Robinson's Difference Engine #1
The author's small scale model of a difference engine from standard Meccano parts. This model was recently on display at the Computer History Museum in Mountain View, California. It will be exhibited at the 2005 NZFMM convention in Auckland, New Zealand March 26-27, 2005.
Click for a larger version
This model operates on principles very similar to Babbage's original designs, though the constraints of using only standard Meccano parts inevitably mean some aspects of the operation are somewhat different. The model can handle decimal numbers with up to four digits, and up to three orders of differences - similar in scope to the fragment of the original Difference Engine #1 which Babbage actually realized in 1832. There is no reason in principle (other than the limited world supply of 2 1/2" gears and ratchet wheels!) why it could not be extended to arbitrary sized numbers and an arbitrary order of differences. Only two basic mechanisms are involved, those for the addition of individual decimal digits, and for the propagation of carries. The rest is repetition. The machine calculates reliably, producing a result about every 4 seconds - somewhat faster than they can be read off and written down. I have no doubt that if the Meccano of the 1920's had existed 100 years earlier, Babbage would have been entirely successful in his quest. It may be amusing one day to attach a Meccano steam engine to drive the mechanism and therefore realize "computing by steam".
The close-up picture below shows the digit addition mechanism in the foreground, with the carry propagation mechanism behind it. These arrangements are repeated vertically to extend the number of digits and horizontally to increase the number of orders of differences.
Click for a larger version
A number of technical problems had to be overcome. First the representation of decimal digits - since standard Meccano is not well endowed with elements possessing a 10 fold symmetry, second finding a mechanism to reliably readout the value of a stored digit and add it to the adjacent column while retaining the original value, and finally organizing the overall drive mechanism such that the required torque could be distributed reliably with no slipping. The handling of carry propagation turned out to be one of the simpler aspects.
The links below provide more details:
Framework
Arrangement of axes and drive system
Decimal digit storage
Digit readout and addition
Carry detection and propagation
Setup and adjustment
Operation
Parts list
Last modified: 14 March 2005 If you experience any problems with this site, please contact the webmaster |
©
2003-2005 Tim Robinson |