Composit parts connections to aluminum frame

Homebuilt Aircraft & Kit Plane Forum

Help Support Homebuilt Aircraft & Kit Plane Forum:

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
I would like to get better understanding about what it takes to connect composite parts to aluminum structure. I'm sure there is correct books to read and videos to watch. What if we build composite wings and need to connect them to steel tube frame? I am curious about connection itself, transition from carbon to metal as example. Do we use rubber bushings? Just like we connect engines to airframe?

In my project at some point I will be connecting composite tail boom to aluminum tube airframe somehow. And yes I will be asking for professional help, but still would like to get some basic understanding of what and how this is usually done.

tempImagepRHHpw.png
 

Pops

Well-Known Member
HBA Supporter
Log Member
Joined
Jan 1, 2013
Messages
9,930
Location
USA.
Not recommending this in what you are doing, but I owned a Piper Cherokee for about 5 years and it has a steel step for stepping on to get up on the wing walk to the door on the right side. One of the first things I did was remove the step that was bolted to the aluminum fuselage and epoxied primed both surfaces after cleaning some light corrosion and then applied a thin piece of sticky back packing tape to the inside surface of the steel step that touches the aluminum skin. No more corrosion.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,494
Location
Saline Michigan
Some basics:
  • This is hardpoint design;
  • Hardpoints add substantial weight to composite structures and so should be minimized;
  • Expected loads derived from flight envelope, landing, ground handling, and construction will be known;
  • Expected loads will be grouped as they are grouped in use, and are analyzed independently in these groups;
  • In composite airplane pieces, minimum FOS is 2.0;
  • In bolted or riveted connections, minimum FOS in most airplane connections is 8.0;
  • No part may exceed first fiber failure stress at max combined load times established FOS;
Once these really basic rules are complied with, bolted or riveted connections do just fine. I recommend Shigley's chapter on connectors. Lots of good analysis.

Now to specifics of connections in composites. Composites tend to have modest bearing strength and do not like having holes drilled in them.

For load bearing like attaching your tail boom, bushings made of metal or Garolite are great. These materials have known and predictable bearing strengths and compression strengths so that they may have bolts tightened on them and distribute loads into the composite structures.

No aluminum or magnesium alloy bushings in direct contact with carbon fiber structures. Steel and titanium alloys work fine. Bushings will nominally be bonded using thickened epoxy or other structural adhesives intended for use with metals and composites.

Stresses in the composites near a bolted joint are kept in range by making the composite structure locally thicker as you approach the bushing. Check that your stresses (from force times FOS) stay below material capability at all places. Usually this means at some radius from the bolt, you need another ply of composite, and at another radius in, you need another, until you reach the bushing. If the local thickness seems too large, record how much weight the bushing plus reinforced composite was for this scheme and then increase bushing OD and run another set of reinforcement calcs. After you have three, you can survey for weight and trends for which way weights are going, and look further in the direction of lowest weights that do the job.

The brackets folded of metal might work fine, with bushings embedded in suitable composite structures, and finished to barely slide into the metal brackets. My biggest worries over them is fatigue of the joint. You would need enough friction between composite and metal from bolt clamp loads to keep the joint from sliding...

My preference is to create a flat flange on the metal side of the joint, with its flat in a vertical plane facing aft. Then the composite aft fuselage would have a similar flange for each joint, and you simply bolt them together. The bolted joints would need more preload in each of them than their live loads to prevent "working" the joint, and this is all covered in Shigley - this will determine your minimum bolt sizes and assembly torques. Then you would also check stresses in all of the failure modes listed by Shigley for bolted joints.

The hard point in both ends of the assembly would have to be properly designed to fully stand the live loads plus the static bolt forces. The metal end may get through this with pretty modest reinforcement near the flanges, but the composite tail boom may go from a couple plies of graphite to a substantial reinforcement near the bushing that is bonded in. Local reinforcement of a composite joint is covered in Marshall's book Composite Basics.

Billski
 

PiperCruisin

Well-Known Member
HBA Supporter
Joined
Jan 17, 2017
Messages
277
Location
Idaho
I don't know if this is really a big deal, not sure what structure you are thinking of. It is done all the time. Think of various composite aircraft like a Lancair. Engine mount to fuselage, landing gear fittings to wing, wing to fuselage attachment points, control surface hinges.

As has been pointed out, a few items to think about (in addition to typical structural concerns):
1. Corrosion. Mainly aluminum and carbon composites.
2. Hardware/concentrated loads. You can bolt composites together (not preferred as Billski pointed out), but should have a potted core and or quasi-isometric layup (some bi-directional at 0-90 and some +/-45).
3. Thermal expansion differences can be an issue.
4. Impact / damage tolerance.

It would be fun to look at all the issues Boeing had in the development of the 787.
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,494
Location
Saline Michigan
It is done all the time ... Think of various composite aircraft like a Lancair. Engine mount to fuselage, landing gear fittings to wing, wing to fuselage attachment points, control surface hinges.
Yes, it is. I explained the engineering involved with doing it rationally and making it reliable. I am certain Eugene does not want the tail falling off his airplane in flight, so a rational and reliable method is appropriate. Some companies have folks who do nothing but bolted joints and fastener engineering.

