Fine-Tuning Cutting and Sanding Operations for Inset Doors

Here's an extended discussion on how to tool up for accurate cutting of inset doors, with minimal need to sand the edges. May 26, 2011

Our current path for installing inset cabinet doors is to trim them on the Altendorf then edge sand to fit. This is a tough way to do anything and I want to streamline this part. I have an idea about how to automate this process but would to hear some opinions, good or bad about the idea. Our doors are processed right now with Leitz insert tooling cutterheads. This tooling produces an amazingly smooth profile after miles and miles of cutting. The stick shaper is a $1000 import with a tractor feed powerfeed. There's a whole lot more material being removed during the during the sticking process than during and edge trim. Sticking removes the ogee shape and cuts a 1/2 inch groove. Trimming takes at best about 1/8.

My plan is to produce a radial arm shaper and use alternate shear tooling with about a five inch diameter cutterhead. The cutting direction would be a climb cut. End grain would be processed first. End grain might also be relieved a click before the actual trimming is accomplished. The shaper itself is mounted to a metal platform and actuated on a linear bearing much like a vertical panel saw. The stroking would either hydraulic or pneumatic.

What I am interested in receiving input on is what the final quality of cut might be. In my mind, this essentially is just a great big router but with better tooling and, because of tooling diameter, more of a scraping than cutting action. I am trying to get as close to no edge sanding as possible.

Can anybody point out errors in my logic from a tooling perspective? How is this different than a CNC machine with respect to tooling marks etc? Recognize that this is not something I personally am going to produce. I happen to be located across the street from a company that engineers and sells linear motion systems. They are the people who will engineer and produce the actual sliding mechanisms and hold down logic.

Click here for higher quality, full size image

Forum Responses
(Cabinetmaking Forum)
From contributor L:
I don't know about the rest of it but climb cutting end grain produces a less than ideal finish. Try it first on your shaper with a feed, take a light cut!

From the original questioner:
I immediately jumped to climb cutting because it usually minimizes tearout but had not thought about it with respect to how smooth the final cut will be. You did, however, give me the idea to set up a linear guide on a shaper to test the idea. That's no more than a two hour experiment and probably worth doing.

From contributor O:
Counter-rotating heads will do the stile ends, with a linear action table between them. Unique has machinery based upon this principle. Also used for coping and curved panel arches. The counter rotating heads cut with the grain. Unique's later designs have the parts stationary, and the cutters move in and out to make the cuts.

From the original questioner:
I looked at Unique Machinery many years ago at a trade show in Anaheim. What they had was perfect for my needs except that the maximum cutting diameter was smaller than my heads (Leitz Profilcut) required. I also like the hopper feed they have for sticking doors.

From contributor F:
Have you made an estimate on how much time you'll actually save and what's going to be involved with building and implementing this custom machine? I'm trying to see where you'll save much time as youíre not really cutting out any steps. Whichever way you decide to trim your panels you'll still have to sand them before finishing anyhow. So you have to figure how many seconds per panel you'll save with a smoother cut? And how many panels it will take over what length of time to make it worthwhile?

From the original questioner:
Those are good points that you raise. I have been thinking about this project for some time now and am not entirely convinced that this will work. If it does, however, I think it will have a very significant impact on our throughput. If you have some patience I would like your input on my logic.

What got me interested in this was the observation that my insert tooling for the sticking process produced mirror like finishes that never needed to be touched with sandpaper. This was not the case when I did my profiling with carbide.

The reason for this crispness has to do with several things, the biggest being the metallurgy of insert tooling. Brazed tooling requires a certain amount of porosity in order to be able to fuse it to steel. You have to have some crevices to produce a membrane between the steel tool blank and the carbide tip. This porosity is invisible to the naked eye but exists throughout the rest of the carbide tip, including the portion that actually does the scraping on the wood. Similar to the way a pothole deteriorates very quickly on a highway, these small potholes dull the carbide very quickly. Just like a shaving razor goes bad quickly, so too does a carbide cutting tool.

Insert tooling, however, is put together with a different alloy that ends up with much less porosity. This is why the tools need to be clamped into position rather than fused into position. The consequence is a much cleaner, sharper cutting action. My hypothesis is that if the insert tool technology does such a good job on the profiling, it might do a similar good job on the trimming.

If this turns out to be the case then I can unlock this task from the kind of skill that edgesanding requires. Of all the specific tasks in my shop I think that edgesanding is the most nuanced. If I can turn this operation into something that requires just a decision rather than intuition then more people will be successful at it. You are possibly right that this won't work. I think, however, that it will be worth the couple of hours it takes to load a linear guide onto a shaper and test the concept.

