In the first four parts, we talked about the guns, the mounts, the projectiles, and the ballistics of the projectile in the bore and in flight. We also discussed how to make the shots effective in terms of actually hitting the target. But there’s another aspect of effectiveness, penetrating the target’s hull and armor.


Making Your Shots Count

The goal of the warship is to subdue or kill enemy ships by degrading their seaworthiness (buoyancy, stability), mobility (speed, maneuverability), and firepower. It can do this directly by damaging the ship structures or indirectly by incapacitating the crew (so sails can’t be adjusted or guns fired).

Attacks on seaworthiness are difficult; you must pierce the hull underwater or close enough to the waterline so water slops in. In theory, hits elsewhere would weaken the hull and could cause it to break. In practice, wooden warships could shrug off an enormous number of hits with solid shot.

A direct attack on firepower is difficult; it means striking a gun or its carriage, both small targets partially protected by the hull. When warships evolved from having many low caliber guns to a few high caliber ones, the guns became a more productive target.

A direct attack on mobility is easier; you can hit the sails, masts, yards or stays of a sailing ship, or the coal bunker, engine, steam lines or screw (or paddle) of a steamship (or the steering mechanism of either).

Still, it appears that with solid shot, you mostly attacked wooden ships indirectly by killing its crew—not necessarily by direct hits; flying splinters could be lethal.

In 1800, the 74-gun British Foudroyant was locked in battle with the 80-gun French Gillaume Tell, so close that “her spare anchor just escaped catching in the mizzen rigging of the Gillaume Tell.” The Foudroyant fired 1200 32-pounder, 1240 24-pounder, 118 18-pounder and 200 12-pounder shot at its adversary, which had already fought the 64-gun Lion and the frigate Penelope earlier that morning. After several hours, the Gillaume Tell was completely dismasted, and forced by rolling to close its lower gunports, and therefore struck its colors. But it wasn’t sunk, and it was towed to port and refitted for British service. (Douglas 210; Wikipedia/HMS Malta (1800)).

Then there was the 1813 capture of USS Chesapeake (36 guns) by HMS Shannon (38). The Chesapeake “was struck by twenty-five 32-pounder shot, twenty-nine 18-pounder shot, three hundred and six grape-shot, and two 9-pounder shot” and for its sins, the Shannon received “thirteen 32-pounder shot, twelve 18-pounder shot, one hundred and nineteen grape-shot, and fourteen bar-shot.” The cannonade was at a distance of a ship’s width (40 feet), yet some shots stuck in the enemy’s sides, without perforating the hull (for exact locations, see Douglas 78ff). Both ships suffered substantial damage to the rigging. The bombardment of the Chesapeake disabled five of its guns, but the casualties among its gun crews probably had a greater effect on its firepower. A total of 47 were killed and 99 wounded on the Chesapeake. (Preble 17, 25ff).

Modes of Fire.

Fire may be classified as direct or indirect. Direct means that you can see the target; indirect that you can’t (and are thus dependent on a spotter). As a practical matter, once ranges exceed the distance to the horizon (OTL, this occurred in the early-twentieth century) indirect fire is forced upon you. Of course, a spotter may be helpful even in guiding direct fire.

Probably the most important innovation insofar as effective indirect fire was concerned was the “aiming goniometer.” (1896; Sweet 179).

Concentration of Fire

In the early-nineteenth century, Captain Philip Broke of HMS Shannon developed methods of concentrating his ship’s fire on a single part of the enemy hull. Essentially, he had three preset aiming points. One, for example, was established by aiming the aftmost gun as far forward as possible at a buoy 300 yards away; once it could strike the buoy, the other guns were aimed to hit it too, and the gun carriages and port-sills marked to show the necessary adjustment. In like manner he established two other aiming points, one to the rear, and last in the middle. (Stevens 39).

Terminal Ballistics

To penetrate a hull (or armor), wreck ship vitals and kill crew, the projectile must do work, which requires energy.

Solid shot penetrates by virtue of its kinetic energy (0.5*m*v2), where m is projectile mass and v is impact velocity. The impact velocity may be determined by the methods of exterior ballistics (part 3). The ballistic limit velocity is the greatest impact velocity the target can withstand without being perforated.

Normally, the velocity of the projectile decreases as it gains range, as a result of air resistance. However, if the shot is fired with a high inclination, then on the descending limb of its trajectory, its speed will increase as a result of gravity. (Beauchant 46). This is an advantage of plunging fire, although offset by the disadvantages of an oblique strike (see below). The maximum possible impact velocity is then the “terminal velocity,” the velocity at which the air resistance equals the gravitational force. The only naval armament this would apply to are the mortars of bomb ketches.

If each “slice” of hull or armor resists the projectile independently of the rest of the armor, then penetration will be proportional to the kinetic energy and thus to the square of the impact velocity (and linearly to the mass). However, if each “slice” resists in proportion to the entire thickness still facing the projectile, the penetration will be linearly proportional to the impact velocity (and to the square root of the mass). (Okun).

Another consideration is the effect of the armor on the projectile. If the projectile suffers “progressive damage that gradually reduces penetration ability,” this too will result in a less-than-quadratic dependence of penetration on impact velocity. This created the incentive to harden projectile noses with chilled cast iron or even steel, and it was also found that the ogive shape was the most effective. (Owen 273). On the other hand, plate quality typically decreased as thickness increased. (Okun).

Explosive shells do damage by virtue of the heat and pressure of the expanding gases, and the kinetic energy of the shell fragments and splinters they carry with them. The damage caused by an explosion is quite dependent on whether the blast is in the open, or confined, and the speed with which the debris is moving.

Oblique Impact. If the shot strikes the enemy hull obliquely, then of course there’s a greater effective thickness to penetrate (Beauchant 49), the effective thickness becoming the actual thickness divided by the cosine of the angle of obliquity (0° for square hits). Moreover, the transfer of the kinetic energy of the shot to the hull is less effective; per a simple physical model, the transferred energy is the product of the kinetic energy and the square of the cosine of the angle of obliquity. (Wikipedia/Sloped Armor). And if the angle exceeds a critical value, the projectile ricochets rather than penetrates.

The curved hulls of the early modern wooden warship, with bilge and tumblehome, thus were able to deflect shot in a way that a “wall-sided” hull couldn’t. Some early ironclads (CSS Virginia) had sloped sides to deflect shot, and deflection is also the raison d’etre for preferring a domed turret to a cylindrical one, and cylindrical to block.

Stout as Oak

In traditional wooden ship construction, the shipwright started with a keel (which itself could be several pieces of wood scarfed together), from which, at intervals, extended ribs, curving upward, made of several curved timbers called floors, futtocks and top timbers. A deck beam connected the tops of each pair of ribs (frames), and the connections were braced by sharply angled pieces (knees). These collectively constitute the framing of the ship. Longitudinal planking was laid over the framing, forming the bottom, sides and deck.

Until the mid-nineteenth century, warships trusted to their wooden planking to fend off cannon balls. While elm was used for the keel and lower strakes, the above-water timbers were usually made of oak, with other woods used for ships constructed outside of Europe (e.g., teak in India). The thicker the hull, the more resistant it was to attack.

In the Seaman’s Grammar (1627) John Smith states, “if you would have a ship built of 400 tuns, she requires a plank of 4 inches; if 300 tuns, 3 inches; small ships 2 inches, but none less.” (13). According to a Dutch shipbuilding rule (Witsen, 1671), the thickness of the inside of the stem (stern) was one inch for every ten feet of the ship’s length (between outsides of stem and stern), and the planking of the hull one quarter of that thickness. So for a 160 foot length, the required planking would be four inches. By another rule (Van Yk, 1697), the thickness was based on the “length class” of the ship, ranging from 2 inches for 40–60 feet to 4.5 inches for 140–160 feet. (Hoving 56, 250ff).

I assume that these rules were for merchantmen, and warships were stouter. The 36-gun Dartmouth (1655), essentially a Dunkirk frigate, had framing timbers of 8 inch “moulding” (thickness). (Kenchington 33). “The hull [of the 1628 Vasa] consists of three layers, and is approximately 40 cm [16 inches] thick.” It’s made of European oak (Quercus robur). (Ljungdahl).

About a decade after the Napoleonic Wars, warship hulls had thicknesses of 10–26 inches, depending on warship size and the place of measurement.(Table 5-1):


Table 5-1: Hull Thickness of British Warships (1828)(Beauchant 132ff)
Thickness, Oak(inches) At Load Line At Upper Deck
10 10-ton brig
11 18-ton brig
12 10-gun brig
14 18-gun brig 28-gun ship
15 frigate
16 60-gun
18 28-gun ship 3rd rate
20 frigate 2nd rate
21 1st rate
23 60-gun
24 3rd rate
26 2nd rate
28 1st rate


(For mid-nineteenth-century French warships, see Dahlgren 183).

