Non-obvious solution
So, I was able to figure out the problem with bolts flying with the butt-end pointing to the right. The first thing I tried, and which did not help,was leveling the uneven and protruding surface of the top of the crosspiece. I did this because I noticed that the T-clamps were slightly off at one side, and then noticed that the field frame was attached to the case in a slight angle which make the bowstring rise up from the slider at the right. The crosspiece in its original shape:
And after fixing it:
Now the little ladder beams, not the crosspiece, rest against the case. This corrected the angle somewhat, but not fully.
Now, what did help was non-obvious: rotating the washers more. It appears that with low pretension any differences in linear …
Still not shooting straight
In my previous blog post I triumphantly declared that I managed to fix the "bolts leave the ballista butt-end pointing to the right" problem. But that turned out not to be the case. So I spent several hours yesterday trying to figure out what was wrong:
- Rotated one of the torsion springs 180 degrees to ensure that both halves act the same. Did not help.
- Switched the arms to the other torsion springs. If one of the arms had had a problem this should have reversed the problem, i.e. bolts should have left the ballista butt-end pointing to the left. But no, butt end was still pointing right.
- Double the number of holes in the top washers to double the precision of washer rotation.
Before the draw the bowstring is exactly at the center of the slider…
Chasing red herrings
I've been working on the cheiroballistra again after a long pause. Why or how, you may ask? This is only possible because family is out of town and I'm on vacation. Basically all that was missing was a new pair of arms. The previous incarnation of arms was "elliptically tillered" whose lack of flexibility caused them to waste quite a bit of the energy (10-20%) fed to them. In addition the crushing of the springs had eventually destroyed all the previous versions of arms with which I was able to do full-power tests.
The new pair of arms, which is now in operation, attempts to fix all these shortcomings:
- On the belly side a strip of reindeer horn is sunk into the wood to prevent the wood from crushing
- On the back side the bar prevents crushing …
Print version of my Journal of the Society of Archer-Antiquaries article
My article is now available on GitHub (28MB):
- Seppänen, Samuli 2018 - The Cheiroballistra - a reconstruction - screen quality.pdf
I tried to upload it to academia.edu which gave a most unhelpful message "Something went wrong". I may need to contact their support to understand what the problem was. Meanwhile please use the GitHub link above.
My new cheiroballistra article
It has been a while since the latest blog post. Work has kept me fully occupied so there has been little progress since late last summer. However, my article in Journal of the Socienty of Archer-Antiquaries volume 61 (2018) was published recently, so I'll publish it here. Unlike (almost?) all ballista articles mine is focused on practical aspects of the reconstruction. Thanks to Rex Harpham for bugging me about writing an article to Journal of the Society of Archer-Antiquaries and to Arthur Credland for actually making it happen.
The quality of this "copy" is not particularly good, but I'll fix that in a couple of weeks when I get access to a scanner. But this version is at least readable.
Regarding location of the release mechanism
Today, while writing an article, I realized something. When the cheiroballistra manuscript discusses the release mechanism it does not actually state that the measurements should be taken from the end of the slider (i.e. beam ΓΔ). The assembly instructions start with installing the handle to the slider and continue by detailing how the claw is attached to the fork. Then 5 dactyls, ΔΟ, is measured on board ΓΔ (slider) and rectangular hole pierced through it. Then the tenon in the fork is lowered to Ο. Also, the distance between the axles of the handle and the trigger ("little dragon"), ΜΠ, is given as 4 dactyls. The distance from the handle to one of pittarion's feet (ΞΡ) is not given, but the reference point is still the handle axle.
The po…
More on crushing damage on the cones
As mentioned in the previous blog post the torsion springs slowly by surely crush the base of the cone. As mentioned there, Dr. Burkhard Meissner shared this finding with me in late 2016 in our email exchange, but back then I did not yet fully realize how grave the problem is. In fact, based on my testing a cone made from elm can take about hundred shots before the crushing weakens the cone enough to create a weak point and break it.
