One of the tasks at which the ShopBot excels is processing sheet materials.  I love that I can slide a sheet of melamine or plywood onto the table and a few minutes later pull off a small pile of parts, ready for edgebanding and assembly.  If you have done your design and programming properly all the machining operations – holes, dados, rabbets, notches, pockets, joinery, etc. – are completed by the machine before the parts are cut to shape and size – as long as they are all on the same face of the sheet.  However, when processing sheet goods it is virtually inevitable that you will need or want to perform operations on the other side of at least some of the parts.  Often these back-side details are referred to as “secondary operations” since they normally are performed on the individual parts after the parts have been machined out of the full sheet.  More colloquially they are called “flip ops” since you have to flip the part over to machine the other side.  In a larger shop these secondary operations might be performed on a different machine, freeing up the primary machine to start cutting the next full sheet.  They might even use a different primary machine such as a beam saw to quickly cut large quantities of rectangular parts and use the ShopBot or a similar CNC machine to perform operations only on those parts that need them.  In any case there is a big difference between cutting parts from a full sheet or an oversized blank, and machining operations on a part that has already been cut to its finished size and shape, primarily when it comes to locating the operations accurately on the part.

A few minutes ago this was a full intact sheet of white melamine. Now it's a stack of parts ready to be edgebanded.

When cutting out of a full sheet or a larger blank, the same machine performs the operations and then cuts the outline of the part without the part moving in between, so even if the sheet is not located precisely on the table, or the machine is not zeroed perfectly to the reference point of the table or the sheet being cut, the operations will be in the correct locations on the part.  With secondary operations there are serious issues of registering the part correctly on the table, and having the machine calibrated accurately, so that the operations are correctly located in relation to the edge or corner of the part.  After all, what’s the point of using CNC equipment to process parts if the parts aren’t going to fit together perfectly, or the holes aren’t going to be precisely in the right place?

So a lot of effort is made to address these issues when secondary operations become part of a production setup.  Let’s look at the two issues I mentioned separately.  First, we have to ensure that each part is positioned exactly in the same place relative to the table of the machine.  On a dedicated secondary machine this should be pretty simple – put stops or fences along the reference edges and slide the part up to the fences or stops.  If you are only making holes or pockets, and are not doing any operations to the edge of the part, that would be all that is needed, but if you might be routing rabbets, dadoes or pockets that extend past the edge of the part you might have a problem with the stops being machined away – and depending on what the stops are made of that could get ugly.  When cutting full sheets you might or might not have a similar problem, depending on how close your parts are to the edge of the sheet – I like to put my parts pretty close to the edge to maximize how much of the sheet I can use and minimize waste.  The stops also need to be removable in order to surface the spoilboard without the stops being in the way.  So one of my requirements is that the stops have to be easily removed or dropped out of the way after the sheet or part has been registered against them.

My original stops consisted of slotted hardwood blocks which were bolted to the edges of the table with threaded inserts. They worked well enough for several years but proved to be too cumbersome for day to day production.

My initial solution to this problem was to install large threaded inserts into the edge of the table, and make some hardwood stops with slots so I could bolt them to the side of the table, then loosen the bolts to drop the stops out of the way or raise them into working position.  This worked pretty well for a good while, but as our throughput increased and we were spending more and more time raising and lowering (or replacing) stops, I needed a better option.  I recently re-built the edges of my table to incorporate pneumatic “pancake” cylinders which can be raised and lowered with the flick of a switch.  I borrowed heavily from Gary Campbell’s article and his expertise while designing and implementing this system.  The result is a robust, repeatable system which is fast and easy to use.  We now use these stops for virtually all the flat work that we cut, from full sheets, to individual parts, to flat solid wood workpieces, often in conjunction with an “L” shaped mask or fence to move a small workpiece away from the table edges, as I have previously shown.

The new pneumatic cylinders are mounted to an extension that I added to the base layer of the table. The aluminum pistons pop up or down at the touch of a button and are extremely solid.

So that takes care of part placement on the table.  With the other issue of machine calibration I also started with something simple, that I could implement quickly, but that turned out to be a little more time-consuming than I could justify as our needs increased.  You might ask, “why not just use the proximity switches that came with the ShopBot?”.  My understanding is that the switches that are currently used are somewhat better than the ones that came with my machine, but my personal experience with my own switches is that I could not rely on them to hold the level of accuracy I needed.  Those switches are held on by only a single small bolt, which made them susceptible to rotating slightly if bumped or just due to machine vibration.  (I’ve since started taking measures to prevent this when setting up a new machine, such as sanding the paint off the area where the switch will be mounted, and using a little epoxy in addition to the bolt when securing the switch in place.)  Also, one of my switches had started to fail, and I replaced it with a home-made version that worked reasonably well, but again could not be relied upon for the repeatable accuracy we require.

My initial method of calibrating the X and Y axes was based on how I was taught to set the hairline on my tablesaw fence – by measuring an actual workpiece after making a cut.  In this case I wrote a routine that routed an “L” shaped dado with its shoulder exactly 1” from the edge of the sheet, then paused while I measured the resulting distances with digital calipers, and entered the actual measurements using an INPUT statement.  The program would adjust the proximity switch values, re-zero the X and Y axes to the switches, and make another cut to verify that the current set up was accurate.  How accurate?  For my purposes, I decided that anything within .01” was acceptable.  Using this system I was typically able to get to within a few thousandths in each direction without too much trouble.  We used this system for several years, relying on the prox switches for general use and full sheet cutting, and using the custom calibration program for flip ops or any other tasks where accuracy is critical.  Eventually, however, the custom calibration routine became too cumbersome, as it required an empty corner of a sheet, or a dedicated calibration blank to be placed on the table, and once the corner of the blank was used, a fresh corner would be needed for the next go-round.  Also, if the operator did not hold the calipers correctly the measurement would be less than accurate.  So a faster, more repeatable, less subjective method was called for.

My 1st generation method of calibrating the machine to the workpiece involved milling an "L" shaped dado and measuring the distance from the shoulder to the edge of the sheet. By adjusting the prox switch values and repeating the process, I could get accuracy to within a few thousands of an inch.

So as part of my recent table re-build, I also incorporated a permanent aluminum XYZ zero block near the 0,0 corner of the table.  I used the “Super Zero” block which I purchased from Morris Dovey several years ago, but since I’m not using the fences that came with it, it’s essentially just a small block of aluminum with a hole.  It’s bolted into a recess in the table base layer, and wired to the control board with a double nut on the end of the bolt underneath the table.  The flat top surface of the block acts as a Z zero plate, and the hole allows the bit to plunge in and zero the X and Y axes to the sides of the hole.  This allows us to re-zero any or all axes quickly, regardless of whether or not the table is in use or even completely covered with a workpiece or a full sheet of melamine.  To accommodate spiral and compression cutters, which need to plunge fairly deeply into the hole to ensure that the bit will contact the edges of the hole at its circumference, regardless of rotational position, I embedded a short piece of aluminum tube (actually a closet rod cut-off) into the hole.

