No CNC Required
Monday 19 September 2011
Robert Ingham shares his technique for producing pinpoint accurate templates from aluminium
With a well-tuned plane and some practice, the fitting of two straight edges that represent a butt joint is well within the capability of any woodworker. However, if the junction between two pieces of wood is a curve or a free-form shape the challenge enters into another zone of complication. This can now be achieved with CNC machinery but it either means investing in the equipment or sub-contracting the process to a specialist.
In the past F&C has covered this very process, most recently with an article by Dave Rigler in F&C 178. Out-sourcing can be a viable option for multiple components and need not cost the earth. The practice is not exclusive to professionals either as there are a growing number of manufacturing houses only too keen to maximise their investment in this kind of technology.
Some time ago a craftsman by the name of Graham Laird who was a professional printer and a keen amateur woodworker, combining his powers of observation and lateral thinking, used offset litho printing plate to solve the aforementioned problem. He showed me the technique and I have exploited it on many occasions. Litho plate is made from high content aluminium sheet which is recycled after the print run that it was used for had been completed. It's possible to get it from a printing company and it's available in a selection of sizes. I Googled 'litho plate' and found companies who recycle the material so it's possible to find it if you can't locate a printer near you.
An A3 sheet of litho plate is 0.3mm thick. It can be cut with scissors or tin-snips but the quality of the edge is quite coarse. However, if it's scored with a scalpel to half the depth of the thickness and then flexed along the cut until it breaks, the edge is very clean and no material is lost in the process.
When the two edges are placed together they fit perfectly. Although aluminium is a very ductile material and is used extensively for extruded sections, it can work-harden if it's repeatedly flexed and will fracture along a line of least resistance.
This line can be created with any tool such as a scriber that will score the surface. After some experimentation I discovered that the score line produced by a scalpel resulted in sufficient penetration to create the fracture line.
The litho plate can then be flexed up and down until the fracture occurs. Geometric shapes can be created with engineer's dividers, (compasses) try squares and set squares. I sharpened one point on my dividers to form a blade rather than a point so that the score line produced was the same as that of a scalpel. For arcs that have a larger radius than the dividers I have made a trammel arm from a strip of wood.
A scalpel blade is inserted into a thin saw cut made with a razor saw at one end and a small pin, which is made from the shank of a 1mm diameter twist drill and forms the pivot point.
To change the radius new holes need to be drilled through the wood strip to accept the pivot point.
Free form shapes can be generated by hand drawing or computer graphics and the paper can then be fixed temporarily to the litho plate and transferred to the plate with a scalpel, pulled in the direction of the flow of the curve. Cut through the paper into the litho plate.
The next challenge is to transfer the profile to the component. The ideal tool for this process is a parallel, ball-race, router cutter. The edge of the litho plate is quite hard and will resist wear because the ball-race causes very little friction.
However, because it is very thin it's difficult to hold in place so the simple solution is to transfer the profile to a thicker, less flexible material, of which MDF is the ideal choice. In theory, the litho plate has to be held onto the MDF in such a way that it can be removed after the profile has been transferred. This can be achieved successfully with dabs of double-sided tape but because the router cutter has a gap between the ball-race and the cutting edges an intermediate spacer has to be employed between the MDF and the litho plate. I use 9mm thick MDF for router templates, which is thick enough to position the ball race so that the cutting edges are effectively used, and 3mm thick hardboard for the intermediate spacer.
Use a pencil to transfer the litho plate profile to the piece of hardboard and cut the shape with a bandsaw on the inside of the line so that the profile is smaller than that of the finished shape. Carry out this process again with the 9mm thick MDF but this time saw on the waste side of the line, leaving about 1mm to be removed with the router cutter.
Fix the litho plate to the spacer with double sided tape so that its edge is inside that of the plate.
This combination can then be taped to the MDF that will form the router template. Although a hand-held router can be used to remove the small amount of overhang, the most effective way to carry out this process is on a router table.
The litho plate and spacer can then be removed carefully and this can be done with a kitchen knife slid in between the two adjoining faces until they are separated.
