Mark Langford's cool plane fab shit

[KBScobravert]

One layer of 5.85 oz fiberglass was applied to the inside surface of the deck, along with a few additional plies in the vicinity of the roll bar. The roll bar is basically a 2" x 2" curved box beam, with about 6 layers of 9 oz cloth on each of the four sides. Trying to squeeze the two legs of the C together with about 50 lbs of force yields maybe .1" of deflection. This mama is STRONG!


The end result. Perfection has been achieved, at least close enough for KR work. The end result weighs 9 pounds, including integral roll bar.

[KBScobravert]

Now we're cookin'! Here's the engine on a temporary "practice" mount that I welded up out of mild steel tubing, mostly as a design/balance study, not to mention the welding practice. I'm fairly convinced that the engine is going to weigh right at 250 pounds "all up" with everything needed to run except a battery and fuel. It's not nearly as pretty as Pat's, but I'll bet mine has a little more power!

I was originally going to put my Weber 40DCOE carburetor on top, but would have been forced to have a 2" tall blister up there. Also, the air on the bottom of the cowling is higher pressure, and I had loads of room down there, so I later decided to move it down below.

The engine's closest point is the starter, which is about 3/4" from the firewall. There's still plenty of room for oil cooler, filter, coil, etc., and lots of room for the exhaust system. The intake system isn't welded up yet. Still waiting on the laser cutter to cut the flanges out for me.

This is a 58" diameter prop that Doug Steen let me borrow. It probably has way too much pitch (74"), but will make a good test prop. With a level attitude, I have 7" of prop clearance with the ground. I'll probably end up with a 56" prop, yielding 8" clearance. My thrust line is 2.5" below the top of the longerons. If all else fails, I have some 2.5" gear leg bracket extensions I'm working on.


This cowling is a stock early KR2, which is about 4 inches shorter than a Revmaster. This one is positioned just about right, but a Revmaster would point in from outside to prop spinner, filling in this gap.

The closest that the cowling gets to the valve covers is about 2.5". The intake manifolds are actually holding the cowling up, so they are a perfect fit. Other than the Weber carb, nothing else interferes with the cowling.


Jim Hill donated this cowling to me when he cracked up his KR2 in a cotton field a few years back. I really wish I'd let Kenny Boyer give me that Revmaster cowling that he made for me and brought to the Gathering. I thought at the time that I'd have to make my own, but it's now apparent that it would have fit with room to spare!
With the engine thus mounted, I was able to do a preliminary weight and balance. With the wings installed and airplane leveled, the pilot's side was 304.5 pounds, passenger's was 307.5, tailwheel was 16.5. I used "stand-in" weights for much of what's missing, but will add a 15 pound exhaust system to the front, counteracted by 8 pounds of aileron counterweights and maybe 3 pounds of rudder counterbalance. My battery is mounted to the rear of the firewall, and could easily be moved rearward (like in front of the spar) but I don't think that'll be necessary at all. Any other weight gains (like paint and wingtips) will probably be close to the CG, so I think all is well. Grand total is now 628.5, so I'm confident that I'll still be under 700 pounds. That sounds like a real porker, until you consider that I expect at least 135 hp at 3500 rpm.
The pictures above were taken without all of the "standin" stuff piled on top of the engine, like spark plug wires, oil cooler, filter, etc. The chain represents the cowling. I think I could have built to this point and only had it weigh 600 pounds, if I could do it over again.
Most of that weight I'm referring to is stupid things that I did while building mine. For example, I was originally going to use the NLF(1)0015, so my spars were not as tall, so to get the same strength I made my spar caps 5/16" thicker. Then when the new airfoil came along, it was back to stock height, so I had to add another quarter inch back to both top and bottom to get back to the original depth. I figure this cost me about 8 pounds altogether. Then I had epoxied the wing templates for the NLF to inner and outer edges of the stub wings, and they had become part of the wing tanks, so I couldn't cut those out, so I cut out some more (with lightening holes) and added them on to the others. More weight. My aft deck has too much foam in it, up to 2" in places. The first one I built was super stiff and only weighed 6 pounds. I should have perfected that method (vacuum bagging) and done the next one that way, but it was a lot easier to just hog out a huge block of foam and glass both sides. This is where I got the idea to do wing skins the way I did them though. The final one weighs about 11 pounds, and it could easily have been 5 if done correctly. The canopy frame used lots of glass to stiffen it. I should have used carbon fiber for the whole frame and it would have been a lot lighter. If I had planned my aileron cable routing better, I'd have simply run an aluminum conduit through the wing tanks, rather than partioning off a wedge shaped area for "future expansion". That would've saved 2 or 3 pounds, and given me 3 or 4 more gallons of fuel capacity. The Cleveland brakes are pretty heavy, so I'd be tempted to use the GPASC or Tracy Obrien's wheels and brakes. I'd also have used the 18/15% airfoil to lighten up the spars and increase fuel tank volume, but I'd have had to add 2" to my spar caps to use it!
In the "other than weight" area, I'd use bent up center spars so the flaps could start at the wing root, spring bar gear fastened to the spar under the fuselage sides (like Grove Aircraft sells) but 6" taller than the Diehl gear, and I'll eventually try elevator balances like Richard Mole designed (the ones on Dana's tail) which Richard says will improve stability and while balancing the elevator (although my intuition makes me wonder about the stability part). I'd also make the fuselage a bay longer and the tail only 76" wide, and a gullwing canopy like Troy/Bobby's (but made from a Dragonfly canopy. And I'd widen the fuselage 6" rather than 4" like I did. I'd also put a 1 or 2 gallon header tank in the front deck that stayed full, kinda like Don Reids's. I may still do that one.

