There is no technical design document that describes the changes made since the 1992 JGR article. (Hey, we've been too busy trying to get to Mars...) Details relevant to science will be put into the analysis papers, and more technical details will be included in engineering papers that are in preparation. The following information is based on material assembled for technical reviews.
MOLA-1 was in perfect condition until the loss of Mars Observer. A plan for rebuilding an identical instrument from spares was developed at GSFC, that would have delivered an instrument in time for a 1994 launch opportunity. A reflight of Mars Observer was not funded, but some opportunity to redesign the instrument was allowed, subject to the availability of parts (some were obsolete and discontinued), maintaining the same external interfaces. Cost was to be kept to a minimum, with a shipment date of Feb. 1996. The main consideration was to reduce possible risks remaining in the original design, Some changes were also proposed to increase resolution and accuracy.
The laser power-on sequence was modified to lower initial fluence and protect optical coatings from damage. Diodes and other components were selected to improve temperature range and performance. Some attention was given to sources of contamination to the laser cavity. Mountings were modified to improve vibration response and simplify installation.
The start-pulse energy sampling was improved, to sample at least 95% of the laser beam. A temperature control was implemented to stabilize the detector. The detector circuit was improved. A higher efficiency power supply was designed with modern components.
The altimeter needed some way to compensate for terrain-induced pulse spreading. To this end, the electronics added a trailing edge detector so that the return pulse width could be measured.
The resolution of the altimeter TIU is determined by the 100 MHz oscillator and the range of the counters. MOLA-1 had a resolution of 1.5 meters. Finer time resolution was obtained by means of three delay lines, so that 0,1,2, or 3 quarters of a clock cycle could be encoded.
The effect of these changes was to increase data output by a total of eight bits per shot, two bits for the encoder, and six bits for pulse width. The receive pulse energy for the second trigger channel was sacrificed so as to maintain the original data format.
The detection circuits consist of four bandpass filters. Therefore the trailing edge detection scheme does not yield a number for return pulse width directly. The pulse width bits are translated via separate lookup tables for each first trigger channel into an equivalent pulse spread in centimeters. This pulse width measures the footprint-scale slope or roughness of the terrain.