I’ve been working on another dining table, this one a “Prairie Style” table for a client out of quartersawn white oak. The design includes a bunch of mortise and tenon joints, which of course I decided to cut on the ShopBot (is there another way?). A dilemma stemmed from my lack of either a tool changer or second Z axis, and the desire to use two bits to create the mortises – a 1/4″ down-spiral for the bulk stock removal, and a 3/16″ down-spiral for the final clean up, mainly to get a tighter radius in the corners. So I was faced with the prospect of either multiple bit changes (2 for each of the 8 parts, plus test cuts), or placing each workpiece back on the table one at a time for the final cuts after doing all the roughing cuts with the first bit. Both options presented disadvantages – the first would wear out my hands and patience, the 2nd would require exact re-positioning of each part on the table for the second operation, and a tedious process of masking and taping the perimeter of each part to ensure solid vacuum hold-down. So I came up with a third option, which involved placing all the parts together on the table, then using each bit to machine all the mortises at one time. Whether you ever want to do this or not is up to you, but the pics will also show my technique for holding solid wood parts of various sizes using only my basic vacuum setup, and no dedicated one-time-use fixtures.
I started with a smoothly surfaced spoilboard, and a simple “L” shaped fence of 1/4″ MDF cabinet back material. Made from scraps left over from cabinet part cutting, it has a non-porous white material on one face, making it perfect for masking uncovered areas of the spoilboard. It’s inexpensive, expendable, cuts easily (won’t damage bits) and creates a complete seal of the spoilboard surface when placed with the white face down. The “L” shaped fence is something I use often to position parts away from the edge of my table where vacuum is the weakest, and to span the space between my wooden stops to make it easier to position small parts.
 
I then tightly stacked the parts to be cut. Accurately machined surfaces allowed the parts to fit snugly to one another, minimizing any vacuum leakage between the workpieces. After closing off unused zones, I covered the remaining open areas of the spoilboard with more scraps of the 1/4″ MDF material.
Taping all the seams and edges with blue masking tape provided a nearly zero-loss vacuum setup, verified by the vacuum gauge showing nearly full vacuum (about 5″ HG with my two Fein vacs) and pushing/pulling on the parts just to be sure. The setup was rock-solid and only took about ten minutes to create.
Now for the real work. I used the tip of a V-bit to record accurate locations for the corners of each set of parts, and drew the parts in those locations in PartWorks. Since the parts varied in thickness, I created separate files for each group of parts and just re-zeroed the Z axis to the top of each group before running it’s file. One file for each bit, for three sets of parts created 6 files. I could have combined the files or created a master file to automate the process, but for a one-time set up it was faster to just run each file manually after re-zeroing the bit.
I made some test cuts in a scrap piece of the same material before committing.
The end result was accurately machined parts, only two bit changes, and a quick setup that used up only some masking tape.
In case you were wondering, here’s how I made the matching tenons.
I have been fortunate to visit a good number of ShopBot equipped shops. This has enabled me to compare many setups and implement my favorite “borrowed” ideas into my own setup. Even though I have seen many ingenious vacuum distribution and valve setups I never really saw one that seemed simple. Our previous system had four 240V vac motors bolted to the bottom of the table. Simple.
 Parts on Table Less Pipe In order to be able to implement a large blower in the future, it was obvious that a valved distribution manifold was in order, even though a less costly vac source would be used short term. My primary objections to the manifolds I have seen are 1) PVC ball valves are hard to operate and seem to get harder with extended use and 2) the plumbing and fittings required to make these valves accessable add to the cost, complexity and restriction of the system. We now have our design criteria for the manifold system, no PVC ball valves and Keep It Simple, Stupid!
 Thru Pipe Glued in and Rounded Over I decided to use 2” knife valves instead of ball valves. Advantages of these valves are: 1) they are equipped with O-ring seals, 2) seals are replaceable, 3) the actuator handles can be extended to allow remote valve placement with up front control. The only disadvantage is that they cost a few bucks more.
Our plenum was designed as a 49 by 97, 4 zone, very typical to many I have seen. We cut the vac holes thru the table with the ShopBot so that a piece of pipe would glue in directly, eliminating flanges or fittings at that location. We used Corian adhesive for this connection. After the adhesive cured the top of the pipe and plenum were radiused using a 3/8” roundover bit.
 Under Table View Showing Valves, Rods and Dump Valve The manifold was constructed using 2 double outlet wyes and 45’s for the branches and a 3” main trunk. A 3” knife valve was installed to allow a quick dump of vacuum for both rapid part changes or emergencies. We plan to add an air cylinder controlled by the software for this valve in the future. The 4 zone valves are installed directly under the vac holes just low enough to allow the extended control rods to pass under the table supports. These extensions were fabricated from ¼” aluminum rods and fastened to the valve using bolt couplers. Holes were drilled thru the 2 center leg supports to allow the rods to pass thru. Holes were also drilled into the table support on the –X end for valve handle placement. After all extension rods were in place, they were marked, removed and had threads cut for the handles. Handles were installed with lock nuts.
 End View Showing 2 zones Open and 2 Closed This system allows a relatively inexpensive and inobtrusive solution for vacuum distribution. Valves are very easy to operate and even when open do not get in the way. Possible one of the nicest features is that under table access is not hampered. We, like most small shop Owners, need every inch of space we can get.
