Forum: Catapult Message Board

A poem from long, long ago, about the Wankel engine:

Twinkle, twinkle little Wankel
How I wonder if your crank'll
Torque enough to turn itself
Or if you'll wind up on the shelf.

From the letters column of, maybe, Road and Track or Sports Cars Illustrated (precursor of Car and Driver).

Thomas

As a member of MrB's team I am somewhat partial to ballistas and can give you some things to think of in ballista design. The most obvious is of course compounding. Since ballistas are quite like bows we get to see what people tried already with bows first, instead of having to break new ground like chucky. What interested me most was some the precursors to the modern compound bow, the dynabo and the Kam-Act.

see compounds and look for Kam-Act (yes they use the K too) and one limb bows for the dynabo.

ballistas are pretty complex beasts to get right right and compounding them will make the problems worse. The reason these two interest me is that a dynabo ballista design would only need one bundle and the Kam-Act shows a good idea of how to separate the compounding pulleys from the ends of the driving arm.

Kev

someone has been thinking abouth this a long time ago, don't think he did more then just draw it...

What are the problems commonly encountered with Torsion machines and how can they be solved?

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?


As for the original question by Walker, "What is the next technological step forward for torsion?", I've posted a bunch of stuff in the past and not going to bore you with them again. Tomorrow is another day though and who knows what this latest fever will seed my brain with...

Ripcord,

Please save that last posting as a blog for future hurlers to reference.
I learned from it.

Thank you,

Walker

Another highly overlooked issue between gravity and torsion is the way in which the sling is employed.

Speaking typically:
A Catapult sling starts in the hanging position, pointing to the 6:00 position and rotates 135 degrees in the split second it takes the arm to move 90 degrees. The payload of which must re-vector from a direct vertical line to a (presumed) 45 degree forward ballistic attitude.
A sling that is the length of the arm will not start that re-vector until the arm has reached the 12:00 and the payload can simply coast into the release angle. (Boo!)
But a sling that is roughly half the length of the arm will start the re-vector when the arm reaches the 10:30 position; or halfway through the 90 degree swing cycle.

However, a trebuchet uses a longer arm swing (and a longer arm usually) which allows a lower RPM for the same curvilinier distance. But the treb doesn't swing the same distance as the catapult and starts with the sling lines and projectile at a position where they can not start their swing-out until clear of the ramp, a position reflecting a 120 degree angle from the tip with a 100% sling length.
This gives the sling 120 degrees of swing to bring the payload around with a 135 degree re-vector and a much longer swing in which to do it.

Note: with the 120 back angle from the tip, if the treb arm moved as fast as the cat arm, the payload would be dragged to the inside of the arc due to centrifugal effect.
This re-vectoring of the projectile is vital to accelerating it to a velocity exceeding that of the arm swing.
Solid pumpkins are esential. (Time for bowlingballs?)

Within the longer statements is/are more data than I can put into this posting with the time limits I have right now. Let me say simply that the Lift-and-revector method used by the torsion catapult is the superior method and can not be enjoyed by gravity engines due to their longer arm swing until those arms learn to swing faster.

All this being said, I really only wanted to explore the method of powering the post-bundle torsion era.
Think: Giant mousetrap spring is torsion.

Walker

Blog is a bit of a mash-up, as I seem to have forgotten how to blog properly, but it's up now.

Torsion problems?

Maybe it's time for bowling balls?

Or medicine balls, I can get eight pounders at the local sports shop. Pie-proof and a better mass ratio.

One problem you also run into is the speed of the elastic reaction of a given power medium. Anything from para-cord to steel springs will only "snap back" at a certain speed. Adding more strands will allow you to move more weight, but not necessarily at a greater speed, which you need to get greater range. The best immediate hope for torsion I've seen is to improve on the Hatra ballista designs. Maybe by trying to get more arc out of the arms, or just going with a larger unit all together. I've refrained from getting into punkin chunkin mainly because I don't have the recources to build and transport an onager with a 12" skien. I don't think multiple skiens like the ones pictured will work as you'll loose too much in the power transfer. Plus the linkage where the throwing arm attaches to the "hub" would be subjected to extreme stresses that would split out most materials. A big part of the problem for punkin chunkin as it relates to torsion, is that eight pounds is just too light a weight to be throwing. You might as well be dry firing those machines. There is a point at where decreasing the weight of the projectile for a given engine will not increase the range of said projectile. Increasing the size of the engine while maintaining the projectile size is doing the same thing, but the hard way

''Hatra ballistae don't seem to offer much advantage either.
I could see it if it were an inswinger ''

I just assumed a Hatra is an inswinger.

Not the garage door roll-up type but about massive specially made coils say maybe

I'm not going to tell you to not try it, but I was in the garage door business and I don't think it will work. If you want a more indepth analysis, e-mail me privately, but I don't think the springs will do what yo want them to, and to mount the springs you'd have to make holes in the base of the throwing arm creating multiple weak points.

In Cats, this translates to being able to move a longer arm which provides faster curvilinier tip speed at the same RPMs. Ergo, better projectile velocity.

I find this very interesting. Having tried different length arms in the same skien I've found the shorter arms to provide greatly increased range over the recomended arm lengths. Marsden says to use an arm 8-9 diameters in length, where I get the best performance out of an arm of about 6 diameters. I'm not saying you're wrong, I'm just getting diferent results, with or without a sling.

I think to truly increase range significantly in a torsion device, a person will have to totally throw away everything we know about torsion and come up with something completely different. Anything from history is based on being able to throw a mostly constant range, 400-500 yards. The only variable being the weight of the projectile, and thusly the job it is to do. (you want to throw a bigger rock, build a bigger catapult, but you'll still be just as far from the city walls.) We on the other hand have a constant projectile but are trying to increase the range without destroying that projectile. That's a very different set of circulstances.

Reading Marsden or any of the other 'experts' should be done with a grain of salt in mind. Their research and experience may have been sufficient for them, at their time, but we are 'now', with a much wider range of experiences (shared in many cases), along with (imho) new insights as well as new materials, manufacturing methods and, possibly most important, a willingness for a certain degree of departure from historical models. This changes things, sometimes in a subtle fashion, sometimes a more dramatic one.

(Pardon me, I'm kind of picking my choice of statements to respond to here.)

Multi-skeins CAN increase rotation rates. There is a limit though, naturally. For a given amount of torque, applied against a given amount of resistance, there can be only 'X' amount of acceleration, hence Payload speed. That is one limit out of two. The other is if the resistance is low enough (think dry fire) then then system will accelerate at some higher, maximum value.

I agree that any torsion material will have a given 'snap-back' or reflex rate. Until one reaches that rate though, once can still possibly suck more energy out of it. If one is on the ragged edge of limit for that material, adding more strands will share the load, allowing each strand to deal with a smaller fraction of the work load, hence more power available.

