Engineers: I'd like input on two downforce measuring ideas

[Jack Olsen]

I've been playing around with different aero stuff on my 911, lately. The frustrating thing is that I have no way to get even a ballpark measurement of the 'net downforce-type effect' (reduction in lift, negative lift, whatever you'd like to call it -- I'm going to call it 'downforce') from different setups.

Of course, there are the correct ways to measure this (wind tunnels, strain gauges, load cells, laser ride-height gauges), and then there are my cheaper-than-cheap ideas. If you're a 'do it right or don't do it' type of guy, then go ahead and move on to the next thread. If not, I'd like some input on two of my latest 'cheapies.'

Here's the goal. If I'm getting 100 pounds of net downforce, I'd like some sort of measuring device that will give me a readout that will correspond somehow to putting 100 pounds of weight on my static rear wing.

One way to do this is measuring ride height. I can compare a run at a set speed with no aero to the same run at the same speed with a wing in place. I have a data logger that can take analog input and clean it up for me.

Now, it'd be about five grand to do it the right way with laser sensors. Rotary or linear potentiometers could be attached to the suspension, but the right sensors for that job cost hundreds. I'd like to do this for the cost of a burrito.

I haven't completely cracked it yet, but here's my latest try:



That's a Sharp GP2D120 infrared sensor. It's used in copying machines and for robotics applications. It bounces infrared signals like a bat uses sonar, and has a measuring range of 1.5-11.5 inches. It takes 5v in and outputs an analog signal from .4v to 3v. It costs $12 for each sensor, which is more than a burrito. But I can run one in front and one in back for $25, which is a lot less than five grand.

Problem one: I wired one into my data logger a minute ago, and just did an experiment where I stepped on and stepped off the doorsill every five seconds. The sensor was mounted underneath the door, pointing at the ground. Obviously, I'm not using the full range the sensor is capable of, since ride height change when I get into or out of the car is less than a centimeter. And 160 pounds of net downforce right above the sensor is a pretty big change.

But here's the good news: on the data file from the logger, I was able to see changes in a graph showing ride height versus time that corresponded to when I put my weight onto the car.

Here's what I'd like to do, though. If it can be done pretty simply, I'd like to find a way to exaggerate the changes in ride height. Imagine using a lever off of one of the points of the suspension to move the sensor a greater distance when ride height changes. The sensor can point at the ground, or I could have it point at any flat surface that would register the change in ride height.

I don't know if I'm describing this clearly, but imagine setting up a teeter-totter with the fulcrum moved so that one kid travels a little and one kid travels a lot, and apply it to my ride height sensor and the suspension components of a 911. I need a solution for both the front and the back of the car. They don't have to be calibrated with each other, though. I can set up equations to compensate for geometry/output differences between the front and rear sensors (or rather, you smart guys will be helping me with that at a later date. )

So, that's problem one: a way to exaggerate ride height changes to make both my front and rear infrared sensors more accurate/precise in their measurements.

Problem two: I thought it would be great to be able to get a much more direct measurement of the downforce effect of different rear wings and wing angles. One idea was to put some kind of electronic or mechanical scale where the rear decklid latch is, and literally measure the force pushing down on it. But this isn't the best point for measurement, since a wing that would cause more drag would show better 'downforce' because pushing back on the wing would mimic pushing down on it, since the decklid pivots on an axis where the front hinges are.

I know that strain gauges are the best way to do this. But they don't sell for burrito prices. On the internet, I found a page on model rocket guys testing their rocket engines with different devices. Some kid built a dirt-cheap test rig where he fired the rocket at a piece of metal inside a PVC tube. Underneath the metal was a veterinarian's syringe filled with water. When the rocket engine pushed the plate, it compressed the syringe, which fed water through a narrow tube to a cheap psi dial from a tire pressure gauge. The change in psi on the gauge could be correlated to engine output -- it was a simple hydraulic scale.

Veterinarian's syringes are less than a buck. Tubing is fifteen cents a foot. I've got cheap dial-type tire gauges lying around the garage. I'm solidly within a burrito budget, here.