As has been pointed out, a few items to think about (in addition to typical structural concerns):
1. Corrosion. Mainly aluminum and carbon composites.
Somebody else thinks so too. Aluminum and carbon fiber shall not touch.

2. Hardware/concentrated loads. You can bolt composites together (not preferred as Billski pointed out), but should have a potted core and or quasi-isometric layup (some bi-directional at 0-90 and some +/-45).
I do not know what PiperCruisin is trying to say here, but things are bolted together in composite structures with bushings and beefed up laminates local to the joint.

Engine mount to firewall is pretty much what I had in mind for what Eugene brought up and how I would prefer to do the connection too. I talked bushings bonded into reinforced sections, PC is talking potted cores. Same approach, maybe different materials. I am not sure what materials he is preferring, but the bolt will have to be tightened enough for preload to exceed live loads in the joint, so materials must be stronger in compression than the forces the bolt puts on it. Shigley suggests a region three times the diameter of the bolt shank is under serious compression. So the elements in compression are eight times the area of the bolt shank. Some folks prefer to analyze for higher stresses in the bolted member, say two diameters, which makes stress in the compressed elements are three times the area of the bolt.

In any event, the bolt loads may not crush or yield the compressed members, bushing, or heavy potted section. I have engineered joints using solid fiberglass, cored sandwich with epoxy filled plywood and phenolic plate, and bushings made of aluminum, steel, and Garolite. The bearing stress from the bolts must be safe for the materials, and it must be bonded into the surrounding composite. The surrounding composite must be beefy enough that stresses in it are safely below its limits too.

3. Thermal expansion differences can be an issue.
This is part of why engine mounts are attached to airframes they way they are. If you have four mount points, typically only two will be triangulated. Think on that.

Then in the region squeezed by the tightened bolt, the joint is typically designed to be at 90% or more of bolt yield, but a joint comprised of materials with significantly different thermal expansion characteristics must have significant strength margins in the compressed elements, and may be designed with standoffs for the bolts. Again, look at engine mounts. This makes much of the compressed elements squeezed between the ends of the bolt and nut of the same thermal expansion as the bolt. This reduces effects on the other compressed elements as temperatures change.

4. Impact / damage tolerance.
What is going to be hitting the aft fuselage that we need to worry over?

It would be fun to look at all the issues Boeing had in the development of the 787.
Thread drift. If this is interesting to you, please start a thread on the topic.

Billski
 
Last edited:

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
OK, let's see if I translated all this information to Russian correctly and created correct mental picture in my head.

I am glad to hear that we do this all the time and nothing wrong with my project. Except it's complicated and you need to know what you doing. I had this example of engine mount brackets in my head and looks like it was perfect example of how this problem getting solved.

On my picture left image is engine mount. And we use rubber bushing to isolate vibration. I have no questions about that because we see it everywhere all the time.

Right image is my understanding of how we do this in composits. We are using special bushing made out of steel or titanium. I am assuming that we can use stainless steel as well. But no copper, brass or bronze. There is no flexing, vibrating or a movement at this joint. It is tight connection. I need to find knowledgeable people who can calculate this type of joints to be big enough and strong enough to do the job. Both on composite side and steel bushing side.

I have seen something similar to this in the medical field. I believe when they do joint replacements they creating similar bushings to be implanted inside of the bones. That is really not much different what we do in here.

I hope that my general understanding is correct.

tempImage3t2aEU.png
 

Map

Well-Known Member
Joined
Dec 29, 2020
Messages
102
Location
California
If the red part in the right picture is the composite your understanding is correct. But from a design standpoint there should not be a thread on the outside of that bushing.

I can do the engineering / analysis of your tail attachment.
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
If the red part in the right picture is the composite your understanding is correct. But from a design standpoint there should not be a thread on the outside of that bushing.

I can do the engineering / analysis of your tail attachment.

Yes, red part on the right is composite arm. And I am glad that I guessed this somewhat correctly. There is no threads on black bushing, that was my way to show a rough surface that I was under the assumption we need to bond steel and composites together.

Thank you for your offer to do needed calculations. I was hoping to find somebody knowledgeable like that who would be willing to do this. I sort of found someone in Russia who offered to give me ready to go recipe of my future composite tail for $1500-2000. Nothing materialized yet because this guys pretty busy during summertime. And no real rush on my end either. I can keep busy with different things for pretty long time yet.

But, before you will say yes, that you can do it, I will post over this weekend few pictures for you to understand what I am about to do. Just so we're on the same page. Thank you

We did a lot of talking here and eventually made conclusion that existing tailboom is too flimsy and too flexible and needs to be replaced with something much more rigid and longer. Presently aircraft is very unstable and unpleasant to fly with horizontal tail volume 0.32 and tail arm 2.2 MAC.

44637383881_8402bfe9a8_o.jpg
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
If the red part in the right picture is the composite your understanding is correct. But from a design standpoint there should not be a thread on the outside of that bushing.