I agree that garage door springs (or similar springs) will not work that well when used in torsion, as opposed to tension/traction applications. True, one could get very nice performance, especially if one can get coils specific to the task, as many have. The problem is that they are intended for a lot of rotation with moderately high torque values, whereas what is needed are much higher torque values through a fairly small rotation sweep.
Ah...a possible answer! "What if" one were to cut such spring into many shorter sections. Then by some process anchor one end of each section while allowing the other end to apply torque to the TA?

Wife home, gotta run...

"What is the next technological step forward for torsion?"

I'll disregard changes in frame design, for the most part, as these changes are normally only to allow for a change elsewhere or to provide more strength overall.

I think there is a lot of room for development in Torsion bundle design. Material selection is only part of it, there is also pattern lay-up, additional structures directly involved with the bundle (spacers, bearings, etc.) even internally. We starting to see some changes already but there is a lot of room for more new stuff.

Throwing Arms...woof! Talk about areas in need of improvement! Sure, material selection is there but that has been done and will continue. More important is the use of that material(s) in new ways.
TA breakage is a common occurrence and this is due primarily to the exceptional power available and the way it is used. New TA designs should allow that power to be transferred to the arm in a way that better avoids self-destruction.
I also think that many new TA designs will allow for even more power to be applied to them, also adding to performance.

I think a better understanding of the math, physics and geometry behind torsion bundles will also allow for improved performance, either with the classic bundles or leading to new bundle designs.

Yea, I've an idea or two of my own. Current life style doesn't allow for practical development though. (Per usual...)

Here you go, some questions for consideration, a list that I've been working on and thinking about for a while now.

What if the the strands of the torsion bundle were made of multiple ropes instead of just one?
What if the outer ends of each strand were individually tensioned?
What if the strands, rather than meeting at a common point at the outer ends of the bundle, were splayed outward away from each other, some distance?
What if a new material were used for the bundle, such as piano wire?
What is the outer ends of the torsion bundle were used not only for the main torquing of the bundle but also for the cocking?
What if each strand of the bundle were of a different length?
Can the center of the torsion bundle be replaced with a support member for the frame?
Can the strands be held in a pre-determined pattern during all phases of set-up and operation?
Can the bundle be designed in such a manner as to allow it to maintain the bulk of its tension yet be separated from the frame as needed?
Does the Throwing Arm need to be a solid piece at the region of the bundle?
Can the Throwing Arm be forked at the bottom and, if so, is there any advantage?
Can the frame of the machine render a design element that contributes to the tension of the bundle, rather than just supporting that tension?
Can multiple torsion bundles allow for greater power and/or better design configurations?
What other methods are available for torquing the bundle?
What torsion bundle materials are best and what are the design criteria for selecting them?
Would there be a significant advantage to torquing the bundle as opposed to cocking the TA, for firing?
For a given strand material and construction type, what is the best radial distance for it to be at, from the center of rotation of the TA?
Should strands around the base of the TA be distributed in a fixed pattern or a determined non-equal pattern?

Sure, I've got some ideas for each of those but....you first.

Torquing up a torsion bundle is often a difficult task and many solutions have been used. Here's an offering for this topic, one that I've never seen used, although it may have been.

It is a form of worm gear called a 'Double Throat Worm Gear'. Many of you are probably familiar with a standard worm gear, which has been widely used in industries for many, many years. The typical 'hose clamp' is usually a worm gear type.
The problem with the standard worm gear is that it isn't all that strong, since few teeth are engaged with each other, putting a lot of strain on any single tooth.
There is the Single Throat Worm Gear, which is one solution to this weakness problem. I'll skip over it though, has it is incorporated in the Double Throat version as well.

The idea is to bring both gears closer to each other, allowing more teeth to mesh together and spread the load over more teeth. This is done by creating a groove in the driven gear, through the tops of the teeth. This allows the shaft of the worm gear (which drives it all) to get closer without collision.
The worm gear is very highly modified and, in fact, wasn't even really possible until just a few decades ago, due to the complexity of the math involved. Anyway...
The worm gear is hour-glass shaped, having a larger tooth radius at the ends then at the center. This allows the teeth at the ends to be able to 'reach out' further and maintain contact for a longer period of time. (Okay, it is really just one tooth, or thread if you prefer, but hopefully you'll get what I mean.)

I've attached a picture of a very poor graphic I made up. It's really bad but hopefully illustrates what I'm talking about without having to 'borrow' it from elsewhere on the net. (I made a much better graphic a number of years ago, for Bob C., but that hard drive died and I don't have a copy anymore.)

These types of gears can have ratios of 20:1 up to 300:1 or more. They can be very strong and withstand a lot of loading. One version or another exists in many steering boxes for trucks, although those have limited travel, using a segment rather than a full circular gear. They also have the design option to NOT reverse direction under a load! That means that a catch pawl is not required, but the gearing can be un-wound easily.

Mind you, these things are not cheap. They are precision pieces that take a lot of engineering to get right and are not simple to manufacture. Still, for those interested, one might investigate junk yards or the like and see what might be available. Even a 'junker' might be perfectly good for torsion use.

Also bear in mind that the worm gear could be driven by more gearing, increasing the ratio even more.

Another option is to use a standard worm gear but mount additional worms to the bull gear, then drive them all together. This too could reduce the strain on any single tooth.

I didn't mean to imply that double throat worm gears were something new, only that they were much newer than a standard worm gear. In fact, that wouldn't be hard to do, considering worm gears have been around hundreds of years, although their true development didn't take place until much later.
So, who invented the worm gear? As best as I can determine, our good old friend DaVinci.

Hydraulics? Sure! Bulldozers and similar tracked vehicles need their tracks re-tensioned often. This is often done using a simple grease gun, hand powered no less. A large diameter ram provides the *oomph* to increase the distance between the idler and power sprockets, thus taking out the slack of the track. The large diameter of the ram means only a few pounds of force per square inch are needed to generate a ton or more of force to move the sprockets. No reason it can't work for torsion bundles, with a little design effort.
In fact, instead of using the more common torsion strand materials, I'd look into using steel cables and fewer than normally found in a bundle.

Questions for consideration:

Q: What if the the strands of the torsion bundle were made of multiple ropes instead of just one?

A: Multi-strand torsion bundles offer a couple of nice options. You could spread the strands over a larger area, at each end of the bundle, rather than having them all come together to make that turn. This optional design choice could open up some of possibilities, such as redirecting the angle each one has with its contact point on the TA. With separate strands, one could lay up the
bundle much easier. Thread 6,10, 15 feet of rope instead of several hundred. Less chance for wear and dirt to get into the strands and a heck of a lot less walking! Another nice option with multiple strands is that, once in place, you could tension each one individually. This would not only take less effort, as opposed to pre-torque the entire bundle as a whole, but you could optionally tension each one to a different amount. More load on those inner strands, that don't travel as far, less on the outer ones.


