit and see if you can see how much the front end may have been lifting from the large rear aero? It might make the data more meaningful.

Interesting information.

Doug

Yes. Static tests will be part of the next round of testing.

My hero
 

Jack,
You have got to be my hero...

If you have enough time and money to sit around and play with this stuff....at what seems to be full time.....you have got to be my hero......:)

But I remember my neighbor complaining about his car
at high speeds because the front end was pulled up by the down force of the rear end so much that the front would float.

If your "duckwing" setup is raising your front end it would appear that you achieved your goal. But now you have to address a new area in keeping your front end down.

The idea setup would be to push the whole car down towards the pavement...dont you think....

Im not a pro at this, just an amature who is thinking out loud..

Chris

Amazing Jack...you need to be working for a race team or something.

I wonder if graphing the torque/hp cycles from your engine and overlaying it would reveal anything?

Indeed, some race cars (mainly formulas, me thinks) are designed to do this exact same thing, float at high speeds. When the front is up, the front (wing, in the case of the formula) is at a lesser attack angle gving less resistance to the drag. You don't need downforce in a straight line unless you drive in the NHRA and have 1000's of HP. Hit the brakes, car comes down, downforce balance is achieved until you get up on the air again.

At least that's what I was told by the formula guys. Of course, the NASCAR boys have a different take on this. Springs that are so weak that he car rides on the stops most of the time and on the ground to prevent air from creating a pressure zone under the car. The formula car has little underside and the air is managed well.

I wonder where the 911 fits in with its relatively small footprint. Probably closer to the stock car. Still, it would be interesting to see if additional front downforce actually resulted in faster lap times rather than having the effect of holding the car back.

I think Jack will know the answer to the question before long.

Jack... do you have any specifics on where to get those sensors? Very interesting!

That's one of the things that makes this data unexpected (and possibly unlikely). If you add a ducktail to the back of a car with no front spoiler, you still increase downforce in the front of the car. Front and rear balance is a funny thing with these cars; it's not as much like a teeter-totter as it appears.

I got the sensor here. But I'd caution against ordering it based on this data. There's still a lot that could explain this as 100% wrong in its findings.

And yesterday I ordered a pair of potentiometer-type ride height sensors from an old Lincoln Continental model that might do the job better and simpler. They were $4 each, so I figure it's worth a shot.

Jack, the ducktail and the wing are not comparable.

You shouldn't take the data from one and use it for the other. The duck is a spoiler, and is reducing lift. The wing is providing net downforce, and is doing so high away from the body independent of the rest of the car.

It would be a mistake to assume that the wing will affect the front end the same way the duck will.

That's true.

But in this case, I didn't have any expections of the rear wing lowering the front end. I just didn't think it would raise it. After all, I was also running a 5-inch splitter up front -- which was removed for the ducktail only tests.

But the data might be misleading. The tail might not be lifting the front at all. I'm going to have to do some more tests before I believe those charts.

Thanks Jack... I thought I recognized that sensor. I remember seeing an article about using something similar in a Lego Mindstorm range sensor (Mindstorms ROCK), which was interesting, because it required a linearizing function, and it walked the end-user through the process of determining that function.

Turns out that the article is actually hosted on that site! How convenient.

I'm in the process of writing my own data analysis software that basically interprets data collected by other systems. I've been throwing around the idea of allowing just such a linearizing function within that software so that it could adjust the output plot of the raw sensor input based on either a formula or a series of data points.

Not really interested in picking up those sensors, was more interested in the specs on it.

Thanks for the link.


...jeff

Ok this is it he final straw Jack, I am selling all my garage stuff.

I am struggling with getting a bleeding coded lug nut from my wheels and then you post someting like that.

For sale: Mildly used garage equiment. Location: Belgium.
Will ship overseas.

Michel

"you have a teeter-totter."
- Do you have a power spectrum analysis you can run the data through?

Also, we need to get you into a grad. program somewhere...

Shiznit Jack. Nice work. Skin would have gone on my wing this weekend but the car came back early and wants to be driven...

Hey, what about me? Can I carry his books?