I can do the engineering / analysis of your tail attachment.
Screen Shot 2021-02-18 at 20.17.25.png47169844_2343427855685717_6312039815449673728_o.jpeg46498173_2339961912698978_4898811028910899200_o.jpgIMG_5280.jpegIMG_5273.jpegIMG_5276.jpegIMG_5274.jpegIMG_5266.jpeg
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
As I see it, new composite tail boom will needs to be connected to this four points. I don't know if aluminum tube between 3 and 4 could be eliminated and composite structure will handle this load? Same question for braces between 1+2 and 4?

IMG_6465.jpegIMG_6466.jpeg
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
Eugene,

The thickness is probably exaggerated some, but I like it.
I was really just looking for a general understanding of how connection supposed to look like.

Now I will try to create sketch of four connections that will work for my airplane. 1 and 2 will be the same. 3 is underneath the engine mount. And at 4 I would like to connect to cut out 120 mm tail tube somehow.

IMG_6465.jpegIMG_6466.jpeg
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
OK, for connection 1 and 2 can we technically use same brackets and same 6 mm bolts? All we need to do is make bushings about three times bigger outside diameter = 18 mm? Wrap them around with multiple layers and starting going multiple directions with many arms? Just like octopus?

IMG_6469.jpegIMG_6468.jpegIMG_5257.jpegIMG_5254.jpegIMG_6465.jpegIMG_6466.jpegIMG_5266.jpeg
 

wsimpso1

Super Moderator
Staff member
Log Member
Joined
Oct 18, 2003
Messages
8,494
Location
Saline Michigan
I was really just looking for a general understanding of how connection supposed to look like.

Now I will try to create sketch of four connections that will work for my airplane. 1 and 2 will be the same. 3 is underneath the engine mount. And at 4 I would like to connect to cut out 120 mm tail tube somehow.
The connection at 4 will carry virtually zero load from vertical tail loads and only carry modest fraction from horizontal tail loads. Once 1 & 2 are stout enough from both metal and composite sides to carry their loads, the connection at 4 becomes pointless, as just about all load will be in 1,2,3.

I just went through your photos. Connection point 4 has a few things going on: The shoulder harness appears to attach at 4; The fuel tank appears to be connected to the large tube and diagonal braces going to 4:
  • If there is an adequate shoulder harness mount elsewhere and a way to mount your new fuel tank without connecting to the big tube going to 4 and its diagonal braces, that big tube can be truncated at the lateral box while the braces going to 4 can also be removed;
  • If the shoulder harness must go back there and/or fuel tank mounting would become difficult or heavy, you leave them in;
Maybe there is something in between. I strongly suspect that the braces going to 4 are there to keep the keel tube from collapsing or deflecting excessively under loads on the tail planes. The vertical brace also connects the keel tube and the top structure in vertical loading and so may continue to have some value. You may be able to omit the diagonal braces.

Billski
 
Last edited:

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
The connection at 4 will carry virtually zero load from vertical tail loads and only carry modest fraction from horizontal tail loads.
Yes, I agree that connection 4 not doing anything for tail balancing load. But maybe doing something four engine support, or balancing two people upfront?

47169844_2343427855685717_6312039815449673728_o.jpegScreen Shot 2021-02-18 at 20.17.25.png
 

Map

Well-Known Member
Joined
Dec 29, 2020
Messages
102
Location
California
Connection 4 is very important to transfer vertical bending from horizontal tail loads to the rest of the structure.

If you want to make the tail boom longer, all the loads increase and all connections will have to be re-evaluated.
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
Connection 4 is very important to transfer vertical bending from horizontal tail loads to the rest of the structure.

If you want to make the tail boom longer, all the loads increase and all connections will have to be re-evaluated.
Tail boom will need to get longer by 2 feet and stronger and horizontal tail will need to get larger by 10% to get to horizontal tail valume up to 0.44 from 0.32 presently. I do need larger tail to compensate high thrust line and CG spread 23% - 36%MAC.

What I have right now is simply unacceptable. 120 x 2mm 2024T3 tube is too flimsy to do the job. Boom is flexing on the load in flight and driving me crazy together with all connections and fasteners.



9C7A6F48-DADE-4444-998F-77CEA7644A90_1_201_a.jpeg
This bolt from connection #3 under engine mount
 

Eugene

Well-Known Member
Lifetime Supporter
Joined
May 26, 2017
Messages
1,787
Location
Merrill, Wisconsin, USA
Connection 4 is very important to transfer vertical bending from horizontal tail loads to the rest of the structure.
I know very little and really shouldn't even open my mouth. But stubborn Russian inside of me really want to say something. So, I will take a risk and try anyway!

There is no argument that connection 4 is very important in present configuration. Because connections 1 and 2 doing nothing for vertical tail load transfer.

But if we going to create carbon tail boom that will be plug-and-play in to 1, 2 and 3. All needed work will be done between this three points. Do we need to incorporate #4 into the structure? I personally feel that we should, but don't really know why.

IMG_5266.jpeg

IMG_6392.jpegIMG_6391.jpegIMG_6388.jpeg
 
Last edited:
Top