Q: What if the outer ends of each strand were individually tensioned?

A: Okay, so what if you tension each strand to a different level? The classic torsion bundle has its strands pretty much grouped together, with some on the outside, others to the inside. Yes, there is that split in the bundle as it passes around the front and back side of the TA, but that just makes the situation a bit worse. This all results in strands going from a given location at one end of the bundle, to probably somewhere else around the TA, then (with any luck) back to a similar, if not exactly the same, location at the other end. For those who attempt to torque their torsion bundles to the max, this means that some strands will be near breaking, while others may have a lot of potential tension still available left in them. It can also mean that those strands that are still a bit loose may be suffering a bit more, getting crushed by strands that have more tension in them.
If strands can be individually tensioned, one could be much more sure of obtaining maximum power from the bundle as a whole. Further, with appropriate tools and consideration, one could measure the tension of each strand as it was being tightened!
If one is careful in the selection of bundle material and knows its mechanical properties well, combined with the dimensional changes it undergoes from lay-up to torqueing plus the added changes due to cocking, one could pre-determine the maximum tension value desired and adjust each strand accordingly.
Now, I haven't forgotten that most torsion bundles are fully capable of collapsing the frame of the machine, without all this attention to fine detail. So why get nit-picky about strand tension? Well, if you want to get maximum performance from a given machine design, here is one method on that road. Also, if the frame is designed to support whatever can be thrown at it, then you would then have the ability to get maximum power from the bundle.


Q: What if the strands, rather than meeting at a common point at the outer ends of the bundle, were splayed outward away from each other, some distance?

A: Consider how the power of the strands is transferred to the TA. The TA rotates through a vertical plane, while the strands cross in front of and behind it. The strands are deflected from their desired path, which would be a straight one, by the torque applied to the bundle as well as the additional torque of cocking the TA down. The tension in the strands attempts to pull the strands back into a straight path. The problem is the severe angle that the strands must work through and by "severe" I mean a very shallow angle. If the strand were able to apply their tension in-line with the plane of rotation of the TA, then almost 100% of their power could be transferred but the shallow angle they actually have means that much less is actually transferred, sometimes well below 50%. (The actual proportion can be figured out with a bit of trig and an understanding of the angular relationship of a given strand to the vertical rotation plane.)
Running the strands directly forward and back could allow for greater power, but doing so would make the normal cocking methods next to impossible. There is also the problem that many materials may not have enough elongation (stretch) in them to cover the distance required for TA rotation. (Think of the arc length distance the TA travels, where the strands come in contact with it.) However, using a material selected for a determined amount of stretch, one could lay-up the strands in a manner that would have them contact the TA at a greater angle than usual.
This would mean that the strands should NOT collapse in towards each other, lest one lose that angle advantage. This method would also require that, in most cases, the strands be individual as opposed to a single long strand laid-up into a bundle. As such, one need not put them into a circular pattern. This could offer another method for focusing the power where needed. Imagine that half the strands are roughly grouped together and come together to pull the TA forward. The other half of the strands could also be roughly grouped and aimed to pull the lower end of the TA towards the rear. There may not be any significant advantage to this grouping but I think the option is worth looking into, especially if the strands do NOT collapse inward towards each other. If they do then they may as well be operating from the usual location at the end of the classic bundle.
This design option is there but is probably, by itself at least, not worth 'going for'. The power already available in most torsion bundles is already pretty awesome and increasing it further would probably mean some structural failure elsewhere, all for a possibly less than significant increase in power. Crunch some numbers though and see how they compare with whatever machine you already have (if any.)


Q: What if a new material were used for the bundle, such as piano wire?

A: Well now, if you want to put more power into an old design, this is one way to do it. Selection of strand material is usually pretty simple, weigh the performance cost against your wallet and get what you can. The breaking strength, elongation rate and working load rating are of primary consideration although other considerations should also be taken into account such as weathering, wear resistance, ease of use, etc. If your design also requires splicing or end termination of a unique sort, then that should also be kept in mind while selecting materials. There are a lot of web sites out there, mostly commercial ones, that can tell you a LOT about a given materials properties. Just bear in mind that these are mostly commercial concerns and 'full honesty' is not always consistent with high volume sales. For the most part though, you can compare several sites that offer the same material and get a feeling for what the honest numbers are.
For a new machine design, careful consideration of the material to be used for the torsion bundle should be high on the priorities list. For a pre-existing machine though, one should look at the specific capabilities of the design then select a material that best suits the machine, as well as whatever one hopes for in performance. Just don't let the 'hope' exceed the design, lest you suffer from catastrophic failure.
For the more exotic materials, such as piano wire, steel cables, nano-tubes, rolled up graphene, carbon fiber or whatever your favorite science fiction author has dreamed up, the selection process will be very similar. Some materials may allow for more compact designs, easier maintenance, more consistent performances or some other feature you desire to have. One great option allowed for by other-than-normal materials can be some large design changes in other parts of the machine. Then again, perhaps a new machine design will demand a new torsion material.
Don't forget that you do not have to limit yourself to just a single material! What if you have a material with a large elongation rate, which torques the TA on a relatively large radius, while using a different material, with a smaller elongation rate, that applies torque to the TA at a smaller radius? This could provide great efficiency, but your math better be up to snuff...


Q: What if the outer ends of the torsion bundle were used not only for the main torquing of the bundle but also for the cocking?

A: Okay, I'm not so keen on this but, it is a possible option. Imagine the torsion bundle mounted a rotating housing of some sort. Un-lock the housing and rotate it, along with the TA, down to the cocked position. Set the trigger/safeties on the TA. Now rotate the bundle housing back into its former position. Off-hand, the only advantage to this sort of system, that I can see, is that one could 'take hold' of the housing with really strong and long levers and torque on it hard, possibly easier than trying to haul down the TA in the usual fashion (whatever that may be for you.)


Q: What if each strand of the bundle was of a different length?

A: I've mentioned this a little bit earlier in here and certainly at other times and places. To put it in a lump here though...
I first considered this idea in connection with classic torsion bundles. With these there is a strong tendency for those strands on the outside of the bundle to continue on the outside and those on the inside to continue on the inside, for the length of the bundle. As the bundle is twisted, those strands on the outside have to stretch further than those on the inside. Once you reach the practical working limit of stretch on the outside strands, the inner may have significant stretch still available, a loss of potential power. How much will depend on the material elongation rate, length of bundle versus its diameter, size of split where the TA seperates the two halves the bundle, etc.
Knowing the particulars of a given bundle design, one can calculate how much difference in stretched length the two types of strands will have. Know the elongation rate of the material, plus that calculated length, one can then determine the over-all lengths needed for the outer, inner and intermediate strands. Then you just have to build the bundle accordingly and be sure those strands stay at their respective radii.


