UberLite -- Ultra Light-Weight design and build

[kennyrayandersen]

Greetings again,
It’s been a while since I posted. Life keeps getting in the way of my TTT/TD dreams. I’m still a year and a shop away from building, but it doesn’t stop me from thinking about it. So, right now I’m working out-of-country and couldn’t build should I want to, so for now I’m in dream, research, and design mode – sorry if that offends anyone who is actually building. I’ll get to that as soon as life permits.
Background:
I do stress analysis of aircraft structures. I’ve been doing that for nearly 25 years. I work with composites and metallic, mostly aluminum, structure. Not that it means anything, but I’ve worked for Boeing, General Dynamics and then later Lockheed, Vought, Bell Helicopter, and even Douglas (back in the day). I CAN hold a job, really, it’s just I’m a freelance job shopper so I go where the demand is high. So, I’ve learned a far bit about how to make structures light-weight, although admittedly, maybe not so cheap (they ain’t giving that graphite away!).

I’m also interested in Microcars – Early Fiats – 500s, 600s and also cars by Messerschmitt, and BMW (Isetta). Sometimes you see little bitty teardrops being towed by little tinny cars and it’s pretty much too cute for words. Mind you the 500 had about 17 HP (early versions had only 13.5 HP) and the 600 had around 28 HP. So, obviously towing a tear drop with so little HP the weight of the teardrop becomes a major issue. Originally, I planned on designing a very light-weight trailer to tow behind a Smart Fortwo, but now with the possibility of FIAT bringing in the new 500, I’m holding off to see whether I can get one of those. Regardless, the point is I like really small cars, and these would require an extremely light-weight TD.

Someone else pointed this out, but a lot of the home-built trailers seem quite robust – even more robust than the commercially produced units. And when a stress guy says robust, what he means is heavy. Obviously no offense toward anyone or any particular TD is intended, and there are more beautiful trailers and tears on this site than you can shake a stick at. What I’m talking about is a branch of the overall trailer family tree that is optimized for weight as anything else (much like a woody is for looks). There are always compromises – engineering is like that. Weight is traded for cost, or schedule, or ease of manufacturing. We may trade structural strength for aesthetics etc. I’m thinking of doing something that is really geared to weight, and not much else (OK maybe some aesthetics – friends don’t let friends build ugly).

I’d like to give some ideas and thoughts about my intended approach to building a really light-weight TD, and if anyone thinks I’m smoking crack – do chime in as I’m long on theory and short on build time.

Requirements:
80 inch bed, Ice chest, camp stove, sink, AC – not much else.