In canon, the timberclads are protected by 48-inch thick bulwarks (1634: The Baltic War, (TBW) Chap. 36). According to Admiral Simpson, “that much timber was invulnerable even to heavy shipboard guns—or what passed for them in 1634, at any rate—at any range beyond sixty or seventy yards.” (Id.). However, the heaviest projectile noted to have struck a timberclad in the Battle of Wismar was a mere twelve-pounder. (Chap. 52). And when Achates ascended the Thames, the biggest guns it expected to encounter were eighteen-pounders, and twelve-pounders were thought more likely. (Chap. 61).

Solid Shot vs. Wooden Hulls

In tests conducted by John Greaves (1600–52) at Woolwich in 1651, a target was set up at a range of 200 yards. The target consisted of three butts, each 19 inches thick (13 of oak and the rest of elm), with fourteen yards between the first pair and eight between the second. Results were as follows:


Table 5-1B
Gun Weight Length Shot Powder Performance




pierced the first two butts and struck in the third.




pierced the first two butts.
whole culverin





three shots passed through the first two butts, of which just one entered gently into the third.





four shots pierced the first two butts, of which one penetrated seven inches into the third, and another pierced it completely.



passed through one butt and entered the second.
demi-culverin2 1/8″ bore



one shot passed through both butts and the other almost passed through the second.


(Greaves 470)

In 1761, Robins (310) reported that at 30 yards, an eighteen-pounder penetrated 14.5–15.5 inches oak with a one pound charge, nearly 33 inches with three pounds, and 37–46 inches with six pounds.

In an 1810 British trial (Frazier), a 24-pounder with only a 4 pound (one-sixth) charge (probably of cylinder powder) fired 21 rounds double-shotted at a 5’2″ thick but of fir, 100 yards away. Ten of the balls penetrated it completely and traveled another 50 yards beyond it; another ten penetrated it partially, to an average depth of four feet, and one ball missed. While fir has perhaps half the resistance of oak, this was still impressive performance, and you can imagine what a 32- or 42-pounder would do to even a ship of the line, especially with a greater charge. (Beauchant 22, 45; Morriss 221).

In 1814, American experiments with a 32-pounder (11 pound charge) resulted in penetration in white oak of 60 inches at 100 yards and 54 inches at 150 yards. (Haswell 1853).

Without, unfortunately, citing his source, Grundner (86) says that “even a 12-pound gun, with a 4.5-pound charge, could penetrate over two feet of solid oak at a distance of a quarter mile.”

Dahlgren conducted experiments in which, at 1000 yards, a 32-pounder penetrated 25 inches of white oak, and a 64-pounder, 37 inches. (Broun 59).

In tests conducted by HMS Excellent in 1838, at 1200 yards, firing at the Prince George, even an 18-pounder (6 lb charge) achieved 25.5 inch average penetration. A 24×9.5 (8 lb charge) and a 32 x 7.5 (6 lb) both passed through 30 inches, a 32 x 9.5 (one-third charge) 34 inches, a 68 x 9 (10–12 lbs) 35, and a 68 x 5.3 carronade (5.5 lb charge) 30. (Royal Engineers 101).

1853, at a range of 1200 yards, their 68-pounder with 16 pound charge penetrated 45 inches into the old 74-gun York, and their 32-pounder with a 10 pound charge achieved 30 inches. (Excellent 2).


Euler and Robins proposed a solid shot penetration formula of the form

z= 0.5 wv2/gR

with w weight (lbs), v striking velocity (fps), g gravity (32 ft/s2) and R resistive force; R is the cross-sectional area of the projectile (sf) times 912,190 for oak, or 475,070 for fir. (Douglas, 124; PC 495). The underlying assumption is that the resistance is constant and proportional to the cross-sectional area.

Poncelet (1829) suggested that penetration was proportional to the logarithm of impact velocity squared. This assumes that there is a combination of constant resistance and resistance proportional to the squared instantaneous velocity. (Bulson 142).

Tests were carried out (1835–1844) at Gavre and Metz and resulted in the following Poncelet-type formula for penetration z (cm) as a function of predicted striking velocity u (m/s), diameter of shot a (cm), and specific gravity of shot:

z=2.306 * a * SpG * log10 (1 + u2/105)

The striking velocity would in turn depend on the projectile mass and caliber, and the range, elevation and muzzle velocity; the latter would depend on the projectile characteristics and the powder charge. The formula would need to be modified if the projectile broke through, as then it isn’t reduced to zero velocity.

Using a French ballistics formula, Dahlgren (177–8) calculated the striking velocities at different ranges for various guns having stated (predicted?) muzzle velocities and then used the formula quoted above, but with “2.306” replaced with “2.2456,” to calculate the penetration of seasoned white oak with 1850 US Navy ordinance at different ranges:


Table 5-2B Predicted Penetrations
Shot Wt Charge Muzzle velocity Range (yards); Penetration (inches)
500 1000 1500 2000
18 6 1720 28.9 17.9 11.0 6.9
24 8 1720 33.5 21.8 14.1 9.3
32 9 1700 38.7 26.5 18.2 12.5
42 10.5 1620 41.7 29.7 21.1 15.1
64 16 1620 49.9 37.3 27.9 20.8


It’s instructive to compare this table with the hull thicknesses of table 5-1, although they are from two decades apart.

Please note that the striking velocity formula assumes that the drag equals (aV2 + bV3) where V is the instantaneous velocity. As explained in part 3, that’s not correct. However, by my calculations (with Benton’s values for “a” and “b”) at about 100 yards it overstates the striking velocity for a 24-pounder by only about 2%.

Beauchant (1828) provides a table (45) for penetration in elm, depending on projectile weight and charge, but maddeningly fails to state the range or striking velocity, or explain whether the numbers are educated guesses or experimental results. But for comparison with Douglas (and table 5-1), he has an 18-pounder with a one-third charge (theoretical muzzle velocity 1300 fps) driving through 25.75 inches of elm, and a 42-pounder with a one-quarter charge penetrating 24.5 inches.

While sources vary (and wood is heterogeneous), the US Army (398) suggested that one “multiply the penetration in oak by 1.3 for the penetration in elm, by 1.8 for white pine, and by 2 for poplar.” While the denser woods appear to be more resistant, the advantage of oak over fir (1.8 or even 2.0) is greater than their relative densities. (1.4)(Simmons 32).

There are a number of statements in the nineteenth-century literature to the effect that the shot ideally just penetrates the hull; this apparently maximized the generation of splinters (Douglas1860, 66ff). Douglas (67) therefore urged reducing the standard powder charge from one-third to one-sixth of shot weight.

This has been confirmed by modern experiments with scaled-down replicas; splinter generation was greater when the hull absorbed 91% of the shot’s kinetic energy than when it removed 54%. (Kahanov). The explanation is that splintering is caused by bending-induced fracture, and higher velocity projectiles punch through by shearing, with little bending. (Cotterell 354).

Mythbusters episode 71 would have us believe that sailors were only killed by direct hits by cannonballs, splinters not being given enough force to be lethal. While I don’t doubt that being hit with a cannonball was worse, splinters could definitely do serious damage. On the Penelope (1854), “one man was struck across the face by a heavy splinter, which buried itself in his brain” (Marquis 599) and at the Battle of Lake Erie, Lieutenant Stokes of Queen Charlotte was struck senseless by a splinter.

Attempts have been made to fire to strike a target underwater, but these were generally ineffectual, the shot either rising and hitting the target above the waterline, or only slightly indenting it below it. (Excellent 3).

Explosive shells vs. Wooden Hulls

As previously explained, explosive shells of a sort were already in use in seventeenth-century warfare. These were, however, fired on high-angle trajectories from mortars, and therefore not very useful for targeting a ship in motion.

Paixhans built a 8,131-pound “bomb cannon” for firing an 86.25 pound spherical shell on a relatively flat trajectory. The fuze ignited when the gun was fired. In 1824, he fired twelve shots at the 80-gun Pacificateur, at a distance of 640 yards. To ensure that the target would survive the test, it was cleared of combustible materials, and dampened. The results were as follows:

shot 1: made breach of 8.5 inches diameter in the ship’s side, which was 29 inches thick, tore off two feet of the inner plank, made a 2–3 foot square hole in the orlop deck, and all told “shattered to atoms about 160 square feet of wood-work.”

shot 2: carried two large pieces of plank off the quarter deck, knocked 3–4 foot, 9.5 inch thick splinter off the main mast, tore off a 130 pound mast band and drove it into the opposite bulwark . . . .

shot 3: tore off a 200 pound oaken knee, and its splinters knocked down 40 figures [wooden test dummies!].

There were nine more shots, one of which made a breach of several feet in height and width in the side of the ship . . . .” (Paixhans 13, 29ff).

Fast forward to the Battle of Sinope (1853). It wasn’t surprising that the Russians won; the Ottomans were badly outgunned. What was noteworthy was that two Ottoman frigates were blown up by Russian shell guns within fifteen minutes (Dahlgren 303), whereas the Russian flagship was “struck by 84 cannon balls without major damage.” (Mikaberidze 837). Soon thereafter, France constructed ironclad floating batteries, and in 1855 France stopped building wooden battleships.