Less than 5 feet of 3″ pipe and under 4 feet of 2″was required for the complete manifold. The 3″ dump valve would only be needed for larger regen pumps that have a long “wind down”.
The final picture shows a front view of the machine with the vac manifold installed. You have to look carefully to see the handles, let alone any of the manifold piping. 
By Fletcher, June 11th, 2010
Our quest to create the perfect solid wood raised panel door on a ShopBot raised many questions. We began by measuring the thickness of our current door panels to determine the needed depth of cut and create a plan for two part jig system to hold the door panels in place while they were being shaped. We also decided that user input and a combination of relative and absolute coordinate references would be the best way to write the program so the user could both size and shape the door panels on the CNC router. Now it was time to start building the jigs, fire up the router, and see how things would really work.
Sizing
Our initial assumption that we could size the blanks directly on the table without the need for a jig proved to be correct. A ¼” double fluted bit did the trick nicely. I wrote the program so that the user could input the width and height, then it would punch a hole at each corner and then cut cut out the doors. This ensured sharp corners. We chose to have the ShopBot make three passes so there was not too much pressure on the panel, bit, or the vacuum. Sizing panels this way might take an extra minute or two but it produced a nice square panel.
Frankenstein’s Monster
We designed a two part jig system using Part Wizard 2. The first part was inserts that we screwed the the door panels to. The second part was the base with clamps to hold the inserts in place during shaping. The base was to be held to the router bed with the vacuum pump. We looked into buying low profile hold down clamps and putting 8 to secure each of the three inserts, but buying 24 clamps was more expensive than expected. So we decided to use a wooden peg and block system instead.
After writing part files to cut the blocks, we got our wood together and used the ShopBot to cut the parts for the jig, then glued and nailed it together. This was a two-day affair, once the solid maple pegs were made, that produced with what must have been at least an 85 pound monster of a jig. Then we screwed one of our panels into the insert and realized that the corners of the panel would likely move up and down during shaping and produce an uneven depth. Also, the screw holes in the back of the panels would have to be filled and sanded, which would be another step in a process we were trying to shorten.
Escape from Castle Frankenstein
After looking at our gigantic gadget, we reconsidered its usefulness. We wondered if the panels could simply be vacuumed to the table directly. We cut only the top part of monster jig (see picture), and designed some “L” shaped corner-chocks to hold panel blanks tight into the corner to prevent any movement along the X or Y-axis’s. We put the panels on the table so the natural bow of the panels was in the middle. This was sucked down by the vacuum and helped hold the edges of the panel securely to the table.
This helped streamline the process considerably by eliminating the need to have two people to get the jig onto the table, to mount the panels on a jig insert, and having to fill screw holes in the back once they were shaped.
From a programming perspective, we chose to have the shaper bit take three passes over the panels. The first and second remove most of the material and the third is a finishing pass to ensure a smooth cove. The ShopBot works from the dead center of the bit, so we needed to know its diameter of the bit to get the cove right. This also meant that a cove needed to be cut into the jig to account for the back half of the bit as it turned, but this was not any real difficulty.
At this point we put the new and much lighter jig on the machine bed, screwed it down, and loaded a panel. We then put the L-chocks as tight against them as we could and started the vacuum pump. After loading the shaper bit, we tried our program. It blew shavings everywhere. There were a few hiccoughs to start with, but they were soon worked out and we produced our first batch of ShopBot-made solid wood raised panels.
Miscellaneous Matters
When our bit supplier came back and saw some of the panels, he was quite impressed. After talking about the process, he suggested that we offset the bit so it only used part of its radius in cutting as the bit is not really designed to cut material directly underneath it. This seemed like a simple fix. Go back to the programming and add an offset to account for the difference between the radius and the part we wanted to use. In doing this,however, I messed up program and the machine kept getting about an 0.125″ off with every pass. The interaction between the offset, absolute and relative coordinate references, and the need to come onto the panel from the corner, was the cause. I finally figured this out by talking with Phillip Fletcher, who has been ShopBotting for many years. By doubling the offset for the first move onto the corner of the panel and removing it for the second and third passes, we solved the problem.
Due to the large amount of shavings and dust produced by the shaper bit, we decided to get a dust skirt. Our only difficulty was that ShopBot’s current version did not fit our machine. So we built one using the CNC to cut the parts. This helped cut down on airborne dust particles and keeping the table clean. One other thing we realized is that we could size the panels in the same jig that we shape them, by modifying the shaping program to account for a ¼” bit and a .375” offset. This meant that we could put the jig on the router table at the beginning of the process and leave it there until the panels were complete. No need to change jigs, only the bit.
Conclusion
Though the door panels may not but perfect, they are better quality and easier to make than our old panels that were made on the spindle shaper. It was a quite a journey from concept to how the process finally worked and required the input of various different people with a lot of trial and error. In the end we developed a low-cost, but efficient jig system, programming capable of accepting user input both for sizing and shaping our door panels, and stream lined process that produced a clean door panel with smooth, sharp lines.
N.B:
For more information regarding the programming or construction of a ShopBot-based solid wood raised panel door creation system please feel free to contact us at 919.683.6123 or vostro420@frontier.com
|
|