Body/cabin
The reason I mention the body first is that I think the body has the potential to carry a lot of the structural loads that the TD is going to see. The standard build techniques using plywood are too heavy for this application. Thin fiberglass facesheets are very strong and can be reinforced locally as required with additional plies. I’m thinking right now of using 2 plies on the outside and inside bottom with one ply generally on the inside walls and roof with local reinforcements as required. The walls and roof would be 1 inch thick foam (rigid blue or pink foam insulation for Home Depot or Lowes) and the floor would use 2 inch foam. It doesn’t seem like too much but read on.
For the last year I have been working on, among other things, an engine cowling for a helicopter. It’s just over 18 feet in length and about 4 foot high with a width of 4 ½ feet. The aft half of it is just ¼ inch thick. The front part, which is qualified for a 160 mph bird strike is just ½ inch thick. The last major part opens up and is used as a maintenance platform (two people can stand on it for maintenance at the same time) and it is just 1 inch thick. The construction of this cowling uses graphite skins with a NOMEX honeycomb core and generally just 2 plies of graphite on each side (most of the aft ¼ inch area has just 1 ply on the inside): these 2 plies result in a total thickness of .016 inches. This cowling is designed for a velocity of just over 250 mph (this would be some kind of Bonneville teardrop land speed record for sure!). The core density is just 2 lb/cu ft.
The DIY version would likely use fiberglass with a foam core instead which is slightly heavier and not quite as strong, but really strong enough for a teardrop and way cheaper than the graphite/NOMEX combination. Also, the home builder is unlikely to have an autoclave at his disposal, so a standard wet layup similar to the one used by many that have posted here would use rather than a prepreg-type material (so this part isn’t so different in the end). The thing that is noteworthy is that this structure can be made to support a LOT of load and if the body is designed in a monocoque fashion using this sandwich construction then not only is it strong, it becomes virtually self-supporting (it is strong by itself independent of the frame). Therefore, the frame strength can be considered independently from the body.
Frame
Since the body is now nearly self-supporting, the light-weight frame that Andrew outlined ( http://tnttt.com/viewtopic.php?t=26891 ) would be perfect for this application. The main consideration that he correctly mentions is that the floor attachment needs to be adequate, but this isn’t that big of a deal, it just needs to be done (I’ll explain later as I start putting the details and attachment joints together). Aluminum could be used to save some weight here, but only desperation will drive me to that as the cost is considerably higher than steel.
Concessions to weight
I figure I’m going to make about 3 main concessions for weight. 1) Since most of these little vintage cars have 10 inch tires, and these tires can be gotten in a whitewall, I’ll probably go with the 10 inch tires rather than the lighter 8 inch ones. The tires and wheels could then look all vintage-like and matching, and again, too cute. 2) Another concession to looks will be the outline. The supperleggera is light-weight and extra small, but I prefer the shape of a more traditional tear like the Cub modernistic though I will probably use a unique profile. It doesn’t add that much weight, and because I’m using fiberglass, I’m not limited by the standard plywood or metal sheet sizes, but rather the fiberglass (which can be purchased in 60 inch widths most anywhere). 3) Trailer width will be 56 inches wide as the wife is a little claustrophobic and I read a survey on this site a couple of years ago that ask people what they would change about their TD build should they do it again and the number one response seemed to be width – this number is driven by the 60 inch fabric width – I want to be able to drape over the top and down the side just a bit.
Axles:
Torsion axles seem like they are about as light-weight as they get.
Weight:
I have a target drive away weight of 500 Lb.
I’ll throw this thread out to the general section because I wanted some general feedback, but will start listing details and sketches over on the building techniques area.
Sorry it’s so long – I’ve been thinking and reading for awhile.

[Trackstriper]

I don't wish to divert from your post but your qualifications have intrigued me.

I'd like to get your opinion about a particular material that might be used for a very light tear, especially if someone was willing to build without insulation in the walls. Your aerospace experience working with aluminum and composites would be valuable in evaluating this material. This product, or a similar material by another manufacturer, is designed for use in the sign industry. It is fairly rigid and reasonably light weight at about 22# for a 4x8 sheet at 10mm thickness. I can buy this locally in sizes up to 5x10.

http://www.nudo.com/nudosign/Alumacorr/Alumacorr.pdf

I just can't help but think that this would make some really great sides for a tear and it would save a lot of fiberglass work and associated finishing. What think ye?

[cleonard]

500 pounds empty weight or 500 pounds all loaded up? If you look at some of the ultralight designs that have been posted on the site I think that 500 pounds empty is certainly doable even with more traditional building materials and techniques. Getting to 500 pounds loaded will be a tough goal.

If you go crazy on the lightweight side at some point the aerodynamics will get important. You want it to stay stuck to the road.

The main loads that you need to worry about are the shock loads from things like a pothole at high speed.

While an autoclave might not happen there is always vacuum bagging.

It does sound like a really cool project. I'm sure that everyone here would love to learn your progress.

[kennyrayandersen]

Trackstriper
That material is interesting, but may have limited use for trailer building and I’ll explain why. First of all polypropylene isn’t particularly rigid. It maintains some flexibility. For a sign that’s probably a good thing as it flexes a little in the wind then returns to flat. For structural applications we need something with a bit of stiffness or we won’t be able to apply much load (it just flexes and doesn’t carry the load we apply). Also, all of the ribs run in one direction so it will be very asymmetrical in its bending properties (some stiffness in one direction, but little in the other). If you weren’t counting on it being structural, you might be able to use it if you orientated it to take advantage of the stiffness that it does have; however, the first problem of the polypropylene not having much stiffness still remains.

Following the engineers propensity to give people way more information than they care to hear, I’ll elaborate. An I-beam has an upper and lower flange and a web that connects them in-between. It has high bending strength because the area (flanges) is as far from the center of the beam as possible. The flanges carry the axial loads (practically speaking) and the web carries the entire shear load. Without the web, the upper and lower flanges couldn’t work together and the bending capacity would be merely equal to the thickness of the flanges rather than the depth of the whole beam – a huge difference!

