For starters...
1. Developing sufficient power.
Okay, I'll stop there, since upon reaching that goal one runs into other problems that are, more or less, typical to other catapults.
I should refine that 'problem' by saying that "sufficient power" results in one of two scenarios, either (1) self-destruction of some sort or (2) Payload range meets expectations. In the case of (2), folks often start looking at how to extend the range and, once again, we're back to (1).
So what different forms does this self-destruction manifest in?
1. Frame collapse.
2. Torsion bundle falure, in various forms.
3. Tensioning difficulties (basically, one can't get the torsion bundle up to expected torque values.)
4. Trigger failures. (I include safeties, haul-downs, etc.)
5. Throwing Arm failures.
6. Pouch or Sling line failures.
7. Payload failures, such as pies or other failures due to high acceleration rates, too-high spin rates, etc.
There are other forms of failure, such as frame hop, damage to unintended 'targets' resulting in collateral damage, etc., but I'll leave that listing alone. On this list also include bad tuning, operator head-space misalighnment, bad karma and pesky neighbors who used to own glass houses.
Each of the 7 listed items above can be broken down into finer specifics, but you get the general idea, eh?
1. Frame collapse. This usually occurs when the power of the torsion bundle becomes greater than the frame can withstand, usually resulting in an implosion of sorts. It can also occure when the Throwing Arm (or twin arms of a ballista) impart too much energy back into the machine and again we see a collapse of some sort.
1A. The typical solution is pretty simple, to build a more robust frame and/or one that is better designed to handle the stresses involved. With few exceptions, usually found among the experienced 'Torsion Techs", most folk greatly underestimate the power potentials of torsion bundles. Even in small models, the forces can be measured in tens or hundreds of pounds of torque. Get anywhere near a 'big' machine and we're talking tons, usually multple tons of torque. This is serious power folks and needs to be dealt with accordingly.
2. Torsion bundle failure, in various forms.
2A. This may include poor material selection, hidden wear and tear resulting in internal breakage, improper lay-up (the pattern used to build up the bundle), rebound effects after Payload release (or if you prefer, at the end of expected TA rotation), etc.
Many of the problems encountered with torsion bundles can be avoided with a little research into how others have solved/avoided them, such as lay-up patterns, termination of rope ends, pre-torque strain values, etc. Probably the hardest problem is the selection of a good material. This research usually results in a balancing act between performance and cost. Hey, some of those ropes are pretty pricey! It is possible though to use a lower cost material while still obtaining high performance values. This takes some decent research and design but certainly do-able.
One can also opt for non-traditional torsion bundle designs, something like the Wankle design refered to by Walker, the use of other materials (piano wire?) or a radical departure in torsion bundle concepts. Hey, that'd be nice to see, eh?
3. Tensioning difficulties (basically, one can't get the torsion bundle up to expected torque values.)
3A. Here I see two fairly common problems, (1) some type of breakage during the main torquing sequence or when the TA is being cocked down, (2) insufficient power applied to torque the bundle to desired levels.
In the first case, redesign or at least look over the torqueing method. In the second case, again the potential power of torsion bundles is often underestimated and inadequte methods are used, to torque things up. Sometimes this is a simple case of "Get a longer lever!" or "Have 'Slim' jump on the end of that lever!" Alas, sometimes 'Slim' has been drinking low-carb beers and isn't all the man (belly?) he used to be or a longer lever isn't available or strong enough.
Another problem may be too much friction between the rotating parts and the frame. This can be solved by some redesign. I think most of the torqueing problems can be overcome with proper forethought and designing accordingly. In other words, understand ahead of time what the potential tension values of your torsion bundle can be and what it will take to overcome them. This might include gear trains, sufficiently long and strong enough levers, hydraulics or other non-human power sources, etc. One may also look at something really different, such as applying tension to each strand individually. (!) One could also look into the lay-up of the bundle and see if there may be a pattern that lends itself to easier torqueing.
4. Trigger failures. (I include safeties, haul-downs, etc.)
4A. Once again (how often am I going to say this?), not understanding the potential power of torsion bundles if often the reason for failures. Trebuchets can have a lot of power, but it isn't all available in a 'lump' at the trigger or haul downs. The real power comes from the fall of the CW and builds up over time. Torsion bundle machines do have all their power in a lump though and must be controlled by the haul downs, triggers, etc.
One also needs to understand that all that power must be released in a controlled fashion. Triggers need to be robust enough for the job while still allowing for controlled release. Another important factor is understanding what can happen with the trigger after it has been released. "For every action there is an opposite and equal reaction." If you have a 1 ton load on the trigger, then not only is the TA going to have some nice power behind it, so is the trigger! Don't let it get loose nor break the frame of the machine.
5. Throwing Arm failures.
5A. If the TA snaps while being hauled down, then there is a good chance there was a weakness in the arm itself, possibly something that occured after the arm was made. If the arm snaps post-release, then a better stopage system needs to be employed, such as the one Bob Carbo uses to great effect. If the TA snaps during acceleration though, well at least you know you have plenty of power available, eh?
It is very possible for a TA to suffer internal (as well as external) damage during use. Unless you have x-ray vision, these defects are almost impossible to know about. Even man-made materials can suffer from material fatigue without visible evidence. Unless you got the finances for a lab to do a critical exam, you do the best you can.
To avoid TA breakage from a 'too powerful' torsion bundle, there are several options, sometimes in combination with each other.
You can make the TA shorter, this reduces the bending stress on the arm but will probably result in shorter hurling distances (not always though!) You can include guy lines on the TA or some other flex-avoidance design, such as a truss style of TA. Better material selection can be a real help, of course.
6. Pouch or Sling line failures.
6A. Well, did you take into account the power of the machine when you chose the Sling line material? How about for the material selection, design and usage of the Pouch? Did you examine both after each use? Okay, I'll stop leaning on ya...
Not only is strength an important factor in Sling line material selection, above a certain power level, so is elasticity. One doesn't want a 'lot' of stretch, but certain amount will greatly reduce the strain on the Sling while not subtracting from Payload performance. How much elasticity though? That will depend on the overall design of the machine. Of course, if the lines are sufficiently strong, you won't need that stretch factor as a design element.
Pouches are, again, a matter of material selection but also of design considerations. There are methods to spread the strain of acceleration over a larger area of the Pouch, thus reducing the odds of it coming apart. Reinforcements in certain areas of the Pouch may also help.
7. Payload failures, such as pies or other failures due to high acceleration rates, too-high spin rates, etc.
7A. Maybe it's time for bowling balls?