Here's my thinking. Imagine there are four 'mounting points' for a wing. They're at the top of the wing's uprights, two forward and two back. My thinking is to make the rear mounting points simply pivot (as though you could change the angle of the wing by lifting the leading edge of the wing, and having the front two mounting points be secured from letting the wing flip up, but also resting the wing's weight on the tops of two secured syringes. There's not very much travel to the syringes when they're filled with water (with air, that's not the case), so there won't be much of an angle change permitted under compression. But I could feed hydraulic pressure from two syringes into a single line that goes up to the cabin, where I could videotape both the psi dial and a GPS unit showing speed, and get a pretty good measurement of downforce from the wing and how it correlates to speed.

Here are my questions on this idea. I'd like to get the greatest range of reading on this system, as well. With my tiny non-engineer's brain, I can't see if one syringe would give me more resolution than two syringes, or if larger syringes (more water capacity in a larger-diameter syringe with a larger-diameter plunger) would give me better 'resolution' than smaller syringes. I guess I'm looking for the clearest readout of pressure from the least amount of physical travel for the syringes' plungers (since plunger travel will slightly increase my wing's angle of attack, and muddy up my results).

One or multiple? Bigger or smaller? Also will keeping the tubing shorter benefit me significantly? Any input on this one?

Thanks in advance for input. I really appreciate the patience you guys show with me.

[Chuck Moreland]

No time to answer, but I have to say I think you are an Engineer hiding in writer's clothing. Your ability to keep coming up with creative solutions is impressive!

[Emission]

This is going to be a cool thread.

[Tyson Schmidt]

Jack, I thought you kicked your smack habit. I'm sad to see your back on the Heroine again.


Yeah, I agree with Chuck. You come up with some pretty ingenius solutions. I hope it works so I can do some testing myself.

[Don Plumley]

What, no solutions in the JC Whitney catalog?

For the IR sensor, you might be better off measuring a change in suspension travel than bouncing it off the road. Not sure how road surface variation or even the speed of the car might impact the measurement.

Are you wanting to magnify the sensor signal by using the mechanical device because of a) sensor accuracy over small amounts of travel (what is the +/- accy rating) or because of your data logger input?

Disclaimer - I'm not an ME/EE but have sold a sensor or two...

[sammyg2]

Here's the simplest way i can think of:
Take some sort of shock absorber like a hood shock or deck lid shock. Any kind of hydraulic or pnuematic cylinder would do.
Carefully drill out one end to relieve the pressure without killing yourself.
Drill and tap a fitting on one end somehow and bolt it to your suspension.
Run a hose from the shock to a pressure guage in your cabin, preferably a very low pressure unit. A compound guage like one used to measure manifold vacuum/boost would be good.
You might have to do some calcumalations to determine displacement and pressure.

the more the suspension is compressed, the more pressure displayed on the guage. If you have lift the pressure would drop and the pressure would go negative (compared to atmospheric).
I would not be linear because of the compressability of air but you should be able to duplicate the readings in a static mode to determine actual force.

[Brad01mc]

Hi!

hmm...after spending 5 min thinking about this...

Part 1

I can't think of a lever solution at the momment, but you might want to play with the sensitivity of the voltage measurements. If you increase the number of decimals, you increase your sensitivity to distance.

You could use a transisitor to increase the voltage differences by amplifying the output signal. As far as the details, I'm at work on a "smoke" break and can't do more at the moment.

http://en.wikipedia.org/wiki/Transistor

And this...

http://en.wikipedia.org/wiki/Image:BJT_Amplifier.png


Brad

[Don Plumley]

Originally posted by Don Plumley
For the IR sensor, you might be better off measuring a change in suspension travel than bouncing it off the road. Not sure how road surface variation or even the speed of the car might impact the measurement.

I thought this through a bit more. A problem with measuring suspension travel/compresion is that you'll need to come up with an average over the same period of time over the same stretch of track to take out the excursions caused by normal suspension travel. That's the positive of measuring ride height directly and not messing with the suspension.

