Class Notes For
August 14, 2000
Table of Contents
A discussion of riveted maille from a historical viewpoint *
Wire Source *
Winding Coils *
Cutting Rings *
A discussion about riveted maille as it applies to the modern maille maker *
Wire Source *
Winding Coils *
Cutting Rings *
Internet Resources *
Theory (conjecture / fact based on documentable evidence)
Technique (applications for the modern maille maker)
The purpose of this class is to learn about the materials, tools, and assembly techniques involved in the construction of riveted maille armour. There is very little in the way of period documentation to help the student of maille. There are also no known surviving maille tools that one may study. There are, however, a few extant maille garments available to examine today. For this reason, I have found it useful to approach the study of maille construction methods as an exercise in trial and error, where an assembly technique is tried, and the result compared against an authentic piece of maille. This class will thus present some data as fact, based on documentable research and examination of authentic specimens. Some data will be presented as conjecture, based on the results of my experimentation. I will vigorously attempt to distinguish between the two.
Because the various weaves of maille are widely known or available, we will not be focusing on how to "knit" maille, either in general or in specific garments. Instead, we will be focusing on what is known about the individual ring elements that make up maille garments, such as the materials of construction and assembly techniques.
There is considerable debate surrounding the medieval maille maker’s source of wire. There is no doubt that ancient man knew how to draw metal into wire. Examples of drawn bronze wire and stone draw-plates dating to around 2000 B.C. have been found. There is an illustration from a Nürnberg craft book dating to 1529 which shows "a long wire being wound from one drum on to another and passing through a multi-hole draw plate fixed between them." Another illustration, dating to 1533, depicts "a man drawing wire through a large multi-hole draw plate with large two-handed pliers." A third illustration, a Dürer watercolor, shows "wire being wound on to a drum of considerable diameter and great length which is being turned by a water wheel."
However, there is also evidence to support wire being fabricated by cutting a strip off of a sheet of metal. Dr. Cyril Stanley Smith, in an article replying to Mr. Burgess’ above sited comments on drawing wire, notes, "The cross sections of most of my European samples had slag distributed in a way that indicated that an approximately circular section had been maintained for some time during the process of reduction, i.e., that the wire had been drawn extensively. The oriental wires, however, all showed evidence of having been cut from sheet or strip, and three of the European ones (Nos. 2, 3, and 5) had been drawn only a little, if at all." He further states, "The European wires may have been drawn through a finishing die, but they had been reduced in cross-sectional area by certainly far less than 50 per cent after the metal had been slit or sheared from a strip. Some of the oriental links showed no distortion whatever of the ends of the fibers, and the metal originally at the corners of the cut rectangular section must have been removed, not compressed inwards. Although it seems surprizing [sic] to us that all wire was not drawn, the evidence on these few links is incontrovertible. It should be recalled that even nails were made by slitting until well into the nineteenth century."
In another of Dr. Smith’s articles, he details his examination of 16 rings from various maille garments of known origin. When speaking of the rings of a 17th century Turkish shirt, he notes that, "they have slag stringers running nearly parallel to each other right across the wire section, and show unmistakably that the wires had been cut from a thin plate or strip and filed, scraped or abraded to the present shape: they are definitely not made of die-drawn wire."
The earliest of the 16 rings that Dr. Smith examined is from a 14th century European coif. Dr. Smith concludes that this ring was made of drawn wire. The earliest mention of wire drawing as related to armour that I have been able to document is a document written around 1260 by Etienne Boileau, Livre des Metiers, which mentions two corporations of wire drawers in Paris.
What we may deduce from this is that maille was undoubtedly made from both drawn and slit wire. We might speculate that perhaps maille wire was originally simply cut from sheets of iron, and that as the technology of drawing wire matured, drawn wire may have become more prevalent. This assumption about the evolution of wire manufacture may be incorrect, however, as the 17th century example made of slit wire, described above, might attest.
