Saturday, May 13, 2017

Load test completed!


The load testing of my Goat wing was completed today, with a little help from some friends.  This was a 3G static load, not counting the approximate 0.5G of the structure itself.  I'll give the load distribution in a table at the bottom of the post for completeness.

Mary helped me move the equipment to the back yard, set up the test rig (again) and the Goat nose/wing.  We lowered the rig onto the upside down Goat, supporting the gantry with two ladders, tied the water knot at head-height, and then raised the gantry plus Goat back into the air.  Raising was difficult with that much more weight hanging in the balance.

We then moved all the distributed weights out to stage them next to their location, checking each and every jug for correct weight to within a half-ounce.  This task took about a half-hour with two people, and was a surprising amount of walking.

With Goat hanging upside down and staged, Trent and his family came over to help.  We loaded weights in tandem, each placing weight at the same time with a little nod when ready to let go.  We alternated between the leading and trailing edges, usually two leading edge stations, then three trailing edge stations.

About half-way through loading from the root toward the tips, we noticed the trailing edge spar had a noticeable bow to it.  Not unreasonable, but definitely visible.  Nothing looked out of place or any weird sounds, so we finished the loading.

Despite the 75/25 leading to trailing edge distribution, the trailing edge spar had much more visible bend between the strut and root connection.  The leading edge spar did not have appreciable bend, likely supported by the nose flying wires.  This matches the FEA runs from before.

The tips outboard of the strut connection likely bowed up, but were pulled back down by the load.  Looking down the spar showed a distinctive gull-shape on each half.

We let the weights hang for about two minutes, taking several pictures and inspecting the structure for any deformations.  We then took weight off, from the tips toward the root.

After a small celebration that a decade of work passed its 3G load test, we looked over the structure for any permanent deformation.  The compression strut where the wing struts attach has a noticeable bow to it now.  This is the same strut where the aileron slide tube attaches.  The trailing edge strut itself had a bow to it now, almost like the tube separating the leading from trailing edge strut was too long by about a half-inch to inch.  More investigation needed on that one to identify if that's the cause, or it's permanent deformation from the test.  Nothing else was identified visually as deformed.  Even the nose carry-through tube, where all the weight was hanging from the strap, looks fine.

Disassembly was uneventful, other than realizing it will be much harder to do once covered.  The gantry is now dismantled and all those milk jugs are empty and awaiting recycling.

It's fitting that Trent was available to help.  He was the one who helped set up and run the FEA.  Trent's comment was "Looks just like it did in ANSYS."  Thanks man!

3G load distribution
Rib     LE      TE
13      15.9     05.3 (tip)
12      23.3     07.8
11      27.6     09.2
10      30.2     10.1
09      32.0     10.7
08      33.3     11.1
07      34.3     11.4
06      35.0     11.7
05      35.5     11.8
04      35.9     12.0
03      36.1     12.0
02      36.3     12.1
01      36.5     12.2 (root)

Sunday, May 7, 2017

Load distribution

I tried to do a proof load, but couldn't figure out a great way to get over a hundred milk jugs hanging from a central point.  A friend recommended trying a beam just like the spars would be, and placing the weight along the beam.  That just wasn't as balanced as I'd hoped, so instead I've decided to trust that this structure is okay and push toward the actual loading.  It didn't budge when both my wife and I hung from it, which is certainly not +1000lb, but it's sturdy.

Yesterday morning while it was raining, I pulled up the AVL model and set up the model to have 50 strips.  These were merged into 25 pairs and the strip forces calculated.  The strip forces give load distribution along the span, and that was scaled up to 3.0G's.  I could always add more mass in a second test based on this same scaling.  I had been saying 3.5G's for a long time, and now that I've done all this work, the extra half-G probably should have been put into this water weight.  Oops.

Anyway, here's the pile of water all ready to be hung.  Every jug was individually measured and matches the load within a few tenths of an ounce.  If you look closely, you can see the masses written on the jugs.  It's a 75% leading edge, 25% trailing edge split, using the actual distribution from the AVL model.

Last night I spent two hours cutting and tying lines between pairs of jugs for looping those over the spars, and bending clothes-hangers into makeshift hooks for the singles.  It's ready to go.

