In the first three parts of this article, I talked about the ships and their guns, and the teachings of ballistics concerning what affects the range and accuracy of projectiles. It's time to look at the projectiles themselves.
Cannon projectiles have been made of stone, cast iron, lead, brass, bronze, copper and even glass. As of the Ring of Fire (RoF), stone or cast iron were the norms; lead was used in small arms.
Stone has the advantage (see part 2) of being compatible with lighter artillery. The problem was that the stone shot had to be hand-carved, which was expensive, whereas iron could be cast. Hence, economical considerations led to the abandonment of stone-throwers in Europe in the mid-seventeenth century, and among the Ottomans in the early-nineteenth century. (Glete, 25; Guilmartin 186). That said, the Dutch navy "presently" includes at least thirteen 1620s vintage ships equipped with stone throwers (pedreros) as well as more modern cannon (sailingwarships.com). As for lead, as George Ripley said (New American Cyclopaedia 1870), lead "is too dear and too scarce for cannonballs."
Molding. By RoF, cannonballs are made by pouring molten iron into molds (cp. de Tousard 349). In canon, Reardon and Torstensson agree that the precision of the bore on Reardon's field guns is overkill. Reardon complains, "those cast iron cannonballs are so sloppy and uneven there's no point. We'd be trying to make a silk purse out of a sow's ear." And Torstensson agrees that he doesn't "have cannonballs to match that precision."
So, what's wrong with seventeenth-century cannonballs? Just because a cannonball is supposed to be a homogeneous sphere of a particular diameter doesn't mean it actually is. It can be over- or undersized, it can be lopsided, and it can have cavities.
"Solid" shot is usually not entirely solid; the outside cools first and contracts, leaving a shrinkage cavity inside. If this cavity is off-center, then that will affect how the shot flies. As a result of the shrinkage cavity, the apparent density of seventeenth-century cannonballs was less than would be expected, given the density of cast iron. For example, a study of an English Civil War cannonball found that it weighed 14.25 pounds and was 4.92 inches diameter, for an average density of 6.13 g/cc (Walton 143); for solid cast iron, 7.2 would be expected.
There will be other imperfections too. Air bubbles in the molten iron can form additional cavities. Williams cast a cannonball in the laboratory and, slicing it in half, found three large defects clustered together; these would have shifted the center of mass away from the center of pressure, thus altering the ball's aerodynamics.
There will be a seam where the halves of the mold met, and a disturbance at the vent sprue where the excess metal escaped. Other problems are created by impurities in the melt. It may be possible to remove some asperities by rolling or hammering the shot while it is heated.
Round shot cast in sand molds were superior in smoothness, sphericity and homogeneity to that cast in iron molds, possibly because the pores of the sand allowed air to escape ((Jeffers 119; Hebert 595). Iron molds came first and one source implies that sand molds weren't introduced until 1824. (Culmann).
Cannonball quality control was primitive. The simplest scheme was to pass the ball through a circular "go" gauge (lunette); balls that were too large to fit were rejected, as they presumably wouldn't fit into the bore.
The Gribeauvalist innovation (circa 1765) was to also insist that the balls not pass through a "no go" gauge. (That doesn't mean that the artillerists of the past uncritically accepted undersized cannonballs, just that they judged whether they were large enough by eye.) The French artillery service wanted to set a spread of six points (one point=7.4 mils) between the two gauges, but were forced (by political pressure) to settle for nine. Of course, all this supposes that the gauges were identical. And so each arsenal had a master gauge, and the working gauges couldn't diverge from it by more than two points.
What about lopsidedness? The "go" lunette was replaced by a tube five-calibers long, fastened obliquely to a support in such a manner that the ball would have to be rolled through the tube.
Shot Towers. The shot towers used in the nineteenth century to cast lead musket balls aren't practical for making cannonballs. Whether the shot be iron or lead, you heat it to a particular temperature above its melting point at the top of the shot tower. Essentially, you are achieving sphericity because the molten droplet falling through air is in free fall—zero gravity—and thus assumes spherical shape. The fall must be long enough for it to solidify without losing that shape.
The "heat content" to be lost will be the latent heat of fusion (over ten times as high for iron as for lead) multiplied by volume (thus proportional to diameter cubed). The heat transfer will be proportional to surface area (thus to squared diameter). Thus, the required time of fall for solidification will be proportional to the diameter. The average velocity will be roughly half the terminal velocity which in turn is proportional to the square root of the diameter, and the necessary height is approximately the fall time multiplied by the average velocity, i.e., proportional to the 1.5th power of the diameter.
Lipscombe suggests that the fall had to allow additional time for the balls to cool below the boiling point of water, so when they landed in water for final cooling, they wouldn't create steam. That of course would make significant the difference between the melting point of the shot and the boiling point of water, and this would be greater for iron. Lipscombe says that "it takes about one-third of the height of the tower for it to solidify and the remaining two-thirds for it to cool down."
The shot tower drops were substantial, typically over 150 feet for making lead musket balls. In 1801 (shot towers were invented in 1782), British lead musket balls had a diameters of 0.68 inches for muskets and 0.89 for wall pieces. (Egerton 30). Consider now how much more of a drop would be needed for an iron cannonball with a diameter of 3.49–6.68 inches (for 6–42 pound balls)(Collins).
Case Shot. This is a generic term for several different kinds of anti-personnel munitions. Smaller projectiles are packed inside a bag (hailshot, grapeshot) or a cylindrical metal (tin) canister; the interstices of the canister were usually filled with sawdust. If the shot was equipped with explosive for spreading the goodies further, it was called a shrapnel shell. The projectiles could be lead or iron balls (hence, "grape"), cubes ("dice shot") or scrap metal. Canister shot was used by the British Navy at least as early as 1625 (Lavery 137).
Typically, the mini-projectiles of grapeshot were larger than those of canister; during the American Civil War, grapeshot might deliver nine lead "golf balls" to the enemy, whereas the canister might hold hundreds of lead "marbles." (McNeese 74).
The positions of the ships relative to the wind can affect the choice of projectile. If ship A is upwind of ship B, then ship B is heeled away from A, exposing more of its lower strakes, but hiding its deck. It's better, then, to try for hull damage. B, on the other hand, should note the greater exposure of A's deck, and perhaps essay a double charge of case shot, to do execution on A's crew. (Douglas 240ff).
Heated shot. The logic behind heated shot is impeccable; warships are very flammable (wood, cotton, tar, hemp); start a fire and the ship's own substance will do the rest of the work for you. That same logic tended to work against the heated shot being fired from a wooden ship. For that reason, the British Navy prohibited the practice.