Present wisdom seems to indicate that the raw material for the wire used to make maille was predominantly iron, not steel. Of the 16 rings mentioned previously by Dr. Smith, only 3 of them contained enough carbon to be considered steel. Those three steel rings had also been heat treated to harden them. However, I have doubts as to whether enough maille has survived through the ages, and whether enough chemical analysis has been done on that surviving maille to draw any solid conclusions as to whether iron or steel was the preferred metal for making maille. Nonetheless, most publications on maille today cite iron as the primary material, and Dr. Smith's analysis above would seem to support that idea. So for the present time, I will assume that iron was the predominant metal used to make maille.
The question arises as to whether or not the medieval maille maker intentionally used iron or steel as a raw material. Certainly medieval metal workers have been familiar with the differences between iron and steel since ancient times. Pattern welded swords dating to the second century show that the mechanical properties and benefits of both metals were understood and put to use, making swords that combined both metals to produce one that was both hard enough to hold an edge and yet tough enough to take the abuse of battle. Thus "it was known that steel was a much more versatile material, but its relationship with iron was not understood. It was thought (by Biringuccio for example) that steel was a purer form of iron, and if iron was left within a charcoal hearth for hours, or even days, it was made purer or more ‘steel-like’ by having the impurities burnt out." Though the medieval smith may not have understood how to make steel, they obviously were able to differentiate between steel and iron. This may have been done by testing various blooms with a chisel to find ones that seemed harder than others, or though spark testing, whereby a piece of metal is held against a grinding wheel, and a judgement made based on the color and shape of the sparks produced.
If indeed the medieval armourer was aware of the differences between iron and steel, why then does it seem that the majority of maille was made from iron, rather than the stronger steel?
It would seem that economics is likely the driving force behind the choice of metals. During the Wallace Collection Arms and Armour Study Day that took place in November of 1999, Dr. Alan Williams presented excellent information that likely has great bearing on the choice of metal for maille making.
Dr. Williams presented data that plotted the size of iron blooms, recovered as archeological finds, against their probable date of origin. What he was able to deduce from this is that there is a definite correlation between the size of the blooms and the economic prosperity of Europe at the time of creation of the bloom.
Thus Dr. Williams was able to explain why it is not until the 14th century that items like one-piece breastplates began to appear in Italy. The reason is that economic conditions did not support the construction of large bloomery furnaces needed to make blooms large enough for such large plates until that time.
According to Dr. Williams, a one meter high furnace approximates 1kg blooms. Furthermore, it takes a ten kilogram bloom to produce a two kilogram plate of steel. The reason for this is because as the plate is heated and worked much of it is lost as scale. Therefore the furnaces producing such large blooms of steel were likely very large, permanent structures, and in fact there are period illustrations of such furnaces much taller than a man.
One of the side-effects to having such a large, tall furnace is that it takes the solid particles iron, carried by molten slag, more time to reach the bottom of the furnace. Consequently, the iron absorbs more carbon than it would in a smaller furnace. Additionally, a larger furnace may burn hotter than a smaller furnace, which again will enhance carbon absorption. To quote Dr. Williams:
"Higher C% is due to both higher temperatures, as you have reaalised, [sic] and, if taller, longer residence time. The particles of solid iron pass down the furnace, absorbing carbon on the way until liquid slag draws them together by surface tension and they sinter into a bloom. The longer they spend sinking downwards, then
the higher their C %, and also the lower the Fe % of the slag."
The upshot is that larger furnaces are more likely to produce blooms of steel, rather than iron blooms usually produced by smaller furnaces.
Assuming that this is true, I believe it may help explain why we find most maille to be made of iron. First of all, only small blooms would be needed to make wire. After all, why waste large blooms, that could be used for larger plates, on mere wire? Well, if we accept that the smaller blooms would have been relegated to things like wire, and if we accept that smaller blooms would be the product from smaller furnaces (less likely to be steel), than this might well explain why most mail was made of iron.
Dr. Wiliam’s response to my above theory was simply, "Precisely."
