Peter Lynn ARCHIVE
Since 2015 I’ve been operating a Wehrhahn “Roland“ horizontal reciprocating log saw, made in Delmenhorst Germany in the early 20th century and originally steam powered (now using a 6hp Lister CS diesel) for the Lynn Historic Woodworking Museum. It was in use by the Hoburg family in northwestern Germany until the 1990s when they kindly allowed us to bring it to NZ as the only known example of this style here.
Roland mills cut on both the outstroke and the instroke, using a variable speed pinion engaging with a rack to feed the log through the saw.
From when it was recommissioned in 2015, especially while milling wider logs, the cut has tended to run-off; wander up and down in a wave pattern by as much as 20mm every metre or so.
Bandsaws can have the same problem, especially when blunt and if the saw guides are loose but horizontal reciprocating saws suffer from this more because sawdust doesn’t clear as easily. (The Wehrhahn’s stroke is just 600mm while some of the logs we cut are more than 1m in diameter).
Many theories and a few certainties:
Sawdust build-up under the rear of the blade- causing the blade to angle downwards? But sometimes it goes up- and sawdust doesn’t seem to accumulate on the top of the blade.
Cutting arc (the frame guide tracks are set at a slight angle to cause the blade to move on a curving path- but which is best, concave or convex?
Deflection or clearances in the saw frame guides causing the cut to angle downward as the cutting forces increase-and upwards when the load comes off? Perhaps this is true, but they have to be very loose to have an effect.
Blades heating and expanding more along the leading (toothed) edge relative to their trailing edge, causing them to either flop up or down? But water cooling doesn’t help.
Too much or too little set? Less than about 25% of the blade’s thickness on each side makes wandering worse, as does too much set.
High blade tension helps- but with too much they crack and break.
1.8mm thick blades wander less than 1.2mm ones.
Narrower blades don’t wander as far off centre as wider ones- presumably because they can correct any angular deviation in a shorter cutting distance.
Making the angle at each end of the blade adjustable exacerbated wandering- allowing some angular float works best.
Blade guides (similar to those used on bandsaws) marginally reduces the amplitude of wandering but not the tendency itself -and they limit cutting width to 900mm.
An active blade guide system that sensed running -off and twisted the blade to correct the deviation was ineffectual- probably because the blades have insufficient torsional rigidity.
Larger tooth size seemed to help, as does sharp edge gullets- because both of these improve sawdust clearance?
Starting each cut slowly and from the narrower end of the log reduces wandering- because starting conditions affect whether wandering establishes or not.
Effects of blades with convex or concave shape? Straight seems best.
Swaging (rather than set) reduces wandering. Because of improved sawdust clearance?
Hardwoods wander more than softwoods- except for some softwoods that don’t cut cleanly
Wandering is much more of a problem with wider logs.
Knots and errant grain can cause wandering.
Higher feed speeds cause more wandering.
Blades wander more as they become blunter.
I eventually found the underlying cause of blade wandering in 2019.
It can best be described by considering a 1metre steel ruler held by each end (simulating a blade). When sufficient force is applied to the centre of such a ruler in the direction of its width, it will buckle upwards or downwards. This is exactly what causes reciprocating saws (and band saws) to run off. The solution therefore is to reduce the amount of force it takes to force the blade through the log to below that which will cause buckling.
This explains why slower feed speed reduces wandering, why softer woods are easier to cut straight, why sharper blades wander less than blunt ones and why thicker blades run truer.
It then took another 2 years to find a practical solution for this particular mill- which is special tooth profile:
First attempts were to increase the tooth rake (angle of the tooth’s leading edge) and top clearance (sometimes called back clearance, the angle between a straight edge laid along the teeth and the top of each tooth). This is by the theory that the blade would then self-feed to some extent- pulling itself into the cut.
Trying this on blades made with Left/Right teeth stalled the mill and almost broke the con rod when they aggressively pulled themselves into the log and jammed.
After many tests, it was found that more rake didn’t provide much useful self-feeding and therefore didn’t reduce wandering but that top clearance did. On this mill, 15 degrees of top clearance is the most that can be used for Left/Right blades without risking jamming and even this requires careful engagement until there is a full width cut to limit the risk. Unfortunately, 15 degrees is not enough to eliminate wandering completely.
The original “4 each way” blades were then tried with increased top clearance, (up to 30 degrees) and these didn’t jam- but also wouldn’t clear sawdust on any log wider than 300mm or so, and therefore still ran off badly.
Eventually (September 2021) a blade was made with the crown tooth form used on hand cross-cut saws. With 45 degree top clearance angles on each pair of teeth separated by a 20mm deep by 20mm wide gullet, they not only feed easily but clear the sawdust even in wide logs. Teeth are prevented from pulling in too deep and jamming by the close proximity of an opposite facing tooth.
Tests on possibly the worst log (for wandering) we’ve ever sawn (some species of hard Maple I think) produced a surface as straight and true as a billiard table- even after 6 hours cutting when the blade was getting noticeably blunt.
Whew, success at last.
I’m very surprised that they didn’t use crown tooth forms on these mills back then, as they were well known for other applications. None of the catalogues I’ve seen offered this style of blade for reciprocating mills
PETER LYNN ASHBURTON, Ashburton, New Zealand, September 2021