[MBAtarga]

Jack,

I can help some with your first option. What you want to do is amplify the signal from the sensor output to get more resolution - that is, see a larger voltage swing for the minor distance changes you expect to detect. You can use an op-amp design to do just that. What you are after is a differential amp configuration. 1 input of the op-amp would be used to "Zero" the output for when the car is static. When the car is at speed, and theoritically your ride height drops/lifts, the output of the sensor will then be compared to the reference voltage, but amplified by the op-amp. I did a search and found some helpful information.
This linkhas some helpful information. I'm sure Warren will have a specific solution if he sees this thread.

[Jack Olsen]

Okay, I did some sitting and staring at both problems. The first one seems pretty straightforward. If I distrbute the measuring task over more syringes, each syringe will have to do less work toward the sum. Along the same lines, the linear travel of a larger-capacity syringe will be shorter than a smaller one. That's obvious to me, now. So, more syringes is good and larger ones are better than smaller. I'm going to rig up one on each side -- possibly one on each corner.

On the ride height sensors. I don't think amplifying the electronic signal is what I'm looking for. This may be a good illustration of my limited comprehension of this stuff, but I can amplify the signal any way I want in the software. It samples at a phenomenal rate. The problem, as I see it, is the 'resolution' of the measuring device itself. It's much better at registering one centimeter of travel than it is at registering a millimeter. I think physically exaggerating the sensor's travel will help with this -- although, again, I might just be seeing the problem wrong.

In the back of the car, I can attach a metal dowel to the bottom of the rear shock tower, where it bolts to the semi-trailing arm. Moving toward the rear of the car, I can create a fulcrum by strapping a piece of aluminum stock to one of my fixed oil lines. The aluminum will have a hole in it to function as a fulcrum. The distance from the fulcrum to the attachment point at the bottom of the shock will be about seven inches. I then have more than 14 inches of room for the dowel to continue to a point up and under the rear quarter panel, where it can either point up (to measure distance to the underside of the sheet metal) or down (to the road). Either way, it looks like I can create about a 2x exaggeration of suspension movement.

The front is more tricky, since the wheels turn. But I think I can attach a steel dowel to the traveling piece of the front strut housing, moving parallel to the strut toward the top of the strut tower. If I can hinge the dowel somehow, all I'll have to do is drill a small hole in the trunk, and have the second dowel (attached with an approx. 90-degree hinge) poke through into the truck itself. The wall of the trunk will act as a fulcrum, and the length of the dowel once it's inside the trunk (roughly parallel to the strut tower brace), will determine my level of amplification. Attached to the end of the dowel, the sensor can then measure the distance up to the trunk lid or down to the floor of the trunk.

Does that make sense?

I think I can do this with some hose clamps and steel dowels. The hinge for the front two pieces is the only part I haven't figured out, yet.

Oh, and in the back I'll have to shield the sensor from the heat of the headers.

[Don Plumley]

Originally posted by Jack Olsen
On the ride height sensors. I don't think amplifying the electronic signal is what I'm looking for. This may be a good illustration of my limited comprehension of this stuff, but I can amplify the signal any way I want in the software. It samples at a phenomenal rate. The problem, as I see it, is the 'resolution' of the measuring device itself. It's much better at registering one centimeter of travel than it is at registering a millimeter. I think physically exaggerating the sensor's travel will help with this -- although, again, I might just be seeing the problem wrong.

It's all about the accuracy of the sensor. If you are measuring 1mm of travel and the sensor has accuracy of +/- 0.5mm, then you need more travel. If the sensor is quite accurate, then amplying the signal electronically is pretty much the same as using a mechanical distance multiplier, just easier to do.

[Jack Olsen]

The sensor itself is definitely the weak link in my system. The data logger can jump through hoops once it has the analog signal. It can eliminate the 'noise' from ordinary road bumps on a set course at a set speed at a set distance -- and basically extract the difference between one lap's input and the next. In the case of my testing wings, the 'lap' involved would be a fixed stretch of highway that I'd repeat with different wing configurations.