At least some maille garments were case hardened and tempered, for specific mention of this is mentioned in the publication Natural Magick Book XIII, in 1558. This suggests that some maille was at least partially converted to steel after assembly, and done so specifically for the mechanical advantages of hardened, but tempered, steel.
Brass and possibly more precious metals were sometimes used to add decorative trim to maille garments, often as edgings.
Once a source of wire has been procured, the next step is to wind it upon a mandrel. If the wire was available in short segments, a tool much like a modern screwdriver, without the head, was likely used as a mandrel.
It is interesting to note that nearly all maille rings are cut from a coil with a right-handed thread. This is not surprising if we visualize the hand-held mandrel described above being used by a right-handed person.
After winding the wire into a coil, the next step is to cut the individual rings from the coil. There are actually a few ways that this can be done.
One way to cut the rings from the coil is to simply make an incision down the side of the coil, parallel to the axis of the coil. This cutting could be done with some kind of snip, nippers, or with a chisel. This produces rings that have no overlap to them, but produces them relatively quickly.
Another way to produce the rings is to cut them from the coil by driving a wedge in between the individual rings on the coil. By using this method, rings may be cut from the coil with the overlap built in. This has the advantage of producing perfectly round rings, but may be slower than the incision method described above, due to the fact that the rings must be cut one at a time.
Obviously if the rings are not cut with the overlap built in, an additional step is necessary to form the overlap. This can be done by hand, carefully working the ring with a hammer or pliers to force the ring into an overlapped condition. Other methods have also been proposed, including forcing the ring through a tapering hole so as to make it smaller and overlap its ends.
In practice, the medieval maille maker may have annealed the rings prior to each step in the production of maille. This would eliminate any work hardening from each of the previous manufacturing processes, and keep the wire as soft as possible.
It is certain, however, that the wire must be annealed before the rings can be flattened. My own experimentation has shown that if one attempts to flatten the rings without annealing them first, the ring at best does not deform as observed in authentic specimens. At worst it will not flatten at all, or it becomes mangled in the attempt. Generally the ends of the wire will tend to split rather than deform smoothly into one another. Alternatively, the overlapped ends will refuse to deform at all, instead preferring to ‘jump’ off of one another when struck.
It is likely that the rings were annealed again after being flattened. This would be done to eliminate the work hardening caused by the flattening process, to make the metal as soft as possible in preparation for the punching operation. This would have the affect of sparing undue stress on the punch tool.
The simplest way to anneal the rings is to simply string them onto a piece of wire and throw them into a forge. After bringing them to a red heat, simply remove them from the fire and allow them to air-cool. The more slowly the rings cool the softer they will be. In actuality, the rings are actually being normalized with this process, and not truly annealed.
In order to prepare the ring for the punching operation, the ring overlap must be flattened. Mr. Burgess suggests that some form of swaging tool may have been used to crush the overlap, and undoubtedly such dies were used in the construction of maille at some point in history, as evidenced by some surviving rings that have been mass produced with writing stamped into them.
However, the simplest way, explained to me by a Mr. Eric Schmid via the Armour Archive, an Internet bulletin board dedicated to arms and armour, is to simply strike the ring with a hammer.
It is important to note that flattening was done after overlapping the rings. If the flattening is done before overlapping the rings, experimentation shows that it leaves very obvious tool marks not present in authentic examples I have examined. Examination of authentic maille overlaps shows that the two ring ends exactly mesh into one another, much like the continents of Africa and South America seem to do. This close fit can only be obtained by crushing the two ends into one another.
The next step is to punch a hole or slit into the ring, so that it can accept a rivet. The rivets used in maille of European origin are almost always of a wedge shape. As such, they require a slit, and not a round hole, to fit into.
The word "punching", as applied to producing the slit for maille utilizing wedge rivets, is actually a misnomer. Punching implies that material is removed during the operation. When producing the slit to accept a wedge rivet, no metal is removed from the ring. Instead, the slit is pierced into the ring overlap, by having a wedge-shaped punch driven through it.