Still, warships were occasionally given this capability. For example, Captain Charles Stewart of USS Constitution decided to equip his new command with a portable sheet iron furnace for heating shot. He intended to use it only in an emergency. (Berube 73).
The first documented use of heated shot by a warship was actually earlier, by the French at the Battle of the Glorious (to the English!) First of June (1794). It should be noted that even if there is no misadventure in handling the heated shot, there are risks in having a shot furnace on board. In that battle, the Scipion's furnaces "were knocked down, and the hot shot in them scattered about the deck. . . ." (James 240).
The first truly successful use of shipborne heated shot appears to have been by Hastings' steamship Karteria, in the Greek War for Independence. She reportedly fired 18,000 hot shells(!), "mainly against shore batteries," in her first year of service. (Dakin 172), Other sources suggest that it was 18,000 shots total, but it was still remarkable that she did this without any injury to the gun crew. (Roberts 76). And he certainly made effective use of both heated shot and carcass (see below) at the Battle of Salona (1827). (Blackwood's 58:510). See also EB11/"Hastings, Frank Abney."
Carcass. This was an incendiary that could be fired from a ship. Incendiary projectiles have a long history in warfare. A relatively recent precedent was the incendiary that set fire to Ronda during the 1486 siege. There is some dispute as to whether this was thrown by a trebuchet (Turnbull 58) or fired by a cannon (Partington 123). Hand-thrown bombs containing "wildfire" ("Spanish pitch, black pitch, saltpetre, sulphur, camphor, turpentine, rock oil and ardent spirit") were employed by the defenders of Famagusta in 1571 (135). Fire pots (clay pots presumed to have contained incendiaries) were recovered from various sixteenth- and seventeenth-century wrecks. (Hamilton).
In one embodiment, the cannon ball was covered with a flammable concoction and wrapped up in a bag. A fancier version was a hollow metal sphere filled with the material and having holes for both igniting the stuff and having it "jet out" against the target. A "spike shot" added spikes to the standard carcass; the theory was that these would stick into the target. (Kinard 125). In the 1796 British carcasses, the incendiary was a little less than 10% of the weight of the projectile, and burned for 3–12 minutes (Beauchant 66).
Double or Triple Shot. You may load two or even three shot into the cannon, and fire them simultaneously. The muzzle velocity is reduced unless you increase the charge to compensate. (Douglas 65). Accuracy was less than with single shot, as the balls would push each other in opposite directions, vertically or horizontally, when leaving the barrel. (67). The shots can easily land more than a hundred yards apart (285).
Bar, Chain and Star Shot. Two balls (double-headed shot) or half-balls (split shot) could be joined together by a chain or bar. The bar itself could be rigid, sliding, or jointed. The theory was that these would be more effective against sails, masts and rigging. Both were in use prior to RoF, as evidenced by a display at the Stockholm Medieval Museum. All dismantling shots were more likely to be effective with a fresh wind than in light airs, as the force of the wind would tend to open up a small hole. (Douglas 253).
Ready-to-fire split shot was found on the shipwreck of the Nuestra Senora de Atocha (1622). The hemispheres were joined to form a ball and the connecting chain was doubled over, splinted, and tied with twine (which would have been burnt away by the fired powder). (Malcom).
Chain shot was reportedly used by Frobisher's ships against the vanguard of the Spanish Armada (Mcdermott 252). For that matter, it was also used by the Protestants in the defense of Magdeburg (1631). Chain shot had a greater spread, perhaps 2–6 times the calibre of the gun firing it. Variations included star shot, in which several "heads" were linked by chain to an iron ring. It certainly could be effective; at the Four Days Battle (1666), during the Second Anglo-Dutch War, the Dutch warships Callantsoog and Reiger used it to destroy the rigging of HMS Swiftsure. The closest modern equivalent of chain shot is the bolo round; lead balls connected by steel wire.
In practice, there were a number of problems with the use of chain shot. For example, the balls could break loose from the chain. And if it didn't, the aerodynamic behavior of the projectile was peculiar, causing both accuracy and range to suffer.
There was experimentation as late as the twentieth century with double-barreled cannon for firing bar or chain shot (the target, by then, was an airship or airplane). The catch, of course, was that if there was a difference in timing, the damage would be to the attacker. Some inventors recognized this, and designed diverging barrels with a single powder change so a single charge would act on both balls (Jeffees USP 24518), but so far as I know none of these contrivances was successful.
There were also attempts to improve range and accuracy by expedients to delay the expansion of the shot. For example, the half-balls could be hollow, with the chain coiled inside, and held together by some fairly flimsy element that would be rent apart by a small, timed explosion.
Cross-barred shot was also intended for cutting rigging, but it was a ball with two opposed spikes.
Round shot vs. specialty shot. In 1603, the English Ordnance Office paid 8 pounds a ton for cast iron round shot, sixpence to two shillings apiece for stone shot, and 2s 6d to 8s each for jointed shot and crossbar shot. (Oppenheim 160). In 1627, the price of round shot was 11 pounds a ton, whereas powder was 5 pounds/barrel. (301).
Some inkling of the relative importance of the ordinary round shot and the various specialty shot may be gleaned from a 1592 inventory of the British Ordnance Office Stores:
Table 4-1: Shot Inventory (Britain, 1592)
"hollow shot armed with fireworks"
lead-covered falcon shot
In 1632, a second rate carried "three lasts of powder, six cwt. of match, 970 round, 100 cross-bar, 70 double cross-bar shot. . . ." (Oppenheim 289).
In land warfare, "as late as 1854, as much as 70 percent of all cannon fire was solid shot." (Bailey 156).
Elongated shot. The original elongated shot was oblong shot, a cylindrical bar with hemispherical ends, thus somewhat similar to bar shot. Douglas recommended using it from nine- or twelve-pounders at close range, to carry away a mast. (61). It had a reduced range relative to round shot of the same caliber when fired with the same powder charge, but Douglas preferred it to bar shot. (63, 304). Had the powder been increased in proportion to the weight of the oblong shot, I suspect it would have ranged further.
Elongated shot have higher sectional density than spherical shot of the same diameter and material. While, as previously discussed, that's disadvantageous from the point of view of interior ballistics, it's advantageous from an exterior or terminal ballistics standpoint. The greater mass means less deceleration from air resistance during flight, and this can be improved upon by giving the shot a more aerodynamic shape (see below). And less deceleration means higher impact velocity, which together with the higher sectional density means more momentum and kinetic energy, hence more damage.
Elongated shot is not aerodynamically stable; a slight disturbance causes it to tumble. Hence, such shot should either be spin-stabilized or fin-stabilized (like an arrow or a rocket) . . . or both.