[KBScobravert]

Originally, I had planned to mount the engine mount "tray" directly to the firewall. These bolts were to protrude through the firewall and pick up a pair of 4130 tubes connecting the top and bottom horizontal spruce members.

The "spools" were made out of 4130 tubing and AN970 washers. Here they are jigged up on the firewall, except that the firewall has a piece of quarter inch plywood between the mount and the acutual firewall, so I wouldn't burn the real thing while welding. Using threaded rod minimizes heat transfer area to the bolts, to minimize the possibility of charring the bolt holes while welding.
While I was almost forced to drill some holes in my firewall to determine some basic Corvair/KR2S measurements, now that I've done it, you don't have to. So the thing to do now is build the mount up on a jig (preferably made from half inch steel) using the measuremetns that I've come up with, and then simply locating the mount properly to the firewall and match drilling the holes.

Here you see another reason why I mounted those bolts to the tray, so that I could jig up the mount on the airplane and ensure that the tray was level in all directions before any welding.

Here, everything is leveled and held in place while the diagonal braces are fitted.

The diagonal braces were drilled at an angle so that a perfect fit with the spools could be achieved.


This is what the Dragonfly guys are doing. One look at this (assuming it's been load tested and it works) convinced me that I could get away without the lower diagonals, which make dodging the exhaust system a little tricky. Note that this Dragonfly system is essentially like what I had planned, with the tray fastening directly to the firewall. But the difference here is that there are no supports below the tray, so all the loads are reacted at the top two diagonals, and at the firewall in the same plane as the tray. If his works, mine surely will! It should be noted that I've yet to lead test mine, but I will shortly.

This is William Wynne's motorized spool welding jig. He just sets the parts on there, tightens the nut, turns it on, and MIG welded the nut to the 4130 tube. While at the Corvair College, we pressed one of these just about inside out, and the weld didn't break, so I guess it's good enough!

So I went to the Corvair College, after William assured us that anybody that wanted to learn how to weld 4130 steel would get some hands-on-experience. I welded several of the joints on this tray, with Terry Bailey and William's supervision. I figured I'd learned enough to go back and finish up my mount, which I left fastened to my airplane, since I didn't have time to completely tack it together and bring it with me. And if I made a big mess of it, it would be good experience for a second "keeper" mount. Well, this didn't sit well with William, who informed me that my attitude was unacceptable. I simply needed to build a mount (right now!), take it home, mount it to the air frame and be DONE with it! So I called my wife to take a few critical dimensions from the firewall, and we went to work. During dinner we finalized a new mount design on a napkin (having seen what the Dragonfly guys were getting away with). Sunday morning I went to work cutting and notching tubing, Pat Panzera welded up a bunch of spools with the MIG welder, and William fired up the torch to get this show on the road!