Re: Poor Man's Ride Height Sensors: some data
 

If I remember correctly Paul Van Valkenburg's vehicle dynamics book went into this. Either way, the cantalever effect is a well known property of a rear wing.

I wonder if you can use the nose lift data to corroborate the downforce data from the rear...

"I wonder if you can use the nose lift data to corroborate the downforce data from the rear..."

Jack,
Maybe some well-placed 50-100lb sacks of Ready-Mix or rice on the car (at rest) can replicate the on-road data to arrive at downforce numbers.

Sherwood

Instead of putting all that rice on the car, he can just go with a double layer wing and maybe some Type R stickers :p

Jack, if that data is correct, and that reaer wing is lifting the front, then maybe the old wing was doing the same to a lesser effect.

That would explain the brake bias issues.

Hey, Tyson, I'm really an amateur at all this, but when you first hit the brakes, doesn't the car tip down and gain a new attitude, camber and all? Doesn't the brake bias have to do with some other factors such as what corner and what suspension set up?

There are THREE things causing the front of your car to be higher:

1) The down force generated by the wing is behind the rear axle, causing the front to cantilever up. The farther the wing is located behind the rear axle, the worse this will be.

2) The DRAG of the rear wing is causing the front to raise up- It's acting like a cantilever also: Imagine if the wing was up at 50 feet above the car. Pushing back on this wing would act like a lever raising the front of the car. The higher the wing, the more leverage the drag will have, the worse the problem will be.

3) The added drag of the wing requires more engine power, transmitted through the rear wheels. The extra torque at the rear wheels required to overcome the extra drag is raising the front of the car, the same way the torque at the rear wheels raises the front when you are under acceleration.

what if you tried attaching the wing to the roof rack you use for transporting your track tires? besides looking really stupid this would move the wing over the middle of the car and put more even downforce on all 4 wheels. i guess the load rating of the roof rack would need to be checked also.

Your now looking for balance.


less rear wing? more front?

Interesting. That's been done on Baja off-roaders and, of course, WOO sprint cars.

However, on a long track race car, the overall areo package has to be considered. There was always a problem at the rear of a car, so the wing sort of takes care of two things at once.

I went out to Mojave yesterday to do some more tests on the wing. With the exception of one big disappointment, I was able to get a lot of (hopefully) good data.

The disappointment? Well, as this thread mentions initially, I wasn't getting any readings from my rear ride height sensor. I took a look, and discovered two things. A broken wire was preventing any data from getting through, and -- even when that was fixed -- the sensor was pointing at an angle at its reflective surface, and had to be re-positioned.

I re-positioned it. I tested it. Everything looked good. I even re-checked the connection and the angle before starting the testing, once I got to Mojave. Still, there was no rear ride height data when I got back. I checked it out again, and it had moved on its mount again. A little adhesive will fix the problem, but... no rear ride height data, which is frustrating.

Still, I got a lot of stuff.

Let's start with the most important question: is the wing working? Well, the lap time, section time lateral g-force data all suggest it is. With a 2-3 second improvement in lap times, I guess you could say any further investigation is overkill.

Okay, so it's overkill.

First off, I wanted to see if it would make sense to make the wing more complex. Right now, it maintains the same angle of attack across its entire span. My previous testing showed me that the angle of the air coming toward the wing is not horizontal behind the car; it's actually at about 11 degrees. But this wing extends so far that it's possible the air hitting the far edges of the wing might be horizontal, or at least closer to it. So I suspended a wire in front of the wing with wool tufts on it to see if the angle of incoming air changed along the span of the wing.

My results are inconclusive on this. I tried two camera positions, but couldn't get a clear enough look to see changes in the airflow angle. I'll have to come up with another way of visualizing it -- maybe just with longer tufts of fabric.

80-mph images:

http://forums.pelicanparts.com/uploa...1159141659.jpg

Next question, is the wing losing adhesion beyond a certain angle of attack? According to the books I've gotten, this isn't altogether a bad thing. At a certain point, a wing stalls and becomes a source of almost all drag and almost no downforce. But downforce continues to increase even as a wing loses adhesion over the underside trailing edge. I've seen graphics with downforce maximizing on a wing with separation on about 40% of its lower surface. There's more drag, but still...