But then, the data logger cost $800 and is something I use every time I track the car. The sensors cost $12, and can go in the trash once I've done my testing.

[Don Plumley]

Okay, according to this datasheet, the unspecified tolerance is 0.3mm. How much compression are you trying to measure? There's a nice graph on the datasheet that shows the distance/voltage correlation.

[Tim Hancock]

I think the differences in downforce are not going to show up very well in a suspension travel measurement. IOW, I am betting that the subtle differences in downforce are not going to cause any easily repeatable/measurable suspension ride hieght changes. I think it would be fairly easy to come up with some sort of temporary wing mnts that incorporate scales or air cylinders to monitor down force changes due to different aero setups. If you really wanted to keep it simple, maybe you could incorporate a spring under each rear mount then attach a bolt or something with a friction slip ring of some sort that would register the amount of maximum movement down against the spring after each trial run.

If you are just comparing, you really dont need to know the actual force involved, you just need to know which wing deflected the most.

[randywebb]

Having done some interesting things with syringes (in the lab, not in my arm) I can tell you that they have a fair amt. of friction, and are not designed to make multiple movements up and down. They will fail pretty quickly. Maybe they'll last long enough tho.

[john70t]

Great ideas. Downforce and drag (load sensor reading at same speed/road) have to be balanced.
What about the lipstick camera and a measuring stick?

[Jack Olsen]

Originally posted by randywebb
Having done some interesting things with syringes (in the lab, not in my arm) I can tell you that they have a fair amt. of friction, and are not designed to make multiple movements up and down. They will fail pretty quickly. Maybe they'll last long enough tho.

The guy who did the rocket engine tester suggested using a light application of KY Jelly as a lubricant on them. But you're right, they're less than a buck each and could be easily replaced if I needed to do more than a day's testing.

[Rob Channell]

Great ideas, Jack. I like the idea of searching for a cheaper solution so we can obtain SOME data over waiting for the money and time to build a perfect system to get ALL the data.

I am not sure, but it does seem like road surface irregularities would affect your downforce measurements. I think Tim's idea of using an air tube type mount to measure pressure in the wing mount might work better. How about a bicycle pump (one of the hand types) with an adapter fitting to go to a cheap air gauge. It might leak too much, but maybe either it would work for short term measurements or you could develop a way to reset it by adding air back in from a pressure tank to reset it back to a nominal height.

You could also use hydraulics, I suppose. Maybe install a cheap brake or clutch master cylinder inline and run the hydraulic line to a gauge. Make sure it doesn't leak or your paint will suffer.

For amplifying the suspension movement, how about a bellcrank. Picture a 90 degree piece of metal that rotates about the center of the angle. Make one side longer than the other like the capitol letter L. Connect the suspension up/down motion to the end of the bottom part of the L and connect something for the sensor to bounce off of to the top of the L. The actual motion measured by the sensor will be from side-to-side rather than up and down. The amount of mechanical motion is proportional to the ratio between the two lengths.

Decent animation here (not mine)
http://www.flying-pig.co.uk/mechanisms/pages/bellcrank.html
Now that I reread this thread it sounds like this is what you are talking about for the front.

Jack, not to be a thread jacker(OK, maybe I am), but does your data acquisition system output the data live on a serial port or can it only be downloaded after the fact or perhaps both?

Did I mention I like threads like this?

[9dreizig]

Stupid question Jack, but don't you have a roof rack?? I would think it would be much easier to rig something up to put on the top of your car as a test fixture. Then a simple scale could be used to read down force, thus eliminating the whole mechanicl fulcrum problem. Even a fish scale might work and you could use one camera for force and the other for your speedometer.
On a side note ( and not to hijack this thread) has anyone looked into building their own scales for corner balancing ??

[dad911]

You can amplify the distance travelled with cable & pulleys. One pulley will double the distance a cable will travel, for the same movement of the load (think block & tackle) with a spring (I'd probably try surgical tubing) to take up the slack.