Part 1 - Construction
As I was working on the window cutouts one day, with thoughts of taking to the air--much like the fiberglass dust floating all about me--I got to thinking: "This airplane has a huge windshield. I'll bet the glareshield is equally large. Funny, I don't recall seeing anything in the builder's manual about the glareshield. Now that I think about it, if I install the windshield now, building the glareshield after-the-fact could really be challenging." Okay, enough of me thinking aloud. But it's a good thing I did. I shot an email off to the factory, asking why the manual didn't mention the glareshield and inquiring if it wouldn't be prudent to complete this task before the windshield installation?
Answer: The glareshield chapter was accidentally omitted from the manual (not sure if it was just my manual, or all of them), and yes indeed, the glareshield should be built before installing the windshield. Coincidentally (or not) they couldn't find any documentation or drawings for the glareshield, so they sent me some written instructions which basically amounted to (1) fit a piece of 1/4" foam in place, (2) glass both sides, (3) remove it, and (4) glass it into place after the windshield is in place. Sounded pretty simple, but in reality this turned out to be the most challenging part I've built yet. Here's why:
First of all, this piece was BIG. Secondly, it was highly contoured. Not just that, but "compound" contours. I built a template out of cardboard in an attempt to design a flat pattern. The first big problem I encountered was the portion that overhangs the instrument panel, what I call the "visor." As you would expect, the glareshield is sloped, and when using a rigid material like cardboard or foam core, the slope tends to carry all the way through the part. I wanted the visor to over-hang the instrument panel by 3" and be parallel to the aircraft waterline. Trying to bend the cardboard (or foam core) around the curvature of the instrument panel bulkhead AND keep the last 3 or so inches parallel to water line just wasn't happening.
Another problem was simply "bending" the foam core. The 1/4" polyurethane foam that I had on hand was really at it's limit when formed around the instrument panel bulkhead. I was afraid that it would break despite the factory's assurance that it wouldn't. Since this foam was really intended for some other use (fuel sump, I think) I decided not to risk it and ordered a piece of low-density 1/4" Divinycell foam instead. According to the Wick's catalog, this stuff "bends with a heat gun," so it seemed like a better alternative. When the piece finally arrived (they failed to mention that it was on back-order), UPS managed to crack it. While I was awaiting a replacement piece, I did a little experimenting and decided that a 2-piece core section would be the best way to go. A week or so later UPS showed up with a good hunk of foam.
I built a template out of left-over 3/4" foam to match the contour of the instrument panel, then formed a 6" wide piece of Divinycell over it with the help of a heat gun. This piece became my visor. I doubled up the aft edge, then shaped it to a nice radius, giving the piece more rigidity, and less of a head strike danger in the event of a *gasp* crash landing. I glassed this piece with S2 glass and temporarily installed it on the instrument panel bulkhead. I found some good quality mounting tape worked well for this.
Then, after getting an acceptable fit with my cardboard template, I cut and slowly shaped the big piece of foam, again using the heat gun where necessary, to mate up to the visor. I created a splice out of the now defunct 3/4" thick visor template, and used hot glue to join the 2 pieces. Next I filled in any gaps with epoxy/micro, did some final sanding and shaping, then glassed to entire upper surface. I wasn't able to glass all the way to the edges of the part, however, because it extended under the windshield cutout and I just couldn't get in there to do a decent job. Once the lay-up cured, I removed the glareshield and did a final glassing around the edges, sanding the overlapping plies smooth to make a nearly invisible transition.
Now it was time for the bottom surface. I designed and cut out some ribs to stiffen the whole thing up, bonded them in place and glassed the surface in sections. Overall I think it came out pretty good. I'm sure, once the windshield is installed, I'll have to do some additional trimming to make it fit just right, but that's expected.
To finish things off, I primed the upper surface with Poly Fiber's UV Smooth Prime. From my experience so far, I've been impressed with Poly Fiber's composite finishing products. I'm considering using their Top Gloss water-borne polyurethane for my airplane's final topcoat, but that will be a future decision.
Part 2 - Systems
"Systems on a glareshield?" you may ask. Well... yeah. 2 of them in fact: Lighting and defrost air. After a little bit of head-scratching, I decided that I needed two nozzles to direct hot air through the glareshield, each approximately 7" wide. I created a drawing of what I thought the nozzles should look like and emailed it to my Dad. From that he made a positive prototype, then a negative mold out of silicone and sent it to me. In each half of the mold I layed up a single ply of S2 glass, then joined them after they had cured. The process was repeated to create a second nozzle. In hindsight it probably would have been easier to adapt some sort of nozzle from a shop vac, but mine came out very light weight, and fit the IML curvature of the glareshield just right.
Before the nozzles could be installed, I had to create slots in the glareshield for the hot air to escape. The first step was to locally remove the IML ply and foam core and fill the voids with potting compound. This provided a smooth, solid surface through which to cut the slots, and looked much more professional than having the foam core exposed. I rough-cut the slots using a rotary cutter and then finished them off with hand files.