Q: Can the center of the torsion bundle be replaced with a support member for the frame?

A: Certainly! It has been done with two machines already, that I'm aware of. The basic idea is to support the torsion bundle primarily through an internal mechanism, as opposed to an external framework. By putting the support directly in line and centered with the tension of the torsion bundle, one is able to gain a much higher efficiency of structure. In almost all other cases, the strain of trying to support the torsion bundle tension must pass through a right angle bend or joint of some sort, usually more than two. Almost all frame failures occur at these points. Even better, instead of adding a support for the frame, use it for direct support of the torsion bundle itself. (Yea, that was the idea from beginning, just wanted to see if you were paying attention.)
Assume that a tube is used for this internal support. Tubes are one of the best shapes to resist compression. Solids would be even better but offer fewer design options. With new designs come new problems and this one is no different. Where do you put the TA? How do the bundle strands 'turn the corner' at the ends of the tube and, if individual strands are used, how and where do they terminate? These and other questions all need answering and it is up to the designer to do so, while being alert for problems of a new nature, being unique to this type of design. It also offers, with some careful consideration, the option to keep the torsion bundle at or nearly at full tension while also being able to be removed from the frame of the machine! Now why you would want to do so, well that is up to you. (Note: You may want to use this center support design option in conjunction with the 'grease gun tensioner' mentioned elsewhere, on the CATMESS.)


Q: Can the strands be held in a pre-determined pattern during all phases of set-up and operation?

A: Yes. Why bother though? A quick look at a typical torsion bundle will show that, especially after it has been fired a few times, the strands are not in a nice pattern. They may start at the outside, dive into the bundle, twist around each other and who knows where the other end comes out. Where strands cross each other there is a high abrasion rate, leading to early failure. It can get ugly in there. Plus, the strands are usually laid-up in a given pattern with some tension on them. When they move away from the pattern, some strands may become loser while others tighter. This can lead to early breakage, loss of power, inconsistent performance and other ills.
Using some method by which to keep the strands in their respective positions can not only avoid those ills but work towards increasing the potential energy of the bundle. If the strands are put into specific places, with a method of keeping them there during operation, then the strands can be tensioned in a manner that brings each to its own best level of strain without great fear of it breaking due to misalignment. Note that keeping the strands in a fixed pattern is just one option available, yet when combined with some of the other design options mention herein, the possibilities are nice to think about.
Note that there is another option to keeping strands in a fixed pattern. If each strand were to be placed inside a flexible tube, of sufficient strength to resist crushing (a very tall order), then even if the strands move about or not, they are kept from crushing each other. It can also reduce the internal friction of the bundle, as long as an appropriate material for the tubes is selected with respect the strand material. Alas, strands do break, yet by this tube method, one should become almost immediately aware of the breakage and which strand is at fault.


Q: Can the bundle be designed in such a manner as to allow it to maintain the bulk of its tension yet be separated from the frame as needed?

A: Surely, but to what end? Actually, I can't think of a good reason for this one, except possibly.... Your in the middle of competition when suddenly your torsion bundle fails due to broken strands. If you have another bundle standing by, already tensioned, just swap them and your back in the game!


Q: Does the Throwing Arm need to be a solid piece at the region of the bundle?

A: Perhaps not obviously, the answer is no. Here is another design option that has already been put in use, at least once that I know of. One advantage is putting hole in that area of the TA, to allow passage for a pipe, as mentioned previously. There is no reason to have strands passing through the TA, they can't develop any torque from that position. Having material there though does strengthen that region of the TA that sees a great deal of stress. Modern materials can carry such a load while still being hollow. This is an option that offers several solutions, mostly to problems that haven't come up, yet.


Q: Can the Throwing Arm be forked at the bottom and, if so, is there any advantage?

A: Well, yea, of course it can be forked. However, the only real advantage I can see is if the torsion bundle is split into 2 parts, each one fastened to one leg of the fork. This could give more leverage for the torsion bundle, over the resistance of the TA. Pretty nice, if you can prevent the legs from snapping off in the process. There are better ways to increase the torque efficiency though, methods that also increase the strength of that area of the TA.


Q: Can the frame of the machine render a design element that contributes to the tension of the bundle, rather than just supporting that tension?

A: Consider the fact that in most (all?) cases, the frame IS providing the tension in the torsion bundle. After all, without the frame, what is the torsion bundle anchored to that allows it to be strained in the first place? The difference between the two is that the torsion bundle is expected to be elastic while the frame is supposed to be rigid. As long as the power reaches the TA, hopefully in an efficient manner, it doesn't matter which part does which task. This opens a world of possible design options. For example...
What if the side rails of the frame were really, REALLY stiff? (If it isn't there will be a fair amount of energy loss, at best.) Now, instead of trying to twist the bundle, separate the two side rails instead. This could be done from one point, two points, a lot points, whatever the design calls for. The separation could be done through wedges, hydraulics, screws or a bunch of weather balloons...hey, use your imagination. The 'trick' to this sort of thing is to get that energy transferred to the TA, this means some type of linkage (or does it?) such as...rope? Actually, I think a solid form of linkage, such as aluminum bars, would be better as they have a lower elongation rate, they're stiffer and would transfer the energy better.
A nice thing about the linkage option is that one could design the TA to accept the links at various points. Repositioning the tension member to a place other than around the TA opens up more than just a single design option.


Q: Can multiple torsion bundles allow for greater power and/or better design configurations?

A: Yes, more power could be made available. This doesn't mean it can all be used or won't present other problems that prove overly difficult to solve, but yes, there could be more power. Adding more torsion bundles may allow for a better design, depending on your definition. If more power is better, your set, if easier torqueing, set-up, maintenance is the goal...maybe not. There have been a lot of multi-bundle designs offered up, including my own, and a few of them have been made on a small scale.
Given the current state of the art though, the additional power provided by these design options is probably still beyond our ability to make good use of it. Throwing arms break. A lot. Come up with a stronger, more reliable TA, then start thinking about this option.


Q: What other methods are available for torquing the bundle?