I took off the wheel from the nose section for getting some additional height, and the aileron crank arm for setting up the sling without that in the way.  The pick point looks good.  I can either push the slings out toward the strut connection points or inward toward the wheel support struts.  Outward is more similar to the loading point from the pilot sitting in the nose, but riskier if the slings try to slide inward.  Inward is easier and directly transfers load to the wheel struts which then transfers to the strut connection points, but less representative of how the pilot's weight transfers to the struts.  Either work in my assessment and will depend on feeling out the setup when hanging.

So it's Sunday now and it's quite windy and gusty outside.  This just isn't conducive to a large-scale load test with literally a decade of work hanging in the balance.  Everything is back staged in the garage waiting for better weather and my availability.  Hoping for next weekend, or it'll have to wait until June...

Tuesday, May 2, 2017

Load. Mass that is...

Soooo many water jugs.  Count 'em if you'd like.  150x.

Looking toward proof loading, perhaps tomorrow night.  No clue how to hang this much weight off a common strap.

Sunday, April 30, 2017

Load testing rig fab

Well that was quick.  An early morning run to Home Depot, a couple hours with the circular saw and drill, and a LOT of sweat.

Just sitting on the grass, this is a bit tippy.  It will need some sandbags placed on the feet to keep the gantry from tipping over.  I had a swing-set of roughly the same design as a kid, but those legs were put into concrete in holes.  A long 2x4 screwed to the bottom of the gantry would work too.

It took two people and some very awkward rope work to flip Goat and pull it up most of the way to the beam.  Next time I will remove the wheel to have another half-foot.

The toughest part now is how to secure the structure to the overhead beam.  I have that yellow strap looped over the beam and looped around the wing carry-through tube, but can't quite get the water knot tied.  There's a foot overlap, but another foot from removing the wheel would make it workable.  How to get the structure all the way up there and tie it may be a trick.  A friend suggested taking one end of the gantry off and lifting it up after tying the knot.  It may also work to tip the gantry over sideways.  It'll be easier with more people.

I want to space those pick-points as wide as possible so it is closer to the ends of the tube instead of near the middle.  The wheel trusses will help transfer the load around if picking from the middle.  Just a lot of bending on that tube...

While it was up, here's how I intend to hang weight.  Each of those gallon jugs is 8lb of water.  Looping the string once around the spar provides some friction, and actually looping it over the rib attachment would be smart too.  I'll need to do the math for splitting up the load into rib-spacing increments.

Since there will be so much riding on this test, I will also do a proof loading of the gantry.  3.5G plus the roughly half-G of weight is 350lb*3.5 + 100lb is 1325lb.  Splitting into 8lb increments is 165 milk jugs.  Woof, that's a lot of jugs.

The weather Sunday is supposed to be gorgeous!  Load testing party at my place on Sunday :-)

Saturday, April 29, 2017

Load test rig design

Among other tasks, which I'll get to in a minute, I finally made motions toward a load testing rig.  The Airchair group didn't have any input, so I'm heading into this without others' experience.

I pulled the Goat drawing profiles to make a low-fidelity model of the wing and nose sections.  Flipping this upside down and with an angle of attack, this let me design a loading rig.  I happened to have a pressure-treated 2x10 that is going to be the main beam.  Beyond that, it was a matter of figuring out how to transfer all the load to the ground in the simplest assembly using 2x4's and screws.

Here's what I have...

The legs are made from 10ft 2x4's.  The supports look good enough to me on paper, but I might make adjustments when building the real thing.  There should be zero doubt that it's sturdy enough before getting a green light.

With the 10ft legs and even leaving on the wheel, that leaves about 2ft of clearance from the ground for adding weights.  In my attic are a few hundred empty gallon milk jugs that are about a foot tall, so is a foot of clearance enough?  From previous finite element analysis, the shape of the wing under load should actually lift the tips away from the ground.

Another funny is how to distribute the load to the ground.  I don't have a large enough concrete/asphalt pad, so this rig will be resting on grass.  It's almost 1500lb sitting on the feet.  Adding another brace to help distribute along more surface area might be prudent, or a piece of plywood under either end.

What else?  A bit of work on the nose that started with fabricating the nose skid from some soft Home Depot aluminum.  This is a full 1" width instead of 3/4" called out in the drawings, because it seemed to fit nicely and was what I already had.

After making the nose skid, I really wanted to rivet it, which couldn't be done before covering the nose.  So I covered the nose...

I thought there would be enough material to cover the nose section with one piece, but ended up having to slice off the back and do that separately.