Some surviving maille, generally thought to be Middle Eastern in origin, shows the use of true cylindrical rivets instead of the more common wedge rivet. This style of rivet requires a round hole, and not a slit, to be punched through the ring overlap. This may have actually been a punching, and not a piercing operation.
Like the flattening operation, there are many ways the punching operation could have been performed, ranging from very primitive to highly mechanized. Since no surviving tools are available to examine, we must try different methods and compare the result to the authentic product.
The simplest method for producing the slit is to do it by hand, using a hand-held punch, and driving it into a punch block that has a recess to accept the point of the punch after it passes through the ring overlap. This method takes considerable practice in order to get consistent, quality results, but it is quite workable.
A more complex arrangement would be to mount the punch and punch receptacle into a set of hand-held tongs. I have constructed such a tool, and it is a vast improvement over the hand-held punch method, conveying increased speed and more consistent quality with very little practice. Certainly the manufacture of tongs are not beyond the skill of the medieval smith; they have likely been in use since man first encountered the need to handle hot metal, as many period illustrations will attest. The mounting of a punch and punch receptacle into a pair of tongs does require a bit of somewhat precise machining, but is probably not beyond the abilities of the medieval craftsman.
In my opinion it is likely that both punching/piercing operations were used, and probably many variations were also used.
The final step of completing a maille link is to set a rivet in it, effectively making it solid. As indicated in the previous section about punching the rivet hole, most European maille utilized a small, thin wedge of metal as the rivet. This rivet was almost exclusively made of iron, even when set into brass rings.
Producing the rivet is quite simple. One can either start with a narrow strip of metal, or a piece of round wire hammered into a narrow, flat strip. This strip is then cut into many small triangle shapes. This can be done with some form of shearing or nipping hand tool, or more simply, by using a chisel.
Once the rivet has been produced, it is driven into the slit prepared for it in the flattened ring overlap. Practice has shown that as the rivet is driven into the slit, it actually tears its way in place. In fact, the rivet can be thought of as completing the punching operation. In this respect the rivet actually functions like a nail, being wedged firmly in the ring overlap.
At this point, the tip of the rivet is protruding from the other side of the ring flat. This will be peened over to function like a traditional rivet. It is notable that most maille garments have a definite inside and outside, with the peened rivet points pointing towards the outside, away from the wearer.
Like the flattening and punching operations, the setting and clinching of the rivets can be done very simply, or with the assistance of specially made tools.
The easiest way to set the rivet is to simply drive it in place with a hammer, supporting the ring with some sort of plate with a hole in it, to give the tip of the rivet some place to go. The tip can be upset by flipping the ring over and peening it with a hammer. It is conceivable that the medieval maille maker used some kind of rivet setting tongs to set the rivets as well. There are period illustrations that depict maille makers at work with a pair of tongs in the garment they are working on. While the detail is not sufficient to see exactly what they are doing, there are few steps that could be performed during the knitting procedure that would involve tongs, except perhaps setting the rivets. The flattening and punching operations would almost certainly already be completed prior to knitting.
The one significant luxury that the modern maille maker has over the medieval one is the ability to obtain nearly unlimited quantities of wire with extremely consistent shape and chemical composition. It is quite handy to buy the wire by the spool. This gives you a price break by buying in quantity, and it allows you to construct a winding jig that can produce very long coils.
If you are attempting to create an authentic replica of medieval maille, you will want to buy wire as close to pure iron as possible. This can be difficult to find, so you may instead opt to purchase non-plated, low-carbon steel wire. You may certainly use other, more exotic materials, especially if you would like to avoid rust problems. Just bear in mind that whatever wire you choose you will have to be able to drive a punch through it without destroying the punch on a regular basis.
It will serve no purpose to use plated wire, because the plating will almost certainly be burnt off during the annealing process. In fact, some wire coatings, such as the zinc used to produce galvanized wire, will give off noxious fumes when heated.