These two photos are from the still-intact elliptically tillered arm:
As can be seen, the crushing damage is considerable. The damage on the arm slowly builds up, and as the crushed part gets thinner it gets more and more stressed during pullback. The end result is this - the cones are the semi-elliptic ones tha…
Elastic versus stiff arms
Since the last blog posting I've had several shooting sessions. The first two related sessions from a couple of weeks ago were done with the so-called elliptically tillered arms I talked about earlier. During those sessions I shot the weapon for about one hundred times, counting all the control shots and unrecorded shots between sessions. Here are the results for each bolt weight:
I attempted to increase washer rotation to 150 degrees, but I was then unable to cock the weapon with stomach pressure alone. I then took a 6mm thick spring steel rod, rested it against the feet of the pittarion and used that to assist the draw. This enabled me to cock the weapon at 150 degree rotation. This increased the power somewhat, but on the second cocking …
Medium power shooting session
Today I finished a low- to medium-power shooting session. The idea was to gather a bit of data about
- Bolt weight on efficiency
- Effect of power stroke on energy output
I had two chronographing bolts available: one weighing 87.78 gram and the other 43.71 grams. Washer rotation was pretty close to 90 degrees when measured from case. When the arms rest against the curves they point about 20 degrees outwards. So real washer rotation was approximately 70 degrees.
The results of the session were not spectacular energy-vise, but that was not the point today:
This data alone tells us a couple of things. The relative efficiency of the lighter bolt (compared to the heavier bolt) drops as power stroke is increased:
It is also clear that increasing the power…
Correcting bolt flight
In my previous blog post I documented my worst misfire session ever. The fact is that those misfires took me by surprise: before I could pretty much count on the fact that if the arms rotate at the same pace, the bolt will fly straight. But that was with linearly tapering arms, which are way more easy to make in matching pairs than elliptically tillered ones.
As mentioned earlier, my plan was to add weight to the faster-moving arm to slow it down and thus correct the bolt flight. And indeed that worked. Below I will outline the process for the benefit of others - the same approach can be used for bows and crossbows as well. However, getting the bolt to leave the weapon at 45 degree angle is difficult without a ballista :).
Anyways, I ended u…
Tale of amusing misfires
Past weekend I finished the final remaining tweaks to the cheiroballistra. Compared to the last session I had made a couple of changes:
- Added a ~3 mm layer of wood on top of the slider (to ensure that bowstring does not strike the bolt too high)
- Fixed angles of the field frames (one was tilted backwards)
- Sank the claw into the slider
- Made the contact between the claw and trigger tighter
- Finished the new elliptically tillered arms with stronger hoops (increases strength of the cones significantly)
So, assuming that everything would "just work" I shot the cheiroballistra a couple of times with the goal of improving previous session's 124 joule record. Instead I had a good round of "most amusing" misfires. The first shot broke my trusty ~48 gram t…
Slow progress recently
Somebody on previous blog post's talk page was asking if the project is still alive. The answer is "yes", but life (work, family, renovating the apartment) has prevented me from working on anything in a long while. Finding time for writing a blog post has been challenging. Anyways, here's a brief update of what I probably have not blogged about yet.
During previous shooting sessions I experienced an unusually high rate of dry-fires. I believe the the problem was caused by one of the field frames being slightly tilted backwards. I corrected that by adjusting the wedges in the pi-brackets. This is actually a very good reason for having such a wedge system.
Another issue was (and still is) that the bowstring is slightly too long. Based on earli…
Purpose of the Pittarion
In the previous shooting session I used a forged quadrangular handle that was locked into a simple pin. What was missing then was the pittarion, a goal-post shaped component at the end of the slider, something that I had in my previous slider.