The parts for the new XYZ Zero block are ready to be installed into a recess in the base layer. After being bolted down, the top face of the aluminum block was surfaced flush with the table using an aluminum cutting end mill.

Now we can initiate an automatic re-calibration of all three axes at any time, and have the confidence of knowing the cutter is accurately located in reference to the pneumatic stops, and therefor to the edges of the workpiece.  Our accuracy has increased and setup times and re-work decreased tremendously.  My only regret is not doing it a long time ago.  Check out these YouTube videos to see the new stops and XYZ Zero Block in action.

Video link: Pneumatic stops on a ShopBot PRTAlpha

Video Link: Automatic X, Y and Z Zeroing on the ShopBot

The overall setup.

So here are some of the ‘enhancements’ I’ve come up with for my current machine over the 6 years I’ve now owned it.  The first major upgrade was the vacuum table which I have covered previously and which is still working hard every week.  Soon after that I added an air drill, one of the earliest models ShopBot offered.  Unlike the current PRS model air drills, which ride piggy-back on the spindle, the PRT drill is a stand-alone unit mounted to the Y carriage with a heavy steel bracket, which proved somewhat unwieldy and hard to keep in adjustment.  I replaced ShopBot’s steel mounting system with a Baltic Birch plywood chassis (remember, I’m a woodworker) made to snugly fit the Y carriage that allows me to easily raise the drill unit up out of the way when working on thicker material with the main Z axis.  The drill has proven itself indispensable for processing closet and cabinet parts; outfitted with a 5mm brad-point boring bit it has probably drilled about a quarter-million holes by now.

The Air Drill lives in a plywood housing and can be raised up out of the way when not in use.

As our volume of sheet materials increased, the Porter-Cable routers on which I had relied for years turned into a bottleneck, requiring frequent maintenance as the bearings would last only a few months, and limiting the cutting speeds we could achieve.  Although I could only afford the cost and electrical requirements of the smallest spindle available, (the HSD 2.2hp), it has been one of the most significant improvements I’ve made to the machine.  Cutting speeds and quality improved noticeably, the “waterline” that I used to get on the edges of hardwoods when making multiple passes with the Porter-Cable router all but disappeared, and the increased speed range without loss of torque allows greater flexibility for cutting strategies and leads to longer tool life, especially in difficult to cut material.  Nothing lasts forever though, and after 4 years of hard use, the spindle started showing signs of being at the end of life, and was replaced a few months ago with an identical model.

The new geared motor with the larger 25 tooth pinion is on the right.

At about the same time as the spindle upgrade I replaced the X and Y motors with 7.2-1 geared motors, which along with their larger pinion gears provide increased torque and resolution.  Paired with the new spindle, I finally felt that I had a machine that was capable of cutting sheets of melamine all day long, at a reasonable speed (5-6 inches per second) without worrying about the machine bogging down, or having to make compromises on cutting quality.

Testing combinations of spindle speeds and feed speeds with the new spindle and motors installed.

One basic task that many overlook is filing the rails smooth and clean.  The rails were ground straight at the factory and they seem pretty smooth until you look really closely (try a magnifying glass).  Realize that every little bump and imperfection will be picked up by the rollers and transferred to the workpiece.  You can CAREFULLY, while maintaining the original angle, draw a file along the length of the rail to smooth out those tiny bumps and polish the roller bearing surfaces.  Once it’s done the first time you should see an improvement in cut quality, after that it becomes more of a maintenance issue to be performed any time a rail gets nicked or damaged, or if you notice a build up of resin or pitting or rust.  Adding felt wiping blocks should minimize the need for rail maintenance – something that’s on my list for the not-to-distant future.  If you have the newer hardened rails this probably doesn’t apply.

This rusty rail was ripe for some attention.

A little TLC and it's like new if not better.

While re-working the wiring between the ShopBot and the control box for the new motors (my control box is mounted on a nearby wall), I decided it was time to get away from the arched plastic cable carrier that had been one of ShopBot’s signature features since the first few machines were made.  Those plastic arches worked OK but had a habit of snapping in the middle after a few thousand flexes, and mine had been cobbled together with duct tape after being replaced once or twice.  I figured I would order some of the “e-chain” which now comes standard with the PRS machines, and which at least a few PRT owners have retro-fitted to their machines.  Of course, it was a Friday evening when I got started on the wiring, and not wanting to wait until the next week for parts to arrive, I started looking around the shop for something temporary that would work for a week or so.  Four years later, my “temporary” solution is now my permanent solution, as it worked so well it never needed to be replaced.

The bones of my home-made e-chain are a strip of polyethylene and a few blocks of wood, mounted to a piece of 1/2" plywood.

This shop-rigged e-chain equivalent has been working well for several years now. Those are nylon zip-ties holding the wires and air line in place.

Along the way I replaced the recycled tool cart that had been holding up the control computer with a custom-designed gantry which hangs from a ceiling-mounted track, allowing it to be moved out of the way or into working position quickly and easily while keeping wires up off the floor and out of harm’s way.  Plus it looks cool and shows off what I can do to visitors and potential clients.

This rolling computer gantry keeps the electronics safe and handy.

So is the machine “finished”?  Not hardly.  There are still several more upgrades on my list, some of which I plan to detail here in future columns.  The first and biggest is to add a “Retro” PRS-style Z axis as my main Z axis, and replace the aging air drill with the original PRT Z axis, giving me a two-Z axis machine.  This will allow more control over drilling depths and speeds, and the ability to use a router bit instead of just a drill bit on the 2^nd Z axis will create a de-facto two-tool tool changer.  I’m really looking forward to that.  I also plan to enlarge the vacuum table slightly to better handle oversized melamine sheets, add pneumatic stops to speed up panel processing, and to replace the PVC plumbing with a plywood manifold and valves to accommodate more powerful vacuum motors.  It seems like for every project that gets crossed off the list, one or two more get added (I have a lot of lists like that).  So stay tuned…



Two of the completed doors.

I had to start with some basic decisions about how to make the curved strips and how to handle the joinery where the mullions overlap.  I decided that bent laminations would be stronger and easier to machine cleanly than milling the curves out of solid stock.  But the joinery almost had me stumped.  I happened to be looking through a router bit catalog and noticed a router bit set for making the exact same joint with straight stock.  That set would not work with my curved workpieces, but it enlightened me to the concept of using a fancy half-lap joint to get the pieces to mate cleanly, with the profiles matching up in all corners of the completed joint.  I drew the overlapping shapes in my CAD program and zoomed in on the area in question to try to visualize what exactly had to be machined.   It took a while, but I finally figured out that by bisecting the joint vertically and horizontally, I could rout out the top-right and bottom-left quadrants of one piece, and the opposite corners of the other piece.  In fact, by holding one piece upside down during machining, the two halves could be machined exactly the same way, and when the upside down piece was flipped right-side up, the parts would mate.  I had my plan.