[KBScobravert]

No, he's not wearing a leather welding apron, or long sleeves, but he definitely knows what he's doing

Here's what a Revmaster cowling looks like under a Corvair on a KR2S, with prop hub about 29.5" from the firewall (just like a stock KR2).

Obviously, the bottom of the mount is a little low in this shot, since I hadn't cut the forward extensions of the tray yet (which will become cowling mounts).

Here's why I let William let me talk HIM into doing the welding for my mount! Thanks a lot for the help!
Update, as of December 1, 2001

Here's the finished product. Please ignore the fact that the bottom of the mount misses the firewall compeletely! That's what happens when you try to recreate your firewall/engine geometry when you're 800 miles away from the airplane! This little mistake cost me about 2 more pounds to reinforce the "new" attachment perch. The bottom line is that if you mount your engine so that the bottom of the aluminum case is 7.5" up from the bottom of the stock KR2S firewall (which puts the thrust line up 13.875" from the bottom) you'll just barely miss the mechanical fuel pump if you use a Revmaster (or presumably stock KR2S) cowling. Other than that, I put the upper diagonals over close to the longerons, and centered on the 3.5" x .625" spruce cross members, backed up by .125" aluminum backing plates, and rested the bottom members on the lower edge of the firewall at the intersection of the two spruce cross members that are at a 101 degree angle to each other.

Somebody asked what the dimensions are for my mount. These pictures, along with a Corvair engine block, should answer those questions. Here's the front view. All tubing is .049" wall 3/4" diameter or square.


Top view. The end of the tape measure is touching the firewall. This results in 3/8" of clearance between the nose of my starter and the firewall.

[KBScobravert]

Side view. The airplane wasn't level when I took this picture.
 

Lower mount detail. The tubing that the 3/8" bolt goes through is 1/2" o.d. .058 wall.


Joint details. Gusset is .063" thick 4130. The round things are just big washers.


Urethane shock mount bushings are inexpensive and last forever. Obviously, hardware is temporary. Many of the details of this mount are borrowed from William's Piet design, but the geometry was a joint effort cooked up over dinner at the Corvair College. Note that it is untested at this point.

These are the reinforcements for the cabin side of the firewall, to tie the two spruce members together better. The plans called for two firewall reinforcements consisting of a large 1/8" aluminum angle that weighs 3.5 pounds for both top and bottom. My two little bent-up plates are all I need, and weigh less than 3 ounces for both. Since my engine mount points extend to locations near the longerons, there's no need to stiffen the firewall out in the middle. Hmm, that would be a total of 7 pounds saved. All I used at the top was some 2" square aluminum plates to keep the mount from pulling out. I also reinforced the corners with several layers of glass tape laid up over flox fillets to assure a perfect connection between memebers, in addition to the T-88 glue bond.


This lower mount reinforcement is a 1/8" thick piece of 6061-T6 bent at a 101 degree angle with a 3/8" radius, with a weight of 38 grams, in case you were wondering. It's located behind the firewall (cabin side) and serves as the connection point for the lower engine mount points.

[KBScobravert]

Cowling Construction
Originally written July 14, 2001, revised May 19, 2002 (last picture at bottom)

This is the beginnings of my cowling plug, which I made from a mold that I pulled off of one designed for the Revmaster and the KR2. It's made out of carbon fiber because for about a week or so I subscribed to William Wynne's attitude to "just get it done and go flying". I thought I'd just make this thing work and be done with it. But, since the KR2S is an inch and a half wider, and my particular installation is 3 inches taller than the stock KR2, I had to do a little surgery on it. The yellow stuff is two-part urethane foam, and the voids were caused by trying to squeegee the stuff into the gap. The other side turned out much better, since I turned the gap horizontal and didn't touch the foam once it was poured into place.
The interesting thing about this picture is the spinner. I was concerned about the flow from spinner to cowling, and have looked at several spinners trying to find a match. Troy Petteway let me borrow three of his, and this one is perfect! It's a 1-340-100 from Wag-Aero, (aka Aero-Fabricators), and it's their Cessna 150 spinner.

