The Existing Test: Now that we know what we want from a test, let's take a look at
the four thrust test on each of these points (grading Good, Bad, or So-so):
Reproducibility: Bad. The four thrust test fails miserably on this key point. Since it is
administered by a person swinging a blade, it will be thrust with different force by different
people, and even by the same person on the same day. While it has been argued that one can
calibrate marshals in a small region so that they all hit with approximately the same impact,
there is no practical way to achieve this across the Society. Furthermore, since each one is
using a different blade, there is a different amount of energy absorbed by flex in each one,
and indeed that amount changes over the life of the blade. Most significant may be the
variation in backing material. In this case, the punch test is usually performed with the
armour on the ground. Clearly, there is tremendous variation in the firmness of the ground,
and I will not belabor that issue further. All these factors combine to mean that the punch
test will never work exactly the same way twice. Further, since energy is delivered
subjectively (by the marshal) rather than objectively (by a measurement), it introduces the
possibility of the marshal, consciously or subconsciously, affecting the results of the test.
And there is no way to prevent that.
Accurate representation: Bad. While we are testing the armour for the exact failure we are
trying to prevent, the variation of the backing substance (the ground) undermines this effort
entirely.
Simple: So-so. While the procedure is simple enough, it relies on the marshal delivering the
proper amount of energy to the sample, with no objective calibration method.
Inexpensive: Good. Most people have or can obtain a broken foil.
Portable: Good. What's one more blade in your bag?
Now, this is not the greatest report card in the world, but we used the four thrust test
because it was all we had. However, I believe we now have better options, and we should
strongly consider them.
Option 1 - The Drop Test: This is the test that was developed by Don Dylan ap
Maelgwn and myself and was used in our report entitled Energy Tests that was
attempting to determine the resistance of armour to schlager penetration. However, there
were several by-products of that experiment. One of them was a test that is much closer to
the ideal than the four thrust test. I have included a copy of that report for your
convenience.
(Dylan and Thomas' Punch Test)
The test worked as follows: A broken foil blade (with a known surface area of the impact point) is mounted with the tang in a coffee can. The armour sample is placed over the open end of an empty coffee can, and secured with a screw clamp (a.k.a. hose clamp, radiator clamp, etc. This was a large clamp, available in home stores.) The blade was then brought up to a given weight by placing loose nuts (the nut-and-bolt kind, not the walnut kind) into the can. The blade was then dropped (point down, obviously) onto the sample down a copper tube. The tip of the blade started exactly 30 cm (1 ft) from the sample. The copper tube kept the blade on track, and prevented any flex in the blade, although little was likely to occur. (The exact parameters and construction of this apparatus will be discussed in a later section.) Comparing this test to our criteria:
Reproducibility: Good. Because the force of gravity is used to impart energy to the drop, and then to the sample, it will be accurately reproduced anywhere on the planet. (Lunar fencing is beyond the scope of this discussion.) Of course, the accuracy is limited to the accuracy of the measurements made, but that's still pretty good. Since the equation is linear (refer to Energy Tests for a full discussion of the physics involved) an error will only produce a linear error (as opposed to the square of the error, or some such). There is no variation in the backing of the sample, since air is used, and while not the same everywhere, is close enough for our purpose. The only possible variation is in how tightly the sample is stretched.Accurate representation: Good, with some explaining. The first thought is "It's good. You're hitting the sample with a broken blade that's moving." But then the question is "Is hitting the armour with nothing behind it the same as hitting it when there's a body in it?" This is a very important question. Consider Experiment 2 of Energy Tests, wherein the intrepid scientists attempted to answer that very question by placing the armour sample over a Leg o' Lamb(. The data indicated that there was a correlation between the results between the backed and unbacked samples. This means that the unbacked sample can give us the same data regarding puncture resistance as a backed sample, provided we recalibrate the test properly. (It is also a lot more convenient that carrying a Leg o' Lamb( around with you. Admitedly, the 'blood stars' caused by this are a lot more interesting than the 'grass stars' caused by the four thrust test, but spoilage makes this point moot.)
Simple: Good. Easy to assemble, and easy to execute.