In the case of the rigid EPS (expanded poly styrene) foam insulation, it will carry load because it is stiff. The foam is like the web in the I-beam and the fiberglass facesheets on the surface are like the upper and lower flanges of the I-beam: they carry the axial (and in-plane shear) loads; so, together they work together and are very strong. The bonus is that it also insulates, so it’s a two-for.

If the plastic in the Alumacorr was stiff, like epoxy impregnated paper (NOMEX) or the rigid foam, then it would carry shear load and allow the facesheets to share the load. But, in the Alumacorr case, that sill wouldn’t solve all of the problems as the core only has stiffness in one direction. Let’s say you could get the manufacturer to make the panel with the foam core and aluminum facesheets – now you’re talking! You could use it on the sides and floor, but it wouldn’t curve without failing the core or buckling the inside radius facesheet, so you’d be limited to a trailer that has only flats – no curved panels. The thicker the core the stiffer and stronger it is.

Some folks put 1 facesheet on the outside radius of the foam and then bend it (this is easier if the foam isn’t so thick), then put on the inner facesheet, but in a teardrop that would mean you are trying to glass the inside – not impossible, but I’m thinking bit messy. I will either hotwire or machine the foam curved before laying the facesheets on.

cleonard
Yes, that's loaded up and it might be a little optimistic...but... After doing the analysis for the cowling I mentioned, I think there is plenty of possibilities to save weight. Cabinets for instance -- I'm going to use the same technique -- two plies on top (compression -- less stable) and one ply on the bottom (tension) -- strong enough and much lighter than say 1/2 in plywood.

As far as aerodynamics goes it would probably fly at freeway speeds (most of these are just big airfoils!), except that the car should disturb most of the air and ruin the lift; plus, I'll make the bottom have a bit of curvature as well which will tend to reduce the lift -- sketches to follow..

[asianflava]

You may want to do a search for Steve Wolverton's Cowper, I think it came in under the 500lb mark.

The big thing (or small thing) is that it is much smaller than what you are planning.

[aggie79]

For build #2 or #3 (should I finish #1 first?), I have been thinking about a technique similar to what you are describing. I have even purchased Burt Rutan's book on composite construction as a primer.

There are many challenges to this technique - well documented on this site. I have thought of using this material for the sidewalls and bulkheads: http://www.plywoodcompany.com/application/common/store/itemdetails.aspx?categoryid=161&itemid=2317. Unlike the BORG insulation board, this has a consistent thickness, is more rigid, relatively lightweight at 1 lb per sf, and much cheaper than aircraft composite boards. The real question is whether or not the hardboard skin would delaminate from the eposy resin.

Haven't worked out in my mind the hatch construction, but have thought about this construction method:

1. Epoxy glass both sides of the side panels and cut/route them to shape of profile.
2. Build a plywood buck for roof profile. Temporarily affix side panels.
3. Use two layers of BORG insulation for roof. Roof panel would rest on top of and extend to edge of sidewalls. Glass one side of each. Kerf other sides so that they can bend around profile. Install first layer, glass side down, kerf side up, over buck. Epoxy bond to sidewalls. Fill kerf with epoxy/microballoons. Epoxy second roof layer, kerf side down. Apply first layer glass to top.
4. Fabric tape and glass top to side walls on exterior.
5. Glass second layer to top and sidewalls.
6. Remove buck.
7. Install bulkheads.
8. Build floor (details not worked out on construction - maybe the same as sidewalls.)
9. Roll tear over and bond floor to tear.

Sounds easy on paper, but with build #1 I'm finding things are a quite a bit more complex than my sketches.

[angib]

OK, it's worth me publicising some off-forum work I've been doing for a few people on a small light trailer. We've already used up the 'ultralight' description so something lighter has to be super-ultralight, or something like that!

The first step, which has a very limited audience, was a wooden replica of the German Piccolo trailer - this is smaller than KennyRay wants (no galley) and is narrower than even small teardrops at 44".



This really is a light trailer, though not as light as I'd hoped - a detailed weight estimate (ie, piece by piece of wood) comes out at 263 lb unladen including margins.