A: Oh my, so many options! Levers, gears, hydraulics, hamster wheels, bicycles...the list is almost endless. Much depends on how 'pure' you wish to be. By that I mean, is external power allowable, such as gasoline engines or electricity, or do you wish to be a purest and use only animal power (including humans)? If you wish to use a nuclear power plant, go figure how on your own. If people power is the way to go for you though, then life gets a lot more interesting.
Simple spur gearing seems to be a popular method and although they provide a nice mechanical advantage, other gearing systems can offer more along with other nice features. Worm gears can offer very high gear ratios, 300 or more in many cases. Plus, in most cases, they offer an anti-backup feature that is a natural part of their design. The high gear ratio works in both directions, meaning that to go backwards, one has to work against not only that high ratio but multiply it by any friction in the system. Do not confuse worm gears with some types of bevel gears. They look similar but behave very differently. Worm gears comes in many types and, for the topic of discussion, I would suggest looking at Double Throated worm gears. I suggest this because, due to their design, they are much stronger and can carry a much higher load per tooth than the simpler worm gear.
A simple lever is pretty nice but they often fail under use. Got a stronger lever, but no place to stand and use it? This is where a ratchet system comes into play. Select the right tooth spacing then, for every *oomph* you apply, the ratchet will bump over one (or more) notches and hold things for you until the next effort. Of course, this usually means re-setting the lever for the next effort...um...unless you have another ratchet system for it as well. Making a suitably non-bendable lever should be a simple task. After all, you just made a really nice Throwing Arm that doesn't break, didn't you?
Another torquing option goes back to the multiple strand option, see above in various places for some details. A really nice thing about individual strand tensioning is that the power required to do so is only a fraction of that required for an entire bundle. As such, other options for 'torquing up' come into play which do not require as much effort at a single time. For example, eye bolts could be used at either one or both ends of each strand. For that matter, one could twist each strand! You could use a lever to pull on each strand, insert some spacers, then twist. Some fun stuff to think about with this.
Wedges should not be overlooked. Threads are a modified form of wedge (or vice-versa, depending on your up-bringing...) Although wedges normally have a much higher angle of incidence, they don't have to. Either way, they can be inserted and then forced inward, forcing two plates apart, as an example. Put a set of plates and wedges and either one or both ends of the bundle and you have yet another tensioning method. How about screws that drive the wedges?

No matter what method one uses to torque the torsion bundle, one should keep friction in mind and ways to reduce it. The very large forces involved are enough to contend with, adding to them by large fractions is to be avoided. Looking at a classic torsion bundle, we see the rope bundle, anchored at each end, and those ends usually resting on plates of some sort against the outer faces of the side frame rails. Most designs call for rotating those plates, causing them to slide over the wooden rails. Depending on the coefficient of friction for the two materials in question, the torque needed to rotate the bundle can be significantly higher, although once moving the effort will lessen by some amount. In almost all cases, the static friction is much higher than the dynamic friction. A fast example? Sure...
Dynamic coefficient of friction for wood on steel is between 0.2 and 0.6. (Exact details of the materials, temperature, humidity, etc. will determine the exact number.) Lets pick 0.4 and work from there. Assume the bundle has 20 strands (one rope, passed back and forth for 20 strands around the TA) and each strand in under 1,500 pounds of tension. 20 x 1,500 = 30,000 pounds. 30,000 x 0.4 = 12,000 pounds. So, very roughly, this is the force one needs to overcome to continue rotating the bundle. Bear in mind that to START moving things, the effort will be much higher (static friction) but if one uses some lubricant, like grease, the effort can be greatly reduced. One example reduces the coefficient from 0.4 down 0.05!
Other methods for easing the effort of rotating the bundle ends exist, such as selecting washers that, when combined, reduce the friction coefficient. By far though, the best method is to use bearings of some sort. Ball bearings are pretty cheap and easy to come by. One could even use non-standard balls that may work as well, such as billiard balls. Bear in mind you'll still need to create a track or groove for the balls to follow and the support structure needs to be stiff enough not to collapse around them.
Tapered roller bearings would work best, as the taper of each roller keeps it moving in a circular path and their length provides a larger support area, as compared to simple balls. Finding these kinds of bearings of sufficient size for the bigger torsion bundles is probably not an easy task and can be costly. If you have the resources, making these isn't that hard, if you have access to a lathe or can bribe a buddy with one. You'll have to figure out the taper needed and how to keep them in place, etc. but this is a task not beyond many hurling hobbyists. Even mild steel will work, no need to find something 'exotic' for this project! How ever you do it though, reducing the friction in a torsion bundle will make torquing the bundle a lot easier!


Q: What torsion bundle materials are best and what are the design criteria for selecting them?

A: Impossible to answer, as each machine is different, the goals may be different, resources different, etc. The criteria will be the same though, even though the list may be re-ordered. Here is a listing of some options to consider.

Cost, availability, workability (can you tie it easily, untie it, splice it, etc.), flexibility (some materials are very nice but can lose a lot of their strength when turned around a corner), elongation rate, working load limit, breaking load limit, wear resistance, weather resistance (including UV rays!), moisture retention, cyclic loading (ability to be stressed many times without losing significant properties).

The elongation rate is one of the most important mechanical properties. This property is readily available from the manufacturer, often through their web site. This property lets you know how far the material with stretch without damage, while being able to rebound back to its original length (or very nearly so.) Some materials are pretty stiff, such as steel, while others have a very high elongation rate, such as a typical rubber band. In most cases, the stiffer the material the more *oomph* you get in return. Sure, rubber stretches a lot and would be much easier to twist, but the force needed to stretch it is pretty low and that is the same force you would get back. So, want to go with something stiffer? Sounds good, so how about steel strands? Yes, you could put a LOT of energy into such a bundle but...would you be able to twist it? Probably not very well. So we look for something else in between these two extremes.
To help in the selection one needs to take into account how much stretch is really needed. After all, do you really need a mile of stretch? Probably not. The idea is to be able to pre-torque the bundle up to a certain level, then raise that level a bit more during the cocking phase. There is a subtle business going on here though that one should be aware of.
As the machine fires and the bundle relaxes, the available power will drop off. If it drops too much, there will be almost no power at all at the end of the TA rotation. Wimpy hurling results. On the other hand, if the power level stays high this can make for difficult torquing and cocking. Well nobody said this was easy, did they?
To keep from over-straining the material over time, let's assume we only load it to 90% of its working load rating. Further, we'll pre-torque the bundle to 85%. So, during cocking we're raising the load on the bundle by an additional 5%. But how far is that in stretch distance? For this information you'll have to refer to the manufacturers information.
To find out how far a strand stretches during cocking, draw a picture of a right triangle. Assuming your TA rotates 90 degrees, take the circumference of the bundle and divide by 4 (360/4 = 90), that distance becomes the short leg of the triangle. The distance from the center of the bundle out to one end becomes the second leg of the triangle. Now you only need to figure out the hypotenuse and that will give the additional length created by cocking the TA. A squared + B squared = C squared. Pythagorean to the rescue!

Example: If the half-bundle length is 48" and the bundle circumference is 48" then...

48 squared + (48/4) squared = C squared
2304 + 144 = C squared
sqrt 2448 = C
C = 49.48

This means that cocking the TA 90 degrees stretches each half-strand by almost 1-1/2" or 3" for the full length. Stretching an 8 foot long piece of rope by 3" is tough enough but remember that is for just one strand, there are multiple strands in the bundle and they have to be all stretched at the same time. The effort needed for one strand isn't multiplied by the number of strands though, not in a classic torsion bundle. The inner strands of the bundle have a smaller circumference, so do not have to stretch as far. Still, even 'whimpy' materials may require tons of torque to cock down the TA or even to just pre-torque the bundle, depending on the overall size of the machine.


