On short notice, I took a week off of working on the Defiant project. A friend of mine was coming to Phoenix to do a commercial glider add-on, so I invited myself along on the adventure. It was about a week of really fun flying. I highly recommend it.
After my soaring sabbatical, it was time to get back to work on Brian’s Defiant. As luck would have it, I had a friend decide to come to Phoenix to help work on the airplane. For those of you who followed along with Betty’s build, you will recognize Greg.
Greg is a friend of mine from St. Paul, MN and a Long Ez builder/owner/pilot. His Long EZ has been flying for 32 years. For the last several years of building Betty, Greg came over at least one day a week and many times two or three times a week to help me get Betty flying. His building experience and dedication are a big part of the reason why Betty is flying.
On the Defiant, we left off with the lower cowl mold halves laid up. When Greg got here, we started by building a wood structure around the mold halves. This not only gave it extra rigidity, but also allowed us a couple bolt holes to repeatably re-align the two halves together.
Then it was time to release the molds from the plug. Unfortunately, this didn’t go as smoothly as I would have liked. We ended up pulling quite a bit of the primer off the plug when they separated. Good news is there was a good coat of PVA on the plug so the primer came off the mold easily. I am not quite sure why the primer came off, but at this point no harm, no foul.
After a bit of clean up, we did the usual wax and then clay along the center seam. In addition to that, there was one area where we put 1/16″ sheet wax. The sheet wax is there to create a joggle, or offset. This will allow the top cowl to come down over top of the aft flanges of the bottom cowl and will make an integral mounting flange in the initial layup. It is hard to see in the photos, but it is the rectangular yellow opaque area on the aft vertical surface. Then it was 5 coats of PVA.
Since I am still trying to learn resin infusion and vacuum bagging, we decided to try resin infusion with the first part from this mold which will be a fiberglass tool cowl. Greg and I took our time cutting 1 ply of deck cloth and 4 plies of BID. These were arranged dry into the mold. Then it was all of the infusion consumables including peel ply, flow media, and then bagging film.
Unfortunately, the best we could get after hours of trying to fix our setup was 10″ of HG vacuum. We had major leaks. We tried everything we could think of. We went around the bag perimeter multiple times looking for leaks. We painted the outside of the mold with flex seal thinking there may be a dry spot with pin holes. We trimmed the centerline flange for access and bagged the exterior of the centerline flange/mold split. We put tacky tape on suspect areas on the outside. We put tacky tape along the sharp edges of the flow media. We added bags on top of the bag. We got a stethoscope out and listened for leaks. We blew smoke at the mold looking to see if we could find where the leak was.
None of that was successful. So we took the bag off the mold. Our plan was then to do a wet layup and try one new bag over top of that as a wet layup vacuum bag. We didn’t expect it to perform much better than the first bag (unless it was in fact the flow media that punctured the bag), but 10″ was at least some kind of vacuum and would give some of the benefits of bagging to the part. But this round of bagging went sideways too. The seal was non-existent and didn’t pull any vacuum on the part.
To make matters worse, I was frustrated enough with the time spent and lack of fixing the problem that I convinced myself to take a shortcut with the wet layup. The mold was already packed with our 5 plies very nicely. I was concerned about getting the plies out and back in the mold without distorting the fabric to the point where we would end up short of material in some dimension. So I reasoned if I could wet out a 5 ply stack on the table reasonably well, we could do that in the mold. (BTW, never do this with a flying part… it is very bad technique, heavy, and you cannot ensure the each layer of the laminate is wet out appropriately.)
As all of you have done composites work know, this was a major mistake. It was not only probably more time consuming in the end, but ended up producing an abysmal part. There are dry areas. There are inter-laminate bubbles. There are places where the cloth didn’t lay against the mold and it is not representative of the shape. etc.
There are 3 good things to come out of that mistake. The first is that the perimeter of the part is wet out sufficiently well to use the tool cowl as intended. It will allow me to precisely trim to fit the firewall and wing roots, etc and transfer these features to the mold. The second good thing is it will give us a part to use to ensure things like the exhaust are located inside the cowl with plenty of spacing, etc. Lastly, I was able to remove the part from the mold without splitting the two halves. The fiberglass tool cowl is a bit more flexible than the final carbon cowl will be, but I think I could get away with bonding the two halves of the mold together and still get a part out. This will remove one major area that is likely the major cause of our leaks.
All in all there was a lot of learning that went on with this part, mostly what not to do. I will end up with a good part out of this mold eventually, but for now we are going to switch gears to something else.
The lower cowl geometry doesn’t lend itself well to making it a single piece mold. The centerline scoop/tunnel/boat tail has geometry that would “capture” the plug, and subsequent parts, in the mold if it was laid up as a single piece. You can still do a single piece mold, but there are some draw backs.
If I did a single piece mold, I would have to break apart the plug to get it out and will no longer have it useable if we wanted to redo the mold or modify our shape. Also, you don’t know how the mold surface turned out until the plug is out. So you are destroying the plug and committing to making repairs on the mold surface if there are problems with the mold. Additionally, you are relying upon the final part that is laid up in the mold to be flexible enough that you can get it out of the mold without having to cut it into pieces or destroy the mold. For these reasons, it is a safer approach to make the lower cowl mold in two pieces.