This is the view from the third story window of our house. You can see that there is a definite discontinuity between the profile of the cowling and the fuselage. I need to fix that...

...and you can see from this that ugly is a definite possiblity here. The yellow stuff on the side is another layer of 2 part foam that I added to the "slab" sides to round them out a bit. The more I looked at it though, the more it looked like it needed more than a little rounding. The Revmaster is just not designed for the S or the Corvair. The lines just don't fit the plane, and there is clearance for things that don't exist on the Corvair engine.
The overhang at the bottom is intentional on my part, since I plan to build in a NACA duct into foam glued and glassed to the bottom of the fuselage for exhaust and cooling air to exit (I got this idea from Troy Petteway's latest incarnation). There's a drawing of NACA duct dimensions in Tony Bengelis' yellow book, on page 200. I'll post a drawing of it after I CAD it for a full size plot.


As you can see here, there's a good bit of extra room around the Corvair for the bottom of the Revmaster cowling. In fact, there's a lot of wasted space down there.

I needed a little clearance in the cowling for my intake manifold, although if I'd been smart I'd have simply milled the "integral" Corvair manifold off about 3/4" and made this a lot less noticeable. I may still do that [later I did exactly that].

[KBScobravert]

I thought about eyeballing it, but while at Walmart I found a styrofoam doughnut looking thing (presumably for Christmas wreaths) hat looked perfect for the job.

I microed a section of it in place...

...and with a little work it became a "bump".

The more I looked at that bulge on the bottom, the less I liked it, considering that I didn't need ANY of it to mount my Weber carb underneath...so I exorcised it!

Doesn't this look a lot better? Well maybe not yet...

...but it's gettin' there!

Next I started looking at the inlets, which are about 3.5" x 6" long, and offset from the spinner a half inch or so. I seems that the spinner should force air directly into the cowling duct, so I'll move the inlet over towards the spinner so that the opening and the spinner align. A quick check of the spinner/head relationship revealed that I only needed 5.75" wide inlets, so I'll shorten them and make them taller. Thorpe says allow .35 square inches per horsepower, so I'm shooting for 47 square inches (135 hp) total area, or 24" per side.

[KBScobravert]

My inlets will need to be more squarish, and located higher up the cowling so as not to impinge on the front side of the cylinders, so hey, I might as well redo those too! I'm going to fill them in and completely redo them. Let's see, I think I have about one quarter of the original cowling left. It's a good thing I made it out of $200 worth of carbon fiber! While I'm at it, I might as well redo the spinner area and make my gap small and uniform, hence, the 3/8" urethane foam stuck on top.

Here's a back-to-back comparison. I think I'm improving it...

Inlets are slightly larger than the Revmaster, measuring around 42 square inches combined, compared to the Revmaster's 36 square inches, but then my Corvair engine will be putting out almost twice the power of the typical VW engine. Thorpe says use .325" per horsepower, which would put me around 45, but Paser is getting away with 30 square inches running a 160 hp Lycoming. I think I can eventually end up there too, but it's better to start big and then after testing and breakin, fine tune them to their smallest dimension. Troy did this using foam and duct tape until he got his inlets down to 20 square inches on his 2100cc VW engine. Note the ramps which will immediately direct the flow up over the front cylinders.

You know, as nice looking as the above cowling was looking, it still just wasn't good enough! What bothered me most about it was that I had managed to build the inlets such that the leading edge of the top was further aft than the leading edge of the bottom, mostly due to the much steeper contour of the top. Here you can see that my inlets were actually raked BACK at the top (the right, in this picture), rather than toward the BOTTOM for improved cooling during climb.