Using CAD to figure out what had to be machined.

I saved all the narrow scraps of maple that were generated from the rest of the job, and ripped them into strips about 1/16″ thick.  After measuring the door frames, I drew the shapes of the forms that would be needed for the glue-ups, and nested several sets of forms onto a piece of leftover 3/4″ plywood.  The glued strips were clamped into the forms and left for a day or two, then wrapped with stretch-wrap to maintain their shape until all the glue-ups were completed.

Cutting the forms out of 3/4 inch plywood

Each curved strip is glued up of 6 laminations

The glued up assemblies were kept in shape with stretch wrap until machining day.

Once the curved shapes were ready for machining, I loaded a scrap piece of melamine on the table and routed out the area where the workpiece would sit.  This allowed me to precisely place one of the male forms on the table  and use the female form with some cam clamps to firmly hold the workpiece in the correct position.  The routed area also created a flat  reference surface for the workpiece to rest on, allowing me to start by  surfacing the tops and bottoms of each strip to create clean surfaces and a uniform thickness.  I then used a profile bit to shape the shoulders of each mullion.  It didn’t matter that the bit also cut away part  of the form.

Some leftover melamine becomes the table fixture.

A workpiece ready for the first surfacing.

After the top face is cleaned up, the piece is turned over and surfaced to final thickness.

The same setup holds the pieces while a profile bit follows the curve.

I machined the half-laps with a 1/16″ diameter straight bit.  I had to order a “long” 1/16″ bit (with a 1/2″ cutting length), and I was sure worried about it breaking when I saw it.  But with such small areas to be machined I was able to use small step-downs, slow feed speeds and ramp settings as gentle as I could get them, and managed to complete all the cuts without breaking any bits.

One half of a half-lap joint.

I thought I’d have to do some hand chiseling in the corners, but to my surprise the overlapping pieces fit together almost perfectly.  They practically snapped into place.

The completed half-laps.

One of the biggest challenges was machining the ends of the mullions to fit the inside profile of the frame.  I drew the profile and used multiple copies of the shape of the router bit to approximate the shape that the ends of the mullions would need.  Then I used the offset and depth of each copy of the router bit shape to create a pass on the router.  It was a little tedious but it worked.  The parts fit together and into the frames so well they hardly needed glue.

Calculating the end cuts using CAD

The ends had to be machined to fit the inside profile of the frame.

Big sigh of relief when the pieces fit perfectly the first time.

Bottom line: this was a successful project.  I started early one morning with the glued-up parts and a vague idea of how to proceed, and ended the day with four custom doors ready for finishing.  Using only scraps, I was able to replicate what my supplier would have charged a high price for, and the customer was thrilled with the result.  Adding up the time spent ripping and gluing, planning, cutting and fitting though, and I came to a realization:  maybe $1,400 wasn’t such a high price after all.

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.

And, in case you were wondering, here’s how I made the matching tenons.

Occasionally I get to enjoy the satisfaction of seeing a long-term project come to fruition.  I recently completed an odyssey that began over 5 years ago when a large black oak tree fell over across the street from my house, and ended recently with the completion of a twelve-sided dining table.

I had only vague notions of tables made of big slabs of oak when I saw the pristine log laying by the road, headed for the firewood pile unless I intervened.   After some consultations with people who knew a lot more about drying wood than I ever will, I built a small kiln around the stack of lumber and dried the wood slowly over a year.  The pieces air-dried another year before casual conversations with a client and a friend produced buyers for half of the eight large slabs, leaving me four “leftover” pieces along with a small pile of fairly massive quarter-sawn oak timbers.

I had plenty of time while the first half of the tree was going through the shop to play with the shapes of the slabs in my CAD program.  I had no idea it would turn into one of the most ShopBot-intensive projects I’ve ever gotten involved with.  Hours and hours of CAD and toolpathing time with DesignCad and PartWorks produced well over 60 toolpaths to machine all the various components, including threaded nuts and dowels for the base structure and a matching lazy-susan which was almost as much work as the rest of the table.

I spent a lot of time designing the structure to accommodate wood movement.  The twelve top segments are cinched together with a steel cable and turnbuckle that is threaded through a cableway which was routed and drilled through the underside of each piece.  I’ve already watched the gaps open and close slightly as the seasons change, although the fit was nearly perfect when the parts first came off the machine.  If necessary, I can dis-assemble the table for cleaning or re-machining if the gaps get out of hand.

more pictures…

The Cabinet entry sheet in Cabinet Parts Pro.

Keep in mind there are dozens of software packages that are focused on creating cabinet parts, and I’m not an expert on any of them, although I’m getting pretty proficient with the one I chose.  I’m only mentioning a few of the many products that are available by name, and I’m not endorsing any one over another.  So if I fail to mention one that works for you (or one you sell), forgive me.  I’m just recounting my personal experiences here, and you should take this (and anything else I say or write) as my personal opinion.  Feel free to e-mail me, and if I’ve made any major omissions or errors, I’ll mention it in a future installment.  A quick Google search for “cabinet making software” or something similar will get you a long list of options to investigate, if you are in the market.

A view from e-Cabinet Systems, showing an assembly in Cabinet Editor view, compliments of Gary Campbell

There are several aspects of cabinetmaking software, and you may or may not need any or all of them, depending on your situation.  The most basic versions allow you to enter cabinet sizes and generate nested sheet patterns and .sbp files.  Cabinet Parts Pro and the entry level of CabinetVision fall into this category.  You use a fill-in page (Cabinet Parts Pro) or go through an Assembly Wizard (CabinetVision) to tell the software what joinery options you use.  For non-cabinetmakers who just want to make a set of cabinets for their own kitchen or bath, or for a shop that will just be cutting parts for other shops, this is a good way to get started and may be all you ever need.  Some cabinetmakers use inexpensive rendering software to generate 3D renderings for client presentation, then enter cabinet sizes into Cabinet Parts Pro to get ShopBot files.  For the budget-minded, this is a good way to go.

This rendering was created in KCDw by Erminio Marrese

Moving up in complexity and price, the software will allow you to create rooms filled with cabinets, and generate shop and/or client presentation drawings including plans, elevations and 3D renderings.  E-Cabinet Systems, KCDw, CabnetWare, and the higher versions of CabinetVision (and many others) all provide this.  With some of these packages nesting and machine output are included, with others you have to purchase an additional module or import the output into another software (such as PartWorks) to do the nesting and/or file generation.  Some packages have pre-made libraries of cabinets that you can choose from to create a room, others allow customization of individual cabinets, such as adding dividers and shelves, and specifying whether each opening has doors, drawers, shelves, hanging rods (for closets), etc.  The beauty of an all-inclusive package is that it is seamless from design to machining (sometimes called screen-to-machine), without having to output to another program to generate machine files.  Parts are accounted for automatically, saving countless hours and freeing up time to work on the layout and aesthetics of the design instead of laboring over cutlists, nesting layouts and other tasks which are best suited to being automated.