In order to accomplish this, the only thing I needed to do was add a tooling flange down the centerline of the part. Of course, that means jigging the flange to our fully prepared surface. In order to protect the finish, I laid down a layer of vinyl tape first, then bondoed the tool flange on. You could also use hot glue for this.
From here, the rest of the molding process is mostly the same as what you’ve already seen me describe in previous posts. Tooling flanges, wax, PVA, and layup. Once you have one half of the mold done and cured, you can remove the centerline tooling flange backer board form. The completed mold half is left in place, and the centerline flange on the existing mold becomes the tooling flange form for the second side. This way they exactly match.
On the second half of the mold, I want a joggle or channel to facilitate sealing the centerline of the mold when assembled. This channel will allow me to put a rubber weather strip in between the halves and hopefully get an air tight seal. I wanted a 1/8″ deep channel. Turns out that my tacky tape was the perfect easy thing on hand to make the form for the sealing channel. I adhered tacky tape to the centerline flange, then I put vinyl tape over the top of the tacky tape as a releasable surface. This should give me a 1/8″ channel or depression in the second side of the mold to facilitate sealing the molds to each other. Again, the mold making process was the same as for the first half. Tooling flange forms, waxing, PVA and layup.
One thing you will notice is that one of the mold halves appears white, the other black. A few weeks ago, Dave Anders told me about a mold making technique he learned from Dave Ronnenburg, the designer of the Berkut. Dave uses essentially an “aircraft structure” approach to making molds where he has laminate-core-laminate. Dave uses micro for his core. This makes for very light, inexpensive mold structure. So on one side I tried that approach (white micro visible), the other side I did the woven cloth and chopped strand mat (see through the glass to black primer). I did this to see if I like micro core better and check if it saves time.
It turns out that both mold halves took me within 15 minutes of each other. The micro core mold took a large amount of micro-balloons, but was much less epoxy than the chopped strand mat required. Because of my temperatures, I didn’t dare mix up a large batch of micro, so I did it in 5 or 6 smaller batches. This certainly added some time to the process. I also have learned some better technique for chopped strand mat and how much epoxy it requires.
On these mold halves, I don’t see a clear winner between the two methods. However, if I could do a single mix (or maybe 2) of micro to get what I need to for the whole mold, I think it would be significantly faster and lighter to do a micro core. I can’t wait to get them off the plug and see if one is stiffer than the other.
Vacuum bagging is a skill I have yet to master, as was proven when we attempted bagging the old design upper tool cowl. so why not take it one step further and try to learn vacuum bagging and vacuum resin infusion all at the same time? I like to take big bites of the elephant, so why not.
There are some advantages to resin infusion. 1) it’s less messy and sticky. 2) it is more repeatable. 3) you have all the time in the world to arrange, cut, and straighten the plies of the layup schedule before going on the clock when you mix the epoxy. 4) you test the vacuum bag before mixing epoxy and if it doesn’t seal, you can replace the bag or start over with little lost but some bagging film. 5) you don’t have to wet each layer out individually.
There are also some drawbacks, the biggest of which to me is the steeper learning curve. Infusion also requires an infusion resin which is very watery and not suitable for a wet layup. So you will need another resin system/kit. Also you need some additional consumables like flow media to make it work.
Those pros and cons being said, I am hoping to infuse my final cowl parts and get a really good surface finish. Therefore I need to learn this new technique. I also was getting a little bored of working on the cowl plugs so I decided to spend the morning yesterday taking my first shot at resin infusion with my induction inlet mold.
Prep was the same as for any layup. Five coats of wax, five coats of PVA, cut cloth, cut consumables the consumables were peel ply, flow media, and bagging film.
For the cloth, I wanted one ply glass, two carbon, and a ply of glass. With the circular nature of the part, it is hard to get flat cloth to lay into the mold nicely. So I ordered some glass and carbon fiber sleeve. It is really neat stuff, almost like a Chinese finger trap. It is woven in such a way that you can expand or shrink its diameter within certain ranges. It really makes slick work of tubular or circular parts like this. You can find it at composites supply shops online such as Composite Envisions.
I then stuffed the mold with the glass, carbon, glass, peel ply, and flow media. From there it was tacky tape around the perimeter and make, pleat, and seal the bagging film. On one side there is a vacuum line, and on the other is the resin supply line clamped off.
To my surprise, it actually pulled pretty good vacuum. About 26” HG. I was pretty satisfied. So we mixed the Proset INF resin. We dropped the supply tube into the cup of epoxy and opened the valve. Low and behold the resin started traveling up and infusing the part. It was very fun to see.
Of course about that time we also started having leaks in the stretchelon bagging film. I think the cut ends of the flow media was puncturing the bag. It took me about an hour, and half a roll of tacky tape to find the leaks, but eventually I got the bag to seal again. It looked horrendous and I don’t suggest resealing this way.
By the end of my workshop session, the resin was cured and I de-bagged and released the part. Which wasn’t easy, especially with all the tacky tape on the bag itself.