Here's the weight breakdown:
Frame - 26 lb
Chassis components (wheels, tyres, hubs, half-axles, coupler, jack) - 71 lb
Body (wood/plywood only, incl fenders) - 118 lb
Body fittings (windows, handles, hinges, hardware, mattress) - 38 lb

You can see that the stuff that comes only in 'regular trailer heavy' (like even the lightest tongue jack, at 13 lb) is a big proportion of the total. And you can also see why only 8" wheels are used as going to bigger wheels would add too much weight.

Eagle-eyed readers will note that I've abandoned the super-light chassis' A-frame and full torsion axle - by going to half-axles, a single tongue can be used, plus the cross-member now holds up the middle of the cabin floor.

I'm also working on a parallel design that is meant mainly for really small tow vehicles, like motorcycles, but is a 'full' 48" wide and has conventional door(s):



I'll add it to my web site (the Design Library) as soon as it's done.

Andrew

[Mark McD]

Aggie,
My thoughts
1.) Build the bulkheads in before you glass the roof: use them as bonding points and stiffeners so everything stays square:if anything hardens up out of true you're screwed.
2.) The website says the skin can be veneered so I wouldn't imagine epoxy resin would harm,let alone,delaminate the substrate.
3.) Epoxy the inside of the sides panels only-before cutting. Cutting through two layers of epoxy and cloth will be like trying to hack through a kevlar vest. Plan on having a few packs of good jigsaw blades on hand.And a dust mask.
4.) I doubt if kerfing a panel that is epoxy/glass coated flat will allow it to bend to the radii found on tears. Not to say it hasn't been done,it probably has, it just seems for the homebuilder to be a big undertaking for little result. I would think it would be easier to build the roof shell on a form epoxy/glass one side, flip it, glass the other. "Easy" being a relative term.It's not easy. You get one shot.Now,the standard one-off practice of making fiberglass structures with layers of foam/stiffeners/glass/resin would be so much easier, not to mention cheaper, and probably result in strengths close to what you are looking to achieve. Put in your spars and foam and glass over it.Use biaxel tape on the side/roof seam.
Don't get me wrong, I'm not trying to dissuade you, I'm just bouncing ideas.
(This coming from a guy who wants to build a ferris wheel in his back yard. NAAH,then I'd have to get rid of the paintball arena.lol.)

[kennyrayandersen]

Hey I noticed you are in Ft. Worth, I’m a little west of town between Aledo and Weatherford – hopefully you’ll have #1 done by the time I get back to TX – that gives you until the end of Jan 2010! I’d like to come and take a look at it.

I don’t believe that there is any lighter way to build than using composites. The blue/pink rigid foam core weighs about 2 pounds per cubic foot. In a 4X8 sheet there are 32 sq ft, which is less than 5 ½ Lb. I’ll have to dig out the glass figures, but I’m pretty sure that it’s lighter, stiffer, and stronger than the hardboard. Aircraft-grade foam is entirely out of the question. As you’ve probably found out it is REALLY expensive. The home depot foam is consistent enough in thickness that it should make it too ugly. If you paint the trailer white you won’t notice any imperfections either, or the fact that you might not have every little tow covered. If you paint it black it may look like a sack of walnuts going down the road. It would be nice to have a female mold so that the OML (outer mold line) would be nice and smooth, but I don’t want to spend so much time on the tooling.

The link you sent for the panels would have a couple of issues – the first, as you suggested is that the hard board is horrible when loaded at all transverse to the thickness. To be honest though, you could probably get away with using them for the sides and bottom and if you trimmed them in 1/8 inch hardwood, and painted them they would probably work for shelves too. The other problem is that thy only come flat, so you are limited to the sides and top. It would obviously have to be sealed REALLY well. As you noted it goes to complete crap when wet. That product would also be difficult to screw into. I thought about cutting the kerfs and filling them like you were suggesting with the foam, but I think that would be a lot of work in the end, so I’m working on something a little less labor intensive.