Fin Stabilization. Das Feuerwerkbuch (fifteenth century) presents an oft-derided illustration of a soldier shooting a fire arrow from his musket. (Walton 150). The notebook of the Elizabethan gunner Edmund Parker goes a step further, suggesting firing arrows from cannon. One approach was to fire multiple arrows at one time; he calculated that a demi-culverin could fire a dozen "massive" arrows, with shafts an inch thick and 16–18 inches long. Alternatively, a piece with 4 1/8" bore and a pound of good powder could fire what amounted to a ballista arrow, four feet long and with a two foot forked metal head, to "cut the shrouds, masts, yards or sails." (Walton 253ff). It isn't clear whether Parker actually implemented these suggestions.
The problem was that gases could rush through the gaps between the fletches and arrowhead blades rather than push the projectile out. The solution was the tampion, wadding rammed down after the powder; it created the necessary seal. The July 19, 1589 inventory for the Tower of London shows 570 dozen "tampions for muskett arrows." (Id.)
For a projectile to have "static stability," its center of pressure (where the aerodynamic forces effectively act) must be at least one caliber (Barrowman) to the rear of the center of gravity. This can be achieved by shaping the projectile so most of its weight is forward ("dart stabilization") or by using fins to create aerodynamic forces (drag or lift) at the tail. The further back the fins are, and the greater their span, the more they shift back the center of pressure. If the fin-stabilized projectile yaws so its axis is not tangent to the trajectory, aerodynamic forces push its tail so it's "back on track." This is the same effect which causes a weathervane to turn to point in the direction the wind is blowing.
Finned projectiles may be fired from smoothbore cannon, and so are of considerable interest, at least until rifled cannon are in production.
So let's look at the choices in more detail. We can use an essentially cylindrical projectile that, at the rear, has fins that stick out, i.e., the fin wing span is greater than the diameter of the main body. If so, then the projectile is effectively sub-caliber (the fins just fit inside the bore) and you need a full-caliber sabot of some kind to seal off the propulsive gases and keep the projectile centered. Certain anti-tank rounds fall into this category. "Fin-stabilized projectiles are very often sub-caliber." (Asfaw, 9).
An alternative would be to engineer the cylindrical projectile so it was full caliber, but the fins fold or spring out once it clears the bore. Another is to have a "tail stick," attached to the back of the full caliber cylinder, to which the equal diameter fins are attached. And finally, we can weight the main body forward, perhaps by giving it a full-caliber teardrop shape, and then the fins are mounted on the tapered aftbody.
Instead of a plurality of fins, you can put a cone on the tail. However, these seem to be used just on practice rounds. While range is reduced, that's attributable to the holes in the cones.
Artillery use of finned projectiles evolved slowly. Hunt developed a "life saving projectile"; it was finned and fired from a cannon. Accuracy data (1878) is available and it wasn't significantly more accurate than either the same projectile without fins, or the fin-free Lyle projectile already in service. Hunt projectiles tumbled in flight even though the static stability margin was two calibers.
In WW I, the Italians had a trench mortar that hurled finned shells. (Thompson 156). These weren't used with sabots; they were just dropped in and fired by hitting a percussion fuze.
Finned mortar shells were used by multiple powers in WW II; unlike their rifled counterparts, they don't turn over when fired at above 60 degrees elevation. (Smith).
In WW II, there were the experimental "Peenumunde" arrow shells fired from smoothbore railway guns. Muzzle velocity was very high, 5000 fps, and range was 94 miles. Fins were 31 cm across; the main body was 12 cm diameter; there was a discarding sabot.
During the Vietnam War, the "Swift Boats" made use of a muzzle-loaded, smoothbore 81 mm mortar, mounted on a tripod. Admiral Simpson, other Vietnam vets, or the Grantville wargaming group may be aware of their use to "flat-fire" finned projectiles.
In trigger (lanyard) fire mode, this mortar could be used for low elevation fire, ranging to over 1,000 yards. (Stoner). The ammunition it used included the M43, M362 and M374 high explosive rounds, all finned. These were highly elongated, with the aftbody perhaps twice as long as the forebody, and the span of the fins equaling the maximum body diameter. The body had a full caliber obturating band so there was no need for a sabot. (Cooke).
More recently, the wide-finned, subcaliber "long rod penetrator," made of tungsten or depleted uranium, fired at 1400–1800 meters/second from a smoothbore gun, has appeared as an antitank munition. (Denny 159).
Finned projectiles occupy a fairly small niche in modern warfare; they are handheld rocket, mortar and, sometimes with discarding sabots, tank weapons. So I tend to doubt that they are superior to rifled projectiles for all artillery purposes. But in the short term, they provide a "fast track" to improved accuracy, it being cheaper to replace projectiles than guns.
Spin Stabilization. For spin-stabilization, the shot is fired from a rifled cannon (see part 1), and typically engages the rifling by means of a driving band. The length/diameter ratio may be as high as 5:1. It's possible to put both a driving band and fins on a single projectile.
Strictly speaking, it's also possible to achieve spin stabilization without rifling. For example, bent fins can be used to cause a finned projectile to rotate. The Zalinsky dynamite gun projectiles had a long tail with canted vanes on it. These would be expected to stabilize both like a fin by vaning and also by inducing rotation. Mean range errors looked reasonable to me—19–88 yards at ranges 800–2000 yards. Strangely, the worst errors were at the lower ranges, possibly because of problems with the pressure control of muzzle velocity.
Armor-piercing shot. Paliser shot (1867) had chill-hardened cast-iron ogive ends, to penetrate wrought iron armor of early ironclads. When defenders adopted steel armor, the points had to be made of forged steel—initially carbon steel, later nickel-chrome or tungsten steel. AP shot were ultimately replaced by AP shells.
Shells are hollow shot, round or elongated, filled with explosive (or something equally nasty). The explosives are set off by a fuze; fuzes are discussed in part 5.
As of the RoF, shells were mainly used in land warfare, and were fired on high angle trajectories from mortars, which were short bore artillery designed for this purpose.
In 1682, the French Navy used its new galiote a bombe (bomb ketch) to bombard Algiers (quite successfully, I might add). The first bomb ketches had forward-pointing side-by-side mortars, and you essentially aimed the ship to aim the mortars. Later British bomb ketches had mortars on rotating platforms. In the eighteenth century, bomb ketches were the only ships making regular use of explosive shells.
The first significant use of flat-trajectory shell guns was at the Battle of Sinope (1853), where the Russians used them against the Ottomans' wooden ships.
EB11/Ammunition distinguishes between "common" shell, filled with gunpowder, and high explosive shell. Common shells have already appeared in canon; the USE ironclads fired a ten-inch cast-iron shell with a filling of black powder. 1634: The Baltic War, Chap. 38. As of 1853, the established bursting charge for a British 32-pound shell was just one pound, and a 10-inch shell would hold 5.5 pounds. (Experiment 9).