Because of the advantage of long lengths of wire available to us, the modern maille maker can make a more efficient winding mandrel than the simple hand-held mandrel used in the past. I will not go into detail describing fixtures to wind coils, but any of the common fixtures used by butted maille makers will do. See the bibliography of this paper for a website that contains plans on how to produce such a winding fixture.
Suffice it to say that the modern maille maker can wind coils of wire considerably longer and faster than the medieval maille maker could. Alternatively, pre-wound coils can be purchased from spring manufacturers or hobbyists.
Like the medieval maille maker, we have a few ways in which to separate the individual rings from the coil. The easiest way to cut the rings is exactly the way butted maille makers do it today; make an incision down one side of the coil, parallel to the axis of the coil. Unlike butted maille, however, the quality of the cut does not have much impact on the quality of the finished product, and therefore speed, rather than quality of cut, is more important. The fastest way that I have found to cut rings in this manner is to use shearing cutters, such as aviation snips. Of course pinching cutters, such as mini bolt cutters, or rotary cutters, such as Dremel™ cut-off disks are also acceptable.
Unfortunately, cutting the rings in this method, while fast, means that an additional step of overlapping the rings must be done. Generally an overlap of .2" is sufficient. Creating a smaller overlap makes it difficult to pierce the rivet hole without splitting the metal in the ring overlap area, and creating a larger overlap tends to give the flattened rings a "bunny ears" look to them, as the ring ends tend to "squirt" out from under each other.
One way to overlap the rings is to simply hammer or squeeze them with pliers until you get the desired overlap. While simple, this method is tedious and mechanically inconsistent. It will take much practice in order to produce rings that are consistently overlapped and round.
Various mechanical dies have been proposed that force the ring into an overlapped state. One such tool is the "funnel" die, whereby a non-overlapped ring is driven through a conical hole in a block of metal. Other such tools have been proposed, such as tongs that squeeze the ring to a desired finished diameter. Additionally, the non-overlapped rings can be threaded onto a mandrel whose O.D. matches the desired I.D. of the finished ring. This mandrel, loaded with rings, can then be laid on the face of an anvil, and by rolling and hammering, the rings can be forced to have their I.D.s match the diameter of the mandrel.
My preferred method of getting the overlap is to simply cut the rings from the coil with the overlap built in. This has a few advantages over forming the overlap as an independent operation.
First of all, it eliminates a step. It does this, however, at the expense of forcing you to cut each ring from the coil one at a time, so any timesavings are dubious.
Secondly, it ensures that each ring is perfectly round. Usually when rings are forced into an overlapped state after cutting, the finished ring ends up with a slightly oval shape, unless the overlapping machinery is designed to force the ring to maintain a round shape.
The round shape becomes important during the flattening process. It is important that the ring ends exactly overlap one another. This happens best when the ring is perfectly round, so that the radius of curvature of each of the ends matches one another and are concentric. If the radiuses do not mach one another, or if the ring ends are not concentric with one another (quite common with rings that end up oval shaped after overlapping), then the ring ends tend to cross each other like an "X" rather than lying directly on top of one another. At best this causes poor deformation during the flattening process, and at worst the ring ends will simply jump off of one another when you attempt to crush them.
It should be noted, however, that much authentic maille surviving today is composed of rings that are not perfectly round. While the cause of this out-of-roundness is unknown, it is possible that it is a result of the manufacturing process. If so, it would tend to discount cutting the rings from the coil with the overlap built in, with the resulting perfectly round rings.
Cutting the rings with the overlap built in can be done a couple of ways. It can be done by driving a chisel between the coils of the spring, thus cutting the ring wherever you desire and ensuring an overlap. Another way to do it is to grind a pocket into the faces of a pinching style cutting tool, such as side dykes. By grinding a small pocket between the two blades, you can effectively "jump" over the first coil in the spring while making an incision down one side of the spring.