The lack of the pittarion did indeed make cocking the weapon significantly harder: the handle consistently jumped up to a 90 degree angle when it contacted the pin. This required me to clap down the handle at full draw, which made it more difficult to focus on feeding as much energy into the weapon as possible. So pittarion definitely did serve a useful purpose. So I decided to bring it back but with some improvements, all of which are described in the new Pittarion article. In short, the pittarion …
New record shots
Yesterday I finally had shooting session with the almost completely reconstructed cheiroballistra. Here's some background information
- Springs were tuned to ~350hz (up from ~320hz)
- The distance between torsion springs was correct
- The (nylon) bowstring was significantly shorter than previously
- Anchor points for the handle were at same places as in the previous case
- Cones were now correctly 11 dactyls long
- Draw length was reduced by ~3 dactyls (~6cm)
- Stomach pressure alone was used for cocking
- A new slightly longer quadrangular handle was used
- Draw length further reduced by ~1 dactyl (~2cm)
In previous cheiroballistra incarnation cones were incorrectly 14 dactyls long. This was only possible because the little arch and ladder were too long (see above).…
"Ballista ready"
In the words of the classic video game Stronghold I can finally proclaim "Ballista ready". Here it is from the top:
From front:
And detail from front at a slight angle:
I unfortunately did not have time to shoot it yet, but will in a couple of days. Even so, I've already learned a couple things.
First, current pretension level is definitely about as high as I can with nylon. Even at mere ~45 degree washer rotation rotating the arms is quite challenging. This is partly because cones are now of correct length, 11 dactyls, and partly because of higher pretension. This is both good and bad:
- The force-draw curve should be about as good as it can ever be. Torsion spring stack should minimal torsion spring stack.
- There's the risk that tuning the two to…
Plans for comparing nylon, sinew and horsehair
Once the nylon cheiroballistra is powerful enough - hopefully in the current configuration - I will move on to comparing authentic spring materials like sinew and horsehair with nylon. My plan is to have a "mini-onager", probably in 1:2 scale, except that it probably would not have a sling like the real one. The nylon version will act as a reference point: it will tuned to store (and release) the same amount of energy as the cheiroballistra in proportion, that is, (1/2)^3 * 1/2 = 0,0625 times that of the cheiroballistra. This translates to about 6-9 joules, assuming the cheiroballistra's energy output is 100-150 joules
After calibration the other spring cord materials will be tested to see if they can produce comparable energy levels (6-9 …
Little archy speedbumps
After yesterday's good day of work the cheiroballistra is now almost ready for "final" round of shots:
- Both torsion springs have been tuned to ~350hz. In practice the initial frequency of the cords was around 360-375hz. To prevent excessive damage to the cords they were not pulled through the jaws of the (unmodified) locking pliers using the winch. Instead the locking pliers were released after a reasonable, probably around 200hz+ tension was applied to the cord. On release of the pliers the cord snapped into its place. This may have had a minor effect on the amount of pretension in the previous cord - how much remains to be seen.
- The slider was scraped and ground slightly make it slide more easily in the female dovetail, assuming both parts are f…
Happy stretching
The cheiroballistra is now truly close to being "finished". The biggest job was making the new field frames which I touched upon in my previous blog post. The field frames are now ready:
Some additional work is required to adapt the field frames to the little ladder and little arch. I've also modified two of the old 14 dactyl cones to make new 11 dactyl cones; only assembly of the throwing arms is still work in progress. I had to replace one of the old washers as its welding seams looked suspicious; fortunately I had made a replacement the last time I visited the forge. The slider is a bit too thick right now - even when it and the case have been stored in a dry place the fit is a bit too tight. A tiny amount of planing and sanding should s…
Various failed experiments
During my work on the cheiroballistra I've ran into a bunch of manufacturing techniques and technical solutions that seemed reasonable at the time, but proved to not work properly in the end. This blog post outlines some of them, and why they failed.
At one point I thought about making the tenons of the little ladder beams from a separate pieces of steel:
There is no particular reason why this would not work. It just turned out that when the little ladder is made from thick enough (~6mm) steel there's no reason not to just forge the tenons from the beams themselves. It the beams were made from thinner steel then separate tenons would make more sense.