Q: Would there be a significant advantage to torquing the bundle as opposed to cocking the TA, for firing?

A: Possibly. It would allow for setting the trigger and any safeties with less stress on the TA, making it a bit safer to work around. Beyond that, I don't know.


Q: For a given strand material and construction type, what is the best radial distance for it to be at, from the center of rotation of the TA?

A: For this answer, go back up to the material selection question. Recall that the elongation rate, bundle diameter and cocking rotation distance (angle) are all related to the power available. So, for a given material there will be a optimum bundle radius at the TA.


Q: Should strands around the base of the TA be distributed in a fixed pattern or a determined non-equal pattern?

A: Yes, to either a fixed pattern or non-equal pattern. This goes back to the observation that the strands in a bundle move under use or even just to the great tension they are put under. To get maximum performance from each strand they need to be positioned in a specific path and then tensioned appropriatly. I will grant though, the tremendous power available from even a poorly made torsion bundle is often so great, no 'fine tuning' of the bundle is really needed.


As you may have noted, many of the Q&A's are closely interrelated. This is no surprise as most dynamic systems are effected, as a whole, by any change in a discrete element. Another reason may be due to my own opinions, but I'm not going to worry about that. They are mine after all...
In the end, I think that good overall performance can be had without a lot of new design work or in-depth fine tuning. For exceptional performance, one needs to dig in and find those new designs. There are usually only 4 basic items that cause bad performance from a torsion bundle powered machine.

1. Bundle failure due to breakage.
2. Frame failure.
3. Throwing Arm failure.
4. Bad tuning.

(No, I didn't forget Sling failures. Not all torsion machines use them.)

If the bundle material breaks before you reach your intended power level, add more strands.
If the frame fails, build a stronger frame and/or joints.
If the TA fails, build a stronger one.

Okay, pretty obvious answers but they are simple and don't require an entirely new design.

I'll try to get to as many of these as I can...
>What if the the strands of the torsion bundle were made of multiple ropes instead of just one?
It would have the effect of using a strand the same size as the "larger" one. It would reduce the power curve of the skien because of the increased number of curves that the individual strands have to pull around. More thinner strands seem to do better.
>What if the outer ends of each strand were individually tensioned?
This was reportedly done in period, wedges were used to individually tighten each strand as it was wound into the skien. But these were also skiens made from animal sinew, so I don't know if that makes a difference. I've never done this, and don't see the reason for it, as it makes the skien too tight to put the proper number of turns on it.
>What if the strands, rather than meeting at a common point at the outer ends of the bundle, were splayed outward away from each other, some distance?
Like an funnel shape? I don't get this and it sounds too complex for garage engineers.
>What if a new material were used for the bundle, such as piano wire?
One problem is that the materials need to have a "proper" amount of stretch to work. I experimented with plumber's tape and found that once all tightened up for firing, it had too short a power curve and you couldn't cock the arm all the way back. I suspect something like that would happen with metal wire, but it may be worth trying.
>What is the outer ends of the torsion bundle were used not only for the main torquing of the bundle but also for the cocking?
Depends on the skien material. For the nylons, you need to keep a certain amount of tension on it. When ever I get my onager out, the first few shots right after tensioning are basically "warm up" shots and not full power. After a few warm ups, I add a couple clicks and I'm ready to go. Tensioning as cocking, to me anyway, would seem to repeat this process, but at a time when you need full power on the first shot.
>What if each strand of the bundle were of a different length?
They already are. The strands are looped over a rod, as each layer is added, the outer layers are slightly longer than the inner ones.
>Can the center of the torsion bundle be replaced with a support member for the frame?
I can only see this working if you put a hole through the base of the TA. And holes in the TA are a bad thing, not to mention you are reducing the number of strands inthe bundle and wrapping said reinforcing bar in the skien and gripping it. If you encase the skien, like with a pipe, you have to cut so much away to allow for the TA movement, that I don't think it remains feasable
>Can the strands be held in a pre-determined pattern during all phases of set-up and operation?
I use a jig to pre-thread my skiens, so I suppose any pattern you can think of as you're winding can be implemented. But for what good? I think you're better off with standard side by side coiling.
>Can the bundle be designed in such a manner as to allow it to maintain the bulk of its tension yet be separated from the frame as needed?
Maintain tension? No. Maintain shape? Yes. I use zip ties to hold the skien together whenever I remove it from the frame. If I had another skien and needed to chang it out, I could do so in a matter of muinutes.
>Does the Throwing Arm need to be a solid piece at the region of the bundle?
Given the materials we have to work with, I'd say yes. I suppose if you can figure out how to hot roll steel into a pipe that is tapered, it could reasonably be expected to withstand the forces at work. But that is well beyond what most of us are capable of.
>Can the Throwing Arm be forked at the bottom and, if so, is there any advantage?
I've only seen one forked TA in a History channel animation on a treb. And I have no idea where they got that idea from. It might work on a treb, but I'm thinking not on a torsion device.
>Can the frame of the machine render a design element that contributes to the tension of the bundle, rather than just supporting that tension?
The frame holds the tension in place by providing a way to mount a right angle of friction to stop the skien from unwinding. You'll need to elaborate on this a little.
>Can multiple torsion bundles allow for greater power and/or better design configurations?
Possibly, but I don't know quite how.
>What other methods are available for torquing the bundle?
Besides a long stick? You already mention a worm drive, which would be my next guess.
>What torsion bundle materials are best and what are the design criteria for selecting them?
For me its Mil-spec parachute cord. Its a finely woven nylon outer shell with a core of 7 straight strands. Allways worked for me with no failures.
>Would there be a significant advantage to torquing the bundle as opposed to cocking the TA, for firing?
See the "outer ends of the torsion bundle" question above.
>For a given strand material and construction type, what is the best radial distance for it to be at, from the center of rotation of the TA?
I must have missed that day in geometry class, can you elaborate?
>Should strands around the base of the TA be distributed in a fixed pattern or a determined non-equal pattern?
Since I use a jig and "reel" the cord onto it in a flat consecutive manner, then they are in a fixed pattern when installed. They may move around some, sometimes one or two of the outer strands may come off the bottom of the TA, but for the most part, once tension is applied, they don't migrate within the skien itself.

It is little surprise that different Hurlers will have different opinions. We all look at these things from different perspectives and backgrounds, as well as experience with torsion bundles. In some cases, we'll have differing thoughts that seem incompatible, but unless blood starts flowing, I consider this a good thing.

>What if the the strands of the torsion bundle were made of multiple ropes instead of just one?
B: It would have the effect of using a strand the same size as the "larger" one. It would reduce the power curve of the skien because of the increased number of curves that the individual strands have to pull around. More thinner strands seem to do better.

