« Previous Page

These beautiful CAD images I have brazenly copied from an excellent paper by Ildar Kayumov and Alexander Minchev.

arch strut 2

Ildar 1


Here is their paper and what it has to say about some interesting “new” ballista finds:    23_Kayumov_Minchev_ROMEC

By my reckoning, that short little winch handle would need to be connected to some fancy internal gearing to ever generate enough pull to be useful.  (“…to match the enormity of the task”, and all that.)  I like the idea of it though. Phoenix certainly has her ears perked in that direction when it comes to winches.

However, of greater interest to me here is the interpretation given the angled struts and their attachment points on the Kamarion.  Given the bending issues we have experienced in the past with the arch in Firefly’s Kamarion wanting to collapse backwards towards the winch,  Mr. Kayumov’s plan makes a lot of sense.

His use of simple rivets to attach the angled struts to the Kamarion may at first glance appear to be the simplest and most straight forward way to anchor these parts together. However, I believe the Romans went a little deeper on this and sought to make the machine something that could be easily disassembled.  A clue to the existence of this more advanced, take-apart design, lies in the loops and tangs on the field frames and Kamarion of the Orsova finds.   Those features make manifest the ancient designers train of thought:  a modular machine that could be quickly knocked apart or put back together again. (Think: packing the ballista on mules, shipboard transport,  yomping the whole thing up to the top of that ridge, assembling it up in a tower  etc.)  Presumably wedges were used to complete the mechanical lock up of all the joints.*  If this hunch about a take-apart machine is right, then it becomes highly desirable to use something other than rivets to connect the angled strut to the Kamarion.  Hence the following:


The wedge has a shallow angle on it and, when driven into place, causes the yoke to grip the Kamarion with great force.  The locking pin is driven through the “sight” hole in the Kamarion and through a matching hole in the wedge.  This arrangement assures there is no lateral slippage of the yoke along the Kamarion, as well as locking the wedge in position so it can’t back out.

Firefly has successfully used this type of wedge and locking pin system to connect all the critical joints needed for her rapid disassembly. (That process takes me, on my own, about 20 minutes)  In all the shooting we have ever done, none of Firefly’s joints have ever loosened a jot.  Very stable she is.

Anyway,  that is my new, favorite explanation for those two holes in the Kamarion of the Orsova artifacts.

*Note, some scholars have suggested the wedges were made of hardwood.  My experience suggests that a metal wedge, driven into the joint and captured by some kind of retaining pin, is the ideal solution.  I fear wooden wedges would not hold up over time.  They would be smacked around by the vibration of a powerful machine and susceptible to all the usual vagaries of wood — splitting, shrinking etc.

Because Phoenix will be another “iron frame” ballista,  there is no reason not to favor all things ferrous if they offer an advantage.


While I very much admire the artwork in those preceding Cad drawings of an Iron frame ballista, there are a few points in the design that I need to scrutinize before applying to Phoenix. I will list these areas below, and hope the authors of that fine paper, and anyone else for that matter, will understand where I’m coming from.  Just filling in all the minutiae before taking those first pecks at the Phoenix project.

(1) I have already noted that the winch handle that is shown, accurate representation of one of the Elenovo artifacts that it is, would not have had nearly enough leverage with a winch drum of the apparent diameter shown. Firefly’s levers for her hand winch are a whopping 5 1/2 feet long and her drums are 3 inches in diameter.   There is an extra set of pulleys to half the load,  and even with all that it takes close to 70 lbs. of down pressure on the end of the levers to drive that drum as it nears the full draw weight of 5,000 lbs. This is not to say the original crank artifact that was dug up is not long enough to draw back a powerful ballista, it is just that such a winch handle would need much more mechanical advantage than shown (i.e. a gear box, or as the authors of the paper suggest, a bloody big cheater bar for that vertical handle), otherwise it would be little more than a toy in terms of power.