From contributor L:
I agree that inserted carbide is far better than brazed. We've been users since the early '70s. We use it on the molder as double back straight knives, gives a beautiful slick finish. I think I'd test a simple setup designed to use conventional cutting, mainly because if it works the setup will be the same for the end grain cuts. It would be possible to design the machine to do both but it complicates it. I know the trick to getting really good cuts is to hold the work so it can't do any moving, vibrating or dancing while being cut.

Maintaining ideal chip load is also a consideration, so it must be mechanically controlled or have an excellent operator. The most consistent movement is hydraulic, also the most expensive. Pneumatics have the springiness that can be a problem with varying load but by using a massive enough assembly and large enough cylinders with exhaust one-way restrictors it would most likely work for this application. I don't see a lot of variable resistance here.

The mass has two functions: to offset the variable loading that may occur due to hard/soft areas of the wood and the second is to absorb vibration energy. Your 1" plate steel should do the mass thing fine. The moving plate plus a couple of clamp cylinders should work well for keeping everything rigidly held. What is the stroke required? If necessary can it be designed to stroke from either direction so you can have both conventional and climb?

Will you use foot pedal cycle start and relay logic or a PLC? With a PLC you could program in a short entry, shift the work back (pneumatically) and move the cutting assembly to the opposite end and run a climb cut. Complicates the design but covers all. The amount of back shifting depends on the maximum cut depth allowed. If this machine is designed to also profile the outside of doors a larger back shift should be built in. If I take your drawing correctly you intend to use ball bushings for guidance - probably fine but I kind of lean towards the THK type. Easier to locate and attach and very solid. Your drawing implies the load is carried only at the end supports of the rails.

From the original questioner:
I hadn't thought of setting up a machine so that it could cut in either direction. That's worth considering. The first thing I am going to do is emulate the machine with some temporary linear guides on an existing shaper. In this case the material will do the traveling and the cutterhead remain constant. I am only removing about 1-3 mm of material so the cutting bite will be minimal. If this produces a satisfactory cut on the endgrain I will proceed to the next stage of development which will be to engage the engineers across the street to developing a linear platform that actuates. I don't know the most efficacious way to do this but these people do. If they say hydraulics then that's what it will be.

If I can produce a sand free solution that can be produced by inexperienced people then it will be worth the pain. My primary contribution to the concept will be definition of the problem to be solved and funding to make it happen. This is not a particular time sink. As so far as cost goes it's just another of a million line items.

From contributor F:
Well I'll have to show my lack of experience as I have not switched to the higher end of insert tooling, I'm still working with my old Freeborn brazed sets. So youíre saying that your doors require no edge sanding after machining? Or is it just minimized to a bit of touch up sanding? I guess I can see that working well on painted finishes, but on stained pieces I would think you would have to have that minimal amount of sanding for the stains to look correct?

If that's the case then I can see where this idea may be worth further investigation. Do you already have the tooling? I may even do an experiment on my shaper for the heck of it. I wonder how a 3" insert head (I happen to have handy) would cut compared to the larger head youíre describing? I've never tried it on end grain before.

From the original questioner:
The inside profile of our doors (ovolo, ogee, etc) require absolutely no hand sanding whatsoever. This part, after miles of stock, requires no tuneup whatsoever. This is what lead me to consider using this type of tooling on the perimeter trim down cut. Like you, I am anxious about the endgrain. I have a three inch insert head with alternate shear knives that is not currently in use. I will test drive that one on a shaper to evaluate the outcome.

From contributor O:

I had a 140 mm diameter, Z4, insert rebate cutter, 30mm shaft, 8000 rpm in our Felder sliding table shaper and tested some end cuts on tight grain VG fir and white oak. The knives still have about 70% life in them. The climb cut was a little smoother than the conventional cut. They were all clean but I would still wipe them with fine paper or a sanding sponge before finishing. The machine has a little to do with it also. We have found the Martin shaper and tenoner that have more rigid shafts will make a little better cut than the Felder. You could do this on a sliding table shaper. The Felder I have will do about 50Ē and they have longer ones but I think edge sanding would be easier. If it were me I would work on precision machining and assembly of the doors to eliminate the slider part of the operation. I always wondered how the euro edge sanders that have split tables like a jointer would work for fitting out inset doors.