It ended up nice in the end though.  Obviously it's not painted, and still needs to have reinforcement tapes applied.  I was worried the nose would be the most difficult part to cover given how many edges and bolt heads are involved.  It was really not that bad.

It did claim one casualty though ... my trusty old covering iron gave up and broke where the handle attached to the shoe.

Friends who can weld to the rescue!

Wish I could say it was as good as new, but the weld held long enough to finish off the first layer of covering and the handle broke just next to the weld, doh.  It was the right fix at the right schedule though, and I want to say thanks to Trent!

So next up is the load testing rig, sorting out the math for the distributed loading, and then having a load-test party with several close friends.  Assuming it passes, it'll be off to the races for final preps prior to covering the wings.

Sunday, April 16, 2017

Thoughts about load testing

Various forums online are for or against a load test.  It is my understanding that every single home-built aircraft in Europe must pass a load test prior to getting approval to fly, and those airplanes aren't crashing because of the test.  And a well-managed test to a load condition within the expected operating range should do no more harm than flying normally.  I'd MUCH rather something break or deform on the ground where I have a chance of catching and correcting the error, versus in-flight.  Practically, there are enough Goats flying (ten?) to give some reasonable confidence in the structural design.  But, and I have to keep telling myself this, it's such inexpensive insurance.

For the positive load test, I will be using some heavy-duty sling straps looped around the nose top tube structure where the seat attaches, simulating my body mass being the load.  These straps will be set up to hold the assembled nose and wing sections at a slight angle to simulate an angle of attack.  And then a series of jugs of water will be split along the leading and trailing edge spar tubes and hung using string or maybe bent coat-hangers.  Hanging weights will get more definition after building the hanging rig.

Available to me (four hours away) is an old wooden swing set with plenty of structural margin to hold the mass and plenty of height too.  It's a lot of work to haul Goat that far, and back again.  Or, it's a lot of work to build a big and temporary test rig in the backyard.  I'm torn.  It'll probably be easier to build a copy of the swing set gantry in the back yard here.  Maybe I can give it to a colleague with a youngster afterwards...

Mounting ribs

Ribs, so many ribs.

It's not so much fun repeating a task over and over again.  Once figuring out how to do it, the process does get faster the second, third, and fourth time, however.  This gets easier when coming across weird problems, like this rib that needed to attach right at an aileron hinge bolt.

I finished epoxying the ribs to the right wing and am just six trailing edges away from finishing the left wing too.  It takes about fifteen minutes to prep a few ribs, sanding with dull 220 to give some tooth, cleaning with mineral spirits (for any leftover masking tape adhesive) and isopropyl alcohol, and clamping the fiberglass tape in place.  For the first several, I epoxied the aluminum then wrapped the fiberglass to pull the epoxy up through the fabric like is proper.  The glass is such a loose weave and wets out so nicely that I began to pre-stage the material and then wet out generously to see that the epoxy contacted the aluminum surface.  I can't tell a difference in the bond quality, and this is a non-loaded bond, as the fabric will tend to hold the ribs in place anyway.  It takes another fifteen minutes to mix a small batch of epoxy, wet out the four staged rib attachments, ensure they're nicely soaked in, and clean up.  After four hours or overnight, I come back to trim the excess fiberglass tape.  A quick sanding with a block flattens any raised surface texture, being careful not to cut the underlying fibers.

I also filled in a couple mis-drilled holes with spare bolts.  As I understand, it's improper to leave a hole, and putting a bolt to fill the hole is an acceptable repair.  In both of my cases, there are other bolts nearby that will bump out the covering, so it isn't a big deal to cover over another bolt head / nut.

At some point, I replaced the aft cabane tube.  The upper cabane compression strut needed to be replaced, or so I thought.  The old ones were in perfectly good shape, just needed to be mounted.  Instead of fishmouthing new ones, the old ones were gently wedged into position, mounting end marked, a new hole placed there, and new washers riveted in place.  To install that strut, some minor pressure to pull the aft cabane tube away lets the compression strut to be slid over the washer mounts either end, and the pressure released.  Fits great, and was a lot less work than making new ones.

This Goat project is getting close to being done.  I don't have any additional planned metal work.  The leading edge shells need to be mounted and sanded and glassed, and that should finish off the required composite work.  I could fair the struts with foam and fiberglass, though that is optional for a maiden.  Then it's really just load testing and covering (in that order).

Locations of visitors to this page