There is certainly room for improvement in the part and my technique. The surface finish is not very good with one small area where the cloth bridged and many pin holes. I know it can be better, and I will keep practicing until I get it there. This part would be useable with a few hours sanding time and some finishing work. It is a strong and light part. But I should be able to make them where they come out of the mold with little to no requirement to do any secondary processing to use them. If I am going to infuse the cowls, I certainly need to be capable of better surface finish too.
All in all, it was a very educational day in the Garaggio, and taking the first steps in the infusion learning curve will get us that much closer to being able to use it as part of our regular bag of tricks.
The last few days, when I have been looking at the rear cowl, I have thought that the trailing edge shape reminded me of something, but I couldn’t figure out what. Turns out, it is the character Sid from the cartoon movie, Ice Age. While I am fairly certain this destines Brian’s Defiant to henceforth be named Sid, I couldn’t help but share my revelation. It gave me a pretty good chuckle.
The last few days in the Garaggio have been pretty exciting. Since Jodi’s working visit, I did add one more tooling flange just forward of the cowl to firewall intersection. On the first version of the cowl, it was challenging getting the plies of glass to end reasonably well and have the bag sealing tacky tape all in a 2″ area. The 4″ wide flange will be great to separate the plies of cloth and bagging materials and give more room to work. It will also make the mold dimensionally stable.
From there, on Wednesday and Thursday morning it was all the usual steps to ensure the mold releases from the plug. I rubbed it down with 5 coats of Part All mold release wax. I may or may not have mentioned it before, but I really hate waxing. It really isn’t that hard, and it certainly signifies that we are ready to make a major step in progress. But somehow, it always takes me way longer to do than it should, and its a lack of enthusiasm causing it. Never the less, a full 5 coats was put on the plug and tooling flanges. Spraying PVA Thursday morning was a breeze and pleasure compared to the waxing. With both done, the plug was ready for a mold layup.
This time, I was a bit smarter than last. I waited until Kevin had the time to come help. Good thing I did as this layup took a good 4 hours with the two of us. We did make things a bit easier on ourselves by using the “extra slow” hardener. I used West System Epoxy as it is cheap, I had a ton of it on hand, and it is available locally. A lot of you will ask again why not polyester, and I haven’t slowed down long enough on this project to give it a try and learn how to use it. I will eventually as I have some.
The layup schedule for the mold was:
1 ply deck cloth
2 plies uni/bid
1 ply chopped strand mat
1 ply deck cloth
Usually, I would use bid for the whole second and third plies. However, I had a ton of “scrap” uni laying around that I wanted to use up. My scrap pieces were 30 and 15 inches wide and 5 feet or so long. These pieces were folded and old, so not suitable for a flying part or anything truly structural. They worked fine, but bid is easier to work with. The bid wets out easier too.
Of course, we had some of the same issues with wetting out the chopped strand mat. It is a pain in the butt to work with, takes a lot of time to wet out with Epoxy, and is prone to air bubbles that are difficult to get out. Even so, it did build up the thickness, and knowing how thirsty it was, I could have Kevin mix up more epoxy each round than I did in the first top cowl mold.
I did make an error of omission during the layup. I forgot one of the plies of BID/UNI from the original layup schedule. So the mold ended up a bit thinner than desired. Surprisingly it is actually pretty rigid with the whole perimeter tooling flange. If this were a production mold that I was intending on pulling many parts from, I would go back and add more stiffness by adding plies. However this mold will pull just a few parts, so it will be fine as is.
Today, I released the mold from the plug. It was again surprisingly easy to do. I am so impressed with the mold release wax and PVA combination. It is a bit more labor intensive than I would like, but it sure does a great job of releasing. It took no more than 30 minutes to have the mold off.
From there I cleaned it up. That involved getting rid of stray clay and a thorough wash to get rid of the PVA. Cleaned, I then started on the path of polishing the surface. It was pretty good as it came off the plug, but I did do a round of each 400, 800, and 1000 grit sand paper on the DA. I think that should do it.
I likely will add a wooden structure around the mold for two reasons. One, it will help with a bit of rigidity since I missed a ply in the original layup schedule. The other thing is the wood structure will give me feet for the mold. Right now to put this mold on saw horses, I have to rest the mold on areas that will be a part of the final cowl. This would slightly deform the cowl shape, which is obviously not right. A wood structure will allow the shape to be unaffected. Other than that, the top cowl mold is ready to be used.
Early next week I will be prepping to make a top tool cowl out of fiberglass. I debated going right to the final carbon cowl, but there are a few advantages to making one out of fiberglass first. I should also be able to get a mold made for the lower cowl. It is significantly smaller than the top cowl, and should be simpler to make.
Some of the materials used to make the mold described in this Post:
The below images are amazon links, click on them to learn more.
On Monday, I got to the point in contour sanding and surfacing on the cowl where I was struggling to find flaws that needed to be fixed. Don’t get me wrong, there were still minor flaws, but most of them small enough that I would be liable to go backwards instead of making improvements if I added filler. With the cowl all in black primer and polished to 800 grit, it looks pretty darn good.
Next up is preparing to make the mold. To my fortune, and her chagrin, one of our Reno ‘September Family’, Jodi, got her flight cancelled in Phoenix. That means she got to spend some time in the Garaggio, and I got a workshop helper.