Clad in Iron (or Steel)

It is a matter of some academic controversy whether the Korean “turtle ships” of the sixteenth century were iron-plated. Ch’oe, Traditional Ships of Korea (94), firmly asserts that the geobukseon, first introduced in the naval battle of Dangpo (1592), was an ironclad version of the traditional Korean battleship. His principal source for the structure of the geobukseon is the Yi Chungmugong Jeonso (1795), written in praise of the Korean commander at Dangpo, Admiral Yi.

The doubt arises from three points. First, why would they bother, when the attacking Japanese warships had few cannon? Second, would they have been able to produce sufficient iron? Third, why, in his own writing, does Admiral Yi only mention iron spikes, not iron plating?

Ironclad construction. An ironclad could have a wooden hull (La Gloire, 1858) or an iron one (HMS Warrior, 1860). Iron-hulled ships also have framing, both transverse ( ribs of metal) and longitudinal (stringers paralleling the keel). The bottoms and sides are formed by metal plates that are riveted to the frame and to adjacent plates. In 1911, the plates were typically 20–40 feet long, 5–7 feet wide, and 1/4 to 1 inch thick (EB11/Shipbuilding). If the plates were steel, they were usually mild steel. Of course, this is just the “shell” plate. Generally speaking, in merchant ships, the transverse framing was continuous, whereas in warships, most of the longitudinal framing was. Warships also had numerous transverse bulkheads, and consequently their transverse frames are fewer and thinner than in merchantmen.

An iron warship would have armor plate protecting the shell; in 1911, the armor plate had wood backing (teak was popular) separating it from the shell. There might also be an inner skin of iron as a further splinter shield. (Farrow 123). Additional armor might also be disposed inside the hull to protect vital areas.

Armor fabrication. As of the Ring of Fire (RoF), the down-timers have the ability to make cast iron, wrought iron, and certain forms of steel. Cast (pig) iron is high in carbon (over 2.1%) and brittle. Wrought (bar) iron is low in carbon (under 0.l%), tough and malleable.

Steel is intermediate in carbon content. The steels available in the early-seventeenth century include blister steel, crucible steel and, in Asia, wootz steel. A steel is called a carbon steel if carbon is the only significant alloying element, and they can be further differentiated by the amount of carbon; mild steel is 0.16–0.29%. Those with other alloying elements are named accordingly; e.g., nickel steel.

The first experiments on the ability of wrought iron to resist shot were carried out by Robert and Edward Stevens and communicated to the U.S. government in 1841. They reported that iron was sixteen times as resistant as oak. (Holley 624).

But some experiments at Woolwich in 1845 were nonetheless discouraging. The thin (e.g., 5/8″) plates tested were “shivered to fragments,” and produced “numerous dangerous splinters.” By 1859, it was recognized that a wooden backing increased the value of the armor. (Beeler 34).

All else being equal, the thicker the plate, the better. However, it might not be possible to make plate greater than a particular thickness. For example, in 1854–1858, the maximum plate thickness was 4.5 inches, and even then quality was irregular. The problem was that the rolling mills couldn’t handle this thickness, and thus the plate had to be shaped by hammering. (Garrison 346).

If so, lamination (fastening several thin plates together) is necessary, and this was done with American Civil War (ACW) ironclads. For example, the Monitor’s turret had eight thicknesses of one-inch iron. (Id.) The laminates could be just of metal sheets, or of metal alternating with wood. (EB11). However, laminated armor was very vulnerable to perforation by pointed projectiles. (Browne 5).

This author believes that the thickest wrought iron armor was that of HMS Inflexible (1881); two layers totaling 24 inches. (Cotterell 365).

Ordinary cast iron, which was much cheaper than wrought iron, was tested, but found wanting because of its brittleness. A few round shots, at point blank range, destroyed a four-foot-thick block. (Holley 629). Nor was it practical to attach cast iron face plates to a wrought iron foundation; the bond failed. (NAVORD).

Wrought iron was easy enough to form into plates, or to roll into cylinders to make turrets, but did not lend itself to formation of hemispherical caps. The difficulty of manufacturing increased rapidly with the thickness and area of the piece.

Gruson’s chilled cast iron was introduced in 1868. Since it was cast, it could be formed into any desired shape, in great masses, and with variations in thickness. It was harder than wrought iron and thus more capable of deflecting oblique shot. Together, these features made it an excellent armor material for a turret. Unlike wrought iron, it was not easily perforated, but rather yielded by fracture. It was thus most vulnerable to repeated fire on the same part of the armor. However, increasing its thickness increased its resistance more than was the case with wrought iron. Unfortunately, to provide sufficient resistance to fracture, it had to be used in thicknesses that made it unsuitable for naval use.

In 1876, tests demonstrated that steel was superior to wrought iron insofar as resisting perforation was concerned by a ratio of about 1.3:1. However, it was susceptible to through-cracking and consequently in 1877 Cammell and Ellis introduced compound armor, in which a steel face plate was melted (Cammell) or cemented (Ellis) onto a wrought iron foundation. In 1887, Tressider found it possible to harden the steel face by “chilling the heated surface of a plate by means of jets of water under pressure.” (EB11/Armour Plating).

By 1880–1890, homogeneous steel plate could be made adequately resistant to through-cracking and was considered equivalent to composite plate of equal thickness for resisting perforation; perhaps 25% more effective than wrought iron. (NAVORD). In 1891, Harvey reported how to face-harden an all steel plate, in a manner described briefly in EB11. The same process also toughened the back of the plate; carbon content was highest at the face and decreased as you went deeper. (NAVORD). The Harvey and Tresidder processes were combined (“Harveyized armor”) and found to give steel twice the resistance of wrought iron. (EB11).

The first nickel steel armor plate was manufactured in 1889. Experimentation with addition of chromium began in the 1880s, and the problems of making large nickelchrome steel ingots were resolved in 1892–3. (NAVORD). Nickel steel armor rendered compound armor obsolete, making possible reduction in armor thickness and thus, increased speed or armored area. (Breyer 39). In one place, EB11 says that “modern” armor contains 2–5% nickel and 1–2% chromium, and in another, 3–4% nickel, 1–2% chromium, 0.25–0.35% carbon, and 0.3–0.75 manganese.

For thicker plates (5+ inches), “Krupp cemented” (KC) armor was favored, and was about 15% more effective than Harveyized armor. As the name implies, the plate was passed into a cementation furnace, bringing its face in contact with “specially prepared carbon” and the temperature gradually increased to cementation temperature, then decreased. It’s then reheated and plunged into an oil bath, reheated to a lesser temperature and quenched, reheated and bent so shape, and differentially heated in a furnace (front made hotter than back), and both sides douched with jets of cold water on pressure. The EB11 is deliberately coy as to “details of temperature, etc.” so these will have to be reconstructed by trial-and-error (cp NAVORD).

Steel may be cast or forged; casting is cheaper and more suitable to producing complicated forms (turrets), but forged steel is freer from flaws and thus more resistant.

The “Baltic War” ironclads of canon had armor formed by overlapping steel planks obtained by running up-time railroad rails through a rolling mill; the plates were bolted to a wood structure, probably oak. (TBW Chap. 38).

Armor Distribution

When armor was first introduced, most naval engagements were at close range, and the guns projected shot on nearly flat trajectories. Hence, it was the sides of the ships that needed protection. The only early exception was on ships of the “Monitor” type, which had very low freeboard, and therefore needed either deck armor or an armored “breastwork”. (Brassey 108ff).

When engagement ranges increased, even more conventional ships had to face plunging fire, and therefore could benefit from deck armor. OTL ship designers didn’t have to worry about aerial bombardment until the early-twentieth century. The situation in the 1632 universe is quite different; warships were attacked from the air at the Battle of Wismar. For that matter, mortars could be used to fire “deck-piercers.” (Smith 44).

As guns became more powerful, armor had to be made thicker, and compromises had to be made. (Figure that every square foot of one-inch-thick plate is about 40 pounds—Friedman 13).

Weight could be reduced by confining the armor longitudinally to a “central citadel,” of perhaps one-third to one-half its length, leaving the ends unprotected. This still protected the armament, because the “broadside ironclads” were replaced by ones having just a few big guns in turrets or barbettes. The propulsion system could be placed below the waterline, and the ends could be finely divided by bulwarks and packed with cork so that many hits would be needed to substantially reduce their buoyancy. Warships could also have armored “conning towers.”

On the early “broadside ironclads”, the armor ran from “the upper deck down to 5 or 6 feet below the waterline” (Brassey 106). Even that was a compromise; on a warship of 75 foot beam, a 10 degree roll would be sufficient to expose the unarmored portion, and if the ship had already been cruising for several months, it would be riding high and a lesser inclination would be enough.