Before the rings can be flattened, they must be softened so that the ring ends deform properly. In practice, the medieval maille maker may have annealed the rings again after flattening and before punching, to eliminate any work hardening that may have occurred as a result of the flattening process. The modern maille maker, with access to modern tool steels, may find that his punches are robust enough to pierce the overlap without the additional annealing. Though I do anneal my rings before flattening, I do not anneal my rings prior to punching and I have not had any problems.
Any method that allows you to heat the rings, in quantity, to red heat is acceptable. If you have access to a forge you can do as the medieval smith did and simply thread the rings onto a piece of wire and toss them in. Once they are red hot, simply remove them from the fire and let them air cool.
If you lack a forge, another alternative is to thread the rings onto a wire and use a propane or MAPP gas torch to heat all the rings to red heat. If you choose this method be sure to evenly heat all the rings.
Another solution that is quite handy because it allows the annealing of thousands of rings at a time is the use of a pottery kiln. Place the rings in some sort of container that will withstand the 1100°F -1500°F temperatures, such as a steel box, and bake them to around 1200°F. The exact temperature necessary to normalize the rings depends on what kind of metal you are using. Check with your wire manufacturer or the Machinery Handbook to find the right temperature. True annealing is a very controlled process, and involves raising and lowering the metal to precise temperatures for precise periods of time. Such precision is not necessary for our riveted maille and it is unlikely the medieval maille maker incorporated it, either. Instead what we are actually doing is called "normalizing" the metal, which sufficiently softens it for our purposes. To normalize the rings using a kiln, simply raise the temperature to the annealing temperature, and then let it slowly cool back to room temperature.
If you opt to use a pottery kiln be aware that kilns can become contaminated by deposits left by vaporized oils or other coatings on the wire. This contamination can effectively ruin the kiln for pottery firing, as it may adversely affect the finish of fired pottery items, especially glazing.
You will notice that after annealing the rings now have a crusty, scaly black finish to them. This is the result of the rings having oxidized due to the high heat and the presence of oxygen. This oxidation actually helps us during the flattening process, as it makes the surface of the wire rough, and makes the overlapped ring ends less prone to skipping off of one another when struck together.
During the flattening process, the maille maker has the option of flattening only the ring overlap, or he can flatten the entire ring. All that is necessary is to flatten the ring overlap. This crushes the ring ends into one another, providing a seamless ring of metal. Additionally the excess material present at the ring overlap causes the overlap to expand somewhat, giving you a larger "target" to hit during the punching or piercing operation for the rivet hole.
I emphasized necessary above because if you are trying to emulate flattened rings or rings made of flattened wire you may also flatten the entire ring instead of just the overlap. Of course if round cross-sectioned rings are what you are after, only flatten the overlap.
The simplest way to form the overlap is to smash them with a hammer on the face of an anvil. If you are only flattening the overlap, you may find it useful to hold the ring in place with something other than your thumb – a penny is quite useful. If you are flattening the whole ring, just smash the whole ring with a flat-faced hammer.
I am certain that quite sophisticated dies were brought to bear on the rings during the flattening process in medieval times. Such tools probably worked much like those involved in the striking of coins. I believe this because some surviving maille rings bear evidence of having been coined, with words or symbols being embossed into them. There is a shirt at the Higgins Armour Museum in Worcester, Massachusetts, where every link has words formed into them. I find it highly improbable that each link was hand-worked with this writing. More probable is that the links were inserted into some kind of die and struck, thereby leaving the impression in them. Whether or not this embossing process was combined with the ring overlap flattening process is speculation, but seems likely to me.
As mentioned earlier, examination of authentic maille samples clearly shows that the rings should be flattened after being overlapped, and not before. Authentic overlapped ring ends fit exactly into one another because they have been crushed into one another. Any attempt to flatten the ring ends prior to overlapping them will not replicate this effect.
As stated previously, the rings will be covered with a black, scaly oxidized finish after annealing. You will want to leave this finish in place for flattening, as the rough surface helps prevent the ring ends from skipping off of one another when struck together. However, we will want to remove this oxidation, as it looks poor and has a tendency to flake off, making the rings messy to work with during the punching and riveting operations.
