Rounding Over Hardwood Block Ends

This week's "stump the chumps" puzzler: how to round over the ends of toy hardwood building blocks in a volume production setting. No answers, but an intriguing question. January 27, 2008

For many years, we have sought a method for rounding the end edges and corners of toy blocks. These are chopped in a variety of lengths from molding, and after this is accomplished, the fresh corners and edges must be rounded. In our minds, we want something like a pencil sharpener. The block (typically 1-3/8 by 2-3/4 in cross section) gets stuck in the sharpener, and when it is removed, the four edges have been rounded or chamfered to a radius of (about) 1/16 inch.

Recently, it has been suggested to us that a tenon machine could be set up to accomplish this, in effect making a zero sized tenon, but rounding the edges. We aren’t really familiar with these machines. Any thoughts or other solutions? We are talking about half a million blocks or so a year, which we currently sand on a stationary belt sander – a slow, dirty, uncertain process that cries out to be automated.

Forum Responses
(Solid Wood Machining Forum)
From contributor D:
I'm not too sure how a tenoner would do that job. It would need sanding heads only. Before going to a full blown tenoner, you might present your problem to Unique Machine in Phoenix. They do end work machine for cabinetry people, and offer sanding heads on tall spindles that can be numerically located for quick change. There are other companies that make similar machines.

As a second, I always thought these small parts were tumbled in a large rotating drum with abrasives to get the desired finish to the corners. Noisy and dusty, but gets the job done. Think of a large clothes dryer one-quarter full of blocks.

And the third, go to the show in Las Vegas and take several of your blocks of wood illustrating before and after. Find some of those grizzled old machine salesmen hanging around and pass your blocks to them with your question. Check out the booth people and abrasive salesmen.

From contributor R:
The tumbling method will work. Pieces of torn sandpaper in a cement mixer is one option. Very old, tried and true method. Remove the fins from the inside of the drum.

From the original questioner:
The tumbling idea is one we have heard many, many times, though we have never seen it implemented. I believe this usually arises because people visualize blocks as like ABC blocks – i.e., small cubes of moderate weight. But unit building blocks are larger and quite heavy. Our largest block, for example, is hard maple about 22” inches long and weighs over 2 lb. These come with the edges rounded in the long dimension quite beautifully – straight from the molder with a radius of (about) 3/32. What is required is that the end edges be rounded identically – or nearly.

Our experiments with tumbling have been completely unsuccessful. If you can imagine tumbling a couple of hundred two foot sections of 2x4, you would have a good visual picture. For one thing, every process we’ve tried leaves tiny scratches on the face of the blocks. For another, the rounding is insignificant short of tumbling for a long time (days), the number of blocks in a potential charge quite small, and the amount of damage to other portions of the blocks too great as the abrasive attacks the softer portions of the block. Maybe someone who has actually done this knows tricks that would evoke an “Aha!” response – but otherwise, we think tumbling is a dead end.

A router table will work – sort of – but climb cutting is required to avoid chip out on one corner or another and it is hardly faster than using a stationary belt sander. Moreover, one objective is speed. A skilled sander can sand the ends of one of these in about 30 seconds. A machine would need to go faster, which calls for some sort of hopper feed or automatic conveyor.

We do know that in the old days, some operators were successful in making these on a rotating table auto shaper. This would accomplish the rounding on a blank by using the shaper head. But this takes some special tricks as the straight sides are quite long and this challenges the ability of the shaper to negotiate the entire length. Moreover, making and sharpening shaper heads with a tiny top and bottom radius is a real challenge.

From contributor D:
You have been around the block a time or two on this. At least when you do solve this, you'll be the expert everyone else can quiz.

As to small roundover tooling, I think the small carbide insert heads will do that economically - Tersa and others. How to hold/feed the parts, while climbing, is the problem. So - back to tenoners. The Bacci and similar can follow the ends and run around the perimeter, and can be set up to climb. Many of those machines are two ended.

Abrasive seems the way to go. There are turning sanders that have follower heads that register on the surface of the part, and as the part turns slowly (5-10 rpm), the sanding belt (3-wheel affair with a forming section that contacts the work), the ends are sanded. The machine has about 4-8 heads, each holding a part, loading and unloading (hoppers) while the whole turret rotates. Turning sanders are what I have heard them called, but there should be a more specific name.