I've also tested various ways to make the field frame rings. Two such experiments, which fail…
Case and slider finished, progress with field frames
Yesterday I made good progress. The case is now finished and the slider was finished earlier:
I've reused the old crescent-shaped piece for pragmatic reasons, even though it is not 100% match with the text as it's attached with screws, not with a rectangular wooden tenon.
The attachment points for the T-clamps had to be reinforced with brass screws - something I'm not overly happy with:
This is largely because of the pine which I used for the case. While the pine is high-quality it still splits easily. However, this tendency is offset by much smaller weight of the case.
I also drilled holes to the new field frame bars:
The straight field frame bars are slightly wider at the middle to increase strength. This is another thing I'm not 100% happy w…
Random thoughts on the most impressive cheiroballistra reconstructions
Today I was writing an email to a fellow ballista reconstructor when I realized I should probably make the email a blog post. So here are two of the most impressive ballista reconstructions/reconstructors right now as far a I know.
Their reconstructions are very interesting as they don't use sinew at all. Nor nylon. They've tested various oiled animal and plant fibres and achieved quite impressive results: bolt velocity up to 86 m/s, range 480+ meters. Their most interesting hypothesis is this: they believe sinew was not typically used as spring cord material, at least not in inswingers. This is because it simply is not elastic enough for springs where
- The spring is relatively thick compared to its length
- The spring needs to be rotated a lot (=…
Composite slider and new triggering mechanism
The renovations in our apartment have finally come to a phase where I have been able to resume work on my cheiroballistra.
A few days back I finished attaching the triggering mechanism part to the composite slider which I created a few months ago. The composite construction can be seen in this image:
The back-end of the slider is of elm for hardness and splitting resistance, whereas the frontend is of pine to keep the weight down. The two pieces are joined together with a half lap splice joint, glue and two vertical wooden dowels. This type of joint should work well, as the slider has to endure mostly compressive forces. The slider is slightly wider than previously, which gives it additional strength. The angle of the dovetail has been adjus…
Two more days at the forge
The cheiroballistra project has stalled for quite a while as we've been renovating our apartment thoroughly since July. Fortunately after Christmas I could spend two days at the forge.
The results of the ballista-related work are these:
On the left there is an old ring for the cone and below it is an old 3mm thick pi-bracket. On the right side, from top, two spare 5mm pi-brackets, a new washer, three thick rings for the cones and eight thick pi-brackets with partially formed tenons. They attempt to fix a few mildly annoying issues with the cheiroballistra:
- The 3mm thick pi-brackets were a bit flimsy, even though they had survived earlier fairly high power tests.
- One of the old washers could not be easily rotated. This was probably caused by so…
New parts
It has been quite a while since I actually published pictures of new cheiroballistra parts, so here it goes. First the new little ladder forged from 6mm spring steel coupled with a better and stronger little arch forged from concrete reinforcement steel. Note how the tenons of the rungs simply project above the beams, instead of being riveted or otherwise attached to the beams. This allow one to bend the little ladder beams outwards when inserting their tenons into the pi-brackets in the field frame bars.
As can be seen from the picture below, the little arch's retaining pins are significantly stronger (5mm) than the previous ones (3mm):
Because the little ladder tenons are larger than previously, the wedge system needs to be adapted slightl…
Progress on the new parts
In the recent weeks I've been working on several new parts for the cheiroballistra.
The new case, which is made of pine, now has the crescent-shaped from the old cheiroballistra. I originally intended to make the new crescent-shaped piece from the crooked piece of spruce I had lying around, but that would have slowed down the project for basically cosmetic reasons. The female dovetails are basically ready, but they have not been glued to the case because the slider is missing.
The new slider is of composite construction, with the front part composed of two vertical layers of pine, and the short rear part made from a single piece of elm. The lamination step was not really necessary, but it reduces the change of the slider bending and getting …
Routledge Open Access week
Routledge is having an open access week. This is of particular interest because there are some ballista articles in there. In particular the Performance of Greek–Roman Artillery by Rossi et al 2015 is extremely interesting. After the open access week is gone, that article probably goes back behind a paywall.