The easiest way to remove the oxidation is to bathe the rings in acid. Before I go any further let me say that this is a potentially dangerous way to clean the oxide off of your rings. You assume all responsibility if you decide to try it.
Hydrochloric acid (also known as Muratic acid) is readily available from hardware stores. It is often used to alter the pH of swimming pools, as well as for cleaning gravel or cement. Obviously it is extremely corrosive, and even the fumes can burn your eyes or respiratory system. Use extreme caution and safety equipment when working with acids or other chemicals, and be sure to have plenty of ventilation.
Place the rings in some sort of acid-safe container (not metal!). Glass works, but I would recommend you avoid it since it is fragile. I have found that the re-sealable plastic containers available at Walmart seem to be impervious to acid, and work well for small batches of rings. It goes without saying that you should not re-use this container for anything other than cleaning your rings. I would also recommend that you test your container for acid resistance before you are in the middle of cleaning your rings.
Place the oxidized rings in the container, and then pour the acid over them, until all the rings are immersed in the acid. The rings should start to bubble slightly and very caustic fumes will be produced (don’t breath them!). Allow the rings to sit for five minutes or so, carefully stirring them once or twice.
To finish the process, I do something that you are never supposed to do with acid. That is, I simply flood the container with water – lots of water – until I am certain that all the acid is gone. You should never add water to acid (only add acid to water) because it can cause intense chemical reactions, resulting in the release of lots of heat and/or acid flying everywhere. Do not flush your acid down drains or sewer systems – it will corrode the pipes. Additionally you may wish to neutralize the acid by sprinkling baking soda into the waste solution.
After the acid bath, the rings will have a light gray color to them. Fear not; they will polish up fine. An easy way to polish them is to use a gun brass tumbler. These are small fishbowl shaped devices that have a built in vibratory motor in them. They are designed to clean and polish spent gun cartridges by tumbling them in ground-up corncobs or other soft medium. For our purposes, we want something a little more aggressive. I use aquarium gravel. The cheap colored kind is fine. Fill the tumbler about ¾ of the way full, and then fill it the rest of the way with rings. I let my rings tumble for about 8 hours.
The easiest way to remove the rings from the tumbler is with a magnet. While the tumbler is running and the rings are all tumbling around in the gravel, run a large magnet around through the gravel. The rings will jump out of the gravel and onto the magnet, where you can scrape them off into a bowl. Repeat this process until you’ve retrieved all of the rings from the tumbler. You will then want to wash the rings in warm soapy water to remove the powdered oxide that invariably the magnet also pulls out of the gravel. Dry the rings thoroughly and quickly so they don’t rust.
The next step is to punch or pierce a hole or slit for the rivet. Obviously round rivets require a hole to be punched in the ring flat, and wedge rivets require a slit. I will discuss punching holes first.
The simplest way is to do it by hand. You will need a block of metal into which you can drill a hole only slightly larger than the punch you will be using to make the rivet hole. You will also need some kind of punch. To make a punch, you will need very hard steel. Alternatively you can make the punch out of a soft, but hardenable steel, and heat-treat it to regain its hardness. Unfortunately not many people (including myself) have access to adequate heat treating facilities, nor the knowledge to use it, so it is easier to simply fashion the punch out of a very hard material, shaping it with a diamond-laced grinding stone. Be sure to keep the metal cool while you grind it into shape. Remember: heat softens metal and we want the punch to stay very hard. If it gets too hot to touch you have let it get too hot.
Many items can be converted to use as a metal punch. I have found that ice picks, awls, and Dremel™ bits can all be readily converted into round hole punches. The trick to any kind of metal punch that is designed to actually remove metal is that the tip must be flat, not pointy. A pointy item driven into a metal object merely distorts the object as it pierces it. As often as not, this tears the object you are attempting to punch (much like driving a pencil through a piece of paper). A flat-tipped punch, however, driven into a close-fit hole, will actually drive out a piece of material from the object being punched (much like a paper punch does). Whatever you use for a punch bit, the trick is to find something slightly larger than what you will be using for a rivet. Punching into the ring overlap and into a punch block will push out a tiny disk of material, leaving a clean hole for the round rivet.