So I positioned a camera on an arm coming off of one of the wing's uprights. It's looking up at the underside of the wing, with equal numbers of tufts on either side of the uprights, so I can compare adhesion on both the inside section of the wing and also the edges, where the angle of incoming air might be closer to horizontal. In the images, you can see the back of the ducktail to the right. To the left is the edge of the wing on the driver's side.

I don't know what the accepted way to measure angle of attack is, at least when you're talking about an airflow that is not horizontal, but is coming down at the wing at an 11-degree angle. So in the images, you see the effective angle (the higher number), and the literal angle in relation to horizontal.

http://forums.pelicanparts.com/uploa...1159142181.jpg

http://forums.pelicanparts.com/uploa...1159142240.jpg

http://forums.pelicanparts.com/uploa...1159142269.jpg

My interpretation of these images is that I'm getting pretty good adhesion, even at the 20-degree angle of attack. Two things could explain this: one is that oncoming air might not be actually coming in at an 11-degree angle (I base that on a measurement from the top of the rear window to the top of the ducktail); and the other explanation might be that the half inch Gurney flap I added to the trailing edge of the wing is actually doing its job and allowing me to run a more aggressive angle of attack by slowing down (spoiling) airflow over the top of the wing surface.

I'll want to do more focused testing, but my initial guess is that I'm getting improved downforce with each of the three settings I set up for the wing. (How much drag I'm getting is still a total unknown.)

As a side note, the thing I don't have an explanation for is the area to the left of the upright in the pictures that's showing consistent turbulence at all three angles. Even at 130 mph, you can see it.

http://forums.pelicanparts.com/uploa...1159142991.jpg

Anyone have an idea of what might be causing this?

Still, the simple lesson -- from a look at lap times and a look at the wool tufts -- suggests the wing is working. And since I didn't get rear ride height information, I have no way to get a more precise look at actual downforce from the wing's different settings.

But it works -- right?

I write that as a question because of the weird data I got that started this thread. I'm getting readings of front lift from my front ride height sensor with the wing in place. Lift is not what I expected, especially since I'm running a 5-inch splitter up front. So in the next post I'm going to talk about the other stuff I did yesterday to get a clearer look at whether or not (and if so, to what extent) the front is lifting.

With no data for rear ride height, I'm going to have to extrapolate a little from front ride height to see what's happening with my wing and splitter at speed. To do this, I found a stretch of new highway out between Mojave and California City where it seemed safe to do some short tests in two opposite directions over the same loop of concrete.

Initially, I ran straightforward acceleration runs, with different amounts of acceleration. The reason for this was that one possible explanation for the raised front ride height is something I hadn't thought about. When the car's wheels accelerate, the car rocks back slightly. This is going to bring the front end up, and it could explain at least part of what I saw in the data from the track tests. Sure enough, there's a pretty clear correlation between acceleration, speed and front ride height.

http://forums.pelicanparts.com/uploa...1159146375.jpg

You can see that under more throttle the car accelerates faster (duh) and the front end comes up proportionately. The curve of the ride height sensor's output comes to mimic the speed curve.

Having tested that, I did five controlled runs -- each in two directions so I could cancel out the effects of the day's (pretty strong) winds. In each run, I brought the car up to 100 mph and kept it there for about 10 seconds. Here is a list of the configurations I ran:

1) No wing or splitter, just the ducktail.

2) No wing, just the splitter in front and the ducktail in back.

3) Wing at 12 degrees, with the front splitter and the ducktail.

4) Wing at 16 degrees, with the front splitter and the ducktail.

5) Wing at the maximum 20 degrees, with the front splitter and the ducktail.

I won't post all 10 graphs based on this, but here's a representative one.

http://forums.pelicanparts.com/uploa...1159143850.jpg

I filtered the ride height data a little to smooth out road bumps and get more useful numbers -- then I grabbed a sample when the car was going 100 mph. The readings from the two directions were different, but consistently higher in one direction than the other, which I attribute to headwinds. I averaged the two numbers (a higher number equals higher ride height) for my results, and this is what I got:

Car at rest: .99 to 1.01

1) Duck alone: 1.225
2) Splitter and duck: 1.182
3) Wing at 12 degrees: 1.265
4) Wing at 16 degrees: 1.308
5) Wing at 20 degrees: 1.356

So, the splitter does work -- it produces a lower front ride height than the duck and front spoiler alone. But the front ride height keeps going up from there, which I think means two things: 1) the rear is getting a lot more downforce than the front (the teeter-totter effect), and 2) the rear downforce is increasing all the way up to the 20-degree setting (and might continue to improve beyond that).

