Advertisers and Sponsors of
Sixgunner.Com™
|
|
THE WARHORSE SERIES: #2 by Ed Wosika, Product Development Guy For THE HANNED LINE Published in the Nov/Dec 2001 issue of the CBA journal The Fouling Shot [ http://www.castbulletassoc.org ]
In my 1903 Springfield, I found a good test bed for an unlikely experiment. Like many older military rifles, it has a larger-than-normal bore (0.3035", in this case), although its groove diameter is reasonable (0.310"). My idea was to take a look at the phenomenon of the "nose crash." By this, I mean that, if the cast bullet’s (CB’s) nose diameter is too small to ride centered atop the tops of the lands in a "slip fit," then it is an unsupported (cantilevered) extension of the drive bands and the spin can cause it to bend sideways, suddenly, until it smashes into either a land or a groove. After that "EVENT," you will be shooting a banana out the muzzle. Mah!! The typical CB exits the muzzle at between 70,000 to 120,000 rpm. That’s smokin’, regardless of the linear velocity! The nose crash occurs when the rotational velocity exceeds some critical value — relative to the alloy strength, nose length, point mass, and nose diameter — and suddenly slumps to the side, plowing sideways, thereafter, into one side of the barrel. If, upon firing, the bullet locks on-line with the nose fairly well centered, then it will not crash. Such shots will all land close together. By contrast, the CBs that started out with a significant cant ("signify-cant," for short) to them are likely to suffer a nose-crash — the faster the muzzle velocity, the greater the percentage of shots that will crash. These CBs will land apart from that group made by the none-crashing CBs. If the CB’s nose crashes onto the top of a land, it will be somewhat bent and will land outside the group; whereas a CB whose nose crashes into a groove will be bent more severely and will land farther from the group. At least, that is how it seems to me. In order to maximize the crash potential, for the purpose of this study, I looked for a long-nosed CB with a nose diameter too small even for most 30 caliber rifles. I found that my Lyman #311332 mould, which casts a long-nosed CB with a 0.298" to 0.299" diameter nose. Perfect! Hey! Finally here was a way to be excited about a really depressing mould! This mould’s bullets were especially ill suited to the Springfield’s 0.3035" rifle bore. Better yet, this Springfield’s barrel has four narrow lands and four wide grooves. The good news was that I wouldn’t have to try too hard to make the CBs nose-crash. The Armored Nose Factor In order to spice up the experiment, I developed another factor: experimental nose armor! I made coils from 28 gauge copper wire wrapped around various sizes of drill shank [see (1) Figure 1: photo of coils; and (2) Graph #1: linear regression of resultant coil OD when using a given drill shank diameter]. The idea was to see if a CB with this copper coil cast-in-place along the surface of its nose would tend to have less serious nose crashes, as indicated by its fliers being located closer to the group, than with the bare-nosed CBs exhibited. Hey! Laugh all you want, but I’ve had crazier ideas than that really panned out. My motto runs something like, "Don’t be afraid to make a fool of yourself, for then you may fail to enjoy it!" Of course, this can also be translated as, "It’s MY experiment and I’ll CRY if I WANT to!" I’ve provided the graph (Graph #1) and the formula for making coils with this thin copper wire so that you can experiment with this for various calibers, if you so choose. The formula will only work with dead-soft 28 gauge copper wire. I wanted coils with around a 0.300" diameter, to fit snuggly into the mould’s nose cavity, so I looked for a drill with a diameter of 0.255" (= 0.300 x 0.8475) diameter, as indicated by the linear regression formula. The closest to that is an "F" size drill, at 0.257" diameter, so that’s what I used. The resulting coils were a good close fit in the nose cavity of the mould, as shown in Figure 2. I got the copper wire from MSC Industrial Supply Company [100 MSC Dr., Jonestown, PA 17038, 800-645-7270, 717-865-5888, Fax: 717-861-5810] — Item #31981616, at $3.07 for a 1/4 pound role (it goes a loooong way, as you only need around 3" per coil). I made good use of a pair of small smooth-jawed needle-nose pliers. They are essential for wrapping and handling these light, springy little demon-coils. You can usually find these at jewelry-making supply shops. Alternatively, purchase a thin-jawed set of needle-nose pliers with serrated jaws, and then grind off the serrations. To make a coil, grip the tip of the wire in the narrow tip of the pliers’ jaws and wrap the wire around the jaws twice. Remove this loop and use the resulting ring as a "handle" to hold the wire against the drill shank with your off thumb. Closely wrap the wire in a helix — I used seven wraps. On the last wrap, rotate your holding thumb forward to hold the last coil. Cut it off (the wire), leaving a 5/8" long tail. Use the small needle nose pliers again to grip the tip of this tail and wrap it into a ring, just like the one you started with. While you are wrapping this ring, keep the wire in tension. Lastly, remove the ring from the pliers and the entire coil from the drill-shank and bend down the ring at each end of the resulting coil so that it ends up near the coil’s center axis. This process takes just over a half-minute per coil, once you get into the