Disclaimer

This blog is for entertainment purposes only, and is not meant to teach you how to build anything. The author is not responsible for any accident, injury, or loss that occurs as a result of reading this blog. Read this blog at your own risk.

Friday, July 24, 2015

Nose and nose gear - part 28

Wheel well mod #4 - nose gear doors (38.6 hrs)

This departure from the plans took a bit of imagination and ingenuity on my part since I had neither a set of instructions nor drawings showing me what to do, I did however have pictures of the gear doors I was trying to replicate courtesy of my friend Beasley (aka BizMan), and that proved to be enough.

For those just "tuning in" for the first time, one last word of caution before I begin. It should be apparent that these doors would never be able to close fully, had it not been for prior modifications that allow the nose wheel to stow completely within the well. These modifications are documented in earlier posts.

It almost goes without saying that, in order to create features that are symmetrical with respect to the longitudinal axis of the plane, one has to first be able to determine where its centerline (CL) lays. 

I had marked such a line on the foam previously, but many layers of fiberglass and peel-ply later, it was a bit hard to see, and then only in a few spots. Using a laser and a silver marker, I recreated it anew on the outer fiberglass shell where it should be easily seen from now on.


Marking the fuselage centerline on the outer skin


My gear doors will be about 3” (7.6 cm) wide and roughly 14” (35.6 cm) long, and will be hinged on either side of the centerline, so the first task after finding CL had to do with laying out two parallel lines 6” (15.2 cm) apart.


Outlining the contour of the gear doors


Then it was time to… cut. 

I become apprehensive every time I have to cut perfectly good fiberglass. I suppose I am worried about cutting in the wrong place. This time however I had been going over the whole process mentally for months, so “I had already been there before” and the cutting felt more rehearsed, and less intimidating.


Cutting the excess glass (Alea iacta est!)

With the corners gone, I used The Dremel tool with a depth stop to remove the blue foam down to ¼" (6.3 mm).


Removing foam

Foam recessed to ¼" depth


The foam I used on the gear doors came from the same sheet I used on the instrument panel, F22, and F28. The doors are actually made from scraps I have been saving for nearly five years, and protected under a cover from sunlight. 

Too much info? 

You are probably right. Let’s move on!

I cut the foam to fit the hole, then masked the surrounding area generously. 


Foam panel fitted to the wheel well


Masking is one of those things that are very time consuming, but the more masking you do, the less frustrating the fiberglassing experience becomes, and the better the final results. 


With masking, more is always better. 


With such a wide "safe zone", glassing the flat panel was almost relaxing.


Usual foam slurrying just before glassing begins

Panel glassed


A day later, I trimmed the long sides of the panel flush with the new wheel well opening.


Panel cut 6" wide


Now, the door hinges will screw to two ⅛” (3.2 mm) thick aluminum anchor plates that will be embedded in the foam and covered by fiberglass, so I needed to remove blue foam to at least a depth of 1” (2.5 cm). Eventually, as you'll see in later images, I ended up removing the foam all the way to the next fiberglass surface, 1.6” (4 cm) down from the outer skin.


Foam removal to 1" depth in progress

More of the same

Foam removal complete

This is where the anchors should go

The hinges would go over the anchors


One problem that I ran into, is that the router ate up some of my "mod #3" fiberglass, so I will have to fix that again.


Routing opened a new hole into the fuselage floor (top left)


Nut plates needed to be riveted to the two thick aluminum anchor plates, and lightning holes added to help the flox to grab on to.


Drilling four AN3 bolts clearance holes on both anchor plates

With a bolt screwed to the nut-plate from the back, and a Cleco clamp holding the nut-plate, the first loop hole is center-punched...

...then drilled to become the first rivet hole.

Cleco fastener installed in the first hole, before removing the clamp, and drilling the second rivet hole.

Lightening holes being drilled

Countersinking a hole (left one already done)

Microstop countersink cutter detail

Squeezing a solid rivet (left one already done)

Nut-plate permanently attached to the anchor plate (back side)

Nut-plate permanently attached to the anchor plate (front side)

Test fitting the new anchor plate

Door hinges attached to the anchor plates

Test fitting the two assemblies


With all the hardware ready to go, the time had come for installing the anchor plates to the wheel well.


Enough foam was removed behind the anchor plates to make sure they would end up flush

Improvised anchor plates holding device

Prepping the job site

Anchor plates floxed and curing

The aftermath (this will need a little cleanup)

Much better now

Testing the hinges once again

Meanwhile, progress continued on the gear doors with  strategic removal of foam, and a temporary hot-glue connection to the hinges for fitting purposes. 


Marking the foam for removal

Don't let the word "foam" fool you, this stuff is incredibly strong. The table saw made things a lot easier. 

Sanding the leftovers was a bitch!

Let's try these babies on!

I used a hot-glue gun to temporarily attach the doors to the hinges

The proper hinge/door relation was captured by the hot-glue

I drilled holes to lock in the geometry, then I removed the hot-glue residue.

Back to test fitting

Test fitting with the gear retracter revealed a few interference issues. The first one with the gear leg cover, easily sanded away...

... the second one with the bolts that lay below. More foam was removed...


