I like the strain gauge idea. They are relatively cheap, easy to apply and very precise. Probably the most reliable setup of all solutions discussed here. You can even leave them on the car permanently for fine tuning your ride (you see them on F1 cars all over the place). You just have to be careful with positioning and alignment (the first one I applied was permanently glued to my thumb - think some fragments are still there after 20 years) and need some basic electrical circuit to connect them to a laptop and some math to do the analyzing.

Spent some time around the tunnels of wind......the problem here is the downforce produced by a monster wing "pulls" on the hinge mechanism,
making a measurement damn near impossible.....

I'd love to hear a solution, as I've pondered the exact same thing Jack is ......... I too will have a 3.8 tail with elevated wing.

Ultimate axle weight at different speeds and wing settings - as well as different wings will tell the true tale. The things that stinks here is all this probably requires a wind tunnel --

I've had an idea about this for a long time but sourcing certain materials has been problematic.

I've been playing around trying to build a simplistic model of a wind tunnel for no other reason than just to see what happens (and for fun ;))

I decided to go with two identical electronic scales. They are about 6"x6" in size and resemble what you could use for weighing envelopes, or drugs if you watch enough cop movies :D

The whole set-up is supposed to accomodate 1:18 scale models that thanks to some on this board I've begun to accumulate. For "wind" I am using an electric fan. It resembles a computer cooling fan but it's bigger, about the size of a CD and more powerful.

I'm still missing a large enough piece of plastic pipe in which the cars would sit. (Looking for clear, not a toilet-type pvc)

I'm not even sure whether it will work, but my idea is to position different cars on the scales w/ front wheels on one and the rear on another. Each model, brand, etc. will weigh different, but what I'm looking for is the change of percentage of weight increase over each axle. This will (hopefully) allow me to see how a 911 w/ a tail vs. w/o a tail is affected, or by how much.

Still haven't figured out how to introduce smoke, or whether it's even worthwhile. Burning incense and a cigarette have been the only ideas I've had so far.

Of course I'm not sure how scientific this will really be, but at least it will be fun for me. For Jack's purpose, he'd have to have a model and butcher a few more to fab different tail configurations.

The problem with strain gaging is that you need to measure all the load paths, or make assumptions about load distribution. To do it correctly you would need to build attachment points for the lid to ensure there are no unmeasured points of contact. This would probably space the lid a little higher than it would otherwise sit. You would use biaxial loadcells and vector sum the measurements to get the true load.

If I were interested in the relative downforce of one wing vs another, I would put an LVDT on the spring plate and plot speed vs displacement over a number of runs on the same course. You'd be better off with a straight run so that lateral accelerations do not affect the suspension and data set.

You could do one better and measure all four corners. This would give you an indication of how the car as a whole is affected. Maybe some tails would give lots of downforce in back but give some uplift in front?

Because the digital scales I've looked at so far don't give a continuous readout. You get on the scale, they switch on, then they take a quick sample, do some averaging (probably), and flash a read-out for a few seconds before switching off. My guess is that it's a battery saving thing.

But I'll look around at different bathroom scales before I try this.

As to measuring at the car springs, maybe I'm missing something. But if I'm looking for sub-1mm changes with a strain gauge, wouldn't the numerous variables of road contour, road crowning, bumps, accel-decel, steering correction, etc, bury the wing effects' data pretty deeply? I mean, I can see it working in a wind tunnel. But who's got these perfectly smooth, perfectly straight, uncrowned roads with no wind at all? I still think the decklid latch would give cleaner data. But that's a non-engineer guy thinking 'intuitively,' which I know is dangerous. ;)

"dangerous" :)

Jack, let me tell ya how that screen-play ought to go ... I've got an intuition about these things, ya know. :cool:

seriously; get Tyson on the problem.

}{arlequin;
A few things to consider...
1) Model size to tunnel diameter. I don't have the information right in front of me, but you want the tunnel diameter to be at something greater then 2.5X the model cross section or else the tunnel walls will interfere with the model.
2) Stabalizing the airflow before it gets to the model. The air coming off of the fan will be fairly turbulent and you want a fairly straight linear flow by the time it gets to the model.
3) Scales accurate enough to measure the change in the model's "weight" and drag given the fact that the model is much smaller then in real life. In real life a 911 is generates varying degrees of lift with an ideal case being little or now lift, so the scales will need to be pretty sensitive to measure little or or lift of a 1:18 model. Basically you want to accurately and repeatedly measure 1/18th of little or no lift.

There was an article in a recent SportsCar (the SCCA member magazine) about a guy using a plastic model as you described to generate a wind-tunnel model of an E-Prod Alfa GTV and test the fender flairs that we was developing for it. I'll see if I can dig it up.

Cstreit; I bet you that you could use your data collection system to accomplish a similar thing to the old tie-wraps on the shocks. Measure your car's ride height (or possibly the shock extension) on a straight stretch of track from some of your recent races. Slap on a tail and collect the same data from the same segment with the aerodynamic changes. While the data will be noisy, you should be able to see a shift in the data if the aero changes resulted in the car having a different ride height. To get to it you'll need to do some sampling and T-Tests (if I remember my stats correctly).

Alternatively, you can now buy a month's access to on-line CFD software from these folks for less then the price of a set of tires. So maybe a bunch of us who are really interested in this could pitch in on the project. It would involve generating a model of a 911, agreeing on a set of modifications to try and a plan for doing it and then chipping in on the cost. Being a competitive sort, I'd say that some generic results could be shared with the wider BBS, while the good stuff would be kept to those who chipped in.

:rolleyes:

"I really appreciate the input from those of you who know what you're talking about."