We do the same type thing running the outside rebate cut on our euro sash only they are deep cuts with large diameter 160 to 180 Z3 insert heads. The finish cut only requires a wipe of sanding on the end grain. We could probably go to the finish room without any sanding but just a habit we have. These windows are a little like inset cabinets; the sash final size has to be plus or minus .5mm in relation to the frame. Through software and precision setups we build all windows and doors to exact size right out of the frame press. You donít want to be trimming these after assembly. The sash get 1mm removed from all sides at the shaper or profiler after assembly. Profilers and single end tenoners that have jump cope climb cutting additional heads eliminate the need for backup chip breakers when doing these types outside cuts or profiling.

Onsrud used to make a lineal profiler that would do the same thing you want and probably better than anything home built. They are big though. Unique machines and the other brand mentioned look like the good solution for the small shop wanting to automate door production and still have versatility.

From the original questioner:
So if I understand you correctly the end grain came out smooth enough to maybe just need a touch with some foam sandpaper? I can live with that much sanding because hand sanding does not materially change the shape or size of the door. Edgesanders take a lot of skill. It's tough to know sometimes how much to lean into the sander or how to land the material onto the sander. Just when you get it right somebody changes the belt on you and what used to be the right amount of pressure is now too much. I have one of those split fence edgesanders but the one we have does not reliable hold position. There is a piece of graphite cloth between the belt and platen and this tends to wear erratically, particularly as we raise and lower the sanding surface.

I don't think we can solve the need for trimming with just better gluing because no matter how straight and true of stock we start with, some of it just gets bowed because of tension imbalances that show up from the machining processes. An ogee profile, for example, removes a tremendous amount of material on just one edge and what starts out as a true board can end up concave or convex, hence the need for tuneup.

But let me ask one more time: You were able to machine a millimeter off the end grain and still get a smooth cut? Was there much blow out as you exited the cut? Did the blow out go away with climb cutting?

From contributor O:
Yes the end grain was very good and would only require light hand sanding. You are right about the edge sander taking skill and itís real easy to round things off. I just wonder what a top line edge sander like a Kuendig with a feeder and set to remove .5mm would work like. I've tried them out a couple times and seemed to work with small pieces but don't know if it would work in real time.

Regarding machining end grain - I was taking exactly 1 mm on the test cuts. Any more would be scary climb cutting on the shaper. Was there much blow out as you exited the cut? No the blow out with climb cutting occurs at the start of cut with end grain.

Did the blow out go away with climb cutting? On the test pieces the blowout left would not come out when you remove the 1 mm when doing the long grain cut. The oak maybe, but not the fir. You would need a backup block to get 100%. On our window sash processing it took a long time to work this out because with slot and tenon corners you have end grain both directions. We used to clamp a back up board to the sash but this was time consuming. We modified an Aigner clamping device for a back up board and that has saved a lot of setup. The ultimate answer for this is the timed jump cope on a window machine or track fed tenoner or a unique type shaper.

CNC router shops sizing big doors and windows usually do the ends first, climb cutting then come back at the long edges. I believe they have to leave more than a millimeter to get rid of the blowout. Whatever machine used, hold down is important. The SAC profiler we are rebuilding has 13 feed wheels for real positive pressure when doing these cuts. Of course we are only talking cabinet doors here.

From the original questioner:
Based upon your preliminary research I am going to now investigate the hold downs and linear motion. As I said before, I am located directly across the street from a company that provides engineering solutions for linear motion and clamping devices. Their specialties are pneumatic and hydraulic systems. They are well versed in the switch logic for activating these systems in syncopation.

I favor moving the cutterhead and keeping the material constant because of the footprint. A 6' cut could probably accomplished with an 8' footprint. Moving a six foot door through the cutterhead would require at minimum 12' of stroke.

I also lean towards hydraulics for actuation because in climb cutting it would be too easy for the cutterhead to defeat pneumatic because air compresses both ways. This is just my layman's knowledge. I would, of course, let the people across the street design the actuation and hold down mechanisms. I am hoping to keep my involvement time-wise to a minimum. My job is to identify the need and provide the funding. Most of these parts are generally available as off the shelf gizmos. Any metal work that is needed can happen with a CAD file and waterjet cutter, also available in the neighborhood.

From contributor L:
Just one last thought on this: I'd look real hard for a readymade solution before trying to reinvent the wheel.

From the original questioner:
You are half right about trying to find off-the-shelf solutions. In a lot of cases this is the way the go. I was pretty amazed at the Unique Machinery coping station as it was quite similar to the one that I had envisioned before I saw theirs. Mine differed on the bump stop arrangement. I was planning to configure this more like the bumpstop on a castle drill. The weakness in the Unique system is that it assumes the operator actually hits the bump before the hold down clamp actuates. In the Castle approach the clamp will not engage unless the bumpstop is fully engaged.

