[nlancaster]

damnit.... didn't know you were giving htem away on the other forum

I was about to cut my double sided maze....

so day

[MadeInTheShed]

Hi Jerry,
Trying out your Dovetailmaker tomorrow. Made the jig to hold the work at 15degrees today, but this looks like another winner.

[MadeInTheShed]

So I played with FingerMaker and Taimaker today and found these just as easy to use as Mazemaker. Will have to start saving ......
I would like to see the same things as I suggested with Mazemaker, but these are only minor issues ( Add an S command, etc).

I didn't thk these were too bad seeing it was using some ply that wasn't totally flat.
Attachment 264824Attachment 264826
Attachment 264828Attachment 264830

[JerryBurks]

So I played with FingerMaker and Taimaker today and found these just as easy to use as Mazemaker. Will have to start saving ......
I would like to see the same things as I suggested with Mazemaker, but these are only minor issues ( Add an S command, etc).

I didn't thk these were too bad seeing it was using some ply that wasn't totally flat.

Thanks for experimenting! As you can see, having the boards totally flat and parallel to the axes is extremely important or it won't fit together.

[JerryBurks]

Thanks for the feedback Alan! The problem with the perimeter is that I am using the same algorithm for the rectangular mazes and the free-form mazes with blocking pattern. The program starts scanning the matrix until it finds the first maze cell (which is the very first one for rectangular maze) and then starts wandering along the edge of these occupied cells like a blind person tapping around with a stick. But it may be possible to exclude the diagonal cells from being found as "empty". I need to think about that.

Actually, I had originally intended the perimeter cut only for the free-form mazes because it is really easy to do a rectangular cut of any shape and size with and without rounded corners with the usual CAM programs. But I can see this may be a problem if somebody does not have a 2d CAM.

Your maze is looking good! I also have to check the branching conditions. I believe I had a feature in there at some point that prevented 4-way crossings. Not sure what happened to it since I stopped working on the software for almost a year before I picked it up again last fall. But I will have another look

[MadeInTheShed]

Im still waiting for my "lolipop" bits to arrive. Ive managed to source some ball bearings of different sizes (got one of the apprentce technicians to crack open some old gearbox ball races). Hopefully I will be able to have a bit more of a play with the maze creating next week. Ive got a week off work next , and once I have finished some projects for SHMBO, I'll get back into it.

Actually, I had originally intended the perimeter cut only for the free-form mazes because it is really easy to do a rectangular cut of any shape and size with and without rounded corners with the usual CAM programs. But I can see this may be a problem if somebody does not have a 2d CAM.

Or, you could learn to write Gcode manually
Nearly 30 years ago, I worked as a CNC lathe/Machining centre operator, then setter, and finally writing the Gcode for 4 lathes and 2 3 axis centres. Back then, we had no CAD software, it was all done with pencil, paper, calculator and a knowledge of trig, and dont get me started on the dramas and amount of work involved with tool nose radius compensation! I would have loved to have had the CAM software available today, it would have made life so much easier.

[JerryBurks]

I use inch size balls and cutters of same size. The software has a function to make the channel larger than the bit. E.g. for a 1/2" or 12.7mm ball I make a 13.2mm channel. You need the clearance to avoid friction with some remaining wood fuzzies.

[pippin88]

The smallest ball end (lollipop) cutter I can find seems to be about 10mm (a little of 3/8ths) - has anyone done any better? I'd ideally like <1/4" so a double sided maze could be done in 3/4" material or less. 1/8" bit would be great, but I don't think they exist...

Edit: Found some a little smaller at 9/32, but cost $59+shipping (Speargun Ball Router Bits | Plunge Bit & 1/2 inch Radius)

[JerryBurks]

The smallest ball end (lollipop) cutter I can find seems to be about 10mm (a little of 3/8ths) - has anyone done any better? I'd ideally like <1/4" so a double sided maze could be done in 3/4" material or less. 1/8" bit would be great, but I don't think they exist...

Edit: Found some a little smaller at 9/32, but cost $59+shipping (Speargun Ball Router Bits | Plunge Bit & 1/2 inch Radius)

http://www.harveytool.com/prod/Under...-270-_127.aspx
I don't know how well these work on wood, but it may be worth a try.

Thanks for the links! I had not seen that yet. Maybe I get one of the Harvey Tools cutters and try it out. Not cheap but a smaller channel maze would be much more practical.

