Quantcast

Joining Composite Tubes

HomeBuiltAirplanes.com

Help Support HomeBuiltAirplanes.com:

SamP

Active Member
Joined
Nov 4, 2016
Messages
28
Hi All,
I'm looking to build a truss structure from fiberglass tubes. I haven't seen a lot on how to join pieces together. Most of the stuff I have seen is about extending a tube using an insert.

In Solidworks, one can miter a joint to make them fit well together. My current thoughts is to do this, making the cut with a dremel, then wrap it in fiberglass cloth. Does anyone have a better idea at joining? I heard you can make custom joints as another technique, but I hear they are expensive. Does anyone have a good source for this? Thanks very much.
 
Last edited:

Norman

Well-Known Member
HBA Supporter
Joined
Nov 28, 2003
Messages
3,069
Location
Grand Junction, Colorado
You might look at videos of carbon fiber bike frames on Youtube. There are several techniques to join composite tubing: all the way from traditional lashing on the bamboo bike up to molded joints with custom made rubber bladders.
 

BoKu

Pundit
HBA Supporter
Joined
Aug 15, 2013
Messages
2,823
Location
Western US
Q: How do you get down off an elephant?

A: You don't get down off an elephant; you get down off a duck.

I've never tried to make a fiberglass truss, but it sounds to me like a rather inefficient use of the material. At issue is that trusses are usually limited by compression buckling, resistance to which is primarily a function of the stiffness of the material, not its strength. And fiberglass, while fairly strong, has a relatively low modulus of elasticity.

--Bob K.
 

Markproa

Well-Known Member
Joined
Feb 9, 2017
Messages
139
Location
Bellingen, NSW. Australia.
unnamed (1).jpg
I built this golden falcon for the Doha Olympic games in 2006. I made a bunch of curving carbon fibre tubes and lashed them together with carbon tow, sort of hi tech boy scout technology. It had to be light enough to be flown around the stadium on four high speed winches and strong enough to carry two men. One guy inside to operate the pneumatic flapping wings and the other an Olympian hanging underneath. It turned out to be very strong and light.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
7,650
Location
Saline Michigan
The reason you can not find info for joining fiberglas tubes is that it is not usually a good idea.

For example, let's just design a simple truss, right out of the Mechanics of Materials text of your choosing. Define the length and depth, weight(s) carried in the middle, reactions at the ends, then the compressive and tensile loads in each of the elements. Then you get to size the tubes so that they will stand the tensile loads and compressive loads (column buckling, Euler's rule, EI dominates). Adjust tube diameter and wall thickness to find the lightest tube in each material, and then compare overall weights. Oh, and you probably do not want to go below about 0.035" wall. Well, the tube sizes that work in steel, aluminum, carbon fiber, and fiberglass will all be different, going from small to large in that list. What you find out is that fiberglass tubes with Exx of about 2.5Mpsi end up a lot bigger and maybe heavier than all of the others (Exx or E's of 10 Mpsi to 30 Mpsi) .

Then you get to how to form the joins...Molded sockets (fit on tube OD) are nice but expensive/difficult/fussy to make, and likely to weigh a significant amount. Glass wrap is pretty sloppy way to do it, which means more material is in the joins than you want. Internal sockets (fit on ID's) have the same issues as the OD fits. Not many good approaches available. Bolting and riveting are tough to make work as the concentrated loads from these methods drive much thicker walled tubing. Once you are looking at the weight of the joins, you can revisit the truss configuration and maybe the weight can be optimized in each material by reducing or increasing the number of bays and thus the number of joins and sizes of the tubes...

But you can avoid all of that join fuss by designing it all as one piece. Yep, it can be done. Build on a table (if the truss is to be flat) or on a curved caul plate (many ways, but one is a supported curved plywood surface covered with plastic laminate). First you laminate the side on the caul plate. Then cut your tube cross sections from light plastic foam. and fit them together on the laminate. Round is hard to work with here, maybe half round or rectangular with square corners against the first laminate and rounded corners up in the air. Then laminate over all of this to finish the shape of the truss, and saw cut the excess. This gives you a tube truss with hat section tubes. You can design this in glass and graphite cloths, and you can tailor the amount of cloth and tube sizes along each of the elements to optimize the truss. This will save you weight over the preformed tubes and join methods.

But we can go one better in most airplane structures... Analyzing the truss with foam cores, it still has excess weight in the flanges where we joint the first and second laminations. If instead of having triangular cutouts and doubled layers in a bunch of places, we could just use a solid plate of plastic foam, with the lower and upper lamination joining each other only at the edges. The plastic foams we use are so light that making the structure "solid" is usually lighter than using the cutouts. Now you have a typical sandwich construction composite panel.