Here is the unfinished, yet mildly cunning plan:

1. Make your wood perimeter for the sides having previously planed all of the wood to 1” to match the thickness of the foam panel.
2. Epoxy the foam into the frame after trimming the interior to save weight.
3. Epoxy all of you plugs or other wood inserts into the foam (any place you plan on putting a screw, bolt or hinge through. Also trim out any cutouts like the door, and leave some plywood (flush) where the trailer axle attaches.
4. Lay up the fiberglass on either side
5. Trim glass and wood edging at the same time.
6. Build the floor the same only use 2 inch thick foam
a. Wood around the perimeter
b. Plywood inserts anywhere load is applied
c. Glass over the top and bottom
d. Trim
7. Set the sidewalls on the floor and screw from the bottom up
8. I’ve laid out the profile so that I have basically one radius in the front that goes from the front bottom corner all the way to the aft top ellipse so I can glue up the foam like barrel staves and the hot wire them to the curvature with a simple tool mad with a router and a constant radius. Lots of people are hot wiring foam for planes and whatnot as you’ve noticed.
9. If you use one profile tool to do the sides and the top you can also do a strip of wood for the center section edge.
10. Put in all of the interior walls and shelves – screwing from the outside (you’ll have to flush and fill later so don’t get carried away. Construction of internal shelves and walls is similar to the exterior,
11. Epoxy the wood side pieces to the sides of the center section.
12. Glass the inside of the center section.
13. Run screws from the side into the rest of the center section (by the way there is no need for ribs, except where a hinge or something needs to be attached) – the top should be flush with the side, but you could always leave the sides slightly proud and then come back with a router bit and trim flush to guarantee that.
14. Fill all of the countersink holes you made screwing stuff together (the ones around the perimeter will get covered with glass).
15. Lay your last 1 or 2 plies of glass on the top and allow them to drape over the side which will form a lap joint between the sides and center section. I would start at the top aft where the hatch hinges and continue forward and around from there. After the top dries you can lay the turtle on its back and finish the outer ply. Now the floor, top, and sides are all part of a singular box
16. Complete the joint by using a couple of 2” wide tape plies to the interior corners and joints.
17. Wood inserts are buried in the floor and can be used to attach the body to the frame as well as attaching the axle to the lower outside corners

Maybe there is some more detail, but that’s a quick overview of what I was thinking – I’m QUITE sure there is room for improvement so I’m still thinking about it.

[brian_bp]

...Eagle-eyed readers will note that I've abandoned the super-light chassis' A-frame and full torsion axle - by going to half-axles, a single tongue can be used, plus the cross-member now holds up the middle of the cabin floor.

What is the size of the cross-member, relative to the one incorporated into the corresponding full axle? Unless it is bigger, you might as well just attached the pole tongue to a full axle as U-Haul did (on their travel trailers), and save bracketry weight. The full axle crossmember is typically bent, but could still support the floor with a suitable shaped filler piece of wood to produce a level and flat top surface.

Just an option... I still like the idea of the minimalist A-frame approach better.

[brian_bp]

...This product, or a similar material by another manufacturer, is designed for use in the sign industry. It is fairly rigid and reasonably light weight at about 22# for a 4x8 sheet at 10mm thickness. I can buy this locally in sizes up to 5x10.

http://www.nudo.com/nudosign/Alumacorr/Alumacorr.pdf

I just can't help but think that this would make some really great sides for a tear and it would save a lot of fiberglass work and associated finishing. What think ye?

Hmmm... basically aluminum skins with a Coroplastcore, presumably offered to provide a smoother and more opaque sign surface than just Coroplast (or equivalent) itself.

The inner structure is inherently - and unfortunately - directional because it is made by extrusion. I wonder if one could laminate two layers of Coroplast with their corrugations in perpendicular directions with surface layers of a stiffer material (aluminum like the Alumacorr, or fiberglass) and come up with something of any value?

If Alumacorr is not adequate as a body material, it at least would be an interesting option for cabinet door panels.

If I were using this in any semi-structural application, I assume that I would pick the thickest (10 mm) option, as it is only 20% heavier than the thinnest (5 mm), and would likely be much stiffer.

[aggie79]

Hey I noticed you are in Ft. Worth, I’m a little west of town between Aledo and Weatherford – hopefully you’ll have #1 done by the time I get back to TX – that gives you until the end of Jan 2010! I’d like to come and take a look at it.