groove. When you have placed a finished coil into the forward portion of each mould’s nose cavities (see Figure 2) —using the small, smooth-jawed needle nose pliers, once more — close the mould and pour in the molten alloy. Set your melter around 100 degrees hotter than the "too hot" setting for continuous casting. This helps compensate for the loss of heat while you place the coils. If you start to get wrinkles, cast a few no-coil CBs quickly. The wire, suddenly superheated by the melt, will grow in length and diameter. It will squirm into a perfect fit to the inside of the mould’s nose cavity. The alloy hardens while the coil is still longish. This traps the two little end-rings down in the center of the nose. As the bullet cools, the wire ends up with a good fit under strong tension. I calculated the total volume displaced by the copper, multiplied that by 2800 grains per cubic inch (the approximate density of wheelweight alloy) and subtracted half of that weight (to compensate for the fact that copper is lighter than wheelweights) from the mean weight of the bare-nosed bullet. The coil-nosed CBs weighed in consistently at that weight, indicating that they did not tend to have gas inclusions from bubbling of the extremely thin layer of paste soldering flux I put on the wire (I kept my thumb and forefinger, on the winding hand, just slightly smeared with just enough of the flux to tell that I might have some on, by feel). I believe that this extremely slight fluxing helped to create the near-perfect fit (almost a soldering) of the wire into the nose.
The Experimental Design The bullet alloy I used was mostly wheelweights, but had 20% linotype. It was so cotton-pick’n hard that I could only just barely bump up air-cooled CBs the 4+ thousandths it took to make the nose fit. Water dropped CBs were impossible, even right after casting.
I decided not to oven-harden the bumped CBs. Therefore, I had four groups to test: bare-nosed and coil-nosed water-dropped CBs with a 0.299"; and bare-nosed and coil-nosed air-cooled CBs with a nose bumped up to 0.303+". I intended for the two small-nosed bunches to show me whether the coil helped lessen the severity of the crash. I intended the two bumped-nose bunches for a completely separate experiment: to show me whether there was any aspect of the "speed limit" we so often find, with CBs, that is derived from side-slumping of the land-supported nose (if the coil-nosed CBs gave better accuracy, then the noses are crashing somewhat, even though supported). All bullets were sized, or sized-and-bumped, using the coaxial method I described in Article #1 of this series: with a Noze-First Top Punch [from the Hanned Line, P.O. Box 2387 - Cupertino, CA 95015-2387 || smith@hanned.com || http://www.hanned.com] and a 0.3105" sizing die modified by Robert Stillwell to have the Hanned entry cone (in the die) and nose-centering cavity (in the push-up rod) [Stillwell Tool & Die, 421 Judith Ann Drive, Schertz, TX 78154 || Phone (210) 658-0112 || rstillwe@texas.net || http://rstillwe.home.texas.net]. I also used another Hanned Line product, our own Paco Kelly’s excellent Apache Blue bullet lube — good shtuff, Maynard! I used ten-shot groups, starting with a low-velocity load that produced good groups with the water-dropped small-bare-nosed CBs (13 grains of Unique at 1430 fps). I then increased this load in four increments until most of the shots were filers (18.5 grains of Unique at 1780 fps). I then went back and re-shot a ten-shot group of each of these load increments with the water-hardened small-coil-nosed CBs. I could see a slight, but persistent, tendency for the coil-nosed CBs to produce a denser group, so I proceeded with the bumped-nose CBs in like manner, but using AA-2230 — a powder better suited than Unique to mid-velocity loads. For the starting load (27.5 grains of AA-2230 at 1730 fps), both the bare- and coil-nosed CBs did well, with the best group being with the coil-nosed CBs. However, additional increments of AA-2230 showed no advantage to the coil-nosed CBs, so I abandoned that aspect of the experiment as a no-show. Analysis Determining the "Miss" of Each Shot — In order to compare these targets, I had to deviate SERIOUSLY from the standard "group size" approach, which measures the C-to-C distance between the two most widely dispersed shots and then THROWS AWAY all the data from the other shots in the group. After all, it’s important to differentiate between a tight group with a flier and a large, even, shot spread of the same "group size." Using MINITAB, my statistical package, this is simple. Using PageMaker, I created an X versus Y grid of lines spaced 1/10" apart, with heavier lines to show the even and half inch positions. I then printed it onto a Mylar sheet. For each target, I numbered the bullet holes 1 through 10 (no special order), taped on the grid overlay, and then tabulated the X and Y positions of each shot to the nearest 0.01". I entered the Xraw and Yraw values in side-by-side columns in MINITAB, then used the program’s Compute function as follows, in two new columns (e.g., Xc and Yc). For each X value, I subtracted the mean X value; likewise, each Yc value = Yraw minus mean Yraw. This gave me X and Y values relative to the group’s center of mass. I then created another column for the radius of each shot from the center of mass. This was equal to the square root of the sum of Xc squared plus Yc squared. Given the radius of each shot from the center, I had the "Miss" of each shot — the radius of the shot, from the group’s center of mass, is how much it "missed." Given a "Miss" column for each target, I could then proceed with the analysis. Comparing How Much You Miss, Between Treatment Factors — This was a case of wanting to see if the point style (coiled versus not-coiled) had any effect upon how much farther I missed the point-of-aim (center of the group) as the velocity increased. It seemed clear to me, therefore, that this was going to be an Analysis Of Covariance (ANCOVA). However, I didn’t get much further than that without crashing and burning with MINITAB. So, I sent an e-mail off to my statistics teacher, Dr. Edward Gilroy, including a description of the project and a copy of my MINITAB project file. He put me back on the straight path and I should, now, be able to do this sort of analysis myself, in the future. What’s an ANCOVA? Do they bite? Well, you’ve seen a linear regression that determines — in the case Graph #1 — the relationship between increase in coil O.D. relative to the resulting increase in drill shank O.D., for coils made with 28 gauge copper wire. Linear regression also provides a formula for calculating the response variable (e.g., the coil O.D.) from the input variable (e.g., the drill shank O.D.). In addition, you may have heard of the "t-test" that examines whether a given sample is likely to have come from the same population as a reference sample. Well, an analysis of variance (ANOVA) is a multi-sample t-test that answers the question, "Is there an odd man out among these several samples?". The ANCOVA, then, is a combination of multiple linear regression (for two or more groups of data plotted per test) and an ANOVA. It can answer ALL of the following questions in a single analysis: 1) Is there a change in mean miss as velocity increases?; 2) With respect to the point forms (coil versus no-coil), does there tend to be a significant difference in the mean miss (as measured by a difference in where the regression line for each point type hits the Y axis)?; 3) Is there a significant difference between the slope of the two regression lines (for the bare-nose as opposed to the coil-nose regression lines)?; and 4) What is the formula for predicting the average miss, for a given velocity, for each point type? It’s about as close to doing witchcraft as you can get without suffering unfortunate crispyfying consequences. The upshot of it is synopsized in Graph #2, which shows the two regression lines’ theoretical perfect fit points (Fits) relative to each increment of increasing muzzle velocity (MV) — for the bare nosed-CBs and the coil-nosed CBs. The coil-nosed line has a shallower slope (more accurate = doesn’t increase its miss so much as velocity increases), but (the accompanying ANOVA, not presented here, shows that) the effect is NOT strong enough to be statistically significant. Therefore, there is no statistical reason to regard the two samples (bare-nose misses versus coil-nose misses, for any given velocity) as coming from two different populations (i.e., one being more accurate than the other). To be statistically significant, one usually looks for an effect that is strong enough that it would occur by chance alone only five percent of the time. The ANCOVA showed that the difference between the coil-nose and bare-nose CB’s intercept points (left side of Graph #2) had a 14.5% probability (it could have occurred by chance alone 14.5% of the time). The difference between the slopes of the two FIT lines had a 13.8% probability. Let me be clear: this analysis showed no statistically-significant difference between the accuracy of the two point types. Nevertheless, it came close to being statistically significant on both the intercept test and the slope test. It’s clear that nose-coils give no significant accuracy help with the very sloppy nose-to-land-diameter spread we had in this test. However, these results give me the idea that I should try this experiment again with a CB that has only a slightly sloppy fit and, perhaps, in a barrel that has more lands. In other words, I may have given the new technology too severe a test. Given a situation where the CB’s nose can’t crash so hard, it may show significantly better performance with the coil-nosed CB. We’ll have to see. I might even try something like using a heavier iron wire, rather than the light copper wire.
Conclusions Therefore, we have no resolution on the question as to whether, generally, an armored nose will help to accurize a loose-fitting CB, other than that we can conclude that it won’t be of much help if the nose-to-land slop is high (0.004"), as in the current study. On the other hand, the test of bumped-nose (bore-riding-nose) CBs, of both bare and coil-nosed persuasion, showed no advantage to each other. It appears that the "speed limit" we run into with CBs (somewhere between 1500 and 2200 fps, depending on rifle, bullet, alloy, and powder type), given a good CB-to-land-top fit, has nothing to do with nose crashing. That’s good to know. Therefore, although we have no definitive results, for THIS go-around, we have, I feel, learned some things we didn’t know previously. I believe this nose armor approach is one avenue that warrants further experimentation. Statistical methods can help to point us towards viable approaches, and away from false hopes. However, in the end, it is the consistently-performing discovery that remains The Holy Grail.
|
Navigation Links on
Sixgunner.Com™ |