Anyone who does not believe this works should actually run the exercise. Steel tubes should probably not go below 0.035" wall (for welding), I will defer to others on min wall in aluminum tube, and glass fiber laminates should not go thinner than 0.025, graphite should stay above about 0.015". You may be able to find some places in the design space where unexpected designs are optimal, but generally you will find that traditional welded steel tube trusses and riveted aluminum tube trusses are pretty darned light. Composite panels are kind of heavy, but you might as well either mold or go moldless in composites, but do well schemed out aerodynamic shapes and eliminate most individual parts. In a finished composite fuselage, glass is likely to weigh more than well designed metal tube or sheet metal designs, but graphite is likely to weigh less.

And if you do design and build a fiberglass tube truss, document it and report on the results here so we can learn how you did it.

Billski
 

pictsidhe

Well-Known Member
Joined
Jul 15, 2014
Messages
8,812
Location
North Carolina
View attachment 64740
I built this golden falcon for the Doha Olympic games in 2006. I made a bunch of curving carbon fibre tubes and lashed them together with carbon tow, sort of hi tech boy scout technology. It had to be light enough to be flown around the stadium on four high speed winches and strong enough to carry two men. One guy inside to operate the pneumatic flapping wings and the other an Olympian hanging underneath. It turned out to be very strong and light.
Huh. The Olympic version of falconry seems slightly different to the traditional sport.
 

cheapracer

Well-Known Member
Log Member
Joined
Sep 8, 2013
Messages
6,290
Location
Australian
Hi All,
I'm looking to build a truss structure from fiberglass tubes. I haven't seen a lot on how to join pieces together. Most of the stuff I have seen is about extending a tube using an insert.

In Solidworks, one can miter a joint to make them fit well together. My current thoughts is to do this, making the cut with a dremel, then wrap it in fiberglass cloth. Does anyone have a better idea at joining? I heard you can make custom joints as another technique, but I hear they are expensive. Does anyone have a good source for this? Thanks very much.
I have seen both flat fiberglass tape and fiberglass thread wrapped around the joint. A few wraps then wet, repeat.

The problem with that is it leaves a bulge on top that may or may not be an issue for you. Google bamboo bicycle building.


Forget Solidworks, Google and download Giles Puckett's Tubemiter.exe, the original, simplest and the best.
 

SamP

Active Member
Joined
Nov 4, 2016
Messages
28
Thanks all for your input. I had forgotten about fiberglass sandwich panels, and I'll definitely look into Puckett's software. Thanks again!
 

BBerson

Light Plane Philosopher
HBA Supporter
Joined
Dec 16, 2007
Messages
14,447
Location
Port Townsend WA
Square fiberglass/carbon tubes could be joined with flat sheet gussets. Like a wood truss Pietenpol
 

wanttobuild

Well-Known Member
Joined
Jun 13, 2015
Messages
740
Location
kuttawa, ky
Old thread, I know.
But interesting none the less.
Since this method is on a case by case basis, I submit the Baby Ace C.
This fuselage has the classic cluster formations. I am looking very hard at executing this aircraft in a carbon square tube, rag covered.
The tube is carbon with a dense foam core, at least dense enough to knock off the corners with a 1/8 roundover bit. Core to be left in place.
The main goal is to be able to provide a flat path for uni to do the connections, as much as possible.
The section of the foam can be reduced where needed for extra plies, without creating a bump in the tube.
I would like to join with uni, but may chicken out and glue on some gusset plates with scotch weld.
What am I looking at for carbon cloth? It may have to be two ply to comform to an 1/8 radius.
The core is the key, it locates everything, provides a means to tailor strength and the mold to wrap the reinforcement.
I would appreciate all thoughts and critical assessments.

ALL speculation at this point as I am waiting for Mark to post the mount info and CofG for the AM10. I am interested to see where where I am going to be in relation to the rear cabane.
I don't want to change the classic lines of the Ace C.
 
Last edited:

wanttobuild

Well-Known Member
Joined
Jun 13, 2015
Messages
740
Location
kuttawa, ky
Post #6
The key to the Composite Truss Tube Fuselage.
Maybe a few folks will realize this concept, to be a realistic approach to the elusive, Magic Black Truss.
When you think gusset plates and fasteners, you might as well go to aluminum!
No Crimp UNI is the "welder".
 
Last edited:
Top