Instead of using a hand-held punch and punch block, alternatively the punch and receptacle can be mounted in some form of tongs. This makes the punching process much faster and makes it much easier to get consistent hole placement on the ring flat.
Piercing the slit for wedge rivets is a very similar operation, except the punch does not travel all the way through the overlapped ring flat. Instead, the pointy, wedge shaped punch only barely breaks through the opposite side of the ring.
Some good starting points for making wedge rivet hole punches are drill bit blanks, pin punches, nail sets, or chisels. The shape you are trying to form the punch into should resemble a very small screwdriver blade. The punch should be very narrow and slightly tapering, like a knife blade. Use care not to make the punch too thin, or else it will be likely to break. The other two sides of the punch should form a rather large angle, say approximately 60°. This is what gives the slit its wedge shape. Trial and error will lead you to a punch that is robust, yet produces a small, tapering slit.
As with the round holes, this punch can be used by hand, being driven into a punch block. Use of this punch requires a little finesse, however, because the smith must be able to feel when the tip of the punch has just broken through the other side of the ring flat.
Also as with the round holes, this punch can be mounted in a set of tongs. Mounting this punch in a set of tongs has the added advantage that you can control the precise depth that the punch penetrates by mechanically controlling how far the tongs can close.
The final step is setting the rivet. I have constructed a rivet setting/clinching tool from a pair of pliers. One jaw face is completely flat, while the other has two pockets ground into it. One of the pockets is quite deep. This is used for initially driving the rivet home, and functions as the pocket into which the tip of the rivet goes when it is driven through the ring flat. The second pocket is merely a shallow depression. This functions much like the die in a modern stapler. Once the rivet has been driven into place, the point of the rivet is placed in the second pocket and then the pliers squeezed. This has the affect of crushing the rivet point into this shallow bowl-shaped pocket, causing it to mushroom and permanently set the rivet.
The steps for setting round rivets are somewhat simpler than setting wedge rivets, for there is no need to drive them into place. Rather it is likely that a piece of wire was threaded through the rivet hole, the excess ends were nipped off, and then the ends were upset, either by hand or with some kind of rivet setting tong, as described before.
Burgess, E. Martin. A Reply to Cyril Stanley Smith on Maille-Making Methods. Technology and Culture, I, 2: spring 1960.
Burgess, E. Martin. The Maille Maker’s Technique. The Antiquaries Journal 1953.
Edge, David & Paddock, John Miles. Arms & Armour of the Medieval Knight An Illustrated History of Weaponry in the Middle Ages. Crescent Books: Avenel New Jersey 1996.
Ffoulkes, Charles. The Armourer and His Craft from the Xith to the XVIth Century. Dover Publications, Inc.: New York 1988.
Pfaffenbichler, Matthias. Medieval Craftsman ARMOURERS. University of Toronto Press: Toronto, Buffalo 1992.
Smith, Cyril Stanley. A History of Metallography The Development of Ideas on the Structure of Metals before 1890. The University of Chicago Press: Chicago Illinois 1992.
Smith, Cyril Stanley. A Reply to Mr. Martin Burgess on Maille-Making Methods. Technology and Culture [issue and/or date unknown].
Smith, Cyril Stanley. Methods of Making Chain Maille (14th to 18th Centuries): A Metallographic Note. Technology and Culture, I, 1: winter 1959/60.
Williams, Alan & de Reuck, Anthony. The Royal Armoury at Greenwich 1515-1649 A History of its Technology. The Trustees of the Armouries: Leeds 1995.
For information on the Internet concerning maille there is no better resource than Sara’s Chainmaille Connection. It can be found at:
Another great internet source is the Armour Archive. The Archive is an interactive bulletin board where armourers from all over the world gather to leave and reply to messages to one another to share information about armouring. It can be found at:
My own personal web page can be found at: