Why the Royal Navy’s Argo Clock, Dreyer Table and kindred gear qualify as “computers”
| 1900s problem | Firing a 15‑inch shell across 10–15 km takes ~20 s; in those 20 s both ships are turning, pitching and accelerating. The gunnery officer therefore must solve, in real‑time, a set of coupled differential equations for the future range and bearing of a moving target seen from another moving platform, while also folding‑in gravity, air density, wind, barrel wear, shell drift and the Earth’s rotation. Doing that faster than a human slide‑rule team was the whole point of the mechanical “clocks”. |
|---|
1. The architecture of an Edwardian fire‑control computer
| Sub‑system | Mechanical analogue equivalent |
|---|---|
| Inputs | hand‑cranked own‑ship speed & course; rangefinder range; director‑sight bearing; barometer/thermometer data. Signals were transmitted electrically by synchro repeaters, so data flowed continuously from sensors to the computer. |
| Solvers | • Differential gears added/subtracted speeds and headings |
| • Wheel‑and‑disk integrators continuously integrated range‑rate → present range | |
| • Cams encoded pre‑war ballistic tables to correct for aerodynamic drag, wind, drift, etc. | |
| • Gear trains driven by an electric motor supplied “time”—one shaft revolution equalled one minute of real time, so the whole mechanism formed a real‑time analogue of the gunnery equations. | |
| Memory & feedback | Gear positions were the memory: if you stopped the motor the state of every variable was frozen. Returned‑shot observations were cranked back in, closing the servo loop. |
| Outputs | Continuous shafts for predicted gun elevation and deflection. In ships fitted with Remote Power Control those shafts directly slewed the turrets; otherwise pointers in each turret told local layers how far to traverse/elevate. |
This integrated arrangement is exactly what later generations would call an analogue computer—a device that represents variables by physical quantities (angles, voltages, fluid levels) and continuously transforms them. Contemporary RN manuals already used that language; the Admiralty Fire Control Table is literally described as an “electro‑mechanical analogue computer”. en.wikipedia.org
2. The Argo Clock (Arthur Pollen, 1908‑12)
Pollen’s insight: instead of estimating range‑rate by eye every few minutes, let the machine integrate it continuously. His Argo Clock coupled two disk‑integrators (for range and deflection) to a gyrostabilised plotting table so that true courses—not merely relative motion—were solved in real time. The result spat out a constantly updated firing solution even while both ships manoeuvred. www.navalgazing.netKey “computing” features
- electric motor gave a uniform time base—exactly the analogue of a CPU clock;
- differential gear trains summed own‑ship and target vectors;
- cams added wind, drift and parallax corrections;
- dial faces and synchros served as I/O so operators could overwrite any variable, equivalent to “live patching” registers. Although only six Argo Clocks reached RN ships before Jutland, they proved the concept of a self‑correcting, closed‑loop analogue computer at sea.
3. Dreyer Fire‑Control Table (RN, 1912 → 1920s)
Captain F. C. Dreyer took Pollen’s integrators but simplified the mathematics: instead of true‑course plotting he assumed modest, slowly varying range‑rate and let operators “eyeball‑smooth” a range‑against‑time graph. The Dreyer Table grafted this plotter, a Dumaresq relative‑motion calculator and a range clock together onto one geared work‑bench. RN liked its robustness, and by 1916 almost every dreadnought carried one. en.wikipedia.orgFunctionally it was still a computer: continuous integration on one shaft, real‑time addition & subtraction on differentials, servo feedback to the plot.
4. How these machines anticipate later computers
| Attribute | Early RN clocks | Later analog/digital computers |
|---|---|---|
| Real‑time, closed‑loop | Motor‑driven shafts updated outputs every fraction of a second | Servo/feedback control loops |
| Programmability | Operator changed initial conditions, switched ballistic cams for new shell types | Software / parameter files |
| Parallel continuous math | Each gear train solved one variable, all running simultaneously | Parallel analog channels / digital ALUs |
| Human–machine interface | Synchro dials, hand‑cranks, illuminated pointers | Keyboards, displays |
In short, the Argo Clock and its descendants were purpose‑built, special‑purpose analogue computers—exactly the same category as Vannevar Bush’s Differential Analyzer or the WWII Norden bombsight, only a generation earlier. Naval gunnery happened to be the first mass problem that demanded that kind of real‑time computation, and Britain’s pre‑WWI gunnery engineers delivered it.
5. Limitations and legacy
- Mechanical tolerance: backlash and wear imposed a practical accuracy limit of a few hundred yards at 15 km.
- Operator burden: a Dreyer transmitting station needed ~25 ratings to feed, pace and nurse the machine; the Argo Clock’s complexity scared many gunnery officers.
- Scalability: once radar arrived, purely mechanical integrators could not accept high‑speed electronic inputs, pushing designers toward electro‑mechanical hybrids and eventually all‑digital computers. Nevertheless, the conceptual DNA of the Argo Clock lives on: every modern naval combat system still streams sensor data into a dedicated computer that predicts where the target will be when the projectile gets there—the same problem Arthur Pollen set out to mechanise in 1906.
Further reading
- John Brooks, Dreadnought Gunnery and the Battle of Jutland (separate chapters on Pollen and Dreyer)
- B. Clymer, “The Mechanical Analog Computers of Hannibal Ford and William Newell,” IEEE Annals of the History of Computing (background on integrators and differentials)
- Friedman, Naval Firepower for technical drawings of the AFCT and AFCC.