Still designers found it necessary to limit the vertical extent of the armor. Belt armor extended from a few feet above the average waterline to a few feet below it, covering the areas that would be intermittently submerged and exposed as a result of wind and wave action. The purpose was to safeguard the buoyancy of the ship. In some warships, the belt extended to the main deck (8–10 feet above water), whereas others had a shallow belt, reaching only 2–3 feet above water. To save weight, the lower and upper edges of the belt armor could taper out.

Naturally, it was also important to armor the ship’s battery, to preserve its fighting ability. In some ships, only the main guns were armored, it being figured that the secondary guns wouldn’t be targeted at the long ranges at which armor was most effective; only the enemy’s main guns would be in action and they had to take out their counterparts.

Armored bulkheads are of some importance in limiting the damage caused by a shell that penetrates the main armor and explodes inside the battery. The explosion of a 12-inch shell on the Hiyei put out of action men 80 feet away. (Brassey 351). If each gun has its own compartment, protected internally by one inch bulkheads, then an explosion in a turret won’t kill the crews of all of the guns it carries.

Explosive shells vs. Metal Armor

Unfortunately (at least from the attacker’s point of view), surface explosions mostly went “splat!” on armor. Experiments at Woolwich (1856) showed that “plates of 5-8ths inch thick prevented any shells then known from passing through and exploding inside the ship.” (MurrayTPSB 24; for later experiments, see AlgerHENW). For shells that partially penetrated, most of the force of the explosion took the path of least resistance, back up the cavity carved by the projectile.

Naturally, the more powerful the shell, the greater the thickness it could shatter, but a good rule of thumb was that armor thicker than one-third the caliber of the shell could protect the ship from a high explosive shell. (Brassey 366).

Against somewhat thicker armor, up to say half a caliber, the semi-armor-piercing projectile was more effective. (367). Beyond that, an armor-piercing shell is required. For shell types, see part 4.

Armor Penetration

There are several different modes of penetration. In discing, the plate first “dishes” (bulges backward), and then the bulge breaks free as a disc. In punching (plugging), the projectile shears a disc of diameter slightly greater than that of the projectile, and pushes it backward, thus shearing the disc behind it, and so on until a circular plug is forced out the back of the plate. In petalling, the plate flows away from the point of the projectile, forming concentric ridges, and tensile stresses crack the plate radially, forming “petals” that are peeled away. This exposes the next layer of the plate to the same attack, and so on until the plate is perforated. More than one mechanism might be in play simultaneously, and it appears that which is favored depends on the shape and hardness of the projectile, and its impact speed. (Dean).

I would expect that the shearing work that must be done in punching through armor would be proportional to the circumference of the projectile, and thus its diameter. (Owen 273), but work done pushing the disc onward would be proportional to the square. For petalling, work is done in plastic deformation (proportional to diameter2), moving the material (diameter4), and bending the petals (proportional to diameter). (Pol). so the relationship on diameter is not a simple one, either.

Damage can occur even without complete penetration. The impact can cause cratering (digging a pit on the front of the plate), rendering the affected area more vulnerable to a second shot, or spalling (fragmentation of the back), which might cause injury to the crew or interior. Also, repeated impacts can cause racking—shaking the armor loose from its fastenings. Racking was best achieved by heavy, large diameter projectiles with low velocity. (Owen 274).

Numerous formulae exist and I am only going to quote those that appear in EB11/Armour Plating. I will reduce them all to a standard form.

First, Tressider’s formula (early 1870s) for the penetration of wrought iron:

t = 10-4.44205 * w 0.5 * v1.5 * d-0.5.

with t thickness (inches), w projectile weight (pounds), d diameter (inches), v impact velocity (fps). This formula is modified for Harveyed or Krupp-cemented armor if capped projectiles are used, the d is then multiplied by a further factor of four.

Krupp’s formula (1895) is given for use when face-hardened armor is confronted with uncapped armor-piercing projectiles:

t = 10 -3.1766 * w0.5 * v1 * d -0.5.

For calculating the effect of an oblique hit, EB11 recommends replacing “t” with “t sec theta”, where theta is the angle of obliquity from the normal, and doesn’t exceed 30°.

A third penetration formula, that of De Marre (1890) for mild steel, is given in terms of the impact velocity:

v=10-3.0094 * t0.7 * d0.75 * w-0.5,

but rearranged for thickness we have

t=10-4.299 * w0.714 * v1.429 * d-1.07

The De Marre formula in fact became quite popular, and was extended to new projectile/armor combinations by replacing “v” with “v/C” where “C” was the De Marre coefficient and reflected how well the armor performed relative to French 1890 nickel steel plate. (Okun).

If you have been paying close attention, you may have noticed that all of the iron/steel penetration formulae assume that penetration decreases with increasing projectile diameter, whereas Beauchant said that in wood, penetration was proportional to the diameter. Is wood really so different from metal?

I suspect not. What I think is going on here is that the metal formulae separate out weight and diameter as separate factors, whereas Beauchant was using diameter by itself. Weight and thus kinetic energy would be proportional to the cube of the diameter, so if resistance were proportional to the square, the penetration would be proportional to the diameter.

Non-Metallic Armor

Supposedly, at the Battle of New Orleans, cotton bales were used to protect American ships from British fire. The British eventually decided to put this to the test, although they used wool instead of cotton. At 100 yards, a 68-pounder with a service charge put its shot through 11 feet of wool and 12 feet of solid earth besides. (Barry 216).

Coal-burning steamships have used coal bunkers as a secondary defense. It was determined that “roughly speaking, 2 feet of coal is equivalent to one inch of iron.” (Browne 341).


General Design Issues. The principal problem in designing a shell was controlling when it exploded. You didn’t want it to explode when it was accidentally dropped, or in the gun barrel, or on a mere ricochet, but only when it struck the actual target. If you were attacking an armored target, results would be better if the explosion were delayed until the projectile penetrated the armor.

In modern warfare, fuzes are manufactured in enormous quantities. If a fuze design requires tolerances which are high by then-current standards, the defect rate may be high. Even in the early 1900s, it was desirable that tolerances not be stricter than 0.01 inches. (Young 506ff). Also, one must consider the number and size of elements, and how they are to be fabricated, e.g., a complex shape might need to be machined rather than cast. If the fuze is drastically different from those made in the past, then you might need to make many new gauges, tools and machines in order to make it in volume.

Safety in manufacturing is even more important than safety in use (think about the quantity of explosive stored in a munitions factory). Can a part be misplaced or reversed, and if so, what is the effect? How much handling, especially of a sensitive material like fulminate, is necessary? Is the charge subject to deterioration with age?

If the fuze is replacing an older fuze, will it fit the old fuze hole? If not, the projectile must also be reworked or replaced. Does the fuze impair the aerodynamic properties of the projectile?

I expect that the knowledge of artillery fuzes in Grantville will be limited; even books on artillery of a particular war (e.g., Hazlett) tend to slight the issue. ACW reenactors may have knowledge of the fuzes used in the land artillery; for WW II on, there may be veterans who served in the artillery.

Time fuzes. Until the mid-nineteenth century, only time fuzes were used, and the original time fuze was a slow match or port fire placed in the fuze hole of the shell. It was cut to burn for 14–20 seconds. (Marshall 19).

One method of igniting the fuze is described in a 1561 text. The shell was placed in a sack containing a priming compound, with the fuze hole facing the muzzle. When the mortar was fired, the burning charge would ignite the compound which in turn would ignite the fuze. That was the theory, but it didn’t always work. So another approach was to put the shell in with the fuze hole facing the breech. No problem lighting the fuze, but the fuze might be driven into shell and explode right in the barrel. Not good for artillery crew morale. (Peterson 27).

In the technique known as “firing at two strokes” or “double firing,” the fuze faced the muzzle, and the gunner first lit the shell fuze and then set off the cannon. (64). If the cannon misfired, the crew would need to put out the bomb fuze very quickly if they didn’t want to earn a Darwin Award. Peterson says that the method didn’t come into general use until 1650, while Lacombe (231) and Tunis (90) says that it was the dominant method until then.

Come the eighteenth century, and someone discovered, perhaps by accident, that the flash could ignite the fuze even with the fuze facing the muzzle end. That, in turn, meant that it was possible to fire a shell from a long-muzzled smoothbore.

Improvement in fuzes became a matter of urgency as a result of the development of the rifled musket; the infantry could hit artillery outside the range of canister fire. So, survival of field artillery depended on the development of a time fuze accurate at long range so the shrapnel burst would be at the right position (50–130 yards in front, 15–20 feet above) relative to advancing infantry. (McCaul 46,78).

Time fuzes are less important for the navy, but given that OTL seventeenth- through nineteenth-century naval warfare relies heavily on boarding actions, the ability to decimate the enemy crew—or at least keep them from handling guns and sails—at a long distance could come in handy. And one can also imagine naval bombardment of troops marching on coastal roads, and marines using mortars in cutting-out attacks or assaults on batteries. Time fuzes also may come in handy for smoke and illumination rounds.