Shrapnel shell (1784) is a common shell in which lead bullets are mixed in with the gunpowder. The original version was improved upon by separating "the bullets from the bursting charge by a sheet-iron diaphragm." The charge was reduced so the bullets wouldn't spread as widely, and the shell was weakened by internal grooving to compensate. Also the balls were embedded in a soft matrix, such as melted resin or sulfur, "to prevent them from rattling around and perhaps cracking the shell prematurely." (Peterson 80). The great advantage of the shrapnel shell was that it meant that an anti-personnel load could be delivered at long range. This is probably more useful in land warfare than at sea, but if a warship were supporting an amphibious operation, or trying to fend off pirates before they came into small arms range, it might come in handy. Moreover, World War I and Spanish Civil War experience showed that high explosive shells are just as effective against soldiers as shrapnel shells, and of course more effective against other targets. (Weir 92).
The Korean and Vietnam War 105 mm beehive round is the modern equivalent of a shrapnel shell; when it bursts, it spews out thousands of flechettes, essentially small metal darts.
The first high explosives, such as TNT, were too sensitive for artillery use, absent an expedient such as compressed air propulsion. While TNT could be desensitized with beeswax, by 1911, a high explosive shell typically used picric acid in some form (dunnite, emmensite, lyddite, melinite, picrine). Picric acid was synthesized from indigo in 1779, and from phenol (a coal tar constituent) in 1841. It's a less complicated synthesis than, say, chloramphenicol, which is already in canon. Picric acid is a bit more than 2–3 times as powerful as black powder.
But we can leapfrog it. You see, by winter 1633, Brennerei und Chemiefabrik Schwarza was making small quantities of RDX—the main ingredient of C-4. Offord and Boatright, "The Dr. Gribbleflotz Chronicles, Part 2: Dr. Phil's Amazing Essence Of Fire Tablets" (Grantville Gazette 7). RDX is about 69% more powerful than picric acid. (Akhavan).
The power of the burst is roughly proportional to the square root of the weight of the bursting charge, which was proportional to the projectile weight and therefore to the cube of the diameter. (NAVWEAPS). A typical bursting charge was at least 6.5% (Okun) and not more than 25% of the projectile weight. (Hempstead 893).
To prevent premature detonation of even these stabler high explosives, the walls of the HE shell had to be thicker than those of a common shell, and a strong material had to be used.
In WW II, the rule of thumb was that shells made of steel with a 23-ton yield strength could have a 15% HE fill, whereas those made of a lesser steel (19-ton yield strength) had to be thicker, leaving room for only a 7% fill. (EvansN).
By WW II, HE shells had replaced shrapnel for anti-personnel use. The effect is dependent on the number of shell fragments, the mass of the fragments, the initial fragment velocity (typically 3000+ fps in WW II), and the directionality of the blast. There's a tradeoff, of course, between lots of small fragments and a few big ones. Also, between fragment velocity and fragment size. For a given initial fragment velocity, large fragments travel further.
British WW II studies showed that the optimum anti-personnel fragmentation was achieved with quite small fragments, generated by a bursting charge of 25%, but the metallurgy wasn't then equal to the task of firing such a shell from artillery. (EvansN).
Armor-piercing shells carry only a small bursting charge (under 5%, Hempstead; 2–3% EB11/Ammunition; 1.4–3.5%. Okun). Initially, this was black powder, out of concern for premature detonation (Alger) but that was replaced with high explosive. They carry base fuzes, rather than nose fuzes, to protect the fuze from damage by the armor. These would be delay fuzes so the explosion would occur only after penetration. In the twentieth century, a typical setting would be 0.03–0.07 seconds after impact. (NAVWEAPS).
Some literature also refers to a rather ill-defined intermediate class called "semi-armor piercing shells." These appear to be common shells made of forged steel (thus having greater penetration) and having somewhat larger bursting charges (3.5–6.5% per Okun).
Diving shells. The Japanese type 91 shells had caps that broke away on water impact to reveal an inner blunt head suitable for a stable underwater trajectory, and extra-long fuzes to give them time to reach the unarmored underwater hull. Well, that was the theory. While it was successful against the USS Boise (1942), it could only be used at moderate ranges (so that the velocity on water impact was high enough) and there was a risk that the shell would pass under or even through the target without exploding. (Evans 263ff). Really, they needed a combination proximity/depth fuze.
A shell could be filled with an incendiary concoction instead of explosive. The compositions used were quite similar to those I discussed in connection with carcasses. One variation on the incendiary shell was the illuminating shell (star shell). Another was the smoke shell; the French had one in 1852, and it produced smoke "of a dense and distressing nature," but not dangerous to life. (Experiment 14).
Red hot shot was replaced around 1860 by Martin's Molten Iron Shell, hollow spheres filled with molten iron. For a ship to use this shot, it would need a furnace. (HMS Warrior had one.) You had to wait until at least four minutes after pouring before firing the shot, but the shot stayed hot for an hour. (Lambert 50). It's mentioned in EB11/Ammunition but without reference to when it can be fired. According to experiments conducted in 1866 on HMS Excellent, their effectiveness increases greatly with the size of the shot, the 32-pound shell (holding 16 pounds molten iron) being "of comparatively little use." The 8- and 10-inch shells held 26 and 45 pounds, respectively. "They can be loaded more easily than red hot shot and have a greater incendiary effect."
Extraordinarily, at the siege of Cadiz (1812), the French filled shells with lead, thereby increasing range. (Douglas 61).
Shaped charges. The use of a hollow charge, in order to focus the explosive force, was proposed by Von Baaader (1792). It was used in mining but not very effectively, since gunpowder is a "low" explosive. There was further work with hollow charges in the late-nineteenth century, notably by Monroe (1888 on), who used both a high explosive (dynamite) and a tin can liner. (Wikipedia/Shaped Charge). The focused blast, in turn, will convert any material it encounters into a high-speed projectile.
The principle found practical application in the antitank weapons (bazooka, panzerfaust) of World War II. A biconical head, typically copper, fits into the conical hole in the charge. The charge detonates when the head comes into contact with the target, which puts the charge at just the right distance away for the explosion to convert the copper into a high speed (8–9 km/s at leading edge) stream. (Denny 153ff). The nose has to be longer than the required stand-off distance. (NAVORD 3A10).
Time is needed for the formation of the high-speed jet, so a relatively low striking velocity is desirable. If fired from a gun, that would mean low muzzle velocity and thus low range. However, rockets are not subject to that limitation. Hence, in naval warfare, a shaped charge might be delivered by a fin-stabilized rocket. (Spin can interfere with jet formation.)