Inexpensive: Good. This rig was made with a pair of coffee cans, some nuts that were in the garage, a clamp (under $2.00), an old piece of pipe, and some duct tape. The only unusual piece of equipment involved is a small scale which is only needed when making the apparatus. Portable: So-so. While the gear isn't very big, it will weigh a few pounds.
We have here a test that scored four 'Good's and a 'So-so'. This is a lot better than the four thrust test, which got two 'Good's, two 'Bad's, and a 'So-so', scoring 'Bad' in the two most important categories. I submit that we now have available a test that dramatically out-performs the four thrust test.
The specifications: These are the specifications for the drop test. This will explain what parts of the test are important to have the same, and what parts can be varied. The critical parts:
The drop height and weigh of the drop unit. (See Calibration below.)
The opening of the sample can.
The tension on the sample.
The surface area of the blade.
The pipe
The screw clamp for the sample
Obviously the drop height and weight are directly related to the calibration, and that section will cover those issues. The samples in the tests were mounted over a standard 23 oz. coffee can (that's the bigger one) with a diameter of 13 cm. If this dimension changes, the fabric will have more or less opportunity to absorb the energy delivered by the drop by stretching. This would affect the calibration of the test. Therefore this dimension must remain consistent. The tension on the sample can likewise affect the test. This is the most subjective part of the test. Samples used were placed on the can to be taut, but not overly stretched. If you pushed your finger into the sample, you could displace the sample less than one inch. This is a good guideline for tightness. The surface area of the impact point is a bit difficult to measure being so small, but can be adequately controlled by specifying approximately where the break is in the foil blade. In these tests, the break occurred 22.5 cm from the tip. A range of plus-or-minus an inch is probably adequate. As for the pipe, the exact form is unimportant, but the fact that the blade is guided and prevented from bending is. The screw clamp should remain consistent, since the method of securing the sample will directly affect it's ability to stretch.
The non-critical parts:What you use to weight the drop
How you attach the weight to the drop
Overall length of the blade
What's in the bottom of the sample can
The exact form of weigh is irrelevant. Mass is mass. We had used nuts so we could easily add and subtract. But if this test is used, the weight will be fixed. Similarly, the method of attaching the weight to the blade is unimportant, as long as the weight of the unit is controlled. The blade can even be shortened for convenience, as long as it is shortened from the tang, not the tip. Something should be placed in the bottom of the sample can to prevent poking through it, but the marshal can use whatever is handy.
The calibration:
Drop height: 30 cm (1 foot)Drop weight: 1.7 kg (3.74 lbs.)
Once we determine the usefulness of the test, we must address the issue of calibrating it. In this case, we need to calibrate the test so that acceptable armour passes, while unacceptable armour fails. We do that by determining how much energy we are going to deliver to the sample. Too much will cause "good" armour to fail, while too little will cause "bad" armour to pass. When determining the energy, you must calibrate the height and the weight. 30 cm is a very convenient height, and I've chosen to leave it there and focus on the weight. Consider the following data excerpts:
Sample Weight
3 layers trigger 1.6 kg
4 layers trigger 2.2 kg
1 layer denim 1.4 kg
2 layers denim 2.3 kg
Leather #496 front 2.1 kg
Leather #523 front 2.7 kg
These are all the weight required to get the armour to fail. We are calibrating this test against the existing 4 layer/4 oz. standard. Therefore we want the test to penetrate the armour for 3 layers, but not penetrate four layers, or a reasonable 4 oz. leather. (Leather #496 was determined to be 4.5 oz. leather, and leather #523 was determined to be 4.1 oz. leather.) Therefore it is my belief that a weight of 1.7 kg will certainly penetrate 3 layers, while not penetrating 4 layer samples or variations in 4 oz. leather. In our testing, a change in weigh of .1 kg (100 grams) made a significant difference. Therefore, if three layers failed at 1.6 kg, and a sample must pass at 1.7 kg, I believe we can be sure of an adequate margin.
A Construction Guide:
(This is a guide to construction and use. This is intended as a sample that could be added to a Kingdom's Marshall's Handbook so that someone who has never seen the test could build and use it.)
The drop test is a replacement for the punch test and is designed to test the armour's ability to resist penetration with a broken blade.