Having made the MDF template the next task is to fix it to the component. In many cases this can be done again with double sided tape or fixed from behind with screws, positioned so that the holes don't weaken or disfigure the shaped component. As a demonstration of the use of this method of producing accurate male and female templates. I'm going to apply this to the making of a wall hanging mirror with a semi-circular veneered shelf, which is lipped with a contrasting coloured wood. The top of the frame is also a semi-circle and the inner curve is formed with the routed template.
The mirror is an ideal design for batch production so I spent some time making the templates so that the components could be held in place securely and safely with repetitive accuracy. This was achieved with screwed slider blocks and toggle clamps.
I decided to make the curved components from solid wood and chose wenge (Millettia laurentii) to contrast with the paler wood of the shelves. I could have laminated the curves but the striped effect would clash visually with the straight components made from solid wood. The curves started as a straight piece, which had angled blocks glued on so that when they were cut to length the additional blocks formed a square corner providing parallel edges for cramp pressure. This also gave me control of the grain direction and pattern, which needed to be symmetrical with the vertical centre line. Two triangular sections would form a pair of curved components, one for the top of the frame and one for the lippings on the shelf. I cut each angled section, slightly oversize on the bandsaw and then dimensioned them accurately with the sliding fence of my table saw. The 45º angle was set with the stock of my combination square, the square edge against the fence and the angled edge against the main surface of the saw blade. A test cut was made to ensure the accuracy of the setting. A mitre sliding stop was positioned on the sliding table fence and the edges that formed the outer square were cut. The fence was reset to 90º and the internal, end-grain butt joint surfaces were cut. The dimension was transferred with a combination square to the sliding stop. Once again, accuracy was paramount as the components had to fit into a router jig that was made initially with the litho plate technique.
Reinforcing end grain
The butt joints, which had end-grain contacting surfaces were located and reinforced with threaded rod and epoxy resin. I made a drilling jig from an off-cut of the section that formed the straight uprights of the frame.
This consisted of a core block that held a drilled brass insert to centre the hole that would engage the threaded rod and MDF face pieces to position the jig on the ends of the components. The core block hole was drilled with a four-jaw self-centering chuck on my engineering lathe. I used 6mm threaded rod, which is fractionally smaller in diameter than the 6mm hole into which it would be inserted. To ensure accurate alignment of the two adjoining ends when the joint was assembled I made a jig so that the two components could be held together with a sash cramp.
The internal curve was marked out with a pencil using the female litho plate template and the excess was removed with a bandsaw before the joined components were installed in the profiling jig and the internal curve cut to the final shape on my router table. Fifty percent of the cut is made against the grain and this can result in breakout. Because the bulk of the waste was removed with a bandsaw to within 1mm of the line, I made the first pass against the router cutter with a reverse feed. The jig was large enough to hold firmly and the awareness of the feed direction gave me control against the risk of the work being dragged by the cutter. This method eliminated the breakout against the grain. A second pass against the rotation of the cutter ensured the final curve of the internal profile.
The shelf and panel below it were made from veneered MDF. Because the two components were the same width I was able to veneer one piece, which was accurately dimensioned to width so that it would fit into the profiling jig that formed the male profile of the curved lippings.
The excess material was removed with a bandsaw and the final curve was cut on the router table. The section that formed the curved shelf was then cut off on my dimension saw. To avoid breakout on the underside I used my removable spelch board that locates into the groove in the sliding table and fixed a strip down onto the main table on the inside edge of the saw blade, with double sided tape. After the male and female curves had been routed the shelf and lippings were glued together with the expected fit of a straight edged butt joint. Voila! The litho plate template technique.
After the components of the frame had been assembled and glued together, the shelf, which was located with a loose tongue and held in place with screws inserted from the back, was fitted. This junction would be glued together in the final assembly. The rebates that located the mirror and the back were routed before the excess material above the curved top of the frame was removed.
The top curve of the protective MDF back behind the mirror, was formed by being pivoted on a pin against a disc sander. As the pivot hole was concentric with the outer curve of the frame, the back was temporarily screwed in place, the pivot was moved away from the disc by the thickness of the frame and the curve was formed on the disc sander.
As is often the case I had to be very focused on the final sanding of the assembled frame and shelf. Because this process can be quite boring, I kept telling myself to do it with a sense of purpose because I had difficulty waiting for that moment of complete satisfaction. The application of the lacquer that brings the beauty of the grain and colour to life.