Hold my beer and watch this! Or: scientific airfoil testing

So I built a plywood wing in this thread:https://grassrootsmotorsports.com/forum/aerodynamics-presented-nine-lives-racing/make-a-functional-wing-with-free-plywood/152176/page1/

And now I'm going to test it. I'll link this thread in the wing build thread too, but I thought this testing is worthy of it's own thread.

Today I sweated through 2 different t-shirts to make this beautiful contraption:

And I started testing!

I'll give more details tomorrow, but so far results are very promising. I'm seeing coef of lift of about 1 - 1.2 at 30 mph, which is actually matching up with the mathematical models reasonably well.

Yes, I drove this thing on the roads:

Best was 19 lbs indicated downforce (which translates to probably about 28 lbs actual downforce - I'll get into how I figure that) at about 10 degrees angle of attack. At 30 mph 28 lbs downforce is nothing to sneeze at. If it continues to follow the models, that means I could be looking at more than 100 lbs at 60 mph.

Not bad for a free wing!

I like that setup a whole lot better than trying to test this out on top of the jag.

Also, I now have an odd desire to own a pickup truck.

Have tried any calibration with a known weight bag (flour, sand, mulch?)?

There's a bathroom scale involved, isn't there?

In reply to TVR Scott :

Yes, yes there is!

sleepyhead the buffalo said:

I like that setup a whole lot better than trying to test this out on top of the jag.

Also, I now have an odd desire to own a pickup truck.

Have tried any calibration with a known weight bag (flour, sand, mulch?)?

I did borrow this pickup truck.

Interesting you ask about calibration. I alluded to it in my first post. I did a couple runs with my first configuration, which looked like this. Then I thought of calibration (I should have started with calibration).

To calibrate, I grabbed a spare wheel and tire I had laying around. It was 48 lbs. I put it on top of the wing, and saw 25 lbs on the scale. Then I removed it and put it on again. Aaaaaaaaaand only saw 15 lbs. Uh oh. Then I realized I could basically get the scale to read anything from 0 to 45 by moving the weight forward and back on the wing. Clearly not good.

So I made extensions:

With this new configuration with the wing fully behind the front supports, the scale read 25 ish to 35 ish with the 48 lb wheel on top in all positions. I could still use some ideas to make that a bit more consistent, but I think 30 indicated to 48 actual is not a bad approximation for now.

nice. glad to see you are making progress. the calibration problem is many faceted. things like the stiction of your pivot and orientation of the strut to the "load cell" can play havoc with consistency, but you are making progress and anything in addition the weight is a bonus.

keep it up. looking forward to seeing it in action.

the trick with load cells is to test at or near the load cell limits for maximum resolution. in your test case, that means more speed or a longer arm from your pivot to the load cell strut.

we routinely try to maximize how close to the limits we run at the wind tunnel, just for the resolution and chasing small effects.

TVR Scott said:

There's a bathroom scale involved, isn't there?

need a 2nd to calculate drag. lol

stafford1500 said:

the trick with load cells is to test at or near the load cell limits for maximum resolution. in your test case, that means more speed or a longer arm from your pivot to the load cell strut.

we routinely try to maximize how close to the limits we run at the wind tunnel, just for the resolution and chasing small effects.

Assuming you follow the above, you might want to move the ‘force down to scale’ post from being all the way at the back to 25-50%.

Any thoughts about rounding over the end that touches the scale, so the force is always applied “tangentially”

Johnny_at_NineLives said:

TVR Scott said:

There's a bathroom scale involved, isn't there?

need a 2nd to calculate drag. lol

Fish scale would probably work better, although that will complicate the rig

sleepyhead the buffalo said:

Johnny_at_NineLives said:

TVR Scott said:

There's a bathroom scale involved, isn't there?

need a 2nd to calculate drag. lol

Fish scale would probably work better, although that will complicate the rig

I actually do plan to test drag as well, with a fish scale. But that will use a separate setup, with the main pivot at the bottom of the main supports. I don't think I'll be able to take a simultaneous drag and downforce measurement. 

But maybe I can optimize downforce, and then try to get an idea of drag.

sleepyhead the buffalo said:

stafford1500 said:

the trick with load cells is to test at or near the load cell limits for maximum resolution. in your test case, that means more speed or a longer arm from your pivot to the load cell strut.

we routinely try to maximize how close to the limits we run at the wind tunnel, just for the resolution and chasing small effects.

Assuming you follow the above, you might want to move the ‘force down to scale’ post from being all the way at the back to 25-50%.

Any thoughts about rounding over the end that touches the scale, so the force is always applied “tangentially”

25-50% of what? You mean move the down post forward? 

Also, I'm pretty sure the rounding is a good idea, but I'm pretty sure the deflection of the scale is one of the smaller distances that changes as this rig rumbles down the road. I can try to measure the scale deflection, but it's probably like 1/16 inch or less.

Yes, move the wing back, out past the down-post. Use the leverage around your pivot. Keep in mind the forces will be pulling up at the pivot this way. With ~40lbs showing now, you can get closer to the scales limits by extending the pivot with arms that put the wing leading edge ~2wing lengths (chord) behind the pivot and put your down-post ~1wing length back. That should keep you from exceeding the limits of your structure. For reference, you have the leading edge at ~0wing lengths behind your pivot in the pictures above, and that would keep your scale and down post at the same basic location (~1wing length).

On the idea of radiusing the post at the scale, I would be concerned about the point load damaging the surface of the scale if you get too pointy and have large loads. First approximations will be fine with a squared off load contact.

Well, finally got another few hours to devote to testing!

I added a bit of extra support because I didn't like how the cross piece connected to the vertical post on the scale was bending. Also in this photo you can see my camera setup. It is a knock off go pro, but has a watch that allows me to take a still image "remotely". Which is really cool.

Today I did 10 runs. Beginning middle and end all the same, and they were all within a reasonable (to me) range. Most runs at 45 mph, I found with pretty good confidence the wing has a coef of downforce of 1 at 10deg AoA, 1.25 at 12.5deg, and 1.5 at 15deg. I'm really happy with that linearity in that zone. 

I also did one run with the wing at 10 deg AoA at 60 mph, and again confirmed a coef of downforce of very close to 1 (total estimated downforce was 74 lbs!!)

Another fun thing I did was staple these cardboard angles to the back of the wing. They came in a box of something we ordered recently and I kept them for this purpose. Sorry about the knuckles in the photo. The basically gave me a 3/4 inch gurney flap.

That increased the downforce of the wing from coef 1 to 1.4!! Nuts! I didn't try it with more angle but I should do that on a follow up test.

Anyway, I'm really stoked with the free wing's performance so far. Can't wait to add some extra elements...

That's very cool.  Ground-breaking work in the field of plywood-based engineering.

Nicely done.

Keep up this kind of development and you may need to consider different spring rates and tire load ratings...

And, yes wickers are crazy effective if you are not too far from peak downforce. Keep in mind the inflow at the back of the truck cab is going to be different than other applications, so your angle of attack is a relative number for the testing apparatus. On a hatch back car and high mount will be different.