Nelson Bros Engineering Scanner / Optimizers

Lineally Scanned Log Bucking Systems

Simplicity is always best.  And lineal scanning is always the most simple, reliable and least expensive.  So it is natural that NBE would favor a lineally scanned bucking system over a snapshot or transverse scanned system.  We recently had the opportunity to prove our conviction, when we replaced an X-Y scanner and a Table Lookup optimizer with a Profile Scanner and a Real Optimizer.

Lineal Scanner for Log Bucking

Optimization requires a good scan image of the log.  Lineal scanners get better coverage and more accurate profile data than transverse or snapshot scanners, but the log must be stable the while the log passes thru the scanner.  If the log rolls or bounces up and down during the scan, the resulting log image will show crooks or squiggles that are not a true representation of the log’s shape.

Two problems exist with an X-Y scanner.  One is that you only get 4 data points per scan.  Defining a log shape with 4 data points is probably good enough for a Table Lookup optimization, but not for Real Optimization. The second problem is that X-Y scanners require a break in the scan belt.  This break in the belt almost guarantees log instability in the scan zone. 

On the system we replaced, the belt beds had a slight V trough shape that did a pretty good job of stabilizing the log before the scan zone, but when the belt transitioned from the V trough to the crowned head spool the log rolled nearly every time.  Wow!  Just when you want the log to be stable you make it transition from a belt trough over a crowned roll, thru the scan zone over another crowned roll onto another belt trough.  The original Table Lookup optimizer had no hope of bucking for crooks, so all it did was buck for diameter and taper.

The NBE upgrade replaced the X-Y scanner with two JS-20 profile scan heads.  These heads scanned from the sides, so no belt gap was needed.  The two belts were combined and the V trough was continued thru the scan zone.  Log rotation in the scan zone was nearly eliminated.  Cork-screw shaped logs that laid with one end elevated, would still experience some vertical oscillation.  This oscillation sometimes resulted in a more pessimistic solution in the elevated end, but it did not result in a false crook in the log image.  Basically, it worked real well.

Real Graphics, Real Optimization

One of the problems with original system and with any table lookup system is that it does not show a real log image with the all the solution boards.  Lookup systems are configured with sets of empirical parameters such as diameters, tapers, and crook limits instead of board parameters such as wane, width, thickness, length and value.  The lookup solution gives buck cut positions, but no graphic of the log or the stacks of boards.  It is something like flying blind

The NBE log optimization program truly fits lumber into the log image.  At each step down the log’s length, all potential lengths (e.g. 8, 10, , , 20) are evaluated at 2 rotation angles.  The initial assessment attempts 5 different offset/taper sets and 2 cant sizes.  If time permits additional rotation angles and orientation sets are assessed.   After segment optimization is done, all length solutions are evaluated to determine the most valuable combination.  The best part is that the results are displayed graphically on the monitor.

We had much discussion on what to do with the excess length.  We ended up with two modes.  In the Allocated Mode, any excess length is first allocated to the large diameter end, then to the small diameter end and finally any remainder is equally divided among all segments.  Parameters determine the large end and small end allocation.  In the Target Mode, any excess length is allocated to the large and small end.  All middle cuts are cut to the target length.  The mill normally runs in the Allocate Mode, but may run in the Target mode when they want to check the cutting accuracy by measuring log lengths.

The Machinery

The infeed deck singulates and drops the log stems (up to 90+ feet long) onto the belt leading into a Nicholson double ring debarker, the log exits onto the V trough scan belt (must be 240 feet long).  The log travels thru a metal detector and thru the profile scan zone out to a kick zone.  The stems are kicked onto a cutoff belt that feeds a Nicholson SuperSaw.  Using the magic of multiple encoders and end detection photoeye curtains, the VFD stops the stem and the log segments are cut.  Cut placement was always within a inch of the requested length.

It is pretty long.  The entire line must have been 600 feet from the infeed deck to the last log sort.


I have never let a lack of knowledge prevent me from voicing an opinion. 

The soft-stop system (e.g. SuperSaw) appeared to be accurate.  I do not know how much maintenance is required to maintain the accuracy.  The benefit of a soft-stop is that excess length can be allocated to all log segments, thus decreasing the number of boards that will need to be trimmed back due to end defects.  That’s cool

The transverse system with multiple, shifting, buck-saws also allows allocation of excess length, though some exclusion zones may exist.  For 90 foot logs at least 7 shifting saws would be needed.  Transverse system needs to transfer the log stem thru a scanner, then thru multiple staging zones waiting for a solution.

The comparison of lineal bucking to transverse bucking is very similar to that for board edging.  Lineal is less expensive, easier to maintain and more reliable.  I actually do not know if the mechanical system is less expensive, but I am confident that a Lineal Scanner/Optimizer for under $100K would be 1/4th the price of a 90+ foot transverse scanner/optimizer.

NBE Log Bucking Scanner/Optimizer


The Optimizer Computer will be in the computer room. The Optimizer's monitor and keyboard will be mounted near the scanner or the operator to allow for scanner calibration and troubleshooting, plus give the operator a graphical view of the optimizer decisions.

The Supervisor Computer will initially be in the computer room, but can be moved into a maintenance or QC office if desired.  The Supervisor is linked to the Optimizer via an Ethernet network.

The Supervisor Computer normally follows the Optimizer, by reading the data from the 25 log archive, then displaying the solution on the Supervisor Monitor located in the computer room.  The Supervisor computer is not necessary for optimization, thus is can be used for offline functions, such as, reviewing previous logs, printing shift reports or replaying logs from recovery tests.

The SUPER program in the Supervisor Computer can be copied to other computers and be used for training or to monitor the optimizer.


The NBE's Log Bucking Optimizer Solution Screen, 2003.

The most important display is the solution screen shown above.  It gives a graphical display of the solution in real-time, some production and decision text and a list of the previous 25 logs.   This screen is displayed on the optimizer, supervisor and any monitor computers.  The optimizer's monitor is normally positioned within the operator's view.



NBE Home Page


Customer List


Email NBE - (360)236-1644 - March 17, 2003