The more feature-filled programs also keep track of pricing and bidding, and generate assembly drawings and reports, such as door and drawer lists for building or outsourcing doors and drawers, and hardware lists.  Some do a better job than others of generating photo-realistic renderings, or again, require an additional module.  For truly custom work, the more expensive software packages (CabinetVision, MicroVellum, 20-20, and several others) allow full customization of cabinets right down to the individual parts, such as changing the shape of a part (creating an arch or a notch) or adding machining (a wire access hole, for instance).  The most advanced packages allow the creation of custom joinery and the associated machining operations, such as line-boring or dados, and the ability to apply those joinery operations automatically whenever certain conditions are met, that is, whenever specified parts touch each other in specified ways.

This rendering (left) helped sell the job. The client could see exactly what she was getting. The completed kitchen is shown at right.

As I mentioned, pricing varies widely. For a few hundred dollars you can be cutting parts with Cabinet Parts Pro (developed by Ryan Patterson, who now works for ShopBot), and use a free or low-cost architectural rendering program if you need renderings.  E-Cabinets is a free package (for cabinetmakers) that handles the design and renderings as well as many of the other pricing, drawing, and report tasks mentioned above.  It was developed for users of Thermwood CNC machines (and therefore has impressive tool control features such as slower machining for small parts). If you want to be able to output ShopBot files directly from e-Cabinets, you can buy the ShopBot Link for $1,295.  Many ShopBotters have found this to be a great cost effective way to handle design and file generation.  Many of the more expensive programs, like KCDw and CabinetVision, have tiered price points, and essentially let you choose which level of features you need now, with the ability to add features as needs and budget dictate.  A few programs have rental or monthly payment options, allowing you to use the software as long as you pay the monthly rate, and avoid a large, one-time expense.

Once you have complete control over part sizes, shapes and textures, anything is possible.

It’s important to realize that with all but the simplest programs, there will be a significant time investment required to become fluent with all the options, settings, and techniques that will allow you to create those beautiful 3D renderings and the sheets full of accurately cut parts.  The more powerful and flexible the program you go with, the longer it will take to get really good at it.  There is no way around it (except maybe hiring an expert to run it for you), but there are a few ways to help streamline the process and reduce the inevitable frustration that results from seeing a screen full of icons and not knowing what 90% of them do, or knowing exactly what you want to do but not knowing how to tell the software how to do it.  Reading the Help files is a good way to familiarize yourself

These renderings were made in CabinetVision for a friend's new web site.

with how a software package works.  It’s an obvious resource that some people overlook, and typically if you just start reading from the beginning you can get the basics quickly.  I visited and re-visited the Help files within CabinetVision repeatedly over the first few months, and always learned something new that pertained to whatever I was working on at the time.  Most better software makers offer training options of some sort; with e-Cabinets there is a series of training videos you can purchase and view at your leisure. CabinetVision offers regional classes, in-shop training by the day or one-on-one online training by the hour.  Both also have well-utilized forums, similar to the Talk ShopBot Forum, where you can get answers from other users and even company technicians.  Check to see what training options exist for whatever package you are considering, and then take advantage of them.

After using and learning my new software for over a year, I can say it was exactly what I needed to complete the picture of cutting out projects efficiently on the ShopBot.  In addition, the ability to compose and view the project in 3D on the computer gives me huge benefits in construction and installation efficiencies as I can more easily visualize how all the parts and all the cabinets fit together before I cut the first part out.  The 3D renderings are a big hit with my customers and have helped me look more professional and sell more jobs.

Anyway, over the course of last year, I count around 40 projects of various sizes that I machined on my ShopBot, ranging from V-carving a single line of text into a single piece of wood for a fellow that was building a podium for his church to cutting out 450 sheets of melamine for my closet customer over the course of the year.  Technically the closet jobs were several dozen individual projects, but I’m counting them as one so as not to skew the numbers too much.   In between those extremes, I was able to divide the other 38 projects into six broad categories, beginning with my bread and butter cabinetry jobs, of which there were about a dozen.  Now I know for a cabinet shop to only do one cabinet job a month doesn’t seem like much, but remember I work mostly alone, and a large kitchen project can keep me going for several weeks.  For better or worse, most of my cabinetry jobs last year were much smaller than that, and only three were full kitchens.  The others included a home office, three bathroom vanity projects plus a bathroom built-in, a wall-to-wall dining room built-in (my first project using hickory in 25 years of cabinetmaking), some bookcases, a reception desk (I just cut the parts for another cabinetmaker), and modifications to an entertainment center that I first built 15 years ago that the client wanted to update to accommodate a new flat screen TV.  For all these projects the ShopBot was used to cut all the sheet materials, including drilling the holes for hardware and adjustable shelves.

Only a few of my projects get fully assembled in the shop before being delivered. This wall unit was my first ever project made of Hickory.

This 100-year-old slab of Heart Pine was still so wet with pitch inside that the chips stuck to everything, including the rails and wheels, making it a challenge to machine.

The next most significant group of projects I worked on last year were those that consisted primarily of machining blocks of solid wood.  As I mentioned, these are pretty broad categories, and this one ranged from cutting some disks out of a piece of Bubinga to working on my own still-in-progress dining room table, which I’m building out of four large oak slabs gleaned from a huge tree that fell over in my neighborhood a few years ago.  I had already sold and processed the other half of the eight slabs that the giant tree yielded for a desk top and a couple of tables, and I decided I could get creative with the remaining pieces.  That project alone involved almost 3 dozen machining operations, but unfortunately after spending a couple of weeks on it early in the year I had to put it aside to get some paying jobs done, and it’s been in storage ever since.  I am determined to complete it this year – and then write a column about it.  The other solid wood machining projects included routing logos into twenty four-inch-thick Cypress slabs for the backs of some park benches, milling out a two-inch-deep pocket in a large piece of re-claimed Heart Pine for a mantle (the recess was so the slab could fit over an existing two-inch thick mantle) and surfacing a couple of Ambrosia Maple slabs for a reception desk for a State College in north Georgia.  The architect liked those slabs so much he commissioned a small table using a leftover piece of one of the slabs as a gift to the project manager, which necessitated my having to turn some tapered legs and rails.  I finally figured out how to mount my small lathe to the end of my ShopBot table so I could turn the parts with the ShopBot.  Definitely one of those “why didn’t I do this sooner?” moments.