R: I think we're looking at this one a little differently, Brian. I do agree that smaller strands, built up to achieve the same size bundle, do provide more power. I attribute this to the fact that the smaller strands are more efficient in filling in the available space for the bundle. I also think that larger strands make it more difficult to put all the fibers of the strand under the same tension, so the smaller ones are, again, more efficient.

>What if the outer ends of each strand were individually tensioned?
B: This was reportedly done in period, wedges were used to individually tighten each strand as it was wound into the skien. But these were also skiens made from animal sinew, so I don't know if that makes a difference. I've never done this, and don't see the reason for it, as it makes the skien too tight to put the proper number of turns on it.

R: "...proper number of turns..." Not sure how you mean that but, if the strands are tensioned as the bundle is being made, then the compressive action should provide more room for more strands.

>What if the strands, rather than meeting at a common point at the outer ends of the bundle, were splayed outward away from each other, some distance?
B: Like an funnel shape? I don't get this and it sounds too complex for garage engineers.

R: Yes, like a funnel shape. Beyond that...who knows?

>What if a new material were used for the bundle, such as piano wire?
B: One problem is that the materials need to have a "proper" amount of stretch to work. I experimented with plumber's tape and found that once all tightened up for firing, it had too short a power curve and you couldn't cock the arm all the way back. I suspect something like that would happen with metal wire, but it may be worth trying.

R: Any material will have some stretch to it but I agree with you, finding the one with the correct amount of stretch to it is very important.

>What if each strand of the bundle were of a different length?
B: They already are. The strands are looped over a rod, as each layer is added, the outer layers are slightly longer than the inner ones.

R: True, for many bundles but not all. Especially for some of the new designs coming along (or a couple of my older ones.)

>Can the center of the torsion bundle be replaced with a support member for the frame?
B: I can only see this working if you put a hole through the base of the TA. And holes in the TA are a bad thing, not to mention you are reducing the number of strands inthe bundle and wrapping said reinforcing bar in the skien and gripping it. If you encase the skien, like with a pipe, you have to cut so much away to allow for the TA movement, that I don't think it remains feasable

R: Ah, but it has already been done and to good effect. Such a machine was at the WCPC event, I just can't find a picture or name for it now. (Anybody else find it?)

>Can the frame of the machine render a design element that contributes to the tension of the bundle, rather than just supporting that tension?
B: The frame holds the tension in place by providing a way to mount a right angle of friction to stop the skien from unwinding. You'll need to elaborate on this a little.

R: Perhaps my long 'answers' in my previous post, above, will clarify.

>For a given strand material and construction type, what is the best radial distance for it to be at, from the center of rotation of the TA?
B: I must have missed that day in geometry class, can you elaborate?

R: From the center of rotation of the TA to the surface of the bundle is some short length. It is this short length that is the lever arm the bundle acts upon to rotate the TA. For a given strand material, including its elongation rate, there will be an optimum distance for that lever arm. In most cases, we don't consider this fact though and approach the problem of best performance in other ways, which seems to work out pretty good. Still, if one wanted to maximize a design, then that lever arm needs to be considered along with all the other variables. I don't see this happening though on classic torsion bundles but new designs will need new approaches to some old problems.

OK, how about rethinking the epizygis? It could be in the form of a thick plate with individual holes for each wrap of rope (giving the opportunity for differing stand lengths) and located inside the frame. The advantages would be that you wouldn't need holes in the frame members for the bundle to pass through. The epizyges could be held with brackets from the outside, and the bundle could be demounted from the frame for transport, eliminating the need to re-thread the thing when you arrive on the site. In a big machine that's very time-consuming. You'd apply a turn or two to the bundle before mounting, just so it would hold the arm snugly, and tension it with screw pressure at the bracket. The details would need to be worked out but that's true for any design.

I'd put up a picture but my etch-a-sketch skills are lacking.

Oh, I think Chucky II proved that long bundles are the way to go. The cocked and at-rest tension is more equal, so if it's maxed when cocked there will be more available when the sling releases, as opposed to a short bundle where you lose a bigger percentage of tension where sling resistance is highest. That's just the nature of torsion springs.

Thomas

i was thinking abouth a rather narrow, rectagular frame, with the bundle wrapped around it, so it would be very easy to move around and the bundle could be very long.. with vertical struts deviding the bundle into two strands of ropes, one could easely use front and rear side of the bundle to put in two short arms powering a long TA placed higher and central on the frame.... been sitting on that idea for a while, keeping it warm untill i get time to work some of it out...

Thomas, I agree that a long torsion bundle will have a lower drop-off of power during TA rotation. How much/little drop-off is acceptable is up to the designer. Another factor is the selection of strand material and construction of the bundle.

"One of these days" I'm going to make a stab at figuring out how much of a difference there is between a classic torsion bundle with standard materials and a 'high-tech' bundle with some of the cooler materials. There are a number of variables but not too many, yet I suspect that some design/material combination's will yield some rather surprising results, not just in performance but limitations.

Warande, sounds like your talking of a hybrid machine. These are always interesting, as they present new problems to solve and the potential for something to catch all our attention, such as F2K, KA, etc.
I just hope you plan for a long base to stabilize that TA, eh?

Someone mentioned using torsion bars from an auto suspension. I think VW Beetles use these. I'd be interested in hearing the results of an experiment testing one of these. Also, what about kevlar? It can (in theory) be bought as a thread, hench very thin strands, and its stronger than steel. It may be cost prohibitive, but again, I'd be interested in hearing results from this, too.

> New discussion springing from Johan's comment that the Euro-Chunk will have a few new torsion machines competing this year.
> (see: LINK
>
Thanks for weighing in Capt Bob.
I know that Onager can land a shot in Europe, I don't see anyone else vieing for trans-Atlantic hurling.

You reminded me that I see Chucky and Roman Revenge growing larger every year yet I see only minor size change from Onager.
Why is that?
Am I missing something?

Capt Bob, you present an interesting question.
It would be fairly simple to answer in a 'no air-drag' environment, alas, we must work in our energy sucking atmosphere.

Mmm... If we allow that most payloads are going to have very similar drag coefficients, can we simply ignore them and carry on from there? Possibly, at least at the lower velocities, but once the velocities get above a certain point, then that drag coefficient changes things considerably since it goes up by the square of the velocity. I know that Team Carbo 'suffers' from this condition already and probably have since their very beginning.

Can't think today. This is a do-able problem though and perhaps one of the more educated types around here can solve it. I would think that once the base equation is set, a simple spread sheet should show all the answers.

Now where'd I leave my abacus?

Threading the Cat.
I don't know nuthin' but:

Could the length of the bundle be calculated and converted to circumference and the bundle wound externally from the machine and inserted as a whole wrap through the epi-thingie?

If the through-port for the bundle passing to the outside of the frame is large-throated, then why couldn't it work?