Because this is a blog, and in blogs one is allowed to rant, I will take this oppurtunity to rant about these paywall-erecting publishers. Take the article "The Reorganization of the Defences of Romano-British Towns in the Fourth Century" for example. It contains the word "ballista" in it, because it shows up in the site search. However, there is probably no way to know if the article is actually relevant to ballista research in any way…
More thoughts about the crosspiece and rungs
In my blog post from yesterday I talked about the attachment of the little ladder rungs. I'll go a bit more depth here (as if yesterday's discussion was not esoteric enough).
It is a bit puzzling why the cross-piece and rungs had different names, because their descriptions are identical:
"And let the cross-piece(1) ΤΥ be made, and having length of 3 dactyls not counting the tenons, and [having] width of 2½ dactyls. And there shall be rungs(2) ΦΧΨΩ, having length of 3 dactyls not counting the tenons, and [having] width of 2½ dactyls."
The most reasonable explanation seems to be that the components were basically identical, but had different names to make the manuscript less ambiguous, as the cross-piece is referred to in several places:
"...an…
Attaching the rungs
The new little ladder made from 6mm (unhardened) spring steel is shaping up nicely. Today I finished sawing and filing the slots into the ends of the tenons:
The field frames snapped into the slots beautifully, and when there, the little ladder beams bend outwards only very slightly. I now consider the "little ladder beams should be 3 dactyls from each other"-mystery solved in an elegant fashion.
I had played with the idea of using a fancy wedge system to bind the crosspiece and the little ladder beams together. The idea was to allow easy dismantling the entire little ladder, which could prove useful if/when I create a second set of field frames for the sinew torsion springs. In the end I came to my senses and just riveted the thing together…
Translations of the cheiroballistra are complete
A few weeks back I finished translating the cheiroballistra manuscript, first into English, then into my native language, Finnish. I did the Finnish translation to see how well Greek translates in Finnish, and indeed, Finnish was much easier to work with than English, because of richer grammatical rules of Finnish. The second reason was to give the text a second go and see if anything changes. A few corrections were made to the English-language version during the Finnish translation round, but nothing really important changed.
Anyways, here are the translations:
- Translation of Cheiroballistra
- Translation of Cheiroballistra into Finnish
In retrospect the cheiroballistra manuscript was fairly easy to translate even with only two basic Greek cour…
Some thoughts on reconstructing ballistas
After a hectic January at work it seems things have finally calmed down a bit, and I can resume my project. First a few thoughts about recontructing ballistas, or any ancient weapons for that matter.
When I've talked about the cheiroballistra reconstruction project with other people of various backgrounds, it seems that putting oneself into the shoes of the Romans who actually used the cheiroballistras is quite difficult. After pondering about this a bit, I think one needs to approach this problem by asking these two questions:
- What was the purpose of the cheiroballistra?
- What properties did it need to fulfill this purpose?
The answer is clear in my opinion: the cheiroballistra was a weapon designed to kill (possibly) armored people at long ra…
Forces exerted by the field frames upon the little ladder and little arch
NOTE: This blog posting was converted into a real article with complete illustrations.
A few days back I had to really think about how the little ladder and little arch are stressed when the arms are rotated in the torsion springs. The way I reasoned it, there are two main forces involved. The first force is caused by the field frames trying to rotate as the arm is rotated. In an inswinger this means the following:
- The little arch starts bending towards the shooter, the stress being concentrated at the middle (i.e. at the center curve)
- The longer little ladder beam (at front) is squeezed (compressed) fairly directly along the longitudinal axis
- The shorter little ladder beam (at back) is pulled (stretched) fairly directly along the longitudinal …
New little ladder, arch and washer rotation tool
During Christmas holidays I was at the forge making replacements for various ballista parts. As I mentioned earlier, my original little arch, and thus the little ladder, was too long, which caused all sorts of unnecessary fuss and misinterpretations. So those needed to be replaced. Here's a comparison of the length of old and new little ladder beams - the outer beams are the old ones:
As can be seen, the old beams are way too long. The new beams are forged so that tenons' one side is in-line with the outer surface of the beam:
This forging technique, when coupled with the notch attachment mechanism shown below, elegantly solves the problem of crosspieces and rungs being too close (3 dactyls) to each other and preventing insertion of the ten…
Were there wedges on top of washer bars?