The ACW wooden fuze had a hollow, fine-grained hardwood (beech, ash) fuze body filled with a hand- or screw press-compacted fuze composition, with uncompacted gunpowder as a primer. The setback forces exerted by firing a heavy gun could throw out the fuze or cause the filler to break the wood, hence the wooden fuze was only used in mortar projectiles. The wooden fuze had the advantage that it could be cut on the spot to just the right length to achieve a desired flight time-to-burst. The paper fuze was also used with mortars; the paper case was a truncated cone and was filled with the burning composition.

In naval use, wood fuzes swelled, so they no longer fit in the fuze hole, and paper fuzes deteriorated once they got wet. (McCaul 64). This led to the adoption of the Navy (Alger) Water Cap (Seacoast) Fuze (70ff) in time for the Mexican War. A cylindrical paper fuze was placed inside a brass fuze case just before loading the piece. The opening was closed with a threaded metal “water cap” having three or four holes, these communicated, through a zigzag mealed powder train in the cap, with the paper fuze. This effectively waterproofed the fuze. Before Alger’s death (1856), he added a lead safety knob; this was broken off by setback forces. The top of the water cap had a recess that was filled with primer, and this was covered with a lead disk until the fuze was ready for use. The fuze was precut for 5, 10 or 15 seconds and the disk was marked to indicate which. The Alger fuze was used by both the navy and the army and was still in use in 1880. (McCaul 76).

The Bormann fuze was perhaps the preeminent time fuze. The fuze composition was completely sealed in a soft metal case. To prepare the fuze for use, holes were punched into the tin sheet at the base to allow communication with the bursting charge. The fuze was screwed into the projectile and a hole was punched into the top of the fuse at a location that determined the fuze time. This hole communicated with a horizontal, curved channel filled with powder. When the gun was fired, the flame would enter the hole and the powder would burn in both directions. (McCaul 82ff; Kinard 186). While the Bormann fuze was safe and reliable, the maximum burn time was typically five seconds, limiting its range to 1,300 yards with a 12-pound field gun (McCaul 85).

The smaller the projectile, the smaller the fuze it could accommodate, and thus the more difficult it was to provide a long enough powder track to achieve the desired time-to-burst. Moreover, in borderline cases, the fuse spoiled the aerodynamic shape of the projectile. (McCaul 107).

Unfortunately, flame ignition didn’t adapt well to rifled artillery. The reduced windage made it more difficult to ignite the fuze. This lead to the development of concussion-time fuzes; the “concussion” (force of firing) ignited an otherwise conventional time fuse. The only concussion-time fuze used in ACW was Tice’s; “the setback forces would break the supports holding a spring in place”; the relaxation of the spring would expose a cotton-buffered fulminate-filled glass vial to loose shot; these would break the vial and the fulminate would flash ignite the powder in the fuze. (McCaul 90ff). Some fuzes (Shenkl and Sawyer Combination) combined a concussion-time mechanism with a percussion mechanism, in parallel, so there were two independent modes by which the shell might explode.

Time fuzes didn’t always burn consistently, and consequently there was interest in achieving greater precision by a clockwork timing mechanism. Clockwork fuzes worked well enough for “enhanced” fireships carrying explosives; Federico Gianibelli’s “hellburner,” a ship-size shrapnel device, was successfully deployed against the Spanish siege troops at Antwerp in 1585. (Gurstelle 104ff). However, fancier engineering was necessary for clockwork to keep ticking as it flew along a ballistic trajectory. The first successful clockwork shell fuzes appeared just before WW I. The source of power for mechanical timers is likely to be either the spinning projectile’s rotational kinetic energy or the potential elastic energy stored in a spring. (NAVORD 3E6).

In 1926, Rhulemann was working on electric time fuzes for artillery shells; the fuze would be triggered at the time of firing and the delay circuit would set off the charge at the appropriate time. This development “lagged,” and Rhulemann switched to developing an electrical time fuze for a bomb. That was successful, and the Luftwaffe adopted his design in 1937. The fuze had a big charge on a storage condenser and a small charge on a firing condenser. When the bomb was dropped, a circuit closed so that the former could charge the latter. The firing condenser was connected to a vacuum tube. When the voltage across this firing tube was high enough, the firing condenser discharged through the tube and detonated the charge. (WRG).

The transistor was invented in 1947; transistors became available in production quantities in 1955; the first fully transitorized artillery fuze (M514A1E1, M728) was a proximity fuze (Arora).

If you haven’t developed proximity fuzes, and you are firing at a small fast target (torpedo boat, aircraft), the chance of a direct hit is not good, and you must either lay down a heavy volume of fire or use time fuzes (or both). (Smith 53). If the fuze detonated at a fixed time from when it was set, as opposed to when the gun was fired, you would need to maintain a uniform rate of AA fire, or the shells would explode at the wrong range or altitude. (Okun).

Percussion fuzes are triggered, directly or indirectly, by the impact of the shell with a solid object (hopefully the target). The discovery (1799) of mercury fulminate (a very sensitive explosive) made them possible, but it took a half-century more to develop the percussion fuze. The fulminate primer (later replaced by potassium chlorate) might be used to set off the shell’s bursting charge directly, or (particularly after the introduction of explosives more powerful than black powder) there might be an intermediate train of lesser “booster” charges.

The earliest booster charges were essentially just smaller amounts of the same explosive used as the main charge, but compressed. The booster series then worked as a delay element.

Later, it was thought safer to use a small amount of a sensitive explosive to set off a larger amount of an less sensitive one. That worked fine in a common shell, when black powder or (better) picric acid booster was used to set off TNT or Dunnite (Explosive D, ammonium picrate). However, in an armor-piercing shell, the booster had to be small and there were a lot of duds because the TNT or Dunnite wasn’t sensitive enough. This mismatch of booster to main charge was overcome by introducing a new, much more sensitive booster: tetryl (2,4,6-trinitrophenylmethylnitramine). (Okun). Tetryl, in turn, has been replaced by RDX, which is in canon.

Impact could either throw a non-explosive “hammer” against an “anvil” bearing a sensitive explosive, or vice versa. Setting off an armed fuze by impact may seem trivial, but sometimes projectiles failed to strike point-first. If the fuze were designed so a glancing blow would result in detonation, then it might also be triggered by a ricochet.

Percussion fuzes may be located in the base or the nose of the shell. Both common and high explosive shells are usually equipped with nose fuzes. Base fuzes were necessary for armor-piercing projectiles; a fuse hole in front would unduly weaken the shell. (Smith 34). On the other hand, the base fuse had to be protected from the propulsion gases, and hence it wasn’t easy to swap or adjust on short notice.

We may classify percussion fuzes depending on how they are armed: (1) by impact with the target (always a nose fuze), (2) by the shock (setback force) of discharge, (3) by the rotation of the shell in flight, and (4) by the gas pressure of discharge (always a base fuze).

The direct impact type minimizes the time interval from impact to detonation. Typically, a protruding plunger would be forced back, as on the German 1916 SQ (super quick) fuze. (Obviously, you would not want to use this kind of fuze with a muzzleloader. Yet Rollins, USP 34,268, 1862, proposed that this would be fine if a “hollow or concave rammer” were used.)

Many early percussion fuzes relied on the setback forces created by the linear acceleration of the projectile in the bore to arm the fuze.

In the Baltic War, the USE ironclads used a fuze described as follows: “an iron ball, coated in an incendiary compound and supported by a thin wire that was sheared off by the sudden acceleration when it was fired. When it struck, the released ball flew forward in the cylindrical fuse’s firing chamber, against still more incendiary compound, and the resultant flash detonated the shell.” (TBW, Chap. 38). It’s plain from the description that this is a Pettman fuze.

EB11/Ammunition, which is presumably Simpson’s design inspiration, characterized the original Pettman fuze as the first practical percussion fuze, and describes it as one “in which a roughened ball covered with detonating composition was released by the discharge of the gun,” and, on impact, “struck against the interior walls of the fuze.”

Barlow (19) provides more details (not available in Grantville). When the gun is discharged, a lead cup is crushed, and the ball is forced back. Various plugs prevented it from rebounding or touching the sides. It is disengaged by the motion of the shell and when the target is struck, the ball hits the wall of the housing and explodes. To prevent detonation on graze, the ball can be covered with two thin copper hemispheres.

The Pettman fuze was tested in 1860: 0/20 premature bursts; 1/20 duds. (HMS Experiment 41). The British adopted it for land (1861) and sea (1862) service. The Pettman fuse was originally developed for use in spherical shells, which of course had the particular problem that you had no control over what part of the shell hit the target.

A modified version was found to also work with rifled shells (Experiment 46) and adopted (1866). However, 1877 American tests in rifled guns reported frequent failures to explode, and it was deemed inferior to, among others, the Schenkl and Hotchkiss percussion fuses. (Ordnance).

