« Previous Page

Here are some photos of the current state of this project.  Clearly the spring bundles and limbs are just rough mock ups.   The various bolts and screws will all be replaced with more authentic looking rivet heads, while the toothed sideplates will be thinned down to give them a forged appearance.

The tangs on the arched strut that fit into the top loops of the field frames are not yet completed.  I have used a slightly different lug system than is apparent on the originals for the purposes of full power testing.  It will be a simple matter to convert them later.    Apart from that, the field frames and arched strut are full scale copies of the originals.  Some may take exception to the metal bolt groove and sliding trigger mechanism that rides on it.  A wooden slider design for this part of the machine had been in use for over 800 years.  Apparently it had been causing headaches for that length of time too.    Seems like I recall reading somewhere that even the ancients complained the wooden slider was subject to warpage and damage,  while rattling around in its dovetailed slot couldn’t have been conducive to accuracy.  The 12 lb weight penalty incurred by including a metal bolt groove, seems a relatively small price to pay if it improves accuracy and reliability.  Did the Romans actually do this with the Orsova ballista?  Perhaps not.   On the other hand, if it gave the machine the ability to knock down an armoured chieftain at a couple of hundred yards, ending the battle before it even started,  maybe the idea has some merit.   As noted earlier, the purpose of this project is to maximize the performance possible from reconstructions of the original artifacts using plausible materials and techniques available to the Romans.  Nothing more is claimed.  The one exception to the authentic materials rule will be in the area of cordage.  Obtaining sufficient sinew and the learning curve needed to make it into rope of some kind, is beyond the scope of this project at present.  Like most reconstructions of torsion engines, modern alternatives will have to be utilized.   The Gallwey ballista used nylon for its spring bundles.  As a baseline of comparison between the two machines it seems reasonable to use nylon again.  I have a source for small quantities of sinew maybe 12″ long.  Perhaps a small model or test bed could be devised to make a relative comparison between the propulsive qualities of nylon and sinew using those 12″ lengths.  From that it might be possible to extrapolate the performance of the original machine.  All that will be down the road a bit.

The holes in these field frames will allow a spring bundle of 3 inches.  Much of the power developed in a torsion engine is dependent on how much pre-strain can be put into the spring bundle before the limbs even start to retract.   When the limbs on the Gallwey reconstruction are “at rest” against the frame, the amount of torque pre-loaded into its 5 inch springs is sufficient to defeat the attempts of a couple of strong men to push the limb tips back even slightly.  If a similar pre-load can be brought to bear on the 3 inch springs of the Orsova ballista,  with the extra limb rotation of an inswinger, it will no doubt zip along nicely.

As I write this the Orsova ballista project is perhaps two thirds of the way down the path to firing its first shot.  Things have progressed a fair bit from the photo shown in yesterday’s entry, it was taken last March.   I will present some new pics soon so that these posts can catch up with the present state of progress, and then this blog will be performing its true function,  that of a journal.   I was asked when I started this project why I didn’t make some kind of proof of concept model before investing all of this work in it.  The answer to that lies in the fact that we are presented with a wealth of information from the artifacts themselves.  It became clear early on that by simply reproducing the artifacts we would have a viable starting point.   As that part of the project was fixed, no further decisions about the final design needed to be made until the artifacts were reproduced.   By the Spring of 2008  the field frames were finished, exact duplicates of the originals shown in the technical sketches seen below.  These drawings are my sole source to dimension the artifacts and I have adhered to them closely.   They come from, Recent Finds in Ancient Artillery, by Prof. Dietwulf Baatz.