Richard Mole sent me this pressure distribution image of a cowling that looks very similar to a KR cowling. Notice the high lift over the top of the cowling, interfering with flow into the inlet. Although they don't actually show the shape and location of the inlets in this drawing (maybe there actually weren't any in this test), you can also see that the "bluff" part below the inlets (very close to vertical) are a real drag producer. And when the plane is climbing, the face that is presented into the "wind" is practically vertical, and the inlet area that the wind "sees" is decreased by the cosine of the angle of incidence, something like 3 percent at a 14 degree angle. That doesn't sound like much, but that's when you need your cooling most, so a 15 degree angle would maximize cooling during climb at low speed (100 mph), and reduce it slightly for cruise, where airspeed is much higher. Delivering more air to the duct than the engine can use for cooling results in the air coming right back out the duct, which is what I was seeing in the oil streaks coming OUT of the outboard end of Jim Hill's inlets.
Also, they were still larger than I will need. If Paser only needed 30 square inches for 160 hp, I know I won't need even that much, especially considering the efficieny of the NACA duct cooling outlet that I'm planning.

[KBScobravert]

So I started over! But as you can see by the above photo, it didn't take long to regret not having pulled a mold off of that beautiful cowling that I'd had before.

It took me a week to decide whether or not to go with the new round inlets that are are in vogue on newer spam cans, Lancairs, LoPresti planes, etc. I'm sure they are more efficient (research was done by Mississippi State, funded by NASA), but given the geometry of the Corvair engine (very narrow, due to the short stroke), the ducts HAD to be adjacent to the spinner, and very compact. This sends some of the high pressure spinner air into the duct and allowed a very small frontal area with more streamlining back to the "body" of the cowling. I may build a "holey cowl" later on just to see what happens, but for now, I'm trying to improve on the more conventional inlets.

Another thing that had bothered me about the previous cowling was that it was very difficult to get the inlets oriented the same with respect to each other. This time I vowed to try for perfection and drew up a template to sand to.

think this will be an improvement. The inlet areas are reduced again, down to 19 inches each, or 38 square inches total. I'm still confident that this will be more than enough for the Corvair, even considering that the stock Revmaster has 36 square inches of inlet area. After all, Marty Roberts is flying a Revmaster cowling with well over 120 hp. After I'm flying, I'll "choke" it down more, using foam and duct tape, to reduce it to the point where maximum speed is acheived and cooling isn't decreased. And thanks to the template, they are at least identical and symmetrical.

At this point the cowling plug is smooth enough for "prime time" and has one last layer of Featherfil to sand down. That should only take an hour or so with the palm sander, although it'll take 3 days to clean up the resulting mess.

Next I shot a layer of DuPont Centauri acrylic enamel paint on it (just for practice, and to give a slick finish) which I'll wax and coat with PVA (polyvinyl alcohol) and lay up three or four layers of glass to make a mold. Another round of wax and PVA and I'll layup two layers of carbon fiber to make my real cowling, which will weigh right at 4 pounds. Then I'll be able to retrieve my engine mount from within the plug, mount my engine, and get on with this stuff.

[KBScobravert]

[KBScobravert]

[KBScobravert]

Here's the final product. Not too bad, huh? It's just stuck on with duct tape at the moment, although the engine is actually installed, and the fit is amazingly good. It is well blended into the fuselage, although it's hard to see that from this photo. Keep in mind that I left 2.5" of space at the bottom for exit area which will blend into the fuselage with a NACA duct for exiting cooling air and exhaust. The white stuff around the top of the inlets is a flox/micro mixture that I used in the mold to assure full contact with the carbon fiber and the mold. Although it looks kinda ugly in the photo, it's really smooth. The little air intake clearance blisters on top were glassed rather than carbon-fibered because the glass is more flexible and conformed better. Any resemblence betweeen the blisters and front turn signals on an early 60's Beetle is purely coincidental.
Last night I trimmed the two parts down to shape with a Dremel cutoff wheel. Over the weekend I'll work to install the piano hinges that will join top to bottom half (while on the workbench) and then it'll be a matter of fastening the whole thing to the airplane. While I'm at this stage, I think I'll go ahead and install the fiberfrax and stainless steel firewall material, and start doing things for the "final" time, rather than trial-fitting.

Since I was able to use just two layers of carbon fiber, the whole thing weighs under 3.7 pounds, compared to eleven for a Revmaster KR2 cowling.
Below are some pressure distribution plots that Don Reid send me, based on my side view drawing and his pressure distribution (airfoil) software.