What’s curious about the USE usage is that its Pettman fuze was base-mounted, which is not normal for a common shell. As Young notes, base fuzes must be protected from the powder gases by a plate, and therefore cannot be changed or set (if they have variable functions) easily.

The original US percussion fuze was the so-called “West Point style”; it was tested but not used in combat. The fuze had minimal safety features (the percussion cap-bearing slider was held in place only by friction) but the fuze was not placed into the fuze hole of the projectile until just before loading the latter into the gun. (McCaul 87ff).

In ACW, the leading percussion fuze was probably Shenkl’s. This was a double action device: firing the cannon created setback forces that broke a side-mounted safety screw and freed a slider; impact created forces that caused the percussion cap-bearing slider to strike the opposing surface (anvil) and cause a flash. The anvil was flat on one side, concave on the other. In storage and transport, the anvil was screwed concave side-in and thus avoided accidental explosion even if the screw broke prematurely. (McCaul 173ff). The main problem was training-related; the men would occasionally forget to reverse the anvil.

Parrott also developed a percussion fuze, but I figure that anyone knowledgeable enough to recreate its design would also know that it was found to be problematic from a safety standpoint. (see McCaul 179ff).

It’s been suggested that the “Hotchkiss fuze” would work for a great variety of shells. It should be noted that a whole series of percussion fuzes, developed over several decades, bear Hotchkiss’ name. Several different Hotchkiss nose fuses were used in ACW (Hotchkiss USP 35,611, 1862; USP 37,756, 1863). In one ACW design, setback forces moved the slider rearward, “forcing the safety plug out the fuze body and releasing the safety wire,” and opening up a channel of communication between the percussion cap and the bursting charge. (McCaul 182). Lead on the base of the slider was used to inhibit rebound. (187). (The Hotchkiss Combination Fuze was essentially the combination in parallel of a 14-second Bormann fuze with this Hotchkiss Percussion Fuze.)

The version of the Hotchkiss nose fuze used in, say, the Hotchkiss rotating cannon (of the 1870s and thereafter) differed from the ACW version in only minor respects, e.g., it had two safety wires instead of one, and the anvil was given a projecting point to detonate the fulminate on the slider. (199ff; Hotchkiss USP 190,861, 1877).

Another post-ACW development was the Hotchkiss base fuze. This had a firing pin standing up from the base of the fuze, but shielded by an annular slider with a frontal concavity whose lip was higher than the firing pin. On concussion, setback forced the slider backward, exposing the firing pin. The slider had an outer body made of brass and an inner body made of lead; the lead would engage the roughened surface at the base of the firing pin. Upon impact, the slider would carry the firing pin forward with it, striking a fulminate-bearing anvil. (McCaul 200ff). This fuse was abandoned in 1894 as a result of a fatal accident in Chicago. The problem was that during transport, the plunger would creep rearward, arming the fuze. (Hawkins 163). While the Hotchkiss Base Fuze is not shown in EB11, it did make the 1922 Supplement.

EB11 mentioned the possibility of arming the fuse by virtue of the shell’s rotation rather than setback upon firing. Those intrigued by this possibility will have to figure out how this is to be accomplished, but one approach in fact taken OTL was to equip the fuse with spring-loaded side stops that hold the slider in place longitudinally. When the shell spins, the side stops are retracted as a result of centrifugal force and the slider is freed.

This centrifugal arming definitely results, if implemented properly, in a safer fuse. Dropping a shell could conceivably result in a sufficient setback force to arm the older kind of percussion fuse, but “under no circumstances could a shell be accidentally rotated over 2000 rpm.” (Smith 37). Of course, centrifugal arming isn’t practical until we are using rifled guns.

While centrifugal force can thus be advantageous, it can also interfere with the operation of, say, a clockwork time fuze. (Young 499).

The only base fuze described by EB11 is the British one. The combustion gases push in a pressure plate at the base of the fuze, and it carries a spindle forward with it. This action unlocks, in an unexplained manner, a “centrifugal bolt,” which is “withdrawn by the rotation of the shell,” thus arming the fuze. On impact, a needle pellet strikes the detonating cap. (More particulars, Woolwich 166ff).


The ACW Union Navy disliked percussion fuzes because they didn’t allow the shell time to penetrate the enemy hull, and therefore used them only for shore bombardment. (McCaul 64).

With the usual percussion fuzes, there was a slight time interval between impact and detonation, as it took perhaps 0.003 seconds (Okun) for inertial forces to slam the slider into the primer. If that was wasn’t enough delay to permit the projectile to penetrate an armored target, then you need delayed action.

In general, in a delayed action percussion fuze, the impact ignites a primer that in turn ignites a relatively slow burning train of pressed black powder or some other delay composition. When the train is burnt through, it ignites a larger, faster acting detonator charge. The length of the train and the characteristic burn rate of the composition determine the maximum delay. As with an ordinary time fuse, the delayed action percussion fuse may have several settings, each determining a different ignition starting point within the train.

Various additives can speed up (high energy density fuels, oxidizers) or slow down (inert materials like chalk) the burn of a powder train, and small particles tend to burn faster.

While I have found reports of late-nineteenth-century experiments with delayed action nose fuses (Baylay 9), I believe that in WW II the naval delay fuses were base fuses. The base position gave the fuse more protection, so the delay mechanism had time to do its job. (Garzke 501).

One clever (perhaps overly so) design was the Merriam base percussion fuze. Bore gases pressed on pistons, which pushed aside the safety clips, freeing a large ball. On impact, the ball was thrown forward, driving a small ball against a fulminate percussion cap. This in turn ignited a compressed powder ring. However, the impact also threw forward a valve that pressed against the ring and also closed the openings by which the ring communicated with the detonator. Since this valve had a shorter distance to travel, it reached its destination first.

Now, here’s the juicy part: as long as the projectile was forcing its way through the armor, the valve would be closed and the ring would burn just at its edge. But when it came to a stop, or passed into open space beyond the armor, the valve would move back, the face of the ring would ignite, and that in turn would set off the detonator. Nifty, huh? You didn’t need to set the delay to allow for a particular armor thickness, the fuze adjusted automatically to circumstance. (Bruff 340; Smith 36; Merriam USP 475,786). In preliminary experiments with plank targets, the fuze worked as desired. (Merriam 63). However, it appeared to have disappeared into oblivion. the Chief of Ordinance, reporting on trials at Sandy Hook, observed a “distinct difference in the time of bursts with and without the delay,” but noted that even with the delay mechanism cut out “a sensible time was required to produce an explosion.” (Heath).

Lindsay, USP 2,837,999 (1958) proposed that when the shell was impeded, a detonating ball was directed down a spiral passage of decreasing pitch. I have found no evidence this base fuze was ever tested, let alone used.

A more typical delay element was the one proposed by Dunn; this was a hollow tube arranged so that a length of compressed mealed powder had to be burned before it set off rifle powder, providing a delay of up to 0.02 seconds. (Smith 39).

One design problem was that for armor piercing shells, the larger the fuze, the more the shell was weakened, reducing penetration. (Smith 41).

In the early 1900s, the goal was to reliably achieve 40–50 foot penetration (obviously that’s not just armor); it took over a decade to arrive at a practical design, Nichols’ 1928 VD7F variable delay fuze. (Carlisle 74).Unfortunately I have not been able to learn any design details.

The WW II USN Mark 23 base detonating fuze had a spring that was cocked by the deceleration of the shell, and pushed forward the firing pin when the projectile came to a halt. (Okun).

German Navy WW II fuzes could be set for 0.015 or 0.035 second delay. (Okun). The best known American artillery fuze of WW II (M48 series) had a 0.05 or 0.15 second delay element. (Smith 49). The Japanese had the type 91 “diving shell” which, since it was intended to travel underwater, had a base fuze with a delay of 0.4 seconds. (Lundgren 10, 14).

Proximity fuzes may also be of interest for some modes of warfare, but they weren’t developed until WW II, and they require significant technological advances to implement.

Radar proximity fuzes for AA shells, requiring several vacuum tubes, had to fit into a rather small package—they were half the size of a pint milk bottle (NAVORD 3E8)—and survive the acceleration in the barrel (20,000 Gs). (

The most common sort is based on putting a transmitter into the projectile and using the shell as an antenna. If the enemy reflected the radio waves (that would require a metal hull), that would create a detectable and distance-dependent interference pattern. This was filtered and amplified, and when the signal was strong enough (at 75–100 foot range), it triggered a thyratron which fired off a primer by passing a current through a resistance wire. The nose cone had to be non-metallic, so radio waves could pass through it.

Of course, the electronics need a source of power. For example, setback forces break a vial of electrolyte, which fills and activates a battery.

With the old time-fuzed AA shells, it took an average of 2,000 rounds to bring down a plane; with the VT (variable time, a deliberately misleading name) proximity fuze, 500. (

It’s worth noting that it took seven months to progress from the first working shell (June 1941) to the first batch with over 50% reliability (January 1942), And it was still not until January 1943 that it was first used in combat. (Id.) That’s lightning speed for introduction of new military technology.