I and Nick Watts have been having many good discussions on his talk page lately. Nick is currently engaged in making a new ballista based on the Elenovo find (see Kayumov & Minchev 2010), and while doing that he thought of combining ideas from Philon's wedge machine and incorporating them into a "standard" ballista. The basic idea is that the top surface of the washer bars is flat, and a pair of iron wedges is mounted on top of each one. Spring cord is then wound into the torsion springs, and each cord is individually pretensioned using a stretcher. So in the end we have a torsion spring frame full of cords which pass over the crossbars and the wedges on top of them. This idea is pretty ingenious, because it solves two very real and seriou…
Was the handle made of wood?
Earlier this week I finished revising my old translation of the case, slider and the crescent-shaped piece in the cheiroballistra manuscript. Then, when I continued on to the part describing the triggering mechanism, I was caught by surprise. Here's the part:
"Γεγοέτω ἐξ ὕλης σιδηρᾶς χειρολάβη ἡ ΑΒΓΔ"
Translated to English:
"Handle ΑΒΓΔ shall be made of wood as strong as iron"
Here's an alternative translation:
"Handle ΑΒΓΔ shall be made of iron material"
And an "obfuscated" translation used by most scholar:
"Handle ΑΒΓΔ shall be made of iron"
I call this approach obfuscated, because the word "ὕλης" has just been ignored. If "iron handle" was what Heron (or P.H.) ment, then the way he said it in Greek is very odd, and that alone would warrant a comme…
Stretching properties of nylon and sinew
In the past I've talked about the problem of having "too few holes" in the archaeological all-metal kambestria (field frame). The problem is that there are 6 holes in the washers (at least in the Lyon specimen) and four holes in the kambestrion. This does not seem to be granular enough for adjusting the limbs (in nylon torsion springs) so that they are perfectly in sync during pullback.
A week or so ago I stumbled on the "Pretensioning applied to ancient torsion-powered ballistae" article (download link, Bibliography), in which the researchers claim that sinew is considerably more elastic than nylon. If this is the case, it would mean a lot of things:
- Rotating a washer N degrees would increase pretension less in a sinew spring than in a ny…
Published an article on the cheiroballistra
A few months ago I was asked by a friend to write an article about the cheiroballistra to a small archaeological publication. Naturally I said yes, even though at that point I was not sure how much power I would manage to squeeze out of the thing before the deadline, and writing about a peashooter would have been humiliating. Fortunately I managed to produce a reasonable amount power just before the final editing of the publication. Most importantly the previous tests showed that the cheiroballistra is capable of generating Firefly level performance, once all the components are strong enough to take the strain.
The article is called "The Cheiroballistra - Producing a viable weapon based on historical manuscripts and archaeological finds" and…
New design for 11 dactyl cones
Today I updated the Cheiroballistra arms article with the latest information, and most importantly, with new CAD drawing of elliptically tillered 11 dactyl arms:
The design shown in the drawing incorporates all what I've learned about making cheiroballistra arms. It is highly unlikely the design will significantly change after this, because all the "bugs" have been squashed already. The fact that the cones are now short again helps me a lot, because I have tons of elm pieces of just the right size stored away.
The field frames can take 350hz
A few weeks ago I pretensioned the spring cord in one field frame to 350hz. This should give about 20% increase in power with the same amount of washer rotation. While the field frame bars are slightly more bent now, I don't think the bend is excessive. Thus it seems that I don't have to make new field frames to reach "full" power, which obviously saves me a lot of time and effort.