[JerryBurks]

I left the maze software project alone last week and went back to the finger joints...
after somebody on the Vectric forum asked about options to make joints with included angle other than 90 degrees (i.e. perpendicular boards) I realized this could be achieved easily with the software, at least for acute angles.
So I published a new beta version that allows for joint angles between 22.5 and 90 degrees and allow that way for making triangular or star shaped boxes and frames with finger joints.

I am not sure if this is just a crazy idea but in case somebody wants to try here is the download link.

So far I did only a simple test (see below) and obviously there may still be bugs, but that is why it is a beta.

[alan_3301]

http://www.harveytool.com/prod/Under...-270-_127.aspx
I don't know how well these work on wood, but it may be worth a try.

[pippin88]

Nice. I'd glanced at that page before but discounted them because they are for metal and looked like they had heaps of flutes.

However, the smaller sized ones have 2 flutes which could work.
The largest 2 flute they have is 3mm. 7/64" (~2.77mm) balls are not hard to find and should have plenty of clearance.

I've sent a few emails to makers of undercutting end mills to see if they'll ship internationally. Looks like those located in the US shouldn't struggle.

Jerry: What's the ideal ball sizing vs cutter? I believe you specify multiple passes for clearance, so does a 1/2" ball work with a 1/2" cutter? If we use slightly undersize balls can we only do one pass and speed up work time significantly?

[JerryBurks]

...........Jerry: What's the ideal ball sizing vs cutter? I believe you specify multiple passes for clearance, so does a 1/2" ball work with a 1/2" cutter? If we use slightly undersize balls can we only do one pass and speed up work time significantly?

For the ball end finish pass it will always make 1 pass forward and one pass back (with offset to achieve specific channel diameter) and then do the same thing one offset distance higher. I found that method not only allows for widening the channel a little beyond the ball end bit diameter but also makes sure that all wall sides of the channel are cut in "conventional" mode, minimizing fuzz and chipping.

You can do a slot and ball end diameter of same size but it will not allow you to get much of an undercut to hold the ball captive. It will probably fall out on channel branches.

[Wackamazoo]

Jerry,
I know you have created the dxf export for single sided mazes. Could you also add the export for at least the top side for double sided?
Also, how to get apply updated values without generating a new maze pattern?

Thanks

[JerryBurks]

Jerry,
I know you have created the dxf export for single sided mazes. Could you also add the export for at least the top side for double sided?
Also, how to get apply updated values without generating a new maze pattern?
Thanks

That is something I am working on (well, if I would not have to do some business travel at the time). I believe I can generate dxf for 2-sided mazes as well but need some time to work it out.
For parameter updates I need to better distinguish between simple dimensional changes and systematic changes that require a new maze anyway. Just something else to improve.

[alan_3301]

I'm still having fun with this, great program.
I will make 2 more feature requests.
1. Have an option to make the perimeter have square or rounded corners, in addition to the 45 degree bevel that is the only option now.
I like to run a roundover bit to smooth the edges, but I can't go full depth or it would cut into the ball channel near the corners. If there were an arc there, it would be no problem.
2. one more extra boundary parameter for only the X dimension, so that it can be widened enough to integrate handles to the sides of the maze.
Thanks.

[MadeInTheShed]

I'm still having fun with this, great program.
I will make 2 more feature requests.
1. Have an option to make the perimeter have square or rounded corners, in addition to the 45 degree bevel that is the only option now.
I like to run a roundover bit to smooth the edges, but I can't go full depth or it would cut into the ball channel near the corners. If there were an arc there, it would be no problem.
2. one more extra boundary parameter for only the X dimension, so that it can be widened enough to integrate handles to the sides of the maze.
Thanks.

Excuse my ignorance, but what timber is that?

[alan_3301]

No excuse necessary, I wouldn't know what kind of wood it was if it didn't have a sticker on it when I bought it.
It's Bolivian rosewood. It seemed like the hardest wood I had on hand, I was hoping get rid of splintering, but there was still some pieces that chipped off when slotting.

[JerryBurks]

Short Video...

[JerryBurks]

I started looking for a design of a watch display box for a birthday present and it ended up as a rather complex piece of equipment. Actually I had to do this twice, since the first version kind of worked but not smoothly enough. This second version seems to work very well.
CAD design


All parts cut out


Assembled with wooden hinge


Complete