I am not going to claim that my reconstruction will look exactly like the original Orsova ballista.   Clearly the field frames and arched strut shown in the drawings will,  but because all the other parts are missing no one can say for sure what the entire machine looked like.  That being said,  I came to the conclusion early on that there were really only two ways to design this thing.  I could make a machine that appeared consistent with what people expected a late model Roman ballista to look like,  (witness the illustration just visible in the museum exhibit from yesterday’s post) or I could let my imagination roam over the artifacts themselves and see what was suggested by the mechanical logic inherent in them.   Having gazed at pictures of the artifacts for many hours I became convinced that the Orsova ballista was a high technological achievement for the Romans.   The artifacts had all the elements of a flexible modular design that could have great tactical advantages in the kind of fluid warfare that started to emerge at the end of the Roman empire.  The connecting loops on the field frames and the tangs on the arched strut suggested a system that could be fitted together with wedges and then knocked apart quickly for maintenance or transport.   A ballista is a lot of kit to lug around, I find it impossible to believe that the Romans would settle for second best on anything when it came to such a super weapon.  If a modern researcher could dream up the innovations needed to maximize accuracy, power and reliability using these artifacts, it is probably a safe bet to say the resourceful Romans did too.  Occam’s razor tells us that less is more if it is consistent with the facts.  Therefore,  so long as the design of this reconstruction was in keeping with Roman blacksmithing, woodworking and cordage techniques,  I felt somewhat free in how it should look and operate if it could be shown to give superior performance.

My experience generating power and accuracy with my first ballista told me a lot about how to approach this one.  For example,  it is generally understood that vibration is the enemy of accuracy.  Most interpretations of late model Roman ballistas do not show any angled support struts bracing the field frames back to the stock.    “And why not?”, we might ask.  They will shoot okay without them, is the obvious answer.  Very likely they would, but shooting “okay” is not what I am after and I suspect not what the ancients were after either.   My experienced with the Gallwey reconstruction indicated the great importance of rigidity in the relationship between the stock and the box holding the spring bundles.  When I introduced struts to that machine, group size at 50 yards shrank from a foot or so down to the 3″ mentioned earlier.   I believe it is a mistake to be too slavish to the writings that have come down to us from the ancients.   Heron may not have given reference to struts in his technical description of the earlier cheiroballista, however from a mechanical standpoint they certainly seem worth considering.   The “absence of evidence is not evidence of absence” scenario seems to be in full effect here.  As such my reconstruction will experiment with them, suspecting they would be an improvement the ancient engineers would likely not have missed if they improved accuracy.   Again, I mention all of this just by way of example.  An indication of approach as it were.  My whole interest in making these engines might be likened to an extreme sport.  Accuracy and raw power on the shooting range are my main considerations in determining design, just so long as it is reliable and consistent with something Roman technology could have actually produced.  Given the ruthless logic of Roman weapon designs, I would suggest that this pragmatic approach might be closest to the truth.  And if it isn’t, perhaps there will be some shreds of insight available that might not have existed had a safer more scholarly path been followed.  Thanks to the brilliant work of  Michael Lewis and Aitor Iriarte,  a new interpretation of how the Orsova machine may have worked surfaced at just about the same time I was handling my newly made field frames.  Their notion of an inswinger ballista was a revelation that put the Orsova artifacts into sharp focus.  It tied in perfectly with my obsession over accuracy, power and the radical engineering potential of ancient artificers.  The mystery was starting to deepen with the realization that with the inswinger, the Romans had apparently developed a new and powerful way of utilizing the twisted bundles of sinew that powered their ballistas.  Ironically their burst of innovation happened at a time when the empire itself was destined to disappear.

Very cool rendition of an inswinger by Dominic Andrews

And back to the Gallwey for a moment.

Shooting last summer, two flyers and four shots in a nine inch group on the 85 yard range.   Less than stellar compared to “back in the day”, ten years ago.   Note the long cocking lever.  Half power shooting, so things are relatively safe.

The extreme sport aspect of high powered ballisata shooting.

A 3″ x 4″ ash limb fractured after 22 shots at full power.  Heavier 3 3/4″ X 4 1/4″ limbs (fitted out with epoxy soaked kevlar wrappings for safety) solved the problem.  The limb let out a painful crack, and then collapsed against the stock with an enormous thwack.   No splinters or other flying anomalies were detected.  Wife was less than impressed when I relayed the details.

Here are the original artifacts of the Orsova ballista unearthed in the 1970’s.