Medieval crossbows and their block and tackle cocking devices, with multiple pulleys and several falls of rope, had very small diameter drums of about 1″ to give them lots of power, and yet their winch handles appear to be about the same length as the one shown in the artwork. What were the winches of those medieval crossbows rated for in terms of pulling power? not much more than a puny 1500 lbs or so, even for the bigger ones.   (……… Which is only relevant because that ain’t the kind of ballistic territory I’m figuring on for the Phoenix.)

(2) The end caps of the field frames on the original artifacts show 4 through holes that hold the locking pins for the bronze washers. It is very important to have enough holes in the washers to allow small adjustments in their rotation. Firefly’s rotational discrimination in her washers can be advanced in 7 1/2 degree increments. These incremental adjustments are the only ones available for tuning the thrust of the two bundles after the ballista is assembled and in use.  Also of note are the catapult finds from Ampurias that demonstrate this 7 1/2 degrees of adjustment.  Some folks have mentioned that the tuning of the bundles is done with linear stretching, and it is true that when the spring is fabricated by winding it around the crossbars it is best to tension the bundles as equally as possible. But in the real world, things get out of wack no matter how carefully the springs are balanced by the linear stretching that is applied when they are made.   (Yes, I know,  “wedge” machines can offer some adjustments after assembly,  but in my experience it is too rough and imprecise a way of performing delicate negotiations between a pair of ornery springs.  In fairness, though, more experiments with differently angled wedges are needed to make that statement absolutely.)  What is clear is that during the lifetime of the springs the need for fine adjustment of the individual spring power is essential to keep twin armed ballistas performing properly. Again, I respectfully remind anyone who might actually be reading any of this, these things I speak of are not just armchair theories. Firefly has been there and gone through all of this.   For a scheme based on rotational adjustments,  7 1/2 degrees or finer seems about right to achieve balance in the springs.  Very important if you ever want to tame those waggle tails!

(3)   This next one is a bit tricky.  I say that because Firefly has only ever tested one particular length of spring.  With longer springs perhaps she would do a better job of using the 120 degrees (or whatever it is) of extra limb rotation depicted in the artwork.  The longest limb rotation I ever tried with Firefly was 110 degrees and the performance was fairly abysmal.  The formula that I hit upon for getting power out of Firefly was pretty simple and it favored much less limb rotation.  The formula goes like this: if the tension on the bowstring is really taut in the at rest position, power is excellent.  To have this extra tautness in the at rest position it is necessary to have the bundles cranked up very high with rotational pre-tension.   With Firefly’s length of spring, this tautness can only be achieved with limb rotations of about 45 degrees.  That is because when the rotaional pre-tension is up this high, the draw weight starts to stack heavily at around 45 degrees. To put it plainly, she ain’t a going no further.  Big limb rotations and long draw lengths are kind of  a non-starter for Firefly, fortunately however,  that does not necessarily matter when it comes to generating velocity with inswingers.  Our experiments have  shown that it is that final snap of the bowstring as it closes the power stroke that is responsible for a lot of the projectile’s final velocity.  To put it another way: loose, at-rest string with Firefly using 110 degrees of limb rotation  and her shots were only puddling along at a meager 225 fps.  Very taut, at-rest bowstring, with a short 45 degrees of limb rotation, and she was hammering the clouds at 360 fps and more.

………So anyway, our current prejudice on this matter suggests that really fast inswingers  do better with much less limb rotation than is shown in the paper’s artwork.  Deep down though, I intend to remain agnostic on the matter because the case for greater limb rotation needs more exploration than I’ve given it so far.  Go figure….

Back in 2009, April 6, I posted this bit of frippery.


Of course it ain’t exactly frivolous.  Those are actual measurements taken from Firefly as she cranks her arms back and forth.  It’s pretty easy to intuit the speed increase if you ponder the relationship between the draw lengths and the limb angles a bit.  The string and limb basically move in unison, therefore the string has to cover the same amount of distance it did in the last half of the draw as it did in the first half, only it has to do it in the briefer interval of time afforded by the smaller 31 degrees of limb rotation.  With an inswinger, that is where the short rotation zip comes from.  Just that simple.