It’s also amusing to note that the Germans captured 20,000 proximity-fuzed shells at the Battle of the Bulge but didn’t appreciate their significance because their engineers told them that proximity fuzing was impossible—the acceleration of the projectile would destroy, they thought, the necessary vacuum tubes. (Salisbury).

Radio-mediated proximity fuzes may be jammed with a variable frequency transmitter tuned to the beat frequency (between the emitted and reflected waves) sensed by the proximity fuze. (Id.).

Altitude fuzes might be available for shells sooner. These would sense air pressure (i.e., they are really barometers) and thus altitude; that would facilitate placing an air burst above the enemy. Jesse’s Belle has an altimeter. (Flint, 1633, Chapter 11). Because of the motions of the projectile, a fuze barometer would have to be of the aneroid type. Note that the fuze would have to be calibrated to that day’s weather conditions.

Photoelectric. Another kind of aerial proximity fuze detects the reduction in light as a result of the sensor being shaded by the target. This wouldn’t work at night, of course, and it could trigger if the sun moved in and out of the field of view, but it wasn’t vulnerable to jamming. These photoelectic fuzes were also used in WW II.

Magnetic. In WW II, there were torpedoes with magnetic proximity fuzes that sensed the magnetic field around a vessel with a metal hull. Unfortunately, the torpedo’s speed made it difficult for magnetic fuzes to respond quickly enough, and countermeasures (degaussing, towed magnetic decoys) led to their withdrawal by 1942. (Zabecki 1123).

Acoustic. Yet another option was to detect sounds made by the enemy’s engines (taking precautions so that the fuze isn’t triggered by the sound of the firing ship’s engines or by internal noise of the torpedo, or by swimming fish or sea mammals.)

Obviously, that wouldn’t be helpful against a sailing ship target, although the latter might be too low value to waste a torpedo on. To detect a sailing ship, or an enemy whose engines are off, the fuze would have to have active sonar.

Hydrostatic. These detect the slowly fluctuating pressure wave from a ship and must distinguish it from ordinary water waves. They are useful in bottom mines but not torpedoes.

Fuzes for depth bombs sense steady state hydrostatic pressure. The water would press against a diaphragm and, when the pressure was great enough, it would overcome a resistance and actuate the charge.

A moisture fuze was developed by Zalinski (1887). His dynamite shell had holes covered with flaps; water would enter by these holes and complete an electrical circuit. (Hamilton).

Underwater explosions

I do not think that torpedoes will come into common use until enemy ironclads become a significant threat. Against wooden ships, shells do plenty of damage, and they are much cheaper—not just in terms of cost, but also displacement volume—than torpedoes.

However, explosive shells are not very effective against ironclads, and it’s likely to take a fair number of hits with armor-piercing shells to put an ironclad out of action. In contrast, a single torpedo can kill an armored warship, by striking it where it’s weakest—underwater.

In fact, in 1891, a single Whitehead torpedo hit was sufficient to sink the Chilean navy ironclad Blanco Encalada within a few minutes. This torpedo probably had a 60-pound guncotton warhead.


All else being equal, an underwater explosion is likely to cause more damage than a surface one. This can be seen by comparing the above- and below-waterline damage to USS Cole.

Underwater explosions have a multitude of nasty effects. (Webster; Kail; SEH). When an explosion occurs underwater, a shock wave radiates outward, mostly at the speed of sound in water (5000 fps), and carrying about 53–60% of the energy. Some of this energy is lost as the wave moves through the water. The peak pressure at a nearby ship hull of this wave front is roughly proportional to the cube root of the weight of the charge, and inversely proportional to the slant distance between the charge and the hull.

Besides striking the ship, the shock wave also hits the sea surface and bottom, and is partially reflected. The reflected waves may also strike the ship, with a strength dependent on the depths of the explosion and the bottom, creating a pressure load. Depending on the shape of the bottom, the bottom-reflected wave may be focused or spread out.

Interference occurs between the direct and the surface-reflected wave, causing cavitation—a cavity in the water. The cavitation region will close, creating a new pressure wave (cavitation pulse) and consequent “reloading.” Usually, reloading is insignificant if the charge is at a range greater than twice the charge depth.

The explosion also forms a highly compressed, superheated gas bubble, carrying about 40–47% of the released energy. This bubble expands (more slowly than the initial shock wave or the reloading), inertia carrying it beyond the point at which the internal pressure equals the water pressure. The water then forces it to contract. The bubble goes though a series of oscillations of declining amplitude. At the end of the first contraction, 84% of the energy of the explosion has been spent. At the end of the second, 92%. The peak pressure from the first “bubble pulse” is 10–15% that of the shock wave, but it is of much longer duration. Variation in pressure results in variation in the vertical velocity of the ship’s girder along its length, straining it and perhaps even breaking the back of the ship. If the frequency of the ship’s response to the shock wave and bubble pulses matches the resonant frequency, hull girder vibrations build up (“whipping”), making failure more likely.

The gas bubble is buoyant and thus rises toward the surface; and it also will tend to migrate toward the ship. If it collapses against the ship, it forms a high speed jet of water (this is the result of the water at the bottom moving faster, because of pressure difference, than that at the sides or top of the bubble). This jet can puncture the hull, or catapult the ship upward (“kickoff”) and of course the ship ultimately slams back down. Because of these effects, an “underbottom” explosion is, all else being equal, more dangerous.

The various shocks not only stress the hull, they may result in the flight of inadequately secured equipment and in shock damage to secured equipment. And naturally the rupture of the ship hull creates shrapnel that can do damage.

It should be evident from the foregoing recital of potential woes why there was interest in inflicting underwater explosions on the enemy, whether by torpedoes, mines or “diving shells.”


Unless the divers who planted the limpet mines during the Baltic War had pre-RoF experience with underwater explosives work (for the military or for salvage), it’s unlikely that anyone in Grantville will have formulae for predicting the shock wave pressure, let alone the damage.

However, an author of a naval yarn might appreciate some guidance. The Navy Salvage Engineer’s Handbook (SEH) defines a damage factor as the square root of the weight (W) of charge in pounds, TNT equivalent, divided by the slant range (R) in feet between the charge and the hull. if you plot this damage factor (W0.5/R) against the ratio of the charge depth to the slant range (D/R), the required damage factor for a given level of damage is constant up to a D/R of slightly more than 0.5, then declines asymptotically to half (or a bit less) of the original value. At the low D/R end, a damage factor of 0.3 causes light damage (0.5–1″ deflection) to 1/2″ plate and moderate damage (1–2″ permanent deflection) to 1/4″ plate. One of 0.54 causes moderate damage to 3/4″plate.

Early torpedoes carried small explosive loads, and SEH gives a special formula for charges under 10 pounds, that gives the weight of charge (pounds TNT) to rupture a plate:

6,800,000 * SIGMA * t3


where SIGMA is the plating yield strength (psi); t, plating thickness (inches), and


where R is slant range (inches!) and l, stiffener spacing (inches).

The damage factor should also depend on the angle of incidence THETA to the hull, which suggests multiplying it by cosine THETA, but even for glancing blows there is some effect on the ship so for greater accuracy one would have to use a more complex function in which the weight given to the cosine was determined by experiment on a model.

The US Hydrographic Office reported that the explosion of 300 pounds TNT would almost certainly cause pressure hull rupture and therefore presumed lethal damage when the center of detonation is within 25 feet of the pressure hull of a U.S light-hulled submarine (SS284 and previous).” At 25–50 feet, there would still be permanent hull deformation and serious machinery damage. I believe that SS284 had a 7/8th inch thick medium steel hull. (USHO;

The peak pressure from the shock wave is proportional to (W0.33/R) 1.13, except at very close range; the damage factor is considered proportional to the energy density of the wave.

What our underwater explosions are attempting to destroy are likely to be more like a nineteenth-century ironclad than a twentieth-century warship. Consequently, Bucknill’s 1889 formula might be more appropriate. Bucknill first calculates the mean pressure (psi) exerted on the target as


According to Bucknill, a pressure of 12,000 psi is sufficient to “give a fatal blow to a modern ironclad.”

W is the actual charge weight multiplied by the TNT equivalency of the explosive. Bucknill gives 0.25 for gunpowder and SEH gives 0.55. Other TNT equivalencies are ammonium nitrate (0.42), C-4 (1.26), RDX (1.60), and tetyl (1.25).

Q is an adjustment for the angle BETA (degrees) of the charge-target line from the horizontal

(1+ (BETA/90)*(e/100),

where “e” is specific to a particular explosive and is 20 for dynamite and 35 for gunpowder. Q is one if the charge and target are at the same depth.

So far, we have been looking at non-contact explosions. For contact explosions, a traditional formula (Keil 1956) for hull plate rupture is that the explosion from a charge of TNT equivalent weight W (kg) will make a hole of radius R (m) in plate of thickness t (m), if the weight exceeds the critical value 2.72*t, and the radius is then 0.0704 times the square root of W/t. (Rajendran).