Sabots are single or multipart full-caliber devices placed behind or around the projectile for various purposes. Sabots were first used for convenience, associating the charge with the shot (see below), and in connection with fuzed projectiles to keep the fuze facing forward. (Kinard 124). In the early-nineteenth century, sabots were expandable bands at the rear of the projectile that engaged rifling (McCaul 319).
In modern use, it's a device that holds a projectile centered in the bore. Such a sabot is necessary if a projectile's diameter is subcaliber (smaller than the bore diameter). This was most likely to be the case with elongated projectiles; for example, early rifled guns fired studded projectiles. A special advantage of the sabot was that it kept the propulsive gases from escaping; i.e., reduced bore-windage.
For round shot or shells, the sabot was likely to be a wooden (usually elm) disk with a hollow on one side to receive the rear of the shot, and it was strapped to the shot, typically with strips of tin.
If the sabot remains attached to the projectile after it leaves the muzzle, it will impair its aerodynamic characteristics. Stafford, USP 39,180 (1863) proposed a sabot that separated from the shot "immediately after leaving the gun." The shot was an armor-piercing subcaliber bolt carrying a spheroidal metal band that fit the bore. The sabot had a truncated conical front piece with a socket to engage a pin at the rear of the bolt, and a concave disk in the rear, "adapted to be expanded by the explosion of the charge and forced into the grooves of the gun." It thus wasn't intended for use with a smoothbore, but that doesn't mean the discarding sabot idea is applicable only to rifled pieces. The separation of the sabot was effectuated "by the resistance of the atmosphere."
Stafford projectiles were used in the Civil War, albeit in limited quantities. The Navy tried them out, and considered them unsatisfactory. (Bell, 301). A letter from the New Ironsides reported that six were fired, tumbled, and fell 200–400 yards short of the target. In one instance, there was difficulty in forcing home the shot. (Stafford 53). Stafford vigorously disputed these findings, but it sounds to me as though it failed to properly engage the rifling.
Scroll forward a century and we have Barr, USP 2,393,395 (1954). This describes a combination of a finned dart and a mostly teardrop-shaped sabot with a hollow center shaft that fits in front, over the dart. A portion of the sabot is cylindrical to improve the seal against the bore. The idea was that in-bore, friction created by the pressure of the propulsive gases on the sabot would hold it against the projectile, but after they left the muzzle, the two would separate. The patent notes that the sabot may be made of plastic to reduce the combined mass, allowing greater acceleration in-bore. There are other patents on discarding sabots; e.g. Dunlap, USP 3,004,409 (1961—appl. filed 1944!).
Generally speaking, you want the sabot to be as light as possible, because accelerating the sabot is a waste of propulsive force. But the sabot must be strong enough to center the projectile and efficiently transfer the gas pressure to it. And it must discard reliably without interfering with the projectile trajectory. (Carlucci 136). These aren't trivial engineering considerations and I am not sure that they will all be "solved" within the "window of opportunity" provided by the lead time needed to develop rifled cannon and matching projectiles for the 1632 universe.
Cup (rear-mounted) sabots are usually single piece, and are left behind by the projectile, whereas ring (flank-mounted) sabots are usually segmented and discard radially outward.
Projectile Design and Ballistics
The projectile decelerates as a result of drag once it leaves the barrel. The deceleration will be inversely proportional to the mass and thus to the "sectional density" (mass divided by frontal area). All else being equal, the deceleration will be least for the one made of the densest material. Stone will slow down faster than cast iron, and cast iron faster than lead. So stone will have the least range and lead the most.
Since mass is proportional to diameter cubed and frontal area to diameter squared, deceleration is inversely proportional to the diameter. Thus, while a larger caliber shot will start with a lower muzzle velocity than a smaller one (given the same charge), it will eventually overtake the latter and achieve a greater range. (Beauchant 47).
According to Beauchant (28–9), the range of grape shot and double-headed shot (chain shot) is two-thirds that of round shot, whereas double shot (two balls, unconnected) have half the range of single shot. Douglas (65) says that if the absolute weight of the powder charge is unchanged, double shot will receive 71% of the muzzle velocity of the single shot, and require double the elevation to achieve the same range.
If elongated projectiles are used, we need to decide on their shape. Projectile designers could of course simply copy artillery shell profiles from books on World War II artillery, but it would be better if they understood why particular shapes were favored.
Projectile shape is affected by considerations of internal, exterior and terminal ballistics. The designer wants to efficiently transfer the pressure of the expanding gases to the projectile, and minimize wobble, yaw, friction, and abrasion of the bore as the projectile moves down-bore. In flight, aerodynamic drag should be minimized. Finally, shape affects how deeply the target is penetrated.
In discussing shape, it's helpful to imagine the projectile as composed of up to three sections: the forebody, in which its radius is increasing as you move rearward; the midbody, in which the radius is constant (cylindrical form); and the aftbody, in which the radius is decreasing.
The surface of the midbody is called the "bearing surface"; it's responsible for holding the projectile in alignment. Too little midbody and the projectile nose yaws; too much, and friction unnecessarily reduces muzzle velocity. (Rinker 207). In addition, the midbody adds to frictional drag.
Test have been performed on smoothbore artillery from the 16–18th centuries, their average muzzle velocity was 454 meters/second. (Baley 155). [citing Hall 1997, pp. 136–7]
While that is supersonic (Mach 1.33), aerodynamic drag would have progressively slowed the projectile, possibly to subsonic speeds if the range was great enough. "In tests, a 6-pdr ball had a velocity of 137 m/sec at 1,000 m, a 9-pdr ball of 293 m/sec, and an 18-pdr ball a velocity of 256 m/sec." (Id.)
Hence, we have to consider the effect of shape on drag in two or even three speed regimes. In subsonic flight, drag is dominated by frictional drag, which in turn is proportional to wetted area; the ellipsoid shape gives the best ratio of volume to surface area at a given length/diameter ratio.
The sphere, of course, is a special case of the ellipsoid. The drag coefficient is about 0.19 up to about Mach 0.5, then climbs almost linearly to about 0.41 at Mach 1.5, and then eases off for higher mach numbers, to perhaps 0.38 at Mach 3.0. (Guilmartin 296).
The teardrop shape (preferably with length three times the diameter) gives the lowest subsonic drag for a given diameter (Benjamin 19). Unfortunately, a projectile that tapers to a point in the rear is problematic; at the muzzle, it's likely to be deflected by the escaping gases. (Rinker 211).
The model rocketeers of Grantville are probably most familiar with the cone and the tangent ogive. The latter shape, like the secant ogive, is defined by rotating an arc of a circle in 3D space. Other rocket nose cone shapes are based on rotating arcs of ellipses, parabolas, or other curves, such as a "power curve."