This mold making prep is the same as what we did for the first version of the cowl mold. Though this time a touch easier. We made poster board templates of the tooling flanges and cut them out of backer board. The backer board was jigged to the airplane using bondo blobs on the back side. Then the gaps were filled with clay and a nice radius formed in the corners to facilitate the glass flowing from cowl surface to tooling flange. Pretty simple, though takes a little time to get a nice fit on the tooling flanges and nice radius with the clay. Was nice having Jodi to help where more than 2 hands was nice, and to get the filleting done quickly. Thanks Jodi.
Last week Thursday, before our trip to Kanab, I got the layup on the lower aft cowl done. What an exercise in frustration. I have done quite a few layups in my building history, but this one made me feel like it was the first time I had ever seen epoxy or cloth.
The plan was to break the lower cowl into four sections, each with their own plies of glass. There would be left and right above the wing & left and right below the wing. Each of these sections would have some overlap. The above the wing pieces are easy. No so much with below the wing. Long story short, this may have worked had I had 6 pairs of hands, but not by myself.
Trying to get pre-wetout fiberglass plies on plastic to stick to the bottom surface of the cowl plug, all the wile working against gravity is a loosing battle with only 2 hands. I was using vacuum bagging film as my plastic backer for wetting out the glass on the table, which made things worse too. It is really nice smooth plastic, but it is heavier than you would think. So the additional weight of this plastic made each ply heavy enough it was impossible to get it to stick to the cowl plug. Oh yea, it is also 100 plus degrees out and the epoxy is kicking off quick enough that each piece was sticky to the point of difficult to manipulate before I could get a ply in place.
I tried everything I could think of to get it to work, and eventually came to the conclusion that I had to not only cut the plies into smaller pieces, but also get rid of the plastic to get it to stick in place. This made it really difficult to get the plies on in any kind of organized manner or with any concern for ply orientation, fiber straightness, etc. It was certainly NOT an aerospace grade layup. Luckily, this only needs to be a tool.
I eventually succumbed to the fact that the layup wasn’t going the way I wanted and it was best to piecemeal it. So I took about 1 foot square pieces of cloth, stuck them in place with overlap, and stippled on epoxy. It was messy, sticky, time consuming, and aggravating. It was not pretty, but it got the job done. And a fair bit of epoxy all over me.
By the time that layup was done, I was glad to be heading to Kanab for the weekend. I needed the time away from glass work, some R and R, and some cool flying. We know how that turned out. Instead I got more glass work, some R and R (just not the R and R I wanted), and a little bit of the cool flying.
Since getting home from Kanab, it has been more of the same process you saw for the top cowl on the rest of the aft cowl surface. Contour sand the fiberglass. Add filler. Sand. Repeat. Once it is to the stage where you think you have the shape, start spraying with primer. Sand. Find flaws, pinholes, etc. Fill Sand repeat. And then get Joe-ed.
While I was sanding on the lower cowl, I was thinking about how I was going to make a mold with a wing root into it etc. Any way I sliced the problem, I figured it was going to be a complex mold with multiple pieces. You certainly can do multi-piece molds, but it adds all kinds of challenges for vacuum bagging etc. So I took a step back and asked, why am I parting the cowl where I am? The simple answer is, thats where the original cowl parting line was. I was just doing the same thing Brian did.
Turns out that there would be significant advantages to moving the cowl seam to the plane of the top surface of the wing. One, there is a significant improvement in serviceability of the engine just with the top cowl off. You can get at just about anything you need. Two, it makes the molds easier. Top cowl can be one mold, so can bottom cowl. Three, it moves the cowl seam to a less obtrusive location for improved aesthetics. The downside is I have to make a new mold for the top cowl. But If the lower cowl was going to be 3 pieces, I am actually making 1 less mold than with the original plan.
So this week I glued top TCSN2 (tool cowl serial number 2) to the bottom TCSN1 and jigged it back on the airplane. I have the full shape now in primer, and for the first time looking very representative of the final parts. I am liking it more and more. A little more polishing and pin hole filling and we will be ready for tooling flanges and making molds.
To me, this cowl project has taken a long time. I am ready to move on to other parts of the project. I am ready to be able to claim victory over a part. In actuality, at just a bit over 5 weeks, its moved at a pretty good clip. I just can’t wait to have the satisfaction and MVP (major visual progress) of having actual finished cowl parts. That will be a lot of fun to see.
A weekend of relaxation at the Kanab Fly-in gets derailed by a canard delamination.
We have been working on Brian’s Defiant really hard for the last several weeks. I am pretty proud of the amount we have gotten done. I am even behind a post and need to update on the lower aft cowl layup that was done. But this post is about taking the holiday weekend off for some rest and relaxation, or so I thought.
The Kanab Canard Fly-In has been hosted annually for the last 30 years just the other side of the Grand Canyon and AZ/UT state line. It is a very low key fly-in in a small town that is relaxing. For once I am off work for the event. Seemed like a great time to go get some down time, fly Betty, and enjoy time with like-minded builders. Plus they host a “timed event” there so I actually get to race in 2020.
Friday morning, Kevin and I packed Betty up, and launched for Kanab. It is just over 200 miles from Deer Valley to Kanab, so about an hour and ten minutes. We left in the morning to have lower temperatures and better rides. Our route of flight was probably one of the most scenic I have ever flown in a small airplane. We flew up the Dragon Corridor of the Grand Canyon. The sights were beautiful.
