Mike,
There is nothing wrong with your approach, so it comes down to the quality of execution and the details. Let me throw out some thoughts in perhaps random order to consider. Then guys can throw rocks at it if they feel the need. There are lots of ways to accomplish things that work. What works for some won't work for others because they don't have the tools, or can't get the right materials so they substitute, or they insist on doing something with glue that isn't a good idea. That kind of thing.
1. The pushrod is not going to lift a battleship. There is only so much axial force your plane can generate in that pushrod. So, the trick is to have sufficient design margin above that amount and not add any more weight above that. If you don't know the actual loads it will see, then you use tribal knowledge (or Sparky's TLAR) and hope for the best. Tribal knowledge tends to make things bigger/stronger/heavier over time because no one wants to lose a nice plane due to control failure. So, I occasionally see a 3/8" diameter carbon pushrods with really thick walls, and....
2. Most of the pushrod failures that have been documented on Stunthanger that I can recall have not been due to axial quasi-static overload failures. The ones I am thinking about were fatigue failures, and just changing fittings to titanium does not necessarily make those go away. And related, Paul's point I believe was that minimizing the "stick out" of your threaded stud reduces the bending moment from lateral loads (static and dynamic) on the threads right where the stud enters your aluminum threaded bushing. That is a very good idea. Let the carbon do what it is good at. One thing I would be absolutely sure of is that your threaded rod has rolled threads, not cut threads.
3. There is nothing inherently wrong with using aluminum against the steel threaded rod as long as the contact pressure is below the galling pressure for that combination. Personally, I would install the stud in the bushing using Loctite and then install a jam nut on the stud. And use an appropriate torque. I like that your bushing is long and you can get a lot of threads engaged. Know that all threaded interfaces deform to allow load-sharing. In other words, the stud will stretch if you put a jam nut on it and cinch it up to stabilize the joint. (All proper bolted joints stretch the fasteners. They are supposed to.) If you don't like the jam nut idea, then you are going to need Loctite and you may have bigger issues with fatigue.
4. Using 6061 aluminum for the bushing is good, but to know that it is suitable you also need to know the temper. You don't want soft, gummy and low yield strength. Condition T6 is excellent.
5. Shear/tearout of the aluminum threads is pretty unlikely from actual use due to the plane not likely creating huge forces in the controls. If you have 8 or more full threads engaged you should be golden. With four or five you would probably "get by," but why not increase your margin for a fraction of a gram?
6. Note that with your shouldered bushing, if it is bonded in contact with the end of the carbon tube it can't go anywhere when the pushrod is under compression even if the bondline failed. So your bondline isn't really even loaded when the unit is in compression.
7. The regular JB Weld is a surprisingly good adhesive for a consumer grade material. I tested some side-by-side with aerospace epoxies for fun years ago for a government project. However, bonding techniques are critical. Cleanliness and proper solvents are critical. Wetout is critical. You don't want any residue left by your cleaning process. Straight acetone is good. Cleaned several times. The issue with bonding aluminum is that it instantaneously starts building an oxide layer when it is cut. The oxides do not improve your bond. For best results, sand it off and do multiple acetone wipes. (You can scrub the oxides off of larger aluminum joints with Ajax cleanser and water, which works well, then rinse thoroughly with water and check for "water break free" conditions. That tells you it is clean enough.) You can use a bit of heat and then apply adhesive all over the contact area. Do the same to the inside of the carbon tube. Wet everything with adhesive that you expect to stick. Now push the parts together, use heat to control viscosity and material flow as needed. Wipe clean. Air cure. Epoxy does not develop full strength for about 5-7 days, no matter what it says on the bottles. At colder temps, the full-cured strength will be less. Letting it cure until it seems "pretty hard" and then post-curing at higher temps will get you the max strength that the epoxy is capable of. For aerospace materials, the manufacturer will tell you what these cure schedules look like. Think time/temperature charts. For a lot of epoxies to get to their quoted strength you might need an hour or two at around 120F? You would need the data sheet to be exact. That said, JB Weld with just a room temp cure for the geometry you are showing with a well-prepared joint should exceed any tension loads your pushrod will ever see.
8. Another way to look at this is which part in the loadpath is the weakest? Will the threads pull out of the Delrin rod end before they pull out of the aluminum? Will the bond joint fail in tension before the Delrin? Will the 3mm threaded rod fail first? Will the surrounding material of the ball socket fail before any of these? All things that can be calculated or tested, but hey, this is just a hobby and maybe tribal knowledge is good enough?
9. It never hurts to score or "rough up" the two bonding surfaces unless that damages/reduces their own structural integrity. In this case, no. But mostly it just breaks up any cleanliness issues on the surfaces. It does also act as a failsafe should your glueline fail, since it keeps the bushing locked into the tube. But if done right, the glueline will not have an adhesive failure, and so rings and gouges and spiral grooves aren't necessary except for easier adhesive application and venting. But I do it also because it costs little to do the extra prep in case I somehow contaminated the joint and didn't realize it. It has happened to some of the best. (aka the "fried chicken incident.")
10. Epoxy joints are strongest when the bondline thickness is controlled. In a metal-filled epoxy you might think this is already done for you. I don't know their particle size, but it is likely very small and not too helpful. Most epoxies I have used will do best with a .002 to .010 inch bond line. It leaves room for the epoxy. It also helps provide some temperature compensation for dissimilar materials like your carbon and aluminum. Probably not too severe a case here since your plane (hopefully) will not see large swings in storage temperature?
11. Failure analysis of a bondline starts by determining the type of failure. A cohesive failure is when the glue is still sticking to both sides of the joint, but the bonding material has sheared between the two surfaces. That means your prep was good or very good, but the adhesive didn't live up to the loads placed on the joint. An adhesive failure is when the glue comes off the parts clean, usually predominantly on one side and the bulk of the glue stays stuck to the other side. That means the joint prep was inadequate or the wetout of the adhesive on the parts was not done quickly enough, or the material was too old, etc. Most of the time, a failed joint will be a combination of adhesive and cohesive failure. If there is a pattern of total adhesive failures then the training, tools, or materials are lacking. To figure that out you have to watch the process. By watching, you can usually figure this out pretty quickly. I worked with a guy who worked on the Grumman American light plane that first used a FAA certified process for bonding primary structure out of aluminum. It was a rigorous development, but seems to have been robust.
Just a few thoughts and some rationale behind them. But modelers have their special methods that they learn to trust and their reasons for doing things that way. But I don't see anything fundamentally wrong with your approach. So it comes down to the little details in actual execution.
Dave
Topic: Carbon fibre pushrod ends (Read 7084 times)