In the supersonic regime, the best forebody is sharp; a conical shape is possible but the tip may be rounded slightly (meplat). For high supersonic speeds, a projectile might even have a spike—the equivalent for small arms is a spitzer bullet. (Rinker 208) According to wind tunnel experiments, we can conclude the following about drag: Above Mach 1.5, best shape is the "3/4 power" nose, followed by the cone, the parabola, and the ogive. At low supersonic, 1.2–1.5, best shape is the parabola. (Perkins). At transonic, 0.8–1.2, best shapes are the Von Karman ogive and the parabola (Stoney).
The composite shape must, for trajectory stability, have its center of mass (where gravity effectively acts) in front of its center of pressure (where aerodynamic forces effectively act). (Denny 100). Consequently, the aftbody is usually either non-existent (projectile has flat base) or what's called a boat-tail, a short truncated cone which just angles in slightly. The flat base creates a vacuum behind the projectile, and thus a lot of drag. That's alleviated by the boat tail. While it increases range it reduces accuracy (Evans 260). Its advantage was recognized by Whitworth in 1854 (Hazlett 206).
A more exotic method of reducing base drag, invented in the late 1960s, was a base-bleed unit—this slowly burned a propellant, not to impel the projectile forward like a rocket, but merely to eject a stream of gas that would break up the normal strong vortex adjacent to the base into multiple weak vortices. This increased range at the cost of accuracy. (Wikipedia/Base Bleed; Suliman).
Bear in mind that any asymmetry in the projectile shape, say as a result of a manufacturing defect, will result in unbalanced aerodynamic forces on the projectile, deflecting it. Hence, there's such a thing as being too clever; the theoretical advantage of a shape may be outweighed by the practical problem of making it accurately.
There are a variety of reasons why a projectile might be given a flat rather than pointed nose. This is done sometimes with armor-piercing projectiles, or so-called "diving shells" used to attack submarines close to the surface. If so, to improve the aerodynamics, the projectile may be given a "false ogive" (windscreen) that collapses on impact. (Crain 253; Officers 513).
Depending on the shape of the projectile, and its angle of attack (projectile axis versus direction of motion) in flight, it may also experience aerodynamic lift (and lift-induced drag). Lift is a significant factor in increasing the range of arrows (Denny 97ff) and I would expect that it would also affect other finned projectiles. Canards (forebody wings) have been placed on artillery shells for range extension.
A complete round of ammunition is everything you need to fire a weapon once. For a cannon, that would be the projectile, the propellant charge, and the primer (and, for shells, a fuze).
In our period, the charge and primer are usually loose powder that must be ladled out to measure, and all the components are separate. However, in the late-sixteenth century, occasionally the powder was in bags (Peterson 27); multiple bags would be used if you needed more oomph. Loose powder continued to be used until the mid-eighteenth century. Bags offered a greater rate of fire, but less flexibility in adjusting the charge. In France, loose powder was abandoned only after it was shown that there was a "maximum" (critical) charge beyond which range couldn't be increased. (Nosworthy 367).
Another eighteenth-century alternative to loose powder (for priming) was the priming tube. (Peterson 64). What might be termed a caseless fixed round took the form of a cannon ball strapped together with a powder bag to a wooden sabot. (63).
Case (cartridge) ammunition was also developed, and it came in two flavors. If just the primer and propellant charge are in the case, and must be loaded with the projectile, that's called semi-fixed ammunition. And if the projectile was also in the case, that's called fixed ammunition. Fixed ammunition lends itself to automatic loading. Semi-fixed ammo must be manually loaded, but it can be handled more rapidly than bag ammunition.
Cartridges first appeared for small arms. Paper cartridges are used in Gustav Adolf's army. (Westwood 27). Unfortunately, they are easily damaged, so cloth (especially flannel) cartridges were preferred in the eighteenth century. Metal (especially brass) cartridges appeared in the nineteenth century.
At the Battle of Jutland, the British lost three battle cruisers because enemy shell fire caused a secondary explosion of the ship's magazine. The powder in the British silk bags was immediately ignited and exploded. In contrast, the Germans used metal cartridge boxes and cases. These protected the powder from explosion for a time, providing an opportunity for the sprinkler systems to kick in. (Breyer 65).
In the late-nineteenth century navy, fixed ammo was used for the rapid-fire guns and bag ammo for the larger guns.
The weight of the shot for a warship is great enough so that it was typically stowed below the waterline and, given the degree of leakiness of wooden ships, that pretty much guaranteed that the shot would get corroded by seawater. Douglas (97ff) proposed that fifteen double-shotted rounds be kept on deck, in some kind of container that would keep it out of the water sloshing across the deck.
So, what niche can rocket artillery fill the post-RoF navy? Eight rocket launchers were mounted on the speedboat Outlaw, commanded by Eddie Cantrell, and its first salvo destroyed the warship Anthonette in the Battle of Wismar. (Flint and Weber, 1633, Chapters 38 and 46).
Rockets are self-propelled projectiles with very slow-burning propellants. Rockets were first used in combat by Chinese warships in the twelfth century(Denny 28). Their first use in Europe appears to have been by the French in the fifteenth century, in siege warfare. Several down-time books describe rocket construction, including Biringuccio's Pyrotechnics (1540), the anyonymous Book of Cannons and Fireworks (1561), Pavelourt's Brief Instructions on matters of French Artillery (1597), and Lorrain's Pyrotechnics (1630).
The "rockets' red glare" in "The Star-Spangled Banner" refers to the British shipborne Congreve rockets, intended for land attack. Their flight was crudely stabilized by a long (15 foot) stick (the rocket proper was 3.5 feet long). They weren't very effective against Fort McHenry, but Congreve rocket assault had been much more successful against Copenhagen (1807), destroying three-quarters of the city.
In the early-nineteenth century, they were typically launched from boats or by a landing party. They could be given different payloads; carcass rockets were used against towns and shell rockets against troops or shore batteries. Extreme range was 2000–3500 yards. (Beauchant 95ff).
Effective ranges were more like 800–1200 yards, but rockets did have the advantage of a high rate of fire—perhaps ten per minute. (Wise 31).
Grantville's second generation rockets were fired from a jury-rigged katyusha by the Jewish defenders of Prague (Flint, "The Wallenstein Gambit," Ring of Fire). Historically, a katyusha was a Russian WW II truck-mounted multiple rocket launcher, capable of launching 14–48 M-13 rockets simultaneously. The M-13 rocket weighed 42 kg and had a 4.9 kg warhead. (Wikipedia/Katyusha).