Interestingly, the car runs higher than its stock resting height no matter what. This probably means the car is generating more lift than the aero devices are counteracting, but the teeter-totter effect could mask some of the overall downforce -- the rear might be coming down to a greater extent than the front is going up. Obviously, I'll have to get that rear sensor working and run some of these tests again.

But what do those numbers mean? I didn't know, so I dug up the specs on the sensor I'm using and learned that my numbers represent a total movement of about 2.7 centimeters in my sensor. But this is not a direct measurement of the height off the ground, since I mounted the sensor on a lever to exaggerate ride height changes.

So the next step was to do some real world tests with dead weight to get a rough idea of calibration. I had four 50-pound bags of sand left from my stress testing on the wing. So I measured the change to the sensor and to the actual ride height from putting 200 pounds of dead weight on top of the front trunk latch. The difference was 4 cm total, 2 cm for each 100 pounds, and the change in actual ride height was half that: 2 cm from 200 pounds of sand.

Then I moved the sand bags around to the back of the car to see if 200 pounds of weight back there would raise the car. They did. With 200 pounds of sand (and one padded dog bed to distribute the weight evenly prevent any damage to the car), the front came up 1.75 cm, by my measurement.

There may be a lot of slop in these figures, since a 100-pound bag of sand doesn't change ride height at all on my car until you manually push and pull a little on the front end to re-settle the car. But I repeated the test and the numbers looked the same, so I think I'm at least in the ballpark.

So at the most aggressive wing setting, my front end is coming up approximately (and I can't stress the word 'approximately' enough) 2 cm compared to the car sitting still, and 1 cm compared to the car with only an S spoiler and ducktail.

Does this mean the wing is generating more than 200 pounds of downforce? Well, it's possible. But there's still a lot of slop in these numbers, and I'm not seeing what's happening to the rear ride height at all yet, so I won't say it's the case without more testing.

But it's an interesting start.

Some better data on the horizon. I just got these ride height sensors from a mid-90's Lincoln Continental's air ride system. $4 each on Ebay.

http://forums.pelicanparts.com/uploa...1159215051.jpg

Unbelievable, Jack. I can't believe you haven't succumbed to "analysis paralysis" yet. You have too much time on your hands.

I wish I had Jacks "time problem" and the money to support it.

Jack,
The things you post are amazing even to a casual but seriously envolved owner like myself. I watch for your exploits and read them carefully to see what I can learn from you. Even though I'm not a track rat I love to go and watch friends make gains from the changes they make. This whole wing and front lift thing fasinates me. My car is lowered and has both a Carrera wing and front air damn/spoiler and at top speed seems to squat like a slot car. Of course I'm not on a track. But we all learn and gain from the trial and error of track development. Good on you, man. Without getting too silly about it I would like to add my voice to your host of appreciative praises.

Thanks. Here's the initial test of the $4 sensor -- I had no idea if its output would be compatible with my logger. But it is. :cool:

http://forums.pelicanparts.com/uploa...1159219866.jpg

The black line shows the front ride height sensor (the car is sitting still) while the red line responds to me working the disconnected rear ride height sensor like an accordian.

Next weekend (hey, I do have a day job), I'll do another morning drive to test it.

Lincoln Continental...

I always knew those things would come in handy for something, someday...

Okay, I installed the Lincoln Continental sensors and took the car out yesterday to run the same loop. I'm still surprised by it, but changes of 4 degrees in the wing's angle of attack produce consistent, detectable changes in the car's ride height at speed. And interestingly (well, uh, you know, interesting to me ;) ), the front ride height data I was getting before appears to have been correct.