- But what about the rest of us?

I'm not sure how Jack can do this with strain gaues for $30, unless someone is willing to lend him some equipment.

My suggestion would be to:
Use a video camera to measure ride height variations. The 600 lb springs pose a problem? Well, put lighter ones in and drive fast enough so that a noticable difference can be seen....

The ride height variation doesn't have to be measured tire vs wheel well either. If there is some way to attach a rod of sort to a suspension member, you could have a "pointer" that indicates the ride height. Once again, if done properly, the pointer could amplify the suspension movements so a relatively small (but meaningful with the right springs) amount can be measured.

I'm with Chuck that the flow change between different tails is probably different enough that measured the reaction force at just the tail mounting areas will not be accurate. Aerodynamics are complicated enough that there are not a lot of good "rules of thumb". People in the know say that when others start talking about truth's, it shows how little they now.

Oh, and that Einstein guy, he said turbulence was too difficult to calculate so he chose another field...

Harleq/Dave - Besides John's ideas, you need to get the flow in the tunnel laminar, and you need to test it to be sure - you can run ahot wire or sonic anemometer back and forth across the dia. to test that. And -- NOTE WELL -- you will always have wind parallel to the "road" in your tunnel, unlike the real world. You can build a treadmill for the car to roll on tho.

Get a copy of Steve Vogel's book and read the back of it - I posted the cite some time ago. I've never seen a book or even a published paper on how to build a scale, low-speed wind tunnel. In my field (biology) it is passed down by oral tradition from prof. to grad. student (sounds like a martial art or stone masonry or something...). If this gets to be a real project for you, post a new thread on it, and I'll try to get some info from friends who use tunnels currently.

re models - CFD or otherwise... Having worked with and invented many models for many years, I distrust them. Everyone of them has to be verified by real testing. They can surely narrow the tests that need to be done, tho. And don't fall into the trap of pretty colors and 3D output...
It's harmless fun tho ...

Looks like Jack ONLY wants to test what various wings/tails do in terms of downforce ( or subtraction of lift )

An idea I cam up with is thus

http://www.supportware.net/porsche/tailtest.jpg

The green horizontal bar is connected to the decklid and pivots there..

The red bit is directly connected to the car itself. The green bar pivots on the red bit too ( at the black square )..

These will turn a relatively small decklid movement into a large movement of the green bar up or down on the left. The blue triangle is the pointer. The curve at the back of the green bar would have markings to illustrate this movement..

Obviously, all of these pieces must be created in such a way as to not effect aero too much AND not flex.. (6mm sheet alu ? )

Some kind of sensor could be attached between the green bar and the blue bar to give electronic readings or you could mount a lipstick cam to the blue bar to see the markings on the green bar's curve..

I'm not an engineer either....

EDIT.. You could mount one of these on either side of the decklid.. REMOVE the two rubber stops or replace them with harder/softer stops if you find that the decklid is bottoming out or not moving at all.. if the stops are removed, the dcklid shocks will 'hold' the decklid *up* against the latch, until aero forces counteract the force of the shocks, that is where 'known spring rate' stops come into play..

If you collected the data with a laptop you could take thousands of samples over a period of several seconds. With a single statistical crunch all the bumps and wind disturbances would average out and you would be left with just the data you are looking for.

There is going to be no smooth steady state data you would get with any solution so the data averaging method is going to be a must. Being in the back of the car means that any wing you test is going to be in some seriously buffeting winds.

If you want some real data you can use you had better give up on the $30 idea. For less than $200 you should be able to buy a data collection unit that can interface with a laptops serial port and the strain gauges. I am sure you can find a pelican member in your area that will be happy to help you set it up and show you how to crunch your data.

It will be money well spent since I imagine you will be able to use it many times when you are looking to tweak your setup in the future.

With all the engineers out there, I can't believe no one suggested a manometer or a rotameter

Run a tube from inside the car on the top of the rotameter to the top outside of the rear deck lid and measure the pressure change from one configuration to another. Won't give you absolute numbers, but it will tell which tail has more down force.

If someone in LA wants to try it, I'll loan them one or two rotameters to try.

I've used manometers to try to determine the best place for intakes, it's not as easy to get solid #s as you might think.

One pressure point on a large tail will be useless for calculation purposes.

Agreed,

But you can get relative numbers from one tail to another I would think. You could also fit a number of tubes on the tail and get several rotometers, or switch the rotometer between several tubes as you drive.

Relative numbers between different tails would still be useless, not to mention I don't think jack wants to drill dozens of holes in his different tails and then try to plumb it all to a meter.

A pressure solution would only be accurate if you could have a measurement hole for every square inch of the structure in question.

Hmmm...
 

http://forums.pelicanparts.com/uploa...1110439478.jpg
http://forums.pelicanparts.com/uploa...1110439492.jpg

Jack,
I've heard the turbo spoiler creates approx. 300 lbs of downforce. Compared to the setup you have, a turbo wing is kids play. But even with 300 lbs, your 600 lb/sq.in. springs would still deflect 1/2" at high speed, enough to record resistance change with a linear pot.

Sherwood

Photo from Race Car Engineering & Mechanics (Van Valkenburgh)
http://forums.pelicanparts.com/uploa...1110752170.jpg

How about minute changes in tire pressure? Could you install tire pressure sensors and monitor the changes with speed? Each wing is going to react differently with changes in speed and not always linearly. At rest you could measure the pressure at each wheel by loading the car at each wheel or at the rear bumper to determine the pounds of downforce. In the end what you are really trying to measure is the wings ability to squash the car into the pavement anyways.
Rick
'78 930