At Prague, the rockets were "a variation on the old nineteenth-century Hale 24-pounder rotary rocket—2.4 inches in diameter, slightly less than two feet long, with a maximum range of 4000 yards. The propellant as well as the warhead were black powder [twenty pounds, total]. The rockets were fired from a single-level rack, twelve tubes mounted side by side on an adjustable framework fixed into the bed" of a pickup truck. The rockets had contact fuzes, rather than Hale's time fuzes. Maximum flight time was twenty seconds, which implies an average speed of 600 feet/second.
I think that by "rotary" Eric means that the rockets were spin-stabilized. Hale's mid-nineteenth century rockets were spun as a result of the combined action of fins and secondary exhaust nozzles; the US used them in the Mexican War. There was a launcher ("Machine Rocket, War, Naval") designed for seaborne use; it looks to me like an elevatable tube that would mount outboard. EB11/Rocket notes that the Hale rockets were made in two sizes, 9 and 24 pounds, with the former having a maximum effective range of 1200 yards.
War rockets were rendered obsolete by advances in rifled artillery; they were abandoned by European armies in 1867 and the British colonial forces about 1885. However, we know that they made a comeback in World War II, as exemplified on the strategic level by the V-2 and on the tactical level by the katyusha and similar weapon systems.
The principal advantage of a rocket is that the launch device does not feel any recoil. Momentum is conserved in that the backward movement of the exhaust gases compensate for the forward movement of the rocket. Consequently, the launch device can be a lot lighter than what would be the case if it had to absorb the recoil that would be necessary to propel a projectile of the same power to the same range from a conventional cannon. It is because of that lack of recoil that a shoulder-fired rocket launcher, the bazooka, is possible.
Another advantage relates to impact velocity. A rocket can be engineered to achieve a higher impact velocity than an equivalent shell (shell mass equal rocket payload, artillery charge equal rocket propellant). The shell stops accelerating when it leaves the muzzle, so from that time on its speed decreases, thanks to aerodynamic drag. On the other hand, the rocket is still burning fuel after launch, and whenever this propulsive force exceeds the drag force, its speed will increase. For any given time-to-impact, the rocket may be given a fuel fraction sufficient to assure that its impact speed is greater than that of the equivalent shell. (Denny 142).
The great disadvantage of the rocket is that it is inaccurate. While spin-stabilization definitely improved accuracy, variation in burn rate can also be a problem, especially with solid propellants. In the short-term, the simplest solution to the accuracy problem is to fire rockets in large numbers. In the long-term, we can attempt to develop liquid propellants and even perhaps some kind of guidance system.
Flint and Weber, 1633, Chapter 29 says that "Mr. Ferrara and his rocket club have been working on a ship-launched surface-to-surface missile for us."
What Eddie fired in 1633 was not a missile. A missile is a rocket with a guidance system. There are two basic kinds of guidance systems, remote and autonomous. With remote guidance, the missile receives signals from an operator. One possibility is a radio control system; the missile must carry a radio receiver, and jamming is possible. Another possible remote guidance system is by electric wire; the wire remains attached to the missile after launch, and reels out as it flies, and the gunner steers it by signals sent down-wire.
Bear in mind that effective remote guidance requires that the operator be able to visualize the target, and its relationship to the rocket—and the rocket, at least, is moving very quickly.
An autonomous guidance system is carried entirely by the missile. In "open loop," it is just following preprogrammed instructions. It can follow a complex course but there is no adaptation to circumstances. In "closed loop," it can sense the environment and respond to what it senses.
It should be obvious, but all the guidance in the world is of no use unless the missile has means to adjust its flight. This can take a number of forms, such as adjustable tail fins, moveable vanes in the exhaust chamber, a gimballed engine, or side thrusters.
While there is a model rocketry club in Grantville, their knowledge of guidance systems is going to be strictly theoretical because the National Association of Rocketry code bans guidance systems. It may be possible to scavenge the mechanical actuator that pivots the wheel of a radio-controlled toy car and use it to control the rudder of a model rocket.
There may well be radio-controlled and control line model aircraft in Grantville and hobbyists who are familiar with them. These may also have parts that can be scavenged, duplicated or adapted.
Guidance makes sense only if we are engaging at ranges at which the unguided rocket would be hopelessly inaccurate, and then only if the rocket is expensive and powerful enough to warrant going to the trouble of providing the guidance system. And even then we have the problem of packing the guidance system inside the very limited space of the rocket, and the loss of space for payload and propellant (with consequent loss of power or range) that entails.
Torpedoes, in essence, are explosive projectiles that are fired at a target, or carried to it, and caused to explode underwater. That way, you may sink the enemy craft outright, which is much more difficult to do with a single shot hitting the target above water.
EB11/Torpedo recognizes two kinds, spar torpedoes and locomotive torpedoes. The spar torpedo is carried by a boat or submersible; essentially it is an explosive charge at the end of a long spar. The charge may be exploded upon contact, in which case the operator will probably be reminded of the saying, "he who sups with the Devil must do so with a long spoon." (The Confederate submersible Hunley was destroyed by the explosion of its own spar torpedo.)
Or there may be some means to attach it to the target's hull, and leave it there, with a timing device (fuze or clockwork) setting it off after the operator has gotten further away. The limpet mines used in the new time line to sink four ships of the Danish-French squadron blockading Lubeck had timers; they were planted by a diver. (Flint, 1633, chapter 44). Fulton designed a harpoon torpedo.
Historically, a spar torpedo was used to sink the Confederate ironclad Albemarle in 1864. In the 1632 universe, the first use is by Jeff Higgins; a fishing boat sallies out at night and destroys a Spanish galleon outside Amsterdam with one. (Flint, 1633, Chapter 42). Spar torpedoes aren't exactly high-tech, and the Danes copy the idea, using ten galleys with spar torpedoes for an attack, under cover of a smoke screen, on Simpson's squadron outside Copenhagen. This succeeded in damaging the USE ironclad Monitor. (Flint, 1634: The Baltic War, Chapters 60–1).
The secret to success with a spar torpedo was stealth; the attacker had to be able to creep up on the enemy. The attack was usually made at night, with the moon below the horizon or blanketed by heavy clouds. The Spanish had picket boats with torches or lanterns rowing about their galleons, but they weren't likely to spot an attacker and if they did, musket fire wasn't likely to stop it. The best defense against a spar torpedo attack was to rig the ship with electric lights, so an incoming attack boat could be seen from a greater distance, and equip the ship with rapid-fire small guns suitable for sinking such a boat. Simpson's ironclad had suitable guns—mitrailleuses—they just didn't take out all the torpedo galleys in the confusion of the attack.
EB11 also alludes to the "Harvey towing torpedo." The torpedo was towed on a line and, as the towing ship's speed increased, these lines would flare out. The towing ship would run pass the target so it would strike the tow line; the torpedo would then swing around and hit the target. That was the theory, at least. In practice, the tow lines would get fouled on the propellers.
