For a number of years I have been working on designs for fully-automatic telegraph keys. This is largely because of the technical challenge, since a semi-automatic bug is a quite satisfactory means of sending Morse code. Several quite successful double-pendulum (and double-rotor) keys have been made (you can see them elsewhere on this web site), but the single-pendulum design has proven quite elusive. The major difficulty is in changing the rate and direction of the pendulum while a character is being generated.
The bug shown above (and described in this PDF file) is the first one I have made (after six attempts) that will bear public scrutiny. Its function calls to mind a famous quotation by Samuel Johnson (who pre-dated Samuel Morse by a number of years). He said:
a woman's preaching is like a dog's walking on his hind legs. It is not done
well; but you are surprised to find it done at all."
from Boswell's Life of Johnson
More recently, WB9LPU has said:
“Sir, this automatic key’s operation is like a dog’s walking on his hind legs. It is not done well, but you are surprised to find it done at all.”
from WB9LPU’s Key Corner
Actually, the key, while not a speed demon, does produce decent Morse code and has solved several design problems of long standing. The instrument was primarily a "proof-of-concept" exercise, and the concept has been demonstrated to be workable. Below are some pictures of the key and its parts. This PDF file contains a more coherent write-up of the operation and construction. Here is an audio file (.mp3) that will give some idea of how it sounds at this stage of development, and this (rather blurry) YouTube video shows it in action. There is also another audio file that is a Morse code copy of an illustration in the write-up. It can also give a demonstration of why Morse code is an audible, not a visual, language - you can copy stuff by ear much better than by eye!
This is the base of the unit, with most of the upright components mounted. The key is built on a base of bloodwood, a dense hardwood variety. Most of the mechanism is made of aluminum alloy to reduce the inertia of the moving parts.
At the front, with the cover removed, is the mechanism for changing the position of the magnetic reed switches (contained in the upright cylinders). This adjustment will control the length of the dots and dashes.
The mechanism is built around a pair of camshafts that contain a number of actuators that move the pendulum stops, magnets, contacts, tension spring arms, and dashpot connections. The shafts are supported with ball bearings on either end.
This shows the two populated camshafts in place, with the lever in position. Moving the paddle pushes the front ball bearings against the follower arms on the shafts and moves the attached stops and magnets.
This is another view of the partial mechanism. The pendulum parts are absent, and the top plate, which supports both vertical shafts (lever and pendulum) is not yet in place.
All of the parts are now in place. The pendulum, made of carbon fiber, is pivoted on a vertical shaft that is supported above and below by ball bearings. At the other end of the pendulum is a carrier for the magnets that operate the reed switches to produce the dots and dashes.
This is the assembled key. The side view shows the dot stop and contact (red knob) and the dashpot mechanism that prevents the camshaft from returning to its rest position and imparting its energy to the pendulum. Also shown is the vertical coil spring that provides tension to the dot mechanism.
This is the current status of the design (see the PDF file for a more complete description. There are other design "irons in the fire". One of these is a rotary version which is working but is less stable than the linear version.
This rotary version shows promise, but some re-engineering is in progress. It has the advantage of being more compact than the linear version and is capable of a wider range of speed adjustment. Instead of the camshaft mechanism, it uses two vertically-mounted "flippers" to carry out the multiple functions.