This graph shows four different 'laps.' The course was the space between two exits on a remote freeway. You can see by the speed that I run in one direction, exit and turn around, and then run in the other direction. In the place I stop to make changes to the wing, you can see where I jacked the car up to remove the front splitter on the blue lap.

http://forums.pelicanparts.com/uploa...1159630646.jpg

And in this close up section you can see how more aggressive rear wing angles effect both front and rear ride height. (With these new sensors, a higher number means the car is riding at a lower ride height.)

http://forums.pelicanparts.com/uploa...1159630700.jpg

So, the first thing I can see is this: the more aggressive I set the rear wing, the more the car is pushed down in the back, and (I'm still a little surprised by this) the front comes up.

Now, it's important to note that even with no wing or splitter, the front still comes up and the rear still comes down. Here's a graph overlaying the two directions in the loop I ran with just the ducktail in place:

http://forums.pelicanparts.com/uploa...1159631293.jpg

The blue horizontal line is where the front and rear ride height lines meet when the car is sitting still.

The run that brought the front down the most was with no wing pushing down in the back but with the 5-inch splitter in the front.

http://forums.pelicanparts.com/uploa...1159631466.jpg

The front is still higher up than when the car is sitting at rest, except toward the end of the 100-mph stretch of one of the two runs, where you can see the front end coming down lower (the line going up on the graph). I suspect this was because a headwind was making the front splitter more effective on that run.

Running the front splitter and the wing in its least aggressive angle of attack (1 degree relative to level, 12 degrees relative to incoming airflow) resulted in the front end running lower than it would with no splitter or wing at all:

http://forums.pelicanparts.com/uploa...1159631711.jpg

But as you increase the angle of the rear wing, and push the rear end lower, the front begins to come up. Here's the wing at 16 degrees:

http://forums.pelicanparts.com/uploa...1159631768.jpg

And here it is at 20 degrees:

http://forums.pelicanparts.com/uploa...1159631795.jpg

This setting is clearly generating the most negative lift (or downforce, depending on how you want to label it), but it's doing so while also raising up the front end.

I did a graphic to show the relative effects on front and rear ride height of the different settings I ran in this test. The angles in this illustration are wildly exaggerated, but they show the 'teeter-totter' type of effect of the different wing angles and configurations.

The graphic's based on a 1-5 'ranking' of worst-to-best numbers for the ride height (which the data logger is able to average out for a given segment of roadway to make comparisons easier). So 'splitter alone' gives the best number for front ride height, while '20-degree wing plus splitter' produces the best number for the rear (but the worst number for the front).

http://forums.pelicanparts.com/uploa...1159631961.jpg

Here are charts Bill Verburg scanned for front and rear lift with the ducktail (they're from a Robert White article in Panorama):

http://www.pelicanparts.com/pmpre/im...ront_small.jpg

http://www.pelicanparts.com/pmpre/im...rear_small.jpg

My reading of these charts is that with a ducktail at 100 mph there's about 35 pounds of lift at the front axle and about 75 pounds of lift at the rear axle. My splitter is clearly reducing that lift. But then the rear wing is pushing down enough to... actually introduce some mechanical lift? I could be reading the data wrong, but this appears to be the case.

What about "yaw"? :)

Jack, i find this testing to be fascinating. i am amazed that the wing has such and effect on lap times considering what the measurements are showing. i would think that the wing is making the 60% rear weight bias of the 911 worse at higher speeds. The splitter seems to me to be of more benefit than the wing.

Just some inspiration photos:
http://forums.pelicanparts.com/uploa...1159640568.jpg http://forums.pelicanparts.com/uploa...1159640604.jpg

My seat-of-the-pants observation:

Nice R&D project.
Fact: The front end is lifting due to the rear wing thus allowing more air under the front end which exacerbates the lift. Solution? Introduce more front end rake and/or increase the effectiveness of front splitters, air dams and/or aero add-on bits. You'll eventually reach a point where front and rear drag/aero effects balance out, hopefully at a higher threshold then where you began.

Sherwood

ride height would also be affected by fuel tank contents,air temp , air pressure, tire temp , and resulting tire pressure...

any data comparison would require complete logging of all parameters

forgot one ... how fat Jack is at the time of logging ... did he eat 5 big macks, or did he go for 10.... :D