We arrived in Kanab about 11 am amongst about a dozen other canard airplanes already there. We found a nice parking spot and started tending to Betty’s needs to put her to bed and get her cleaned up. While I was unpacking our bags, a Cozy owner, Steve Franseen, owner came up to the airplane and asked, “is this a de-lam?” (De-lamination) To which I responded, no, thats been there for a while, tap tests ok, and is a contouring issue from the paint shop.
A little background: For quite some time, in fact since I repossessed the airplane from the paint shop, there has been a noticeable discontinuity in the left canard tip. I remember seeing pictures during the paint job of a significant amount of filler added to the tip of the canard. All along I thought this surface discontinuity was from that. A delamination will have a hollow sound when you tap test it, and every time I had tap tested this area, I was unable to detect any hollow sound or difference in tone from the surrounding areas.
When I got the last baggage out of the spar, I dismounted from the strake and was immediately shocked at how large the discontinuity had grown. It was at least 3 times its previous size, and significantly taller too. It was inflating like a balloon right in front of my eyes. It tapped hollow, big time. The skin was hot to the touch. I had never felt it so hot, even in AZ summer days.
With the temperature nearing 100 degrees and the sun high overhead for direct harsh solar heating, I wanted to get the airplane into the shade, so I found and coordinated a hangar quickly. With the airplane in the shade, it was out of danger and time to evaluate the damage and our options.
This bubble was so big and had progressed so much so quickly that it needed to be fixed before further flight. It wasn’t the kind of thing I was going to limp it home at a slow speed and moderate altitude to make a repair. Some delaminations can be easily fixed by injecting epoxy between the skin and foam and letting it cure. That is the easiest repair if appropriate.
With our damage and the delam bubbling around an inch above the contour of the canard, we would have to get the skin to sit back down into contour for this to be a viable repair option. I tried pushing on the skin bubble slowly increasing pressure until there was a slight cracking sound. The skin only moved 1/16″. I further tried drilling a hole into the bubble to relieve any air pressure in there. But that didn’t improve the ability to move the skin back into place.
At that point, my options are to either try to heat the skin to the point where the epoxy becomes soft and try to reform the skin to contour and then let cool. Or to replace the skin. Heating the skin can do further damage and I have less control over the new contours of the canard. It also has the potential to do further damage to the foam core. I chose to cut away the skin.
I was pretty disappointed at this point. It was our first time to Kanab, and there are a lot of fun, but low key events planned for the weekend. A guided tour flight of Zion National Park and Bryce Canyon were the Saturday morning plans, and the race on Sunday. It also was embarrassing to have such an avoidable problem and was painting an airplane that is my pride and joy in a light not very becoming. Looks like my weekend of R&R is going to be a weekend of unplanned work, and we will miss the flying events.
I travel with a fly away kit (FAK, click here to see what I tend to travel with). My FAK has the tools in it that I think I could possibly need to make likely minor repairs while on a trip. I never plan on doing layups and skin replacements, so I didn’t have any of the tools or materials I needed with me. Not to worry though, this is where the aviation community really shines. Pilots, especially builders, go out of their way to help each other out.
I texted three buddies at my home airport to see if one of them could pack up the materials I would need from my hangar and fly them up to me. Without hesitation, they agreed. Brian, Travis, and George were going to pack up materials, and George would fly it up to me Saturday morning early.
While they were heading to the airport, I borrowed a dremel and cutoff wheels from the Kanab Airport Manager, Jeff. Of course that was an exercise in frustration because the exploding dremel cutoff wheels only last for a couple inches of cut. So I eventually went to a hand held hack saw blade. I cut away skin until we got into good glass and foam that was free from de-lam. Doing this skin-ectomy only took an hour. I wanted to do it before the boys packed up my materials to make sure I didn’t have any unexpected material/tool needs.
While I was doing surgery on Betty in Kanab, Brian fortuitously ran into another Long Ez friend of mine, Gary, at Deer Valley. Gary was washing his Long Ez, in preparation for a trip to the Kanab fly-in early Saturday morning. Brian was able to coordinate with him to bring my materials up to me, saving George the early morning flight. Talk about good fortune.
When the boys got to the hangar, we did a 45 minute long FaceTime session of me guiding them though my workshop to find the items I needed. Of course, all the while they were ribbing me for my less than clean, somewhat disorganized workshop. I guess I have some work to do there. But they successfully packed up a care package that was compact enough to fit in an airplane. Care package contents:
Pro Set Resin/Hardener
Sand Paper, (40/80/120/180)
5606 Hydraulic fluid (for someone else with gear/brake problem)
Saturday morning, I went out to the airport early to do whatever preparation I could on the canard tip until my materials arrived. I started with what ever tools I could get my hands on in the airport managers tool box. The preparation included digging a trough for a flox corner around the edges. By the time that was done, Gary arrived with my care package.
From there it was sanding to create my scarf joint. Sanding back 1″ per ply in the directions I could. Toward the trailing edge and the tip, I didn’t have the distance to do that kind of a scarf, but those areas are almost a fairing with only a 2 ply skin. So I wasn’t concerned with just lapping the repair onto the remaining skin “flange.” The DA sander and a variety of grits and interface pads made quick work of this.