May call for a trip to Europe's Ligna show. You could detour to the Brio factories in Sweden and perhaps see how they do it.

From the original questioner:
Yes. This sort of idea seems to us to be the best route, though we believe that we will probably have to seek a purpose built production machine rather than adapting, for example, an existing tenoning machine. The big trouble seems to be finding a machine designer in the US. Moreover, this kind of thing requires some experimentation and it's tough to contemplate spending the money when there is less than certain assurance that it will ultimately work. We are reminded that Mark Twain blew his whole fortune trying to fund an automatic typesetter.

But an abrasive machine with cammed followers seems to be the right idea. And you are spot-on in remarking that the feed mechanism is the heart of the problem. What I haven't mentioned is that there are four different cross sections and (altogether) about 80 different shapes. So whatever we buy, it has to be pretty adjustable.

I might add that this problem has been solved. We know of at least one American company and a couple of European ones that apparently have such a machine - but what it is hasn't been something we can ascertain. TCTimber (in New York) was able to do this before they were sold. Somebody out there knows the answer.

From contributor D:
Interesting as to how it all comes down to know-how. Granted it is machine know-how, but know-how all the same. You have done some good research and are close to an answer. Short of corporate spying, you might try the simple thing. Call the people you know have solved this and ask for a machine dealer, name, description. If nothing else, you'll get some interesting responses.

The turning sanders I have seen could be adapted to this use. That may be the place to start with a machine designer. How to budget for the indefinite is tough, but R&D is essential for growth.

From contributor T:
You're right about the CNC solution being slow. However, if you can set up a bunch of these parts to run at one time, the operator can be doing something else quite productively while this is running. It may also eliminate some other material handling processes along the way.

From the original questioner:
Don't mean to be dense, but what CNC solution? Anyway, we have looked pretty hard at CNC routers for trying to accomplish this job - start to finish. It's quite a trick, however, as this material doesn't consist of sheet goods, and trying to use a CNC (router) on all of these small pieces and different shapes gets into myriad hold-down problems, not to mention that the edges require undercutting all around. Not that we haven't discussed this with half of the CNC salesmen on the planet.

At one point we built a router table with four small routers on it and a small square track on the inside that operated a little like the pencil sharpener we've mentioned above. You put the end of the blank in it and ran it around the inside of the track and it (more or less) routed all four edges. But it took quite a lot of skill to keep everything in line, it wouldn't work for climb cutting, and it was hard to adjust. It's in the back with the cement mixer we used to try tumbling just behind the two belt sanders we welded together to try chamfering two edges at once.

We did find an engineer (Russian) who said he could build us a machine, but he needed 25K upfront and wouldn't tell us anything about his approach to the problem. The money would have been okay - but I kept remembering my sister who sent off $20 for a breast enhancement device about thirty years ago and got a plastic hand back in the mail.
I guess what I am really hoping is that one of the old crew from T.C.Timber, or Bird-in-Hand, or one of those companies that now import all their blocks from China will say - "Hey! I know how to do this! We used a Framistan Dibbler from Steptoe & McFarlan Co.!" But - no luck so far.

From contributor M:
If you're looking for a CNC solution, take a look at the Weinig Conturex. What makes this different than other CNC's is that it doesn't use pods or a vacuum table to hold parts, it uses fingers that holds the parts. It self loads, so all you need is to make sure it has enough blocks in the magazine to keep up with the machine. I saw it demo'ed at IWF in 06 and was pretty impressed. I went to the Weinig web site and they appear to have a lot of downloads that you can take a look at. It can do all you're asking and then some. It's kind of like using an M16 on full auto to shoot a sparrow, except you're doing a half million a year.

From contributor S:

At the Atlanta show last August, there was a man there who had a robot display set up. He wasn't getting too much response, as a single-arm robot is not really woodworking-type stuff. But I had some interest in robotics for welding at the time, so I spent about 45 minutes talking to the guy. It's just a thought, but maybe your work could be accomplished with some combination of a robot handling the pieces and running them across "stationary" routers, abrasives (or whatever), much like your employee does now.

From contributor I:
I make knitting accessories and most of my machines are custom made, but then I am in India. In your case, I can think of two ways. 1) Have a series of sanders and guide the blocks over them. 2) Try changing your work flow, use a post forming machine to round the blocks, then without cutting, mould the corners as you usually do and then separate the blocks.