And here is my partial reconstruction of them done in the spring of 2008.

In the museum display the left field frame appears to be an  exact reproduction of the original field frame shown on the right, rather than a mirror image of it.  My version shows how the original field frames would most likely have been positioned relative to one another.

Frowny face means the wheels are turning.

Before starting to record the progess made so far on the reconstruction of the Orsova ballista, it might be useful to show my first ballista built back in 1997.  It was inspired by this pen and ink drawing found in Sir Ralph Payne Gallwey’s book, The Crossbow. 

And here is what I came up with after nine months of furious gestation. 


Granted no ancient ballista probably ever looked like my reconstruction here.  I had seen the Gallwey pen and ink drawing in a dictionary when I was ten years old.  The desire to make one like it was baked into my subconscious at an early age.

Field testing of this first ballista continued from 1998 through 2001.  After that the whole thing was pretty much mothballed, until now.  At the time I put it away the machine had fired 358 shots and other than a broken limb (very exciting) showed no sign of damage. In retrospect I see that my documentation of this first round of testing never really did justice to the amount of labor involved performing it.  Hence the reason for this running blog on the Orsova project as it will ensure a meticulous record of all the myriad  details as they occur.  What records I have from ten years ago show that the muzzle velocity for this first machine averaged 310 feet per second when using 3/4 pound bolts.  That puts the muzzle energy at 1150 foot pounds, about the same as a high end load for a .44 magnum.  After the 5 ” diameter, 32″ long torsion springs were tuned, it would put ten shots into a 3 inch circle at 50 yards.  I never got around to making a perfectly matched set of bolts, so this level of performance was only achieved by shooting the same bolt over and over again.  However, considering the variation caused by having to resight for each shot,  (it proved impossible to cock the machine without disturbing its aim)  I was fairly ecstatic about its accuracy.  The 1″ diameter steel tip used on the bolts for accuracy testing had a flat end to help reduce penetration and thus ease removal from the plywood backstop.  Even at that, the 1″ flat tips would punch a hole through four sheets of 3/4″ exterior plywood and leave a healthy dent in the fifth.  Sharp tipped points would make it through six or seven sheets.  Maximum range (shot at a 45 degree angle) with the 3/4 pound bolt was 480 yards.  Lighter, 1/2 pound bolts made it out to 540 yards,  while 1/4 pound flight style bolts averaged 760 yards.  The maximum pull weight at full draw was 4012 lbs.  The following graph was generated from direct measurements using a 15,000 lb. dynanometer.  Thanks go to my good friend Brian Kern for drawing it up.

It will be interesting to see how a similar draw force curve will look on the Orsova project.  The more so as this new design will operate on a radically different principle than the conventional ballista.  More on this tomorrow.  For now here are a few more snapshots from a decade ago.




Daughter Sarah self-decorated with bolts.


That 760 yard shot.  Bolt just visible a few feet from muzzle.


Scoping out the range


Rear sight detail


Front sight detail


Trigger mechanism




Tightening gear

A barbarian’s worst nightmare



Photos by Prof. Dietwulf Baatz.


Sometimes it all starts with simple photographs.

It was the winter of 2007 when I first laid eyes on these pictures of some artifacts excavated from the Roman fort in Orsova, Romania.  For a reconstructor of ancient weapons, they suggested a design of deep and compelling mystery.  The top photo is the one remaining field frame from a set of two.   They belong to a type of torsion catapult used extensively in the ancient world as an anti-personnel weapon.  The bottom photo is the arched strut that held the two field frames in place.  No other parts of this machine have been unearthed. Dated to the mid fourth century, these artifacts represent what appears to be the final design incarnation of an iron framed ballista from the late Roman Empire.  Caught at a time in history when the technical expertise of the Roman empire was destined to evaporate, this radical improvement over archaic Greek designs was based more on blacksmithing than it was carpentry.   One look at these photographs and I knew I had to attempt a full scale reconstruction of the machine.  This blog will be a record of the progress made on the project.

« Previous Page