cnc shaper

Whatever you can get for cheap would be where I would start, preferably smaller. Gary Martin at Martin Model & Pattern sells a casting kit for small hand shaper that could be converted to powered operation, but it isn't what I would call cheap (I have one, by the way, still unmachined). For me the Lewis is kind of perfect, since it was a kit shaper they were pretty much all different, so If I wind up having to make some mods to it, it's not like I'm ruining a piece of history. One other thing I forgot to mention above, it will need to have an automatic oiling system. With the antique shapers, and a lot of the small hobby shapers, there was an expectation that you would stand over the machine and squirt oil in various places as needed. My Lewis shows the futility of this with scored ram ways, but that's another story. CNC is something you expect to hit the start button and walk away from, so the machine needs to be capable of it. The ram oiling system on the Cincinnati shapers is a good prototype for this, it is what I was considering for the Lewis.
 
Last edited:
To bring back an old thread, I am going to try and model a 42" Cincinnati XM-1 shaper, (1/5 scale), which was probably the most up to modern shaper ever made(1955), sadly it came out too late to make it into production (high ram speed (400 ft per minute), carbide capable). I want to make it look like the XM-1 on the outside with the screw driven triangle shaped ram, but using modern electronics (magnetic rotary encoder for ram position,CNC motor driven table and knee. Call me crazy for trying this! I am going to make a small model of the basic concept to see if my idea even works, just on a napkin at this point. Input?
Cincinnati All Steel 42 inch shaper 1 of 3.jpg
Cincinnati All Steel 42 inch shaper 2 of 3.jpg
 
No reason it wouldn't work. I would use a magnetic linear encoder rather than a rotary for the ram position. I have these on my CNC mill, and for my lathe DRO. Bulletproof and very accurate.
 
I was looking at using a AS5134 programmable high speed magnetic rotary encoder. It can program zero position and has 9 bit multi turn counter. With this I can zero the beginning of the ram start position and count to the end of ram stoke at specific material location, store this count in memory, thus I have zero (beginning stoke) and the high count (end of stoke) . Then oscillate between these two counts (When the ram is at max count travel then reverses count back down to zero). I would still use end travel emergency switches. This is what I am thinking right now, but still in the research phase.

AS5134 High Speed Magnetic Rotary Encoder - ams | DigiKey

ams AS5134 is a contactless magnetic rotary encoder for accurate angular measurement over a full turn of 360°.
www.digikey.com
 
I've actual put quite a bit of thought into this, as I have shaper syndrome, a condition in which you impulsively buy shapers and planers until you no longer have room for anything else in your shop, house or yard.

Here are my thoughts (and I am working on plans to convert a Lewis shaper to CNC right now)
Requirements:
1. CNC machines are meant to be set up, turned on and basically forgotten until you need to change the part. Shapers, especially the older ones require constant attention. Any CNC shaper that is more than a proof of concept will need to have either sealed bearings or automatic oiling for the ram moving parts. I've seen too many shapers and planers with ram ways scored because of lack of lubrication.

2. CNC machines need a tooling system that is easy to change out and consistent, which means some sort of quick change tooling system. I had planned on using a BXA lathe tool post with custom holders and carbide insert tools on my machine.

3. You don't need a clapper box. You also don't need the toolhead to rotate, in fact it should be pinned perpendicular to the table travel. You should be able to just program the tool to move a clearance distance away from the cut on the back stroke. You can also program it to cut any angle within the capability of the tools you have.

4. You don't need ballscrews. Some sort of backlash takeup would be nice on the toolhead, but force on that tool should only ever be in the upward direction, I thought about using some air springs for it to take that up. You should be able to get away with planning the cuts to not change directions and use backlash compensation. For that and a few other reasons that I will get to further down, LinuxCNC is my choice for control software.

5. You don't need encoders on the ram system. Some hall sensors and magnets or switches and a cam should be sufficient. Some adjust ability would be nice if you wanted to make sure you had the maximum time to move the tool in and out of the cut, although I won't know if this is needed until I see how fast I can accelerate the table and workhead. You could also compensate for this somewhat by using more lead in and lead out for the tool to the part at the ends of the cut. If you wanted to automate a bit more, you could use an encoder and do some software work in LinucCNC to set them by positioning the ram and pushing a button. This is another reason I like LinuxCNC, there are existing M codes for digital input and output that can be used for this or turning on an oil pump, if you don't have a mechanically driven one.

The real problem that I see with a CNC shaper is the lack of a CAM tool to build the G code files with. I've been working on something in Excel VBA, but I'm no programmer, so it has been slow going. The CAM tool should be able to make facing cuts (cut starts on the side of the work and goes to the other side or the top of the work), plunge cuts (cut starts on the top of the work and goes to the bottom or out he side), slotting cuts, dovetail cuts, inside and outside radius cuts and other specialized cuts such as gear generation. It should be able to do all these cuts at any angle the tool is capable of. That's a pretty tall order, but I think I've figured out (part of) the math and logic of it, I'm just having trouble with the programming end of it. They still make new shapers in China and India, and I checked once long ago and they sold a CNC shaper from China. Don't know if there was a software package that went with it, though.

Good luck, that's quite an ambitious project.
 
frugalguido said:
I was looking at using a AS5134 programmable high speed magnetic rotary encoder. It can program zero position and has 9 bit multi turn counter. With this I can zero the beginning of the ram start position and count to the end of ram stoke at specific material location, store this count in memory, thus I have zero (beginning stoke) and the high count (end of stoke) . Then oscillate between these two counts (When the ram is at max count travel then reverses count back down to zero). I would still use end travel emergency switches. This is what I am thinking right now, but still in the research phase.

AS5134 High Speed Magnetic Rotary Encoder - ams | DigiKey

ams AS5134 is a contactless magnetic rotary encoder for accurate angular measurement over a full turn of 360°.
www.digikey.com
Click to expand...

Depending on how you set up the system and what drive motors you use, you may not need an additional encoder. For instance, if you are going to use a servo motor to drive the ram, then it would have a built in encoder. Exactly how the ram position is controlled is really dependant on the controller that you use.
 
Would a servo motor encoder have the ability to count and be zeroed?
 
It depends on