A contact explosion is mostly likely to be against the side of the ship, so you don’t gain the advantages of an underbottom explosion. Setting the latter aside, which is more damaging, contact or proximity explosion?

The amplitude of the shock wave should decrease with standoff distance, which suggests that a contact explosion is worse. However, I have seen it argued that the closer the explosion, the less the hull area affected, and what you want is the distance at which the pressure at which the hull fails is exerted over the largest possible area.


The explosive charges for a torpedo were perhaps ten times those for an armor piercing shell, so torpedoes were more worrisome.

The best torpedo defense was to shoot down the torpedo with a rapid-firing gun before it came into blast radius. That failing, a layered protection, with thin flexible membraness dividing the protective area into “concentric” compartments, worked better than simple armor—the latter would have been prohibitively heavy, and rigid armor transmits the shock.

An air-filled compartment would allow the gas to expand freely, reducing the overpressure, but wouldn’t stop fragments. A liquid-filled compartment was less helpful for reducing overpressure, but more so for retarding fragments; the liquid could be fuel or seawater. (Czarnecki).


We have seen ad hoc anti-aircraft fire in canon already. When Hans puts the Belle I on an attack run against the Christiania, Captain Vadgaard realized that there was “no way he could possibly elevate Christiania‘s guns high enough to engage a flying target,” and consequently must rely on musket fire. Helped no doubt by Hans’ decision to launch his rockets at “point blank range,” musket balls do hit both Hans and his plane. (Flint, 1633, chapter 47). Another time was when the ducal fortress in Munich fired all of its cannon, double charged and loaded with canister, at the Monster. “The cannons missed entirely; no single bit of shot from those guns came within a hundred yards of the plane. They were much lower and closer to the cannons than they should have been, and the Monster was a big, slow-moving plane. Still, it wasn’t close enough or slow-moving enough—or, most especially—low enough, to be hit by cannon fire. ”

For defense against airships and aircraft, a ship will need guns that can be elevated to a high angle, whose bearing and elevation can be adjusted rapidly, and which have a high muzzle velocity and rate of fire.

While air resistance has a much greater influence on the movement of a projectile than does gravity, the vertical range of a gun is likely to be only two-thirds to three-quarters of its horizontal range. (Rinker 282). Moreover, wind strength increases with height, so wind deflection is going to be more of a problem.

Tracer rounds may be useful for judging whether the shots are going where they should. These have an incendiary filler in the base; this is ignited when the round is fired. (Rinker 203).

Shells, of course, have an advantage over solid shot, in that you don’t have to hit the aircraft directly, as long as the burst is close enough so that fragments strike it. But then the problem, if you don’t have a proximity fuze, is timing the burst. There was a period in which AA shells used mechanical time fuzes and AA gunners practiced to achieve a fixed rate of fire consistent with the preset for the delay from setting the fuze and firing the gun. (Okun).


Computers are desirable both for compiling ballistics tables and, ultimately, for gunfire control (see part 3). There are a fair number of digital electronic computers in Grantville and those will certainly be put to the first task.

For fire control use, the computer would have to be taken on shipboard and supplied with power, either by battery or by direct power generation. That’s possible, but I doubt that its owner will want to risk it in battle. Instead, it’s likely that the first NTL-made fire control computers will be either analog mechanical or digital fluidic computers. Knowledge of fluidics concepts has percolated out of Grantville by August 1634, see Carrico, “Adagio” (Grantville Gazette 27) and by August 1635 fluidics components are being manufactured in Venice, see Flint, 1635: Eastern Front, Chap. 17. For more background, see Boatright, “The Aqualator” (Grantville Gazette 30).

If the computing elements can be miniaturized sufficiently, computers are also useful for guidance systems for missiles and torpedoes (part 4), and for certain sophisticated fuzes (this part). Fluidic circuits were developed in the eighties for roll rate control of cannon-launched missiles. (Joyce 17).

It has also been proposed that a fluidic device can be used as a generator, ingesting ram air and converting pneumatic energy into electrical energy. Thus, a fuze employing this device would be armed only when the projectile was in flight. (16).

Naval Armament Timeline

Constructing a timeline for naval armament development isn’t easy. We have the following considerations.

Knowhow limitations. There are no replicas of naval guns in RoF Grantville. Hence, we must figure out how to make them on the basis of the maddeningly cryptic references in Grantville literature (likely to say more about what a weapon did than how to make it, and in any event to be skimpy on details), the recollections of those who served in the US military or in reenactments, and the general expertise of Grantville’s scientists and engineers. There is a lot that we will have to figure out the hard way, and a small corps of individuals, initially, with the necessary training. Putting them on one project will slight another.

Technological limitations. Some innovations are practical only when an enabling technology, such as steam power or advanced metallurgy, is in place.

Economic limitations. Even when we know how to make something, there may be trouble finding the resources, buying them, and transporting them to where we need them. Steel is one bottleneck that’s been discussed in canon (the railroads were denied salvaged rail that was needed to build the ironclads), but there will certainly be others. Capital is also probably limited. Again, no nation can expedite every possible project simultaneously.

Additionally, it’s easier to replace projectiles than big guns, and easier to replace big guns than warships.

Social limitations. People are resistant to change and some naval innovations took decades to be fully accepted. In some cases, education is needed to lay the groundwork; you can’t expect mathematical illiterates to use trigonometry to estimate ranges.

Military necessity. There’s no need for armor-piercing projectiles until the likely foe has armored ships. If your ships can safely close to short range and have the advantage there, there’s no need for improving the accuracy of long range fire.

Because of all these considerations, I have only attempted a rough categorization of innovations as being in canon, or available in the near, middle or far future. This article does not establish canon, so feel free to depart from the proposed timeline, but also be prepared to justify what you propose.

I tend to think of “near future” as the mid-1630s, “middle” as the late 1630s, and “far” as 1640s or later. But again, I was deliberately vague. For most innovations, expect them to be initially available in limited quantities and in a crude form with “kinks,” and only move to a mass produced, perfected form over a period of 5–20 years. A “late” innovation might come sooner if someone gave it high enough priority. On the other hand, we can’t make everything a “Manhattan Project.”

The “weird tech” category is for innovations that were only in very limited use in the old time line and might not work as hoped.


Table 5-3
Category In Canon* Near Future Middle Far Future
Gunfounding sand molds (BW), UTC wire-wrapped (BW)* solid casting tube-in-tube, NC wire-wrapped autofrettage
Gun rifled cannon (BW), volley gun (BW), carronades (BW) sliding block breech loader interrupted screw breech loader high-low pressure gun
Laying & Loading elevating screw, powered ammo hoists (BW), metal cartridges clinometer for elevation measurement powered turntables & powered elevating screw hydraulic and electric power
Aiming ring-and-post sight tangent sight; trigonometric methods telescopic and reflector sight; rangefinder; roll gauge; continuous aim methods gyroscopic sight**; Dumaresq; ship stabilizer
Firing & Recoil flintlock (BW); UTC hydraulic recoil (BW) percussion lock; NC hydraulic recoil electric ignition;
Propellant Black powder powder QC (cylinder powder) prismatic powder; ballistite; cordite liquid?
Primer Mercury fulminate (USE), potassium chlorate (French) percussion caps friction tube; experiments with reducing firing interval lead styphnate; electric primer
Shell Load RDX, napalm picric acid shaped charge
Fuzes direct impact improved time, percussion concussion; mechanical time proximity
Fired Projectile common shells (BW); incendiary, finned projectile; high explosive, shrapnel, illuminating & smoke shells armor-piercing shot and shells; discarding sabots base-bleed; diving
Self-Propelled Projectile spin-stabilized (Hale) rocket fin-stabilized rocket; crude torpedoes wire-guided torpedoes missiles, homing torpedoes
Armor thin steel plate (UTC) wrought iron laminate as vertical armor compound armor; turrets and barbettes; deck armor nickel iron; face-hardened; torpedo defenses
Miscellaneous ballistic pendulum; range tables by numerical integration of flight equations gun chronograph; barrel pressure gauge; interior ballistics
Weird Tech multicharge gun, steam gun, pneumatic gun, double-barreled gun, onboard shot furnace, molten iron shell, disappearing gun (sinking turntable), side-to-side slidable deck guns


*Usually as of Baltic War (fought in 1634, USE construction began in late 1633).

**gyro turn indicator in canon Feb. 1635

UTC up-time construction (up-time material in short supply used to build limited number)

NC new construction


At long last, this survey of naval armament and armor—covering both the “state of the art” as of the Ring of Fire, and the lessons to be learned from the “old time line” future—is complete. Hopefully, it provides ample food for thought for those writing about naval battles in the 1632 universe.

My ballistics spreadsheet, bibliography, and some additional notes that didn’t make it into the actual articles, will eventually be posted to the Gazette Extras section of .