Or could the two sides be pre-threaded before stretching them apart to mount on the frame.

It also seems to me that a simple X-box framing could help make the frame stronger.

I don't know anything anout torsion engines yet. I'm just going from my gut feelings. So I know there areprobably lots of reasons why any on my ideas here can not be used.
What I don't accept is that no one does it because no one has done it yet. I don't want to try something that has already been proven to be unusable.

A VW Bug's torsion arm is a little different but serves to demonstrate that there are possibilities out there.

At some point in the past, spring rods, wider seperation in the windings, etc has been discussed but never satisfactorily.
And what about replacing the arm with an endless wheel?

Walker

Here's one for y'all to ponder. (See attached)

"Forward" would be to the right, in the picture.
Red strands pull the upper portion of the arm (log?) to the right, while the green strands pull to the left, below the axis of TA rotation.

I've kept this drawing very simple and left out most of the usual components, which you can fill in as desired. Nothing is to scale and the strand positions can be moved as desired. Perhaps putting the anchored ends on a slant, closer together, whatever. Similar thinking about where the strands cross over the TA. There are a lot of options here, so don't be afraid of whittling away at this concept as you please.

My own basic thinking runs something like this though...

Strand lengths can be made to match the amount of elongation needed while keeping the same amount of force as the other strands. Keep in mind that the longer strands also have to stretch further, due to where they cross the TA.
The distance between the side-frames can be much narrower than is standard for a torsion bundle, making the overall size and weight of the machine smaller.
There are more options for the location of any cross-beams between the side-frames.
There are no large holes needed, anywhere! Be nice to keep that material at (mostly) full strength, for a change.
Each strand can be tensioned individually, although it is possible to tension them all at the same time with some extra mechanism.
Strand replacement, if needed, can be done without disturbing anything else, including the other strands.
Strands can be anchored to the inside of the side-frames but I'm thinking that having them go through and anchored to the outside would provide a stronger anchor for each but also allow more elbow room to tension them.
One could also provide multiple 'set points' on the TA (indicated in the graphic by the pegs set into the TA). Then one could move the strands to different points as indicated by a change in strand material, payload mass, etc.

Now let me offer another tensioning method, very similar to this one but...

Anchor each strand per usual and pre-tension it to some level. To put the strand under full tension, apply a force in the middle of the strand and oriented at right-angles to it. In other words, push or pull sideways on it. This can be done in a number of ways, details of which I leave to your imagination.

Agreed, not a torsion design. I stuck it on this thread though because...

um...

That "um..." got to bugging me and I started thinking about what the difference was between a torsion machine and a tension machine, at least in the context of Onagers.

The power of a bundle comes from the strands trying to rebound back to their original length. In a sling-shot, this movement accelerates the pouch along a linear path that nearly parallels that of the power-bands.
On an Onager though, the re-bound of the strands is at nearly right-angles to the motion of the TA.
Still, it is the re-bound of the strands, in either case, that supplies the power. So it would seem that the only difference is the angle at which the power is transferred through.

We've seen torsion bundles that had very little twist, making the angle of applied power pretty high. Other bundles have more twist to them and so reduce the angle that the strands have with respect to the plane of rotation of the TA.

I'm not saying that either approach is better than the other, not in this discussion at least, but it does present a possible design to further the discussion.
Imagine a bundle of sufficient diameter (large) but short in length with respect to the size of the TA. At least, in comparison to the typical bundles we're used to seeing. Now put a LOT of twist in it, so much that the strands follow a pattern that looks like a screw thread. This mean the strands approach the TA at an angle that is very close to the TA plane of rotation. Just like a sling-shot.

Is it no longer a torsion bundle but a tension bundle instead? If so, what angle defines the transition between the two types?

> That "um..." got to bugging me and I started thinking about what the difference was between a torsion machine and a tension machine, at least in the context of Onagers.
>

To me it looks like the same difference between a sailboat and the wind. A 30 knot wind can move a sailboat at 45 to 50 knots because it crosses the line of direct force.
Traction uses direct force, torsion uses redirected traction (the two ends pulling inward) as well as back-twisty-ness.

Walker

Sailing *sigh* I do miss it.

So if re-direction is the criteria between them, how much re-direction is required or will any amount do?

Of real interest is which system allows for more power?
Re-direction is very lossy but for a tension machine, the strands would have to be either (a) much longer and/or (b) much more elastic. If different materials are used, the comparison is not fair.

I'd be willing to bet that a modified Onager using tension (roughly as indicated in my graphic, above) could at least equal the performance of a classic Onager with torsion power, with both being of a near size to each other. My one stipulation would allow for differing strand materials.

if i look at your design, it looks like a torsion, just with a very short skein..a very short one..

Direct pull (tension) machines, such as Hypertension, have performed very well over the years. There have been some duds of course, which just shows engineering ability, not concept failure.
The basic difference between this 'tension onager' and a Hypertension type is the length of the tension devices, coil springs versus strands of rope. One could also point to the direction the tension is acting in but I consider this a minor point as Hypertension, along with many others like it, mounted the coils such that they were non-parallel, I just take it to a more extreme case. (I don't require a physical axle either.)

It has been interesting to see, over the years, how so much discussion and even argument as taken place of defining what kind of machine "it" is.

I can't believe I missed this thread last time I logged in earlier this summer. Lots of exciting stuff.

I'd just like to say that *I* know what the future of torsion will be like, and I'm quite sure it is all about the pre-tension. It is actually commonly known among those who build historical recreations from the ancient texts and was discussed above several times. Naturally, I had never heard of it until recently. They used it on building the impossible many years back, but I ignored/missed it. It is described on Darius' web site, but mysteriously I missed that too. I did catch it when I was hanging out with some Roman recreationists at the Higgins armory in MA, and what a pain in the ass it was to implement on Mr. B last year. Guess what though. We didn't break our arms and we threw farther than before, and I attribute both to this technique. (Of course, extra arm-stops helped too.) Naturally I shared this technique at length with Bob at the last chunk because I like it when Bob Carbo beats Chucky.

I made a long write-up on how we did it with Mr. B on siege-engine.com last year, and why I think it is so important. Pretensioning was a major pain in the ass. With 2 bundles on a ballista, and new home-made equipment to do it, pretensioning required 24 hours to implement. Due to how much better it is, we intend to continue the trend in the future.

On the flip side, I think Ethos's hollow bundled machine is exceptionally clever, and he uses (or intends to use) the same technique. The extra bonus there is that his controlled strand bundle simplifies the device to the point where it is possible to model it properly. I don't know if you can model a traditional torsion bundle due to how dynamic it is.

Lastly, I have to agree with Kevin (naturally, since he's on my team) that as far as harnessing torsion, the hatra/inswinger is the way to go on the ballista front. I never got my onager models to produce numbers I trusted in order to allow a direct comparison.

Bonus link: Mista Ballista helps demolish a building