And now for that moment of snapola:

Click for vid   20121107141039(2)       Bolt is a Dura Europos style, 340 grams,  359 feet per second, 1500 foot pounds of muzzle energy. Maximum range is 900 yards.

Here is a video of our farthest shot yet, 998 yards.  Click for Vid:   20121107135533(1)   Projectile is a 276 gram Dura bolt moving at 375 fps.



(5)  If this laundry list were to turn itself into an apology for shorter draw lengths, I would remind the reader of the monumental advantages of high power in a small package.  Very Tacticool, as they say these days.   Phoenix will be at least 2 feet shorter than Firefly, and have greater relative power due to her improved structural strength and ability to handle higher draw weights.   The advantages of shorter machines in towers or on shipboard or wherever, would be a welcome advantage derived from the inswinger’s natural form of compound leverage. (See above envelope.)

(6)  I really don’t know where this goofy list is going anymore, so let me just say,  many thanks to Mr. Kayumov and Mr. Minchev for being such unwittingly good sports and appearing in this posting.   You are helping Phoenix to get her gaps filled in. Cheers!


I seemed to be a bit stalled on this sinew vs. nylon thing.  I think it is suffering from estivation.

As games go, this one makes a handy executive decision-maker.

Click for vid: 00033    or    00020   or   00024

Two say yes, and one a maybe.  Not exactly rock solid divination. Perhaps I need to frame questions about Phoenix a little differently.

Beats throwing darts at a dartboard though.

For a long time now I have wanted to do some serious accuracy testing at 100, 200 and 300 yards.   Unfortunately, the place I found for today’s shoot had a backstop loaded with rocks that were hidden in the sand. My bolts and that hillside really did not get along. The hardened steel tips are undamaged by their ordeal, but the bamwood shafts shattered in many places.


The style of shooting I used in today’s jaunt could be best described as standing offhand.   There is no rear support for the machine other than my shoulder.  Where my shoulder goes, the machine points. Here is a video where you can see the fluid nature of Firefly’s aiming potential.  (The weight bearing down on my shoulder is about 20 lbs.) Click for vid:  00014(2)

And here are today’s shots in order of appearance:

(1) The target at 150 yards is a piece of 1/2″ plywood 24 inches square and painted white. The first shot was made with Firefly’s peep sight set at something that was just a wild guess.  At 150 yards the first shot hit ten feet low and to the right three feet.  Richard reckoned the flight time at 1 1/2 seconds.  My mental ticker estimated 2 seconds.  (Must bring a stopwatch next time.)  Muzzle velocity with 520 gram bolt was 305 fps.  (I guess that makes Richard’s brain clock, spot-on)

(2) A sight adjustment had the second shot strike just beneath the bottom edge of the target, and a few inches from it’s right hand edge.   (This is the shot seen in the above video.)

(3) Another sight adjustment and the third shot hit one foot low and six inches to the left of the plywood square.

(4) After another sight adjustment we hit a couple of inches away from the left hand edge of the plywood and in line with it’s lower edge.

(5) Yet another sight adjustment and, bingo!  Click for vid:  00017(1)(1)

(6) With the sixth and final shot I omitted making a change in the sight position and decided to use a little Kentucky windage to try and center the shot better.   It went high and right by a foot.  I figured my allowance had been a tad enthusiastic.  In any event, bolts were shattering on that hillside at an alarming rate, so I called off the proceedings.

Even with all the sight changes that were made, it was possible to see the shots from the machine respond fairly precisely to the adjustments I was putting into the rear sight.  My sense is that she will shoot into a one foot circle at that range.  But that, of course, is something we’ll have to prove with more rigorous testing.

Clearly, a better hillside is in order.

And, finally, Oona shows her stuff when it comes to bolt retrieval.  Click vor vid:  00015(1)

oona 1


« Previous Page