Figure 1. A Bramah cylinder (with key).
21 April 2003
Bramah locks are an important, and relatively little known, part of the history of physical security technology. First manufactured by Joseph Bramah in 1784, these locks were probably the first commercially produced cylinders to offer good security against manipulation, perhaps even measured against today's standards. Joseph Bramah's interests, by the way, went well beyond locks; among his other inventions are the first beer pump, a flush toilet, a hydraulic pump, and a number of industrial machine tools. Perhaps the most remarkable thing about the Bramah lock, however, is that it has been manufactured, with very few design changes, almost continually for over two centuries. The current version of the company is Bramah Security Equipment, London (owned by Jeremy Bramah, one of Joseph's descendents). The locks are not inexpensive (a rim cylinder retails for about US $100) and thus seem to be aimed primarily at the higher ends of the British residential and commercial markets.
In addition to being an interesting design in their own right, Bramahs also played a central role in the evolution of the modern evaluation of lock security and resistance to lock picking. For 50 years the company had a "challenge lock" in their shop window, offering 200 Guineas to the first person to defeat it. A.C. Hobbs made a name for himself when he claimed the prize at the Great Exhibition in 1851, a feat said to have taken him over 50 hours.
I stopped by Bramah's London shop a few weeks ago and picked up a couple of their cylinders. Some photos and comments follow.
Figure 1. A Bramah cylinder (with key).
The basic design could be described as something like a tubular cylinder (the type commonly found on vending machines and U-shaped bicycle locks) whose tumblers (called "sliders") are more like wafers than pins. The current design uses seven sliders. Versions have been produced with as many as 18.
Figure 2. Bramah cylinder face.
Note in Figure 2 the seven sliders radiating out from the center of the keyway hole.
Figure 3. Bramah key (bottom) and standard "137" tubular key (top).
A Bramah key (at bottom of Figure 3) is basically a hollow rod with seven notches cut in the end that mate with corresponding sliders in the lock. The index flange lines up the notches with the correct sliders, ensures that the key is fully inserted before it is turned, and prevents removal of the key when it is only partially turned. The depths of the notches are the bitting of the key; six possible depths appear to be used.
Figure 4. Bitting end of Bramah key.
Seven notches are evenly spaced (every 51.4 degrees) around the key. Note that there is a notch under the index flange, which can't be as deep as the others.
Figure 5. Bramah key (with plastic cap).
Bramah keys are equipped with a protective cap to keep, e.g., pocket lint from collecting in the notches. (Apparently, this practice dates back quite a while; Dickens mentions "a Bramah key with a small plug... to keep the dust out" in the Pickwick Papers.)
Figure 6. Back of Bramah screw-in mortise cylinder.
Bramah cylinders can be disassembled easily; a hex screw on the back (see Figure 6) holds the core in the lock.
Figure 7. Cylinder parts breakdown.
The core (center of Figure 7) contains the sliders and spring and rotates when the correct key is inserted. The silver- colored ring around the center of the core is at the shear line.
Figure 8. Bramah core with key.
The key notches mate with corresponding sliders in the core; when the correct key is pushed into the core, a notch at the side of each slider is lined up at the shear line, which allows the core to rotate with respect to the sliver ring. Interestingly, there is only one spring, which sits around the center pin of the core and is shared by all seven sliders. Only the sliders pushed down the farthest by the key remain under spring pressure when the key is inserted; the others are held in place by friction aginst the walls of the core.
Figure 9. Bramah core (front).
Figure 10. Bramah core (with shear line ring pulled away).
The tabs on the outside of the sliver ring on the core hold the core in place in the shell and form the shear line.
Notches on the inside of the shear line ring (see Figure 10) allow the sliders to freely move up and down when the key is inserted or removed in the neutral position. Notches (at the bitting height) on the outer edge of each slider allow the core to rotate within the ring when correctly keyed.
Figure 11. Bramah slider for deep bitting (with coin).
Figure 12. Bramah slider for shallow bitting
A deep notch at the edge of a slider corresponds to the correct key bitting; a shallow notch is usually also present to frustrate picking (they are not deep enough to fully pass the shear line ring but do allow the core to rotate just slightly). The slider in Figure 11 has a true notch toward the bottom (corresponding to a relatively deeply bitted key) and a false notch toward the top. The slider in Figure 12, on the other hand, has a true notch toward the top (corresponding to a relatively shallowly bitted key) and a false notch toward the bottom. Sliders can also be master keyed, by having multiple true notches.
How secure are these locks? They are certainly well made, and any attempt to manipulate them would have to be correspondingly precise. But they are not magic, and the design would seem to allow for many of the same kinds of attacks that work against other keyed locks. I can imagine several approaches.
Picking of the individual sliders appears to be possible with the same techniques used against tubular pin tumbler locks, using a "variable key" that allows the operator to set each slider to a fixed depth and feel which ones bind. (It does not appear to be feasible to modify a conventional pin tumbler tubular pick for the Bramah, however -- those locks have eight pin positions spaced at 45 degrees, while the Bramah has seven at 51.4 degrees). One complication is that there is only one spring, shared by all seven sliders, and so only the slider being pushed down the most (by the shallowist bitting) remains under spring pressure. That means that testing whether a slider is binding moves it slightly, and the lock would need to be "reset" each time a slider is tested. (A variable key design similar to that of Peterson's tubular pick would remember the current position of each slider and make this less tedious than it sounds). Another possibility is a pick that "grabs" each slider so that it can be pulled back up into the original position after testing.
Decoding by shimming also appears feasible. I found it is possible to insert a .0015 inch shim alongside each slider, and, with proper control of the angle, feel the depth of the true notch. Measuring the depth at which the shim hits the notch effectively decodes that position. Once decoded, it should be (relatively) simple to produce a corresponding key.
Images taken with a Nikon D-100 digital camera with a Nikkor
85mm 1:2.8D tilt/shift macro lens (with Kenko extension tube).
Lit by electronic flash and various reflectors.
All images and text Copyright © 2003 by Matt Blaze. All rights reserved. You may not copy, modify or use these images or text, in whole or in part, for any commercial or non-commercial purpose without permission.
21 April 2003; revised 1 May 2003