Although the location of the anchor plates turned out to be near perfect, I decided to slot the mounting holes on the hinges by 0.050” (1.3 mm) to allow for a little adjustment since I couldn’t get the doors to rest completely flat on the fuselage bottom. 


... a third issue was with the hinge side of the door touching the fuselage before the free hanging side. The mounting holes were slotted.


That did the trick.


Doors completely flat in the front

No interference with the strut cover

Doors flat in the back

With the doors in a “happy place”, my attention went back to the wheel well. I had to plug the bolt holes temporarily, then flox the remaining gaps and intersections, and fiberglass over the top of it.


"Masking is your friend!"

Two plies of BID everywhere

Saran Wrap (in red) rolled onto itself, inserted in the bolt holes, then cut flush.

Floxing every gap and corner

Glassing job completed, curing period started.

After trimming, the next morning.

"That came out pretty good!"

Drilling an initial hole over the Saran Wrap

Enlarging as needed

Removing the Saran Wrap from the bolt hole

Bolt hole is ready to go, no further cleanup necessary.

Using this method protected the threads from the flox


The next morning I was glad to see that the wheel well had come out better than I had anticipated. Still, I had to shorten both ends of the hinges slightly because they would run into the hard flox fillets, and would no longer lay flat against the anchor plates. Easily done on the belt sander.

What I needed to do next was to finish the inside of the actual doors. 

After shaping the foam as necessary to ensure a smooth layup, I added 3 plies of BID and let it cure overnight.


A little shaping was necessary

This is what I ended up with after sanding the edges of the foam smooth

3 BID plies resting overnight

Cured and flipped over the next day


After the doors were completed, the next biggest problem became how to sequence the opening and closing of the doors with the nose gear.

Since others have successfully gone before me in this quest, I saw no reason to “reinvent the nose wheel”, and I borrowed as much as possible from them. Even then, success was all but guaranteed, and failure always looming sure to follow any misstep.

I chose to go with a spring connecting both doors, and I have to thank BizMan for sharing his research on the type of spring, and the proper positioning on the doors to make it work. Without all his work I’d still be drilling holes all over my doors looking for the sweet spot.

I bought the spring at Lowes for less than $5, it is item #543578, and they can look it up on their computer by the sku# 038902002501.


The spring

In case you ever wanted one

"Some assembly required"


I started by shortening it to 6” (15.2 cm), the width of the wheel well. The easiest way to do this is to bend it, and cut it with a Dremel grinding wheel. This is hardened steel, and it’s really tough to cut with anything else. 


I wore goggles and a shield for this

But it turned out to be a non-event


As far as attaching it to the doors, I made two steel cylinders  ½" long (1.3 cm) to be bolted to the doors, and on which the spring would mount.

Now, I have to caution you, what I did here is temporary, and I plan on redoing these cylinders by threading the outer surface at 18 TPI. This will allow the spring to actually thread onto the steel cylinders and not slide off. I am also considering supplemental method of securing it, because I don’t want this steel part coming off in flight and hitting the propeller. That would make for a very bad day.


Largest diameter to easily fit, before the spring started deforming (I made the cylinders slightly larger).

Making the cylinders out of 4130 steel on my CNC mini-lathe

Drilling for a 10-32 thread

10-32 spiral tap in action (I love these taps!)

Cylinders ready for action

Next time I'll thread the cylinders outer surface for 18.5 TPI (18 TPI should work)


The only reason I didn’t thread them yet is that the CNC conversion of my lathe is not complete, and at this time I cannot single point thread. I still need to mount an RPM encoder on the spindle which will feed the rpm values back into the Mach3 controller software. The controller will then be able to vary the Z axis feed rate to match the spindle rpm at that particular thread per inch count.

I will post an update below when I get that done, but for now all I needed was a working proof of concept, and the slick cylinders will be fine.

Another critical dimension was where to mount the spring along the longitudinal axis, or seeing it from the point of view of the gear leg, where to have the gear run into the spring as it closed in such a way not to cause any issues. 

Luckily, there is a notch in the nose wheel casting that seems made just for this purpose, so I wasted no time taking advantage of it, and marked the position line for the spring on both hinges.


Marking the hinges


The exact place to mount the spring on this semicircular line of position would greatly affect the retraction mechanics. If the spring was attached too high, the doors would try to close before most of the wheel was in the well, and if the spring was too low, the wheel might be in the well, but the door could potentially still be partially opened. Of course the length of the spring also affects all of the above, so this is where a little intel beforehand pays off big time.

Looking at photos of BizMan’s installation, I noticed that after trying a few different locations, he settled on a spot that lined up with the row of bolt holes on the hinges, so I did the same thing.


Cylinders in place on the hinges (note how they line up with the existing bolt line)

"Ok, so this is the bid idea"

The "big idea" bolted to the wheel well


The notch on the gear casting hit the spring precisely where I intended, and dragged the door closed beautifully…


Notch hitting the spring

A close up


… but the door would not open fully on its own!





Gear doors opening only partially




This was a big problem, because it meant that the nose gear could not be retracted without smashing into the fiberglass doors.

To make a long story short, it turned out that a shortened spring would recover more promptly from the over-center condition the retracted gear had forced it in, and be able to fully open on its own.


Removing one loop at a time and retesting, I found that three less loops was all it took to make it work.






Gear doors fully opening