To direct airflow and keep foreign objects from rolling into the nozzles, I installed 4 fiberglass vanes in each slot. The shot at left was taken before the vanes had been trimmed flush, of course. Once that was finished, it was just a matter of bonding the nozzles into place.
Now if I could only figure out where my hot air is going to come from. That will largely be determined by my engine selection, which, again, is a future decision. I was hoping to install a diesel engine, but they are maturing so slowly that I'm not sure that will be a realistic option. Then again my progress is pretty slow, so who knows.
With that task complete, it was time to think about a lighting system. I wanted to get the lighting system installed before I painted the glareshield, and the painting had to be done before installation, so it was really a sequencing thing. Even before I started drawing circuit diagrams, I knew that I wanted to use white LEDs for illumination. While I was at it, I decided that adding a second string of red LEDs with a selector switch would be a good idea. LEDs have a lot of advantages: they use very little current, they're nearly indestructible, they last forever, and they're pretty cheap. There are some disadvantages too: they usually don't come with any sort of housing or mounting device, their beam pattern is narrow, and they're a bit tricky to wire up. After looking through the Digi-Key catalog for a while, I decided that using surface mount LEDs would be best. Since they have no leads sticking out the back I could bond them directly to the underside of the glareshield, and without a lens they have a wide (120°) view angle. These type of LEDs don't put out very much light, however, so I ordered 9 white and 8 red along with some aircraft grade wire and appropriately sized resistors.
After receiving my package, I went to work wiring up the system. The good news was that the light output was very even and just about the right intensity to illuminate the instrument panel without being annoying. The bad news was that the volume of wires required to make this work was quite cumbersome. The wiring actually overlapped many LEDs, blocking their output, so I had to bond little spacers between the LEDs and the glareshield surface. Getting all that wiring to fit inside my transparent conduit was also presenting a problem. Of course a larger diameter conduit could have solved that problem, but I was beginning to get frustrated with my design. The nail in the coffin was when a few of the wires pulled loose from the LEDs. The solder joints on those tiny LEDs were very fragile. That being the case, I decided that this design was not robust enough to survive an aircraft environment and I ripped it off.
I think to do the LED lighting system properly, I would need to have some printed circuit boards custom built. That would provide the neatest and most durable solution, but gets much more expensive. For a little more money, I could buy Aero Enhancements' Ultra Vision system, which is a very nice solution to glareshield lighting. At this point that's the direction I'm heading, but I may revisit the lighting again later.
Anticipating the need to run power wires from one side of the instrument panel to the glareshield visor, I installed a conduit.
Part 3 - Finish
I knew that I wanted to 'paint' the glareshield, but I didn't know what material to use. I considered just using some flat black paint, then I went to an auto paint supply store and considered using "bumper paint," but I figured there must be a better solution. One day at work, while I was in the flight deck of a new 767, I got an idea; why not use the same finish as we use on the Boeing glareshields? A little engineering drawing research revealed that the material was indeed a paint, known to us as BAC5755 Type 16 Non-Reflective coating (color 8845 brown, of course). Turns out this stuff is commercially produced by Red Spot Paint and Varnish Company as Nextel Suede coating, so I ordered some. Actually Red Spot is not a retailer, so my Dad ordered it through his company and we split a gallon. Good thing too, as this stuff is expen$ive. For a gallon of the Nextel Suede, plus a gallon of primer, plus reducers, accelerators, and shipping, our order topped out at $300. I also had to buy a decent spray gun, which set me back a couple hundred more. I opted for a conventional (high pressure) gravity fed gun, DeVilbiss model GFG-644.
Before painting, of course, I had to prep the surface. I already had a rough coat of UV Smooth Prime applied, but it needed touch up around the defrost air vents, and I had a few depressions and imperfections to fill. The UV Smooth Prime sanded very easily, and produced a nice, smooth base. My only complaint was that it's so thick it takes forever to run through the paint strainer.
The next day was 'show time.' I mixed up the Nextel Suede coating with the various components necessary to make a working batch, and loaded up the spray gun. A couple of test shots on some scrap cardboard, and I began spraying. It only took a couple of minutes to get a nice, even, consistent coat. I chose a dark grey color, with a few little speckles. It came out great, if I do say so myself. It didn't take much paint to cover the glareshield, so I still have nearly ½ gallon of the coating left. Perhaps I'll paint the instrument panel with this stuff too.
Part 4 - Installation
Installing the completed part was not terrible tricky. I had already trimmed it to the correct size long ago, so it was just a matter of clamping and wedging it into place, then floxing around the edges. I also layed up a 3" wide ply of fiberglass around the joint on the lower side.
This job would have been much easier with the airplane upside-down, but I was afraid that a drop of epoxy resin may leak down the inside of the windshield, or down the painted surface of the glareshield. That would have really pissed me off, so I did it the safe way and worked over my head.
That's it. The glareshield is now permanently installed. I heard that another SQ2000 builder wanted to make his glareshield removable. Nice concept, and sure would make working behind the panel convenient, but I just don't see how that would be possible.