Before restringing the other field frame I'll do practical to compare how much more resistance the 350hz frame offers compared to the 320hz. I have some old otherwise useless limbs which I can use for this testing. Hopefully I have some results of this tomorrow.
Stretching to 350hz
Today I started pretensioning one of the spring bundles to 350hz. This should yield 1.2 times the power of the previous 320hz setup, and 1.96 times the power of the original 250hz springs. I suspect the field frames can take this power increase just fine. This time I'm using unmodified locking pliers, which work just fine with the "tighten with a winch first, then release pliers" (a.k.a. "snap in place") technique. Smooth jaws would only be required if the winch was actually used to forcefully pull the cord from between the jaws, which is what I did when I used the clip of my own construction.
While the torsion spring was dismantled, I noticed that one of the field frame rings had bent inwards on the side. This had happen earlier with the 2…
Forging oval field frame rings, take three
While at the forge I started wondering how to make field frame rings of the type found in the archaeological record. So far I've successfully used two techniques:
- Making steel field-frame rings
- Forging rings using a horizontal ring forging template
Neither of these produce the exact kind of rings as in Lyon (Baatz 1981), Orsova (Baatz 1974; 1978) or Elenovo (Kayumov & Minchev 2010) field frames. The Gornea (Baatz 1974; 1978) field frames are round, so the method I used for forging washer rims would work for those. A few potential techniques for making truly oval rings came to mind after a bit of pondering:
- Cut (e.g. hot-chisel) the ring from steel plate and finish the form with a file.
- Start from a flat steel bar. Cut a longitunal slot at the e…
Not so total rebuild?
I started the work on new ballista the previous weekend at the forge, focusing on the little arch and the little ladder, which were too wide.
Instead of just shortening the little arch I decided to make the curved part shallower to increase its strength, because I suspected it might have been crushed a little. However, doing that while keeping the straight parts correctly aligned proved to very difficult. So I had to let go of the old little arch completely and start planning on making an entirely new one from fresh, round stock. Forging the shape first and then flattening is something I've done plenty and which works very well. On the bright side I did manage to forge nice U-shaped forks with sharp, 90 degree corners for the little arch wi…
Total rebuild
Today, after documenting the stretching process, I inspected the damage on the cheiroballistra caused by the previous dry-fire. One of the strong screws that kept the little ladder beams attached to the rungs had become partially loose, which is probably enough to explain the dry-fire. I suspect the hole for the screw was drilled a bit too large, so that the threads had not sunk deep enough: the other three screws had remained perfectly static. The field-frame on this weakened side had also canted slightly backwards, even though the reason is not entirely clear. The T-clamp that also came off during the dry-fire also came from this side, but it is not clear whether that had contributed to, or was the result of the dry-fire. In any case the …
117 joules reached, more to come
Today I finished the first round of shooting with springs tuned to 320hz. The results were really good and what even more important, showed that lots of improvement is still possible using this spring configuration. These tests were performed with a new 48.85 gram test bolt:
The new bolt is significantly heavier than the bolt used for testing the 250hz springs, which skews the results in favor of today's shooting session. With this new bolt the 250hz springs could probably have yielded about 90-95 joules of energy as opposed to 85 joules.
I did a fairly extensive battery of shots with the 48.85 gram bolt, progressively increasing washer rotation. Washer rotation angles were not measured exactly, but should be roughly correct:
These results ca…
Torsion springs tuned to 320hz
In my earlier blog post I planned on tuning my torsion spring cords to 350hz from the original 250hz. That change would have roughly doubled (1.96x) the force required to pull back the arms, given the same amount of washer rotation.
Later I peddled back and decided to settle for 320hz, which should increase energy to ~1.64 times the original. There were several reasons for the decision. First, I was afraid that I could not rotate the washers at all, because they're not that symmetrical. Second, I did not want to bend the field-frame bars for no good reason. Third, based on the amount of cord in Nick's Firefly and in my cheiroballistra, the "Firefly" level of performance would be 150 joules. My previous best with 250hz was about 85 joules wit…
Replacement washer bars
Today, when I was about to start arming one of my field-frames, I noticed that all of the washer bars had bent:
The bend was not much, around 1mm, but it was obvious that doubling the pressure of the springs would not be possible using these washer bars.