We were working outside per the airport managers request, and rightfully so. The hangar was where multiple meals/banquet would be taking place and we didn’t want to create all the fiberglass dust in the meal location. But that also put the airplane back into the sun. Despite our best efforts to cover the airplane with borrowed canopy covers and tarps, the blue areas still were getting hot. So I sent Kevin and Greg to NAPA to get white paint. I scuffed up the dark stripes on the wings and we painted them white. No more heat worries for the rest the day for the wings. We also did locate a small delam on the right canard tip. This one was small enough to repair with injecting epoxy under the skin.
With the repair area on the left canard tip prepared, it was time to start rebuilding. I did a couple small pour foam areas into the larger gouges. By larger gouges, I mean the pour foam ended up being less than 1/2″ thick and in very localized areas. Since pour foam can continue to grow over time, I wouldn’t have used it in a big area as it is strong enough to change the canard shape. But with thin localized areas, it was fine.
It was time to mix up some epoxy and flox structural filler. The troughs around the perimeter were filled with this. Then the rest of the epoxy was mixed into a micro slurry the consistency of peanut butter. The foam was all slurried filling in the remaining dings and trying to keep the surface as close to proper contour as possible. I also had to do a small repair to the elevator counter weight cavity. There was about a 1 square inch area that needed to be replaced. So two tiny plies of Bid were put in place. I left this to gel for a while to make the next steps easier.
After a bit of time that allowed the flox and micro to get to a stiffer, but still manipulatable consistency, we started laying up our plies of glass. We replaced each ply per the plans schedule. This was uni, bid, uni, uni from elevator tip inboard, 2 plies bid or uni outboard of elevator end. Then per normal repair technique we added a ply of bid over the whole thing. Lastly, to save a cure cycle, and give me a shot of being able to fly Sunday, I mixed up micro and slathered it on the repair area. We had just exactly enough micro. To help the quality of the micro spread, I put peel ply over the fill to help prevent tearing and ridges. When I was satisfied we had the micro spread consistently, we called it done. The layup went really, really well. It was a really long day of work from 8 am to 5:30 pm for Greg, Kevin, and I.
Sunday morning, it was back at the airport at 6 AM. With the 100 degree temperatures during the previous day, the repair layup was cured. Pulling the peel ply revealed a nice repair that needed only minimal contour sanding. A combination of DA sanding to knock off obvious high points and my contour sanding durablocks to refine the shape, were the tools of the morning. Within the hour, I had an airworthy repair that was blended well into the rest of the canard. Maybe I can make the race…
Doing a repair like this, in my opinion, requires a test flight. While I was confident in the repair, in my mind it was only prudent to confirm unchanged flying qualities before leaving the airport area or having another person onboard.
When transitioning roles from mechanic to pilot, I try to make sure I take the time to reset my mentality and to slow things down to an appropriate pace. No one is served by rushing things like this. I preflighted the airplane and also had Greg go over what we had done to make sure we didn’t miss anything. With flight control throws verified, and all areas where we worked given multiple looks it was time to fly.
It was a beautiful morning to fly, and the takeoff was conspicuously normal. I then orbited over the field, slowly increasing speed until a high speed cruise. Handling characteristics were normal. I then dove the airplane slowly increasing speed to Vne. Everything was normal. Last I did a power off stall/pitch buck test. Everything was normal. All handling was unchanged, even the pitch buck speed was the same as usual. I proclaimed the test flight a success and came in for a landing. I shut the airplane down about 7:50. With plenty of time to walk over for the pilot briefing for the race at 8 am.
The race was a ton of fun. In other years, I would have gotten at least one race in by now, and be preparing for Reno. In 2020, this is the first, and likely the last race I will do for the year. Of note, it is the very first time I have gotten to race Betty. Which was a goal of mine since well before Betty first flew. We launched second in the race behind Dick Rutan and ahead of Allen Floyd. It was a beautiful morning to fly and Kevin and I enjoyed the sights, probably even more so after all the work we did over the last two days. Allen ended up passing us during the race in the last 20 miles. We averaged 217 MPH. The official results has Allen faster than us by 6 mph. But he also has at least 30 more horsepower than Betty has. 🙂
There are a lot of things to take away from my experience. First off, Burt said that you can paint your Long Ez whatever color you want, as long as it’s white. I can attest to the fact that the implications of heat build up in the composite structure of a non-white Long Ez is no joke.
I don’t want to give the impression that I painted Betty silver and blue without thought. I did in fact have samples of the paint colors and temperature tested them in Phoenix, in June, on a 110 degree day, at noon. These tests resulted in changing the colors to lighter colors than initially planed. But, as I’m learning now, there are many more factors to look at than just the temperature of a paint chip. The heat transmissivity of a paint sample vs paint on the exact same substrate (glass/foam), in a no wind situation, oriented at harsh sun angles (pointing south), etc are not trivial factors. It requires a much more in depth study than I did.
In the end, analysis of the removed skin shows that my paint certainly transmitted enough heat to melt the polystyrene foam. This created a delamination between the skin/micro slury and the foam core compromising the structure. Therefore I will have to accelerate my plan to repaint the airplane, and when I do it must be lighter colors or using solar heat reflective paint.