These slabs of Ambrosia Maple were some of the most beautiful pieces of wood I’ve ever had the pleasure to work with. Watching them being sliced out of the log was a real treat. Necessity is the mother of invention. I finally figured out how to attach my lathe to the end of my table.

I had made this jig some time ago to drill leg ends for T-nuts and levelers; with minor modifications it made easy work of an otherwise difficult task.

Need some angled tenons on the ends of tapered legs? No problem. The finished table. There is an inscription on the underside, V-carved of course.

The third broad category was what I’d consider craft projects.  These also were mostly solid wood machining but more akin to fabricating intricate parts than machining large single pieces of wood.  The first was a fairly complex pair of multi-tiered jewelry storage carousels for a large bathroom project that I had completed in late 2007.  They turned out pretty nicely, but were way too time-consuming considering I made them as a gift to the client as a thank you for commissioning the bathroom project as well as one of the larger kitchen projects I’ve ever fabricated.  No regrets, but I may try to be a little stingier with my free time in the future.  I also made a wine rack insert for a kitchen drawer, and a couple of fairly quick projects as gifts for my lovely bride of almost 20 years.  While these don’t pay off monetarily, the fringe benefits make them more than worth the trouble.  The other notable intricate wood project from last year was to create four sets of interlocking curved mullions for some glass kitchen cabinet doors for my largest kitchen job of the year.  I used the ShopBot to cut out curved plywood forms which I used both to glue up the bent laminations and also to hold the laminated workpieces to the table for further machining – another project that may be worthy of its own column in the near future.

Nice but not profitable. At least I got to keep the pictures.

These curved mullions were one of the more challenging tasks I tackled last year on the ShopBot. The door frames were outsourced along with the rest of the doors for the project.

The completed kitchen.

Next up are a few quick projects that each took less than half an hour from start to finish.  A template out of ½” MDF for a woodworker friend, cutting a glued up blank into an oval for a mirror frame, a quickly V-carved sign for another buddy who has since bought his own ShopBot, and routing the centers out of four kitchen cabinet doors for glass panels are all good examples.  While I can’t make a living doing these little items, an extra 50 or 75 bucks once in a while doesn’t hurt, and, after being immersed in a large cabinet project for several weeks, it’s nice to be able to start and finish a project in less than an hour and get paid right away.

The fifth group from the 2008 Hall of Fame includes a wide variety of jobs that just don’t fit into a real category.  One was a set of large wall mounted plaques for a local country club.  The job consisted mostly of applying moldings to plywood blanks, but since there were only one or two of each style and the moldings that had been specified were special order items I used the Extruder Virtual Tool to make some of the more complex moldings on the ShopBot.  Although they take some time to rout, using a very small stepover to minimize sanding, they didn’t take that long to draw, toolpath and set up for cutting, and I was working on other tasks while the machine did its thing.  Each profile saved me a set-up fee from the molding supplier.  I ran some of the simpler profiles on the router table and saved the ShopBot for those profiles that I couldn’t make with my collection of router bits.  Another oddball project was a large venthood cover for a kitchen renovation, with curved, sloping surfaces on the sides and front.  The ShopBot cut the curved ribs – which I then covered with bending plywood – as well as the simpler parts that made up the structure of the finished piece.  I also cut several other projects that used a bunch of curved plywood parts, including sets for my kids’ school play, a ten-foot-tall plywood “tree” for the school’s lobby, and a 20-foot long S-curved island for my own kitchen.  One of the great benefits of owning a ShopBot is the ability to design almost anything without worrying about how to cut out complex parts, and when creating projects for myself, I enjoy total design freedom since I don’t have to please a customer’s tastes.  Fortunately my wife appreciates most of my designs, and after waiting twelve years for a new kitchen she wasn’t about to complain when I finally came up with something.

Nesting curved parts like this is always a trade-off between time and material yield. Now that PartWorks 2.0 includes nesting, a lot of that tedium has been eliminated.

The assembled prop piece for the school’s graduation ceremony.

The final few projects in my informal survey don’t even fit into the “other” category.  I spent a week in Florida building a vacuum table for a new ShopBotter and helping him get up to speed with his machine.  It was the first prototype of a new manifold and valve design for multi-zone tables, which I plan to develop and offer to other CNC owners.  (Don’t hold your breath though; it’s taken me six months just to write this column, so the R&D required will likely take place over an extended period).  I’ve cut the parts to replace my existing PVC plumbing and valves with this new system, but since there’s no immediate payback and the transition will mean taking my existing table out of service for a few days, it will have to wait until the time is right.  I also had my first experience cutting “plastic” lumber, machining some curved parts for my son’s Eagle Scout project.  Not surprisingly, like most soft plastics, the stuff machined pretty easily.  Following Gary Campbell’s lead, I made a jig for producing dovetailed drawers with my ShopBot.  The jig itself was time-consuming to make, but once it was set up, I could machine a set of dovetails in a couple of minutes.  Unfortunately, the time and cost of preparing the wood and finishing the assembled drawers means I’ll probably continue to outsource my dovetailed drawers, but the jig may prove useful in the future for other machining operations.

Kitchen in progress. I can’t even begin to explain it.

This twelve-sided, live-edge table will be 7’ across, if I ever find the time to complete it.

A sample Rafix cam and stud. The cam requires a 20mm hole and the stud screws into a 5mm hole.

So in this installment I’ll start by describing the fittings I use and how to draw them on the parts to be machined, talk briefly about nesting the drawn parts onto sheets, and finish up with my process for creating toolpaths from those nested sheets. The fittings I use are called Rafix, and I get them from Hafele, a major supplier to the furniture and woodworking industries. (You can get similar products from many other sources as well.) There are quite a few variations of RTA fittings, some of which require holes to be drilled into the edges of the material to be joined, which is a task not particularly well-suited for the ShopBot. One of the reasons I like the Rafix fittings is that all the machining is done to the face of the panel, allowing me to machine for the fittings and cut the parts out of the sheet in a single operation. The Rafix system consists of two parts, a metal stud that screws into a 5mm diameter hole and a plastic insert with an integral metal cam that presses into a 20mm diameter hole. The 20mm hole is positioned slightly overlapping the edge of the part, allowing the cam to mate in a simple butt-joint with the metal stud. Specifically, the 20mm hole is placed with its center 9.5mm from the edge of the part, and the stud is placed where it will fall centered in the edge of the part with which it mates (based on ¾” thick material). So when laying out the parts I simply draw a 20mm diameter circle in the proper position on the appropriate part, and a 5mm diameter circle on the mating part. With my toolpathing system I could actually draw circles of any diameter on the part as long as the center of each hole is in the right place, but it’s easier to visualize the finished product with the holes drawn accurately sized. I typically use two fittings per joint for parts up to 16” wide, with extra fittings if needed for wider parts. I usually place the fittings 2” in from each corner, but I go a little closer to the corners for parts narrower than 12”.