So, I naturally made stronger washer bars that had a shoulder and all. The total height of the new bars including the 4mm shoulders is 24mm. The old bars were 14-15mm high, so we're speaking of a very significant increase in strength. The new bars are probably an overkill, but I just want the damn thing to stay in one piece for a while, plus the washer bar's height has absolutely no effect on any other aspects of the machine.
I know the text gives 2/3 dactyls (~1.33 cm) for the height of the ba…
Little arch too long
Today, while measuring my Cheiroballistra parts for updated CAD pictures I noticed that I had (unintentionally) made the little arch sigificantly wider than it was supposed to be:
As can be seen, the error had occurred when measuring the base length (ΓΕ) of the little arch. I'm not entirely sure how an error of this magnitude managed to slip in.
This means a couple of things:
- The new, longer 14 dactyl cones ("the text must be corrupted") are too long to steer clear of the case. The manuscript's number for the cones, 11 dactyls (22cm), is just fine.
- I will need to make a new little arch and modify my new unfinished little ladder beams to be the correct length.
- I will need to make new shorter elliptically tillered cones.
- I will probably have to ma…
Replacement parts
Three days at the forge/workshop allowed me to get lots of things done.
Reinforced old little ladder beams
I welded small wedge-shaped pieces to the tenons in the old little ladder beams:
This should allow me to continue shooting with high-power without the need to (immediately) finish the new little ladder.
Reinforced the curved field-frame bars
Small pieces of steel were welded to the inside of the curves:
The curves are now much stronger and can certainly handle the spring cords tuned to 350 hz.
Forged parts for a new little ladder
The main differences to the old ones are:
- Beams are made from 5mm, not 4mm steel
- The beam tenons are slightly narrower (9mm vs. 11mm)
- The ends of the bars (tenons) have been forged flat, not folded
- T-clamps were forged w…
Thoughts on linear pretensioning
I've been discussing torsion spring stacking with Nicholas Watts. For those who have been hiding under rock, he is the maker of the Firefly - an extremely powerful reconstruction of the Orsova ballista.
As mentioned earlier, I believe the only way forward (=more power) is by applying more linear pretensioning. My theory is that in previous tests the washers needed to be rotated too much (up to 270 degrees), which caused the torsion springs to stack too early in the draw. This theory is based on simple geometrics: with zero washer rotation the spring cords are vertical and look like an "I" if you will. Rotating the arm displaces the cords easily, because the cord needs to stretch only a little. As washers are rotated more and more, the cords …
Hard won joules
Today I assembled the cheiroballistra and shot quite a few bolts through the chronograph. I could easily get to 67m/s with 30 gram bolts (66 joules) with 225 degree washer rotation and a draw of 59 cm. I tried rotating the washers 270 degrees, but the stacking at around 10cm to full draw was pretty extreme. Drawing the thing may have been physically possible given enough courage, which I lacked. Having gotten scared, I reduced the washer rotation to about 250 degrees. At this rotation I only managed to pull of one successful shot, from which I can roughly estimate velocities for two different bolt sizes:
The estimated velocities are based on my earlier results.
In any case, I learned several good lessons today:
- The shims and wedges keep the l…
Back in business
It has been a while (over 6 months) since I actively worked on the cheiroballistra. Some weeks ago I finished my new workshop and, after a few detours[1], was able to resume the work on the cheiroballistra. Over the last few days I've fixed a large number of minor issues in the cheiroballistra.
Wooden shims were added between the little ladder tenons and the pi-brackets under the wedges:
The shims should help with two things:
- The slow loosening of the wedges due to little ladder tenon movement
- Slipping of the field-frame bars from the notches in the tenons
A pair of pins was added to each end of the little arch to rigidify the upper part of the field-frames. The holes in the Orsova little arch may be explained with similar pins:
The elliptically…