Locomotive (self-propelled) torpedoes have yet to appear in canon. EB11 describes the uncontrolled Howell and Whitehead torpedoes. Both had tail fins and exploded on contact. In the Howell torpedo, a flywheel was spun up, to act both to drive the propeller shaft and to act as a gyroscope to keep the torpedo on course. Its range was about 400 yards. (Wikipedia/Torpedo).
The Whitehead torpedo was propelled by compressed air that had been stored in the torpedo, and drove a piston engine that in turn drove the propeller. It also had had the first successful depth control mechanism (EB11 says just that a change in pressure actuated a "horizontal rudder"; for particulars see DennyHP). The original torpedo had a single propeller, but the EB11 article refers to twin propellers. If these are contra-rotated, it will defeat the tendency of a single prop torpedo to turn in a circle. In 1895 Whitehead incorporated the Howell flywheel, setting off a patent fight.
Some historical perspective may be helpful. The first Whitehead torpedo had a range of 300 meters at less than six knots. A decade later, it could reach 500 meters, at 18 knots. In the early 1880s, torpedoes carried a seven-pound warhead. (Ireland1997, 41).
Alternative Propulsion Systems. Besides the flywheel (Howell) or compressed air (Whitehead) systems, a torpedo may use an electrical cable (Ericsson 1873), electric battery (WW II), winched wires (Brennan 1878), or combustion. In the last approach, one option is internal combustion; the torpedo burns alcohol or kerosene, and the combustion gases drive a piston or turbine. Or you can use external combustion; the combustion chamber is cooled with water ("wet heater"), which is flashed into steam.
The source of oxygen for combustion was typically compressed air. Since air contains nitrogen, which is not consumed, the torpedo would leave a trail of nitrogen gas bubbles. It was recognized at a very early stage that pure oxygen would be better than compressed air; the weight of the nitrogen would be eliminated and the torpedo would be wakeless. So what's the catch?
Well, if the oxygen torpedo exploded on the host ship, that would do tremendous damage. And that was a real risk; oil and grease in the oxygen pipes could ignite spontaneously. And it is difficult to properly lubricate the moving parts of the torpedo without increasing oil and grease exposure. "The Japanese began experimenting with oxygen-driven torpedoes about 1924, but gave up after numerous explosions and failures." Then, somewhat farcically, in 1927, faulty intelligence led them to believe that the British were well on their way to producing a workable oxygen torpedo and they decided that they needed one too. It took them until 1933 to complete the development of the type 93 "long lance" oxygen torpedo. Part of the solution was beginning the ignition with air and only gradually changing over to oxygen. And they took many precautions against inadvertent oxygenation of the lubricants. (Evans 267).
Guidance. Torpedoes may be divided, like rockets, into three categories: uncontrolled, remotely controlled, and autonomously-controlled.
Radio-controlled torpedoes were proposed by Von Siemens in 1906, but the problem was that the radio signals would have to penetrate the water.
The first wire-guided torpedo was perhaps the Brennan torpedo (1878), which was used for coast defense in 1890–1906. It had two drums of wires, which reeled out as it sped through the water. At the ship end, these wires were connected to winding engines. If the operator caused one wire to wind out faster than the other, causing its drum to rotate faster, this activated the rudder. (Wikipedia/Brennan Torpedo). Note that the wires were also its propulsion system. Its effective range was limited by the wire length provided; it was about 2000 yards. A serious problem with the Brennan torpedo is that its depth control was poor; it "porpoised," and that could mean that it passed under the target without detonating. (DennyHP). I can't help but wonder how much better it would have performed if it were equipped with a Whitehead depth control! Another problem was that the wire was thick and thus "required a mass of wire so large that it was inconvenient and even dangerous aboard a ship." (Weir 123).
For a guided torpedo to be practical, the operator must know where it is at all times. The Brennan torpedo was fitted with an indicator mast so it could be spotted from above. Of course, that meant that the enemy could spot it, too, but the operator knew where and when to look. The other option is to equip the torpedo with sensors and some sort of two-way communication.
Autonomously controlled torpedoes home in on the target, either passively or actively (the latter "ping" the target). Acoustic torpedoes home in on the target's engine noise, whereas wake-homing torpedoes zigzag until they detect the turbulence of the target's wake and then follow it in. Obviously, there's no point in even trying to develop acoustic torpedoes until steamships are commonplace, and I doubt a sailing ship would have enough of a wake to be particularly detectable.
Deployment. How is the torpedo carried and discharged? One option is to carry it above the water, and launch it from a gun port, or, better yet, a tube. The launch tube may be above or below water. The former has the problem that an enemy shot or shell might strike it, detonating the torpedo. That may sound improbable, but that's what happened to the Almirante Oquendo at the Battle of Santiago.
There's a fair amount of mechanical complication associated with a submerged torpedo tube. It must have a breech door at one end and a muzzle door at the other; the two operate together somewhat like an airlock on a spacecraft. The tube is drained of water for loading and filled for firing.
Expelling the torpedo initially was accomplished mechanically, with a push rod, but this could damage the rudder, and so the rod was replaced, first with a compressed air gun, and then with a propellant (gunpowder, cordite) charge. (EB11/Torpedo). While that's EB11's last word on the subject , we know that modern torpedo tubes use compressed air discharge.
A further problem is making sure that the torpedo isn't buffeted unduly by water currents as it leaves the tube. One expedient was to extrude a tube with grooves that would guide the torpedo until it was clear. (Id.)
These discharge tube designs have the disadvantage that the tubes have to be built into the hull, potentially weakening it.
The greatest disadvantages of the early locomotive torpedoes was their short range and low speed. Without the flywheel, the effective range was just 800 yards; the gyroscopic stabilization brought it up to 2000 yards. The speed of the first torpedoes was 10 knots, and by 1911 they had been brought up to thirty knots. The speed (and range, to limit of directional accuracy) could be increased by increasing the pressure to which the propulsive air was compressed; in 1911, 2000 psi was possible. These ranges and speeds are far inferior to those of cannon shells, let alone rockets.
The EB11 essay does not mention another problem with the early uncontrolled torpedo; that, as Admiral Lord Fisher put it, it will "blow up anyone it comes across." While an artillery shell could also hit a friendly ship, a rudder malfunction could cause a torpedo to circle back and "bite the hand that fed it." The initial solution was to condition explosion on a minimum distance or time of run. Modern torpedoes have a safety interlock that disarms the warhead if the azimuth changes more than a set limit after the engine starts. If the engine starts spontaneously while still on board, the sub does a U-turn to disarm it.
This article comes to a close in part 5, "Thrust and Parry."