It has been a whirlwind weekend. It was supposed to be a weekend of rest and relaxation, but it turned into a weekend of repair and repaint. We are very proud of the repair we did in Kanab. I don’t think it would have been any better had I done it in my own workshop. I now have a little more work to do to finish the body work on the canard. I knew I would have to repaint the whole airplane eventually, but looks like Betty has moved herself up the priority list.
Special thanks to those folks who helped make this repair happen. Kevin Vernon-Harris, Greg Struve, Gary Mowad, George Catalano, Travis Foss, and Brian Mitchell. I don’t know how I have been granted so many wonderful friends but I will keep you around for a while.
The last post we started a right side lower aft engine cowl, say that 5 times fast, and got it to a point where the foam shape mostly replicated the left side. Since then, I spent a bit more time adding, sanding, and shaping foam using my handy templates derived of 3D scanning to improve the symmetry. When I got to the point where I was happy with the shape, and couldn’t tell if I was improving symmetry or making things worse, I moved on to the next step, drywall mud.
Same story here, slather on the drywall mud and sand to shape. Foam sands to a pretty coarse contour, the drywall mud has more finesse. The temperatures here in Arizona have give us a little respite, which has been wonderful. It does, however, make the drying time for the drywall mud longer. My impatience had me getting the heat gun out to accelerate the drying. It worked fine, but you need to be careful that you add heat slowly. If you dry the exterior and the underlying mud is still wet, it will crack easily and still won’t sand.
As is to be expected, especially if you have every done any drywall work in your home, it is again many rounds of mud, sand, mud sand, mud sand. If you find a place that is really low in contour, you can always go back to foam. I had the opposite problem in a few areas, I needed to sand through the mud to foam and go deeper into it. This can be a risk because you may not have much foam in that area to work with. Luckily, I had barely enough and while that area has become flimsy, it is the shape I need.
Again, with the mud, you use the templates to get close. From there, I tend to stand back and look at the shape from all different angles trying to compare sides or areas that are a known shape. You need to get high, low, look from the front, the back, sides, underneath to compare all the different areas.
One of the things that I continue to find interesting is that when you get the foam to the place you are ready to move to the next finer medium (drywall mud), you would swear that the shape is pretty darn good. Then you put the drywall mud on, and even before the whole thing is covered, oy vey, awful flaws in the shape. So you work the mud until you get it to the point where it looks as good as you can get it and move on. This time when you wax it for a layup, as soon as you put the wax on you start seeing things you didn’t before, and round and round we go.
Today I got to the point where I couldn’t get any/much better with further rounds of drywall mud. It was the point of diminishing returns. So I decided to move on to the next process. That means a fiberglass layup to make the plug which I’ll refer to as Lower Aft Tooling Cowl Serial Number 1(TCSN1). We do love our acronyms in aviation.
In preparation I did 5 coats of wax on the drywall mud surface, then 3 coats of PVA. Time ran out to get a layup done tonight yet. Kevin and I were working hard on getting all the materials together, cloth cut, plastic cut, etc. But by the time we were ready to start mixing epoxy I decided it was too late. There is nothing worse than trying to do a complex layup, against gravity no less, as it is getting late, you are getting tired, and it is way past dinner time. Maybe I am getting soft in my old age. So I will do the layup first thing in the morning. It is starting to get to the fun stage for this cowl. I’m excited to see how it is going to turn out.
Replicating the lower cowl shape is moving right along. In the last post we discussed how it was going to be more difficult than the top was to duplicate and mirror since it is more complex. One of the benefits of blogging and posting all about this project is the constructive suggestions I receive. Thomas suggested that I use some technology to make the duplication easier.
Turns out, there is an iPhone app that can help. Scandy Pro is a low cost app that uses the facial recognition camera and technology to be able to do a 3D scan. Thats right, some of us have a 3D scanner in our pockets. I have found it to be very useful for multiple projects already. So far, I have found the app can be finicky, difficult to use, and routinely failed to actually generate a useful file. I haven’t learned enough about why yet, but I am guessing if I learn better technique I will have more success with it. That being said, when you are successful with a scan and get a file, it is amazingly accurate.
In preparation for the scan, I dropped plumb bobs and marked the centerline of the airplane. I then placed a box on this line that I would also scan so we would have reference planes to work from. Again kudos to Thomas for the idea. Using the Scandy Pro app, I then scanned the lower left cowl shape and my box.
This process was actually infuriating because of how often a scan would fail. It took dozens of attempts to get a few useable files. I tried doing small sections of the cowl when doing the full thing didn’t work. No matter what I tried, it was more miss than hit. I am sure it is part of the learning curve, but it seems that the app gets lost often. That is to say that it often doesn’t know where in space to place what it is seeing through the camera. It could have even been that as I was moving around trying to scan my shadows and reflections may have been affecting the scan. Much more to be learned here to make it a routinely useable tool.
All that being said, for the low cost, it did eventually turn out a useable scan. I then took the solid model from the scan and made templates to check the accuracy. I was pretty impressed. Now, admittedly a few of the stations didn’t fit perfectly, but they were better than is necessary, and easier to do than making templates manually.