Typical fittings located for a corner joint. The metal studs screw into 5mm holes using a Phillips or #1 square drive bit. The first few can be a little awkward since they are small and the metal is soft, but after a while you can do it with your eyes closed. The cam fittings get knocked into the 20mm holes with a hammer or mallet. They need to be reasonably well oriented with the hole – the flat edge should match up with the edge of the part – but there is some leeway so they do not need to be perfectly aligned.

The assembled joint. The cams are secured to the studs with a turn of the screwdriver and are oriented so that the screwdriver is held at a slight angle to allow clearance for your knuckles. You can use a regular #2 Phillips driver, but a “Posi-drive” screwdriver (which is a special type of Phillips screwdriver) works best if you will be doing a lot of these.

You can tell a Posi-drive screw and driver by the little grooves between the slots in the screw head and the extra little ridges between the four standard ridges of the screwdriver tip. They allow a more positive connection between the screw and driver, which means more torque with less slipping.

If you are not using stand-alone CAD (Computer-Aided Design) software you can draw the parts and the circles for the fittings directly in VCarve (or Partworks). Since I was already proficient in a separate CAD program (I use DesignCad, but any CAD program should be capable of doing the same thing) I still find it easier to do most of my drawings and nesting in DesignCad and then import the nested sheets into VCarve. (I am using VCarve Pro 4.6, but if you are using PartWorks just substitute PartWorks for VCarve Pro for the rest of this article – they work exactly the same way).

Once all the parts are drawn, they have to be arranged (nested) on the sheets from which they will be cut. This is a simple matter of drawing a rectangle to represent the size and shape of the material, usually 48” x 96” or 49” x 97”, and then placing each part on that rectangle. The challenge is to arrange the parts to minimize waste, a process that can be tedious and time consuming when there are a lot of parts of varying sizes. I’ve gotten pretty good at maximizing the yield, at least for the closet parts where there are a lot of repetitive sizes. Some jobs just fit better on the sheets than others. Once the parts are drawn and nested I export the sheets and bring each one into VCarve for toolpathing.

Here’s a sample sheet layout for a small shelf unit with two sides, three fixed shelves (top, bottom, and middle shelf), three rails (one for each fixed shelf) and eight adjustable shelves. The parts are sized so the entire unit can be cut out of a single sheet of plywood or melamine. The little U-shaped vectors on the adjustable shelves will be cut ¼” deep. The resulting recesses fit over the adjustable shelf supports, to keep the shelves from sliding out of position. [Download sample file]

OK, so let’s take each layer and work out the nitty-gritty details. I use layer 1 for the part outlines. As previously detailed I use a roughing pass and a finish pass to cut the parts out. For both passes I use the Profile toolpath with “Outside” selected. For the roughing pass I use a climb cut with an allowance of .02 and a cut depth of .73 (for .75” material). If I’m cutting melamine I set the cutter’s pass depth greater than the material thickness to make the roughing pass in a single cut, but for plywood and MDF I set the pass depth less than the cut depth in order to make the roughing pass in two steps. The finish pass is done with a conventional cut using 0.0 allowance and a cut depth 0.012” greater than the material thickness. For this pass the cutter’s pass depth should be greater than the cut depth so that the finish pass will be cut in a single step. The roughing pass and finish pass toolpaths get saved into a single .sbp file. Even though I mention the part outlines first and use layer 1, the part outlines will be cut last, after all the other machining is complete.

The 2nd layer is for the 20mm holes. For the 20mm holes you could simply use a pocketing toolpath but I wanted a little more control, so I created a small file (RTA Fitting.sbp) which routs one hole just the way I want it. It uses a CC (Cut Circle) command with a spiral plunge to drill a slightly undersized hole, then another CC command at the final size to create the final hole size. It then moves the cutter to the center of the newly machined hole before raising it back up to safe jog height. As outlined in my previous article, I customized a standard VCarve postprocessor so that I can use a Drilling toolpath, but instead of drilling each hole the resulting .sbp file will move the tool to each hole location and run the RTA Fitting.sbp file with a 2D offset.

The sample shelf unit in VCarve/PartWorks. Each layer gets its own toolpath and each toolpath will be saved with its own customized postprocessor (except the outline layer gets two toolpaths and is saved as a single .sbp file). Each resultant .sbp file will be called in order by a master file.

For the 5mm holes (layer 3) I use an auxiliary air drill and another modified postprocessor which drills the holes using the appropriate variables to define the offsets in X and Y between the drill bit and the router bit. If you don’t have an auxiliary drill you can just use a standard Drilling toolpath.

My 4th layer may not apply to many people but for my closet client I have to put notches in the adjustable shelf ends. The notches fit over the shelf supports to prevent the shelves from sliding out of the cabinet. For this I use a Profile toolpath with “On” selected so the bit will follow the vector. I rout the notches ¼” deep in order to ensure a clean top surface since I’m using a compression cutter. Anything less than ¼” deep will lead to chipping of the top surface.

The sheet of parts, ready for edgebanding.

Since I’m using customized postprocessors to output the .sbp files, I have to save each layer as a separate file. So I now have four files for each sheet that I plan to cut. I could just run each file in the appropriate order, but since I usually have a lot of sheets to cut I use a master file which calls the 5mm holes, 20mm holes, adjustable shelf notches and outline files in order for each sheet, with appropriate speed controls, pauses between sheets, and spindle and dust collector switching. My current master file works but I have several improvements in mind so it’s still a work in progress. Hmmm, sounds like an idea for a future column.

The fixed bottom shelf really should have the cams on the underside, to prevent the shelf from falling away from the cams (especially if the unit will be mounted to the wall, as many closet installations are). That means that the 5mm holes for the shelf-to-rail connection need to be on the opposite face from the 20mm holes. Instead of putting the already cut part back on the ShopBot to drill the 5mm holes in the other side, I find it easier to drill both sets of holes normally during the initial machining, then drill the 5mm holes through the part by hand. The resulting unused holes on the underside won’t be seen once the unit is installed. In the picture notice the short piece of dowel over the drill bit to limit the drilling depth, and the piece of scrap material under the workpiece to prevent blowout.

The completed shelf unit. The RTA fittings are surprisingly strong.

It seems too few people realize that the VCarve Pro postprocessor files are text files which can be edited with any text editor program. (To clarify, PartWorks is the same program as VCarve Pro, but it comes with only the ShopBot postprocessors. For the purposes of this article, PartWorks and VCarve Pro are identical, with on exception which I’ll get to.) Our friends at Vectric are understandably hesitant to encourage us to fiddle with the postprocessor (.pp) files, since if you make a mistake it could lead to ruined material, broken tooling, damaged machinery, and/or personal injury. So please don’t tell them that I’m writing this, but if you are reasonably comfortable with editing .sbp files and understand what the various ShopBot commands look like and how they perform you should be able to successfully modify a .pp file to your liking.