Please stop by! (I actually live in Watauga, but no one knows where that is.) #1 will hopefully be done by this fall unless the rest of my tools get stolen. (See this thread: http://www.tnttt.com/viewtopic.php?t=28191.)

Here is the unfinished, yet mildly cunning plan:

You've obviously thought through this further than I. I have also mused about the hot wire process, especially since I like using templates for cutting and fitting wood rather than by measuring alone.

Andrew is the master and he does bring up (as usual) a very good point that a large part of the weight in lightweight design are the wheels & tires, axle stubs, subframe, hitch, etc., so chasing the last bit of weight savings with composites may not yield much weight savings over very light wood construction. But then again, bragging rights are priceless.

It's fun to think about these.

Tom

[angib]

The blue/pink rigid foam core weighs about 2 pounds per cubic foot.

Please search for a recent thread where a forum member kindly tried some structural tests on home-type foams as cores. They consistently failed in shear at the sandwich joint - which is exactly what you would expect - you can run the numbers yourself on shear stress to test this out. Insulation foams are just not structural materials.

Having worked on composites at the top end of the boatbuilding spectrum, I'm aware of the ludicrous price for aviation materials - but take away the aviation certification and exactly the same product is available for half the price!

Here is the unfinished, yet mildly cunning plan:

Lots and lots of work there!

Andrew

[angib]

What is the size of the cross-member, relative to the one incorporated into the corresponding full axle?

The axle cross-member is just 1"x2"x11ga. The rectangular tube cross-member and half-axles weigh 75% of a full torsion axle (Dexter #8 ), plus it avoids the 'you can't weld to an axle' proscription. But, yeah, a full axle could be used too.

Andrew

[kennyrayandersen]

The blue/pink rigid foam core weighs about 2 pounds per cubic foot.

Please search for a recent thread where a forum member kindly tried some structural tests on home-type foams as cores. They consistently failed in shear at the sandwich joint - which is exactly what you would expect - you can run the numbers yourself on shear stress to test this out. Insulation foams are just not structural materials.

Having worked on composites at the top end of the boatbuilding spectrum, I'm aware of the ludicrous price for aviation materials - but take away the aviation certification and exactly the same product is available for half the price!

Here is the unfinished, yet mildly cunning plan:

Lots and lots of work there!

Andrew

Andrew, I'll search for the thread -- you say failure at the joint -- do you mean it failed in core shear (that would be the expected mode), or that the bond between the core and facesheet failed (that would not be expected if the surface had been preped properly)? Also, I remember someone trying to glue some plywood to a foam core -- was that the one you are referring to?

I have a friend using the insulation foam to build a plane. The material isn't so robust, as you pointed out, but it's stiff enough to carry shear load and as long as you keep the allowable stress conservative, and design around the capability that it does have. I think the key is not using it where you ought not. A conservative shear allowable for that foam should be around 30 psi -- a friend of my friend who is building the plane did some tests and it was failing around 90psi. My plane-building buddy is using just 15 psi as the allowable shear stress. I figure since I'm not flying, and no one will actually be in the trailer I can be a bit more aggressive. Generally, trailer manufacturers use 3gs for over-the-road loads and 5gs for off-road loads. By keeping the weight down, the loads stay down as well.

I will definitely do some testing though. A teardrop, if it's laid out well, shouldn't be loading the core with so much shear. The exceptions are where the frame attaches. I figured that if I had an insert 5.5 in. diameter, that would be enough to spread a 500 load out to where the core shear is capable of reacting it (that is 2 inch thick core in the floor). Alternately, I could bury a wood crossbeam into the core as required to carry any normal loads -- I have to lay it out and then run the numbers to see which is lighter.

It will be a lot of work -- you are spot on on that observation; so, I have to consider what has already been pointed out which is the body is only a certain percentage of the overall weight and whether the work is worth the weight saving. After I get it laid out I'll be better able to estimate those weight savings. It might be that I can, or need to use a combination of traditional and uber-light (hows that for a thread title?) techniques.

Even at 1/2 the price structural foams are just too expensive IMO to use in this application. What may be doable is to use a small amount of the strong foam where you need it (local load introduction etc.) and the weaker foam where the strength is not required (out in the field) -- that way you could keep the cost down, but maintain the strength where you really need it. I sure do appreciate your input BTW -- you make me think harder, which is a good thing!