From there, I took these templates and mirrored the exterior curve so I would have the profile of the right side of the cowl. I turned them into drawings, added some thickness, and printed them. These would become patterns to jig into place on the right side. Of course I don’t have a large format printer, so I had to print them in pieces and tape them together. From there I traced them onto foam core board and cut them out. If I would have been really slick, I could have had them laser cut. But I am too impatient to wait for that.
Now, turns out that fitting these patterns in place every 2 inches is tedious and riddled with approximations when there are conflicts. Part of the center tunnel already exists and we have to match into the existing wing roots and pieces of Brian’s original cowl. Each of these conflicts require template modifications that took significant time to cut/fit/cut more/fit/cut more and so on. I decided to take a hybrid approach.
The tried and true method you’ve already seen using foam splines is pretty simple when the geometry is already defined by the adjacent areas. So I will use that for the simple areas, and use patterns for the lesser defined areas where I had to make choices. That would mean the shape above the wing root, forward of the trailing edge, and below the top cowl is simple to spline. It is well defined by the top cowl and fuselage shape. Another area that is easy to spline is forward of the trailing edge of the wing on the bottom surface. Again it is defined by the fuselage shape and the centerline tunnel. There is no difficulty in either of those places.
The place where there is difficulty is blending aft of the trialing edge of the wing to my augmenter boat tail shape and the top cowl. There is a lot of shape change going on over a short distance. So I made 4 patterns along that length. These were glued into place. That way when I pour foam around them I can sand until I see the full pattern and I am a match to the other side.
With that plan in place, I got to work splining and foaming. By lunch time I had the foam in place. Then it is the usual contour sanding to reveal a nice shape. You can see by the end of the day today, I am 75% there. There are some places that need attention yet. Most notably two low spots where the splines transition to the patterns. One on the bottom and one on the side. The splines didn’t transition as well as I would have liked. The good news is I have my templates which will guide me on how much filling I need to do. Not a big deal.
I think it is already looking very good. Some improvements to the symmetry and filling of low spots and we should be ready to do drywall mud very soon.
Today was further proof that the lower cowl certainly is more challenging than the top. Even though I have determined some defining characteristics of the cowl and come up with a plan for the overall approach, it is still a challenge to blend these shapes and features together. It isn’t always intuitive how to blend complex intersections in 3 dimensions. I am a master of gettin into analysis paralysis, so for me, the only thing to do is get started.
That means lay splines and sand to shape. Unfortunately, when it isn’t intuitive to me, it normally means that I start in the wrong place and end up with the wrong shapes. That is what happened with the lower cowl. I had placed foam “control splines” in woefully wrong places. Since I was using these to define my cowl shape, sanding made it worse, not better.
At that point you need to be willing to get the saw out, remove the offending material, and try a new tack. Luckily, I have a good group of “phone a friend’s,” when I get to a mental blockade. This time, I talked with Thomas, and he said something that made sense to me. When he talked about how he would solve the problem he started with, “If this were a CAD problem…” That turned on a light bulb for me.
I have watched tons of youtube videos about how to loft a shape in both Solidworks and Fusion 360. That is essentially what I was trying to do in real life. So if I approached the problem in the way you would when drawing in CAD, it would be easier. This allowed me to better choose control splines and make the shape.
Yesterday’s session ended up with a majority of the left half of the lower cowl splined in foam and sanded to a rough shape that would allow me to evaluate if I was happy with where we were heading. Turns out that it seemed like the right approach and allowed for the best compromises in shape. I was able to keep retreating angles to a minimum by putting more emphasis on the lower surface. It still wasn’t intuitive, but it was better than what I had previously done.
Today I finished splining and sanding the left side of the lower cowl using that approach. It is really simple to do when the right “control splines,” and methodology is identified. I very quickly got to the point where I had enough of the shape roughed out and the foam was of a coarse enough resolution that I couldn’t improve the shape without going on to a material with more resolution.
At this point, I would normally spline in the other side trying to mimic as closely as possible the first side. However, I still had some lingering doubts about the shape at this point. The splines were disorganized making for optical illusions. The differing colors of pink foam/ yellow pour foam as well as areas with voids were making me question the contours. I needed to see a more refined shape.
Therefore, I started adding drywall mud. This would give me a more refined shape, more control and resolution with sanding, and a single color surface. Once I had the first layer of drywall mud on, I was convinced I had the right shape. From there, it is the familiar repetitive process. Add filler, sand to shape, find flaws, add filler, sand, repeat. I did a few rounds of drywall mud today and will have to do a few more yet.
The good news is I am happy with the shape, though there is still a lot of sanding and finessing to do. The not as good news is I now need to replicate it as closely as possible on the other side. All of this with out the benefits of being able to see the individual splines on the left side. Hopefully I have enough time lapse videos to determine how best to mirror the shape to the right side. Eventually templates will drive the shaping of the right side, but it is nice to get as close as possible from step one.
I can’t wait to see how the cowl shape looks with a wing installed. I can’t get over the visual of how tall it is when it is just the fuselage and cowl. I really think it will look sleek when the wing bisects the cowl and you only see part of it at a time. I have to keep reminding myself this is a ‘station wagon,’ not a sports car. It makes sense that I have a big cowl.