First, be very careful when modifying a postprocessor file as every character makes a difference and the results can be unpredictable. Be sure to keep a copy of the original postprocessor file, or, as I do, save the edited file under a new name so the original file is kept unchanged. You can make the modifications in any text editor including the ShopBot text editor. The postprocessor files are found in the “Program Files\VCarve Pro 4.0\PostP” folder on your hard drive. You can name the new file whatever you want – the name of the file will not affect the performance or even be visible while using VCarve Pro. For simplicity’s sake I’ll use as an example the “ShopBot_alpha_Arc_inch_router_control.pp” file – others may vary slightly but should be similar enough to follow along. Open the file and familiarize yourself with the layout and contents. Any line that begins with a “|” or a “+” character is a comment line and is there for information only. So you can see the first paragraph is a list of the authors of and contributors to this postprocessor file and the dates on which their contributions took effect. Go ahead and add your own name, date and the changes you are making so you’ll have a record of your work. Just remember each comment line has to start with a “|” or a “+”

Just after the first section of comments is the line:

POST_NAME = "Shopbot (arcs)(inch)(alpha_control)(*.sbp)"

Whatever type is between the quotation marks in the above line is what will appear in the postprocessor drop-down menu when you go to save a toolpath file in VCarve Pro. I suggest changing it to something short, sweet, and descriptive, such as “Plywood” or “RTA fittings”. You do not need to include the (*.sbp), but don’t forget to leave the quotation marks. (Here is the exception for PartWorks users – you will have to have the name “Shopbot” included somewhere between the quotes.)

The next several paragraphs are used by VCarve to set variables which will be used later in the file. I have no idea what most of them are for but some of them are obvious and if you are sure that they will not be used in the final postprocessor file than you may safely delete them. For instance, if you do not need to set the spindle speed from within the resulting .sbp files, you can delete the following section:

+------------------------------------------------
+ Spindle Speed
+------------------------------------------------
var SPINDLE_SPEED = [S|A||1.0]

If in any doubt about any line or section, just leave it be, it won’t hurt to leave it in.

Farther down the file you’ll find a section that begins with this:

+---------------------------------------------
+ Start of file +
+---------------------------------------------
begin HEADER

followed by a bunch of commands and/or comments. Whatever follows the above will appear at the beginning of each file which is output using this postprocessor. Since I use a master file to call all my sheet cutting and drilling files in the proper order and to control things like cutting speeds and turning the router on and off, I simply delete this entire paragraph for most of my .pp files. But if you want any particular commands or comments to appear at the start of each file, this is the place to make that happen. Notice that each line starts and ends with a quote mark [ “ ] which will not appear in the resultant files but cannot be left out. Note also that this is where the commands are generated that send the tool to the “home” position at the beginning of each toolpath file, so if you want to eliminate that pesky feature (like I did) just delete the line:

"J2,[XH],[YH]"

Here’s a hint: If you replace everything after “begin HEADER” with the simple line:

“C5”

Then the command C5 (Custom Cut 5) will appear at the start of each .sbp file output with this postprocessor. You can then save any standard start-of-file commands as a Custom Cut file (Custom5.sbc) and it will be run as a subroutine by each file before starting the cutting moves.

Next, look for the section that starts with:

begin FIRST_FEED_MOVE 

"M3,[X],[Y],[Z]"

Let me digress here for a moment. One of the benefits of learning how to modify the postprocessor files is the ability to create .sbp files that perform a specific task or function at each vector location in your drawing, instead of the cutting action which VCarve thinks it is creating. Specifically, if you use a Drilling toolpath on a group of circles, you can substitute the code of your choice in lieu of a Move command in the postprocessor file where VCarve thinks it is going to drill a hole. For instance, you can use the CC or CP commands to drill a hole with a spiral plunge, the C# command to invoke a custom cut file, or even the FP command to run a particular file at each hole location on your drawing. This last sentence is the key to my technique for machining holes for RTA fittings with VCarve Pro, which as I said I intend to cover in more detail in my next installment.

So back to our file. The following line is the move command that will be generated for each required move. If you are just tweaking your router or spindle .pp files to adjust the headers and footers you’ll probably want to leave this alone, but to change the function dramatically try replacing this entire line:

"M3,[X],[Y],[Z]"

with:

“FP, RTA fitting.sbp,,,,,2″ (0r substitute the name of the file you want to run at each hole location.) Don’t miss the five commas followed by “2” – that specifies that the named file will be run in 2D offset mode.

Do the same thing where the file says:

begin FEED_MOVE 

M3,[X],[Y],[Z]"

Now the resultant .sbp file will move the machine to each hole center location, and run the indicated ‘RTA fitting.sbp’ file (or the file or commands of your choice)

If you’ve made it this far you can probably decipher the rest of the postprocessor file. For my RTA and drilling postprocessors there are no “arc” moves – or any other type of moves needed so I just delete the entire rest of the file. If you are just tweaking your cutting postprocessor you can adjust what happens at the end of each saved .sbp file in the section just after the line:

begin FOOTER

Again, substituting the original lines of code with a C# (or FP) command will let you run a standard end-of-file routine at the end of each .sbp file.

Once you are reasonably sure you’ve correctly modified your way to a new, improved postprocessor file, save it with a .pp extension in the same directory mentioned above. Then open VCarve Pro (or PartWorks). The software actually scans each postprocessor file upon opening, so if there are any obvious syntax errors you will get an error message right up front, with a line number that corresponds to the line number in the .pp file, which gives you a place to look for the problem. If the software starts up OK, then try saving a toolpath using your new postprocessor. Make sure to check out the first few resulting .sbp files with a text editor to be sure there are no surprises, and run the first few files in preview mode just to be sure. Since I started using my modified postprocessors I rarely if ever have to edit or even check my .sbp files before running them. Here’s another tip: you can delete or move to a new folder all the .pp files for any machine you don’t currently own and don’t expect to own any time soon. Do the same for any stock .pp files you won’t use, such as metric files if you only work in inches. That way when you go to select a postprocessor file in order to save a toolpath you’ll have a custom menu of all your own .pp files with no unnecessary extras. (You can restore the original .pp files by re-installing VCarve Pro.)

Sample files: The attached files are meant as samples only, although you can use and/or modify them as easily as the ones which come with VCarve Pro. For each .pp file, I used VCarve Pro to toolpath a simple drawing consisting of four small circles placed at 2,2; 2,4; 4,2 and 4,4. For the RTA Only file I used a Drilling toolpath with the depth of cut set to zero; the resulting .sbp file runs my “RTA fitting.sbp” file at each hole location.

For the Custom Cut file I used the same four circles but a Profile toolpath and a .5” depth of cut. The resulting .sbp file has a C5 command as the Header and a C6 command as the Footer, with the typical machine output in between. If you use the C#, format you can use Custom Cut files with 2-digit numbers and keep your single digit numbered Custom Cut files for use with the keyboard. [Download sample files]

That was me about a year ago. I’ve been purposely taking baby steps towards using the Shopbot for all my sheet cutting needs as I pretty much have to do it all myself and any time spent on R&D or machine maintenance/improvements is time I can’t spend producing cabinets (read: earning a living). As I wrote last October, I started with the vacuum table, which has been a huge success. Not that I’m not already thinking about improvements and what I might do differently next time, but it works well as-is, and my only regret is not taking the time to install a decent vacuum table sooner. But shortly after writing the installment detailing how I was purchasing pre-ripped and edgebanded melamine strips for my closet client, I made the plunge and tried machining a couple of small jobs entirely on the ‘Bot. Well, to get right to the point, I haven’t looked back. Even though I still lack full-featured software that can take my designs from “screen to machine” I already find it so much easier and more productive to do all my cutting and machining on the ShopBot than on the tablesaw/linebore/etc. that I now have a hard time imagining cutting a full sheet of plywood or melamine on the tablesaw.

So, of the three obstacles I mentioned, the hold-down system is taken care of, and I’m still using a Porter-Cable router – it’s loud and the bearings need regular replacement but it gets the job done. I’m getting much closer to being ready to order a spindle. That leaves the software. I’ve had several inquiries from folks who want to know if I’ve found a program that works for me, but unfortunately I still don’t have a real good answer for them. We all know there are numerous programs out there that will set you back as much or more than the cost of the Shopbot, and quite frankly, if I found one tomorrow that would work for me in every situation it would be worth the price of admission by saving me hours on every job. It’s not so much the cost of the software alone that is holding me back, but the time investment required to evaluate all the options out there and then to get up to speed on whichever one I choose. Just as the purchase price of the Shopbot doesn’t tell the whole story of how much it costs to get one up and running and get familiar enough with it to make it profitable, the time required to decide which software to buy and then learn how to use it would cost me as much or more than the cost to buy it. So for the time being I’m using my trusty DesignCad to manually draw and nest my parts and VCarve Pro to create the toolpaths for each sheet. As with anything, the more I do it the faster I get. I’ve saved a library of drawings so I can quickly bring up a drawing that has all the parts required to build a typical cabinet, arranged in such a way that I can easily stretch all the horizontal parts to correspond to the width of the cabinet, then cut and paste those parts into a drawing that represents blank sheets of material. I find that I can nest those parts onto the sheets as well as if not better than most nesting programs I’ve seen, although I admit it will take me longer than a decent nesting program would take. I created a macro for DesignCad that automatically exports each layer of each sheet into a separate .dxf file for importing into VCarve. The process of creating the toolpaths in VCarve is so repetitive (and so reliable) that I showed my 13-year-old daughter how to do it and now (when she has the time) she takes care of that part of the process for me.

I’ve worked out a pretty good system for creating clean, consistent parts from most sheet materials. I’m mostly cutting ¾” two-sided melamine-coated particleboard, but I use the same techniques for veneered MDF and plywood. First of all, I use a ¼” compression spiral router bit which leaves clean surfaces on both the top and bottom faces of the board. I could easily cut all the way through in one pass, but if I do I’ll have a problem when I get to the last few parts of each sheet. Since they won’t have the hold-down force of the entire sheet to keep them still, they would shift and lose the vacuum seal. So I cut everything in two passes. The first pass leaves a “skin” of .02” – just a little more than the melamine coating. The entire sheet is cut with this first pass before the second pass cuts through the material and .012” into the spoilboard. In addition, I have to take into account that the machine, like all things mechanical, has just a little bit of “give” to it, which can mean that two identical parts might not end up identical depending on several factors. First is that a climb cut can yield different results than a conventional cut, since climb cutting has a tendency to pull the bit away from the part during cutting while a conventional cut tends to pull the bit towards the material. We’re not talking a lot of difference, but it’s enough to notice and it’s not hard to compensate for. Secondly is that in order to minimize waste and sawdust, I want to place the parts as close together as possible on the sheet. So, when cutting the first part the bit is fully engaged in the material for the entire cut, but when cutting the adjacent part the bit is mostly traveling in the kerf that was cut for the first part. This creates a different amount of sideways force on the bit which can also leave a noticeable difference in size between the two parts. So what I do is leave an allowance of .02” on the first pass (the first pass cuts the part .02” oversize on all sides), using a climb cut to make sure the bit is not being pulled towards the part. Then once all the parts have been roughed out slightly oversized with the first pass, the second pass cuts the parts free from the sheet, to the final correct size, using a conventional cutting direction. This technique yields uniform, proper-sized parts since the sideways force on the bit is consistent on all sides of all the parts.

A few more details complete the story. On all but the smallest parts I’ve found that I can space the parts .30” apart – that’s .25” for the bit diameter, .02” allowance for the finish pass on each part, plus an extra .01” to make sure that the first part’s first cut, which is the climb cut, does not encroach on the adjacent part’s .02” allowance. I also place parts no closer than .30” to the edge of the sheet, to be sure to remove the factory edge and not cut into the stops which I’ve bolted to the edge of the table to aid in sheet placement. For smaller parts such as rails which are typically only 4” wide and can be less than 20” long in some cases I leave a full inch between them and the next closest part. This leaves a skeleton of waste material surrounding each small part which helps keep those parts from moving during the final pass. If I had a more powerful vacuum system I probably would not need to do this. It’s worth mentioning that even with a pretty decent dust collection system the kerfs tend to stay full of sawdust even after the second pass is complete. This is due to the fact that the compression bit has only a short upcut segment on the end and is mostly comprised of downcut geometry. Although it leaves more dust on the table to be cleaned up between sheets, it works in my favor to preserve the vacuum seal even towards the end of the second pass. As for cutting speeds I’ve found I can run the router at 6ips for both passes for about the first 15 sheets or so, then once the bit has some wear on it I have to slow the first pass down to 4ips or the router sounds like it wants to bog down a bit. I’m hoping once I upgrade to a spindle I’ll be able to maintain the faster speed if not increase it some. I can usually get 25 sheets or more from a bit, which I then save for less demanding cutting like raw MDF or thinner materials. For veneered MDF I keep the speed at 4ips from the start as the router doesn’t like the faster speed in the denser MDF core. Also, any machining such as pockets, grooves, holes, etc. is done before the parts are cut out from the sheet to ensure that the parts cannot move during the machining process.

A sheet of closet parts ready for edgebanding.  All machining except for edgebanding is done on the Shopbot – grooves, holes, notches, cutouts – whatever.

]]> http://www.shopbotblog.com/index.php/2007/08/almost-there/feed/ 0