Earlier, in the latter part of 2002, Richard had conceived a method of accelerating the scanning process and simultaneously invited Wes Melander and Kevin Keymer to work with him in developing the "Doubler" concept into something workable. In early January 2003, he documented this concept.
In late January 2003, Wes Melander documented their collaborative concept into what appeared to be a workable conceptual schematic. Basically, it promised to substantially increase the speed of scanning for Richard's MK3 board in use at the time. Shortly after, Kevin picked up the continued with the Doubler development and brought it to life as a working system. I assembled for him a MK3 board, to which he grafted a hand wired accelerator board, then sent the hybrid board, dubbed MK3+, back to me for testing. I ran about 150 scans through this hybrid board trouble free and we decided to formally implement the new MK3a board. Over the past month, I have scanned about 100 rolls trouble free.
Previously I was scanning rolls at 2'/minute. With the MK3a board, I have been scanning rolls at 8'/minute, a substantial improvement. It is theoretically possible to achieve upwards of 10'/minute, but then the paper is moving so fast, that the roll-end comes up too quickly, tearing it off the core. At 8'/minute, a typical 88n roll may be scanned in less than 2 minutes.
It is being named the MK3a board on the perception that it is entirely possible that somebody else may well come along with an improved version, at which time it may emerge as the MK3b board.
The following pictures show the fully assembled MK3a board, then pictured in use within my roll scanner. Following these pics, is the MK3a PCB layout, the MK3a PCB overlay, the MK3a Schematic, and its parts list. Many thanks to Richard for his work in integrating the doubler ciruit into this documentation.
Following that is a series of pics of the auxiliary components I have installed in my scanner. The optical encoder is the one recommended by Gene Gerety in his message #969 d/14 June 2002. The inexpensive 12vdc flourescent tube and companion inverter were found by Albert De Boer. On my scanner, the lamp happens to be the same dimension as the inside of the spool frame, making it easy to affix it by friction, held in place by a dab of plastic glue.
The stepper motor and companion controller were discovered by several people within the Rollscanners group. I plan to install one in my scanner when an opportunity emerges. At the moment, my free-running 12vdc pancake gear-head motor and chain drive are working just fine, so I'm a little reluctant to make changes at this time.
The PCI Lava card is for an LPT2 parallel port. This card provides high performance throughput for the mammoth amount of pixel data moving through the system into the computer. Not imperative, but very desireable.
For those who have built a roll transport and wishing to acquire an assembled, complete, tested MK3a board, its 4 cables, a bit-twiddler pot, and a LAVA card, I am prepared to make them available at cost - $135.00 (US), including postage, on the understanding that the files emerging will be made freely available. Not included will be a CIS array, an optical encoder, light source, and stepper motor. Just the board, its 4 cables, a 1-turn bit-twiddling pot and a LAVA card. The only part of the kit I can't test is the cable for the encoder.
MK3a PCB layout.
(Available on request in Corel Draw 9 format.)
MK3a PCB overlay, showing component layout.
(Available on request in hi-resolution)
MK3a Circuit Diagram.
MK3a Parts List
Auxiliary Mk3a Scanner Components
Richard Stibbon's Doubler Concept
Back to Rebirth