I bring this up to point out the benefits of continuous improvement. Trying to get Unique Machinery to customize a machine for me might be viable but it might not. It is for this same reason we handle website development in-house.

Many on this forum will disagree with me about versatility. I personally think that machinery should be very simple and single purpose. I also favor machinery that is right sized for its purpose. I currently have an SCM shaper with a 48" footprint. This seems kind of excessive to me when the majority of my cope cuts are only five inches long. Building your own machinery is a pain but after you learn who the creative machinists are I think the process will become easier.

From contributor U:

I don't know a lot about them but couldn't you set up a double end tenoner to do what you want. I know some of the big door companies use them to do final sizing and squaring of doors while cutting the edge profile.

From the original questioner:
Like you I also don't know a lot about double end tenoners. I think it would be a good idea to look at these to see whether they would do the job or maybe just glean some ideas about them. You can get a lot of information from just synthesizing what has come before you.

The majority of woodworking operations consist of simply clamping something into position, rotating a cutterhead then applying linear motion. Where it gets complicated is when you try to make these machines too versatile.

My coping shaper, for example, has a split fence system, the ability to raise or lower the spindle and the ability to change the belt speed. This would all be great if I was doing custom millwork but I'm just coping doors. I don't use the fence, I have never changed belt speeds and the arbor never changes height. I use stacked cutterheads and simply use interchangeable platform heights to move from one profile to another. This shaper sells for about $12000 new and requires far more footprint than is needed to do the job elegantly. The components for the machine, not including the frame, are probably about $2000. I can, for example, buy a 1 1/4 inch arbor quill off the shelf for under $300. One of the reasons I am fascinated with this project is that I have about a half dozen machines I have wanted to build that all involve shaper arbors.

For example: We build most of our drawers with pocket screw construction and rabbeted 1/2 inch plywood bottoms. We use a Castle drill for cutting the pockets but the grooving and rabbeting is done with yet another table saw and another shaper. Because of the layout of our shop these machines both require their own dust collection. This is an incredible footprint intrinsically but also requires a lot of cart management and real estate to navigate these carts.

Once these parts have been machined and collected together we still have to bring them to a bench and clamp them into position for inserting the screws. This is a lot of strokes. One machine I have designed will rabbet the plywood and groove the drawer sides yet only require an additional three inches of footprint beyond the bench that you do the actual assembly on. There's no travel time involved with this approach. With a few more strokes this bench can have an integral clamp for holding the parts together while you screw them. This can work in a factory setting or can roll under your outfeed table if you don't have the space to have a dedicated work station.

I have a half dozen of these machines that are all very inexpensive and perfect for a company that believes in Lean manufacturing. The first couple of machines will take a few more strokes but after you learn who to do business with the rest are really easy. I'll use the guys across the street to develop the prototype then I plan to shop the machine to Castle Machinery. They already have the lawyers on staff and the technological know-how to bring this to market. My role is just the idea.

From contributor O:
Double end and single end track feed tenoners are good for any kind of door production but too costly and take up a lot of floor space to be effective in a small shop. The Unique machines are basically single end tenoners scaled down for custom cabinet door production.

If I understand correctly you size the door to the opening with the slider then want to take off a millimeter or so all the way around for clearance? If that is the case the simplest low tech method would be to use the shaper with a split fence and feeder. I had a couple old cabinet doors and set the shaper up to remove 1mm using an 80mm diameter Oertli spiral insert head we use for pattern shaping. I did half the fir door conventional cutting and the alder door climb cutting. You can see the chipping from the leading edge of climb cutting. The end grain cuts I did yesterday were better with the larger diameter rebate cutter. The spiral insert leaves faint lines but they disappear with a little sanding. This is interesting with the end grain being cleaner with climb cutting. With a hand router it usually pulls the end grain with a climb cut - must be the larger diameter. We normally don't do any climb cutting with the shaper but removing only 1 mm is not too dangerous.

Click here for higher quality, full size image

Click here for higher quality, full size image

From contributor R:
Not to change the subject, but why do you use such a large, expensive shaper for cutting copes? It is probably way overpowered for the job. It is also over featured and oversized as you stated. Why not use a less expensive, smaller and no frills machine?

From the original questioner:
We use stacked cutterheads with Leitz insert tools. The heads are about five inches in diameter. The top head is an aluminum body but the bottom head, (which was custom made) by Leitz is steel. With this much weight it seemed to make sense to go with a heavier spindle and this was the easiest machine to bring online at the time.