Construction: The simplest way to build the thing is to get the following materials. Important measurements are in bold.
1 coffee can - regular sized 'not-quite-a-pound' can
1 coffee can - 23 oz. (It should have a 13 cm diameter)
1 broken foil blade - the break should be clean, flat with no sharp pieces, and the break should be anywhere between 8 and 10 inches from the tip.
1 pommel
1 screw clamp (hose clamp, radiator clamp) big enough to go over the big coffee can with some room to spare.
1 piece of copper pipe, 14-18 in. long.
Something for weight. You can use loose screws, sand, or whatever you have.
Put a hole in the middle of the bottom of the small coffee can. Put the can on top (tang end) of the broken foil. Use the pommel and some duct tape to stabilize the can. You may find it useful to put a foil bell on the blade before the coffee can, and/or to use a handle on the blade. Now put the whole thing on a scale, and start adding your screws (or whatever) until the whole thing weighs exactly 1.7 kg (3.75 lbs). Once you have your screws in the can, you can tape the whole thing up so they don't come out, but remember to account for the weight of the tape if you're using a lot. Now put a mark on your pipe exactly 30 cm (1 foot) from one end. Since you won't be able to see the tip when it's inside the pipe, you might want to lay the pipe and blade parallel. Put the tip at the mark and the short end of the pipe up toward the tang. Put a mark on the blade where the pipe ends. This mark should be at the top of the pipe when you are ready to drop when you are testing. That's all there is to it. (If you test a lot of armour that fails, you will be punching holes in your big can. You can put something in the bottom to stop the blade if you like.
Performing the test: Set your armour sample on the big can. Clamp it on with the screw clamp. The armour should be taut, but not overly stretched. A good guide is if you can't move the armour more than an inch into the can with your finger. Make sure the clamp is tight. Get someone to hold the pipe resting (not pushing) on the armour. Put the end of the blade into the top of the pipe, and align your mark. It should look like the diagram here. Drop the blade. The blade should fall straight down and onto the sample. (The person holding the pipe should hold it firmly lest they find themselves beaned on the noggin as the weight tips to the side after it hits.) If the blade penetrates the armour (metal comes through) then the armour fails. If not, then it passes.
Option Two - The Spring Test:There is another option in a small tester designed by Garrick Mapmaker that uses a spring to generate the energy. The spring is in a small tube with a dowel and a nail. The armour is clamped on one end of the tube, over a place with a hole. The whole thing is then pushed against a solid surface until the end of the spring is fully in the tube. If the armour is not penetrated, it passes. You already have a working model of this tester, so I will not be so specific with the details. Comparing it to our criteria:
Reproducibility: Good. The test is based on machined springs, and therefore should perform very consistently. The only variable is the give or stretch in the sample, which is something none of the tests can address.
Accurate representation: So-so. This tester applies a static, (sitting still and pushing) rather than dynamic (hitting in motion) impact model. I am not enough of a physicist to say that they are or are not the same.
Simple: So-so. Using the test is simple. Construction is somewhat more complicated.
Inexpensive: So-so. The parts are not terribly expensive, but they are not readily available, and it takes some doing to put together.
Portable: Good. A very small and convenient unit.
This test also seems to be a good step up from the four thrust test. It's problems are in construction, but it has pluses in portability. I would propose that we consider this an acceptable alternate test.
The Proposal: My proposal comes down to this: The four thrust test has many problems. I propose that we eliminate the four thrust test and replace it with the drop test. The drop test would become the new Society wide standard for armour testing. However, that is a large transition to make in a Society the size of ours. A more logical alternative would be to declare that these tests are acceptable substitutes for the four thrust test. This would let those who wished to continue to use the four thrust test while allowing people to begin using the drop or spring test. It also give us a chance to work out any bugs in the new procedures. Once the new tests are established and the norm, the four thrust test could be phased out.
The Next Step: Clearly this decision will not be one made without careful consideration. However, I feel strongly that this proposal should be thoroughly considered and a decision made one way or the other, rather than having it fall by the wayside to be forgotten about. I would offer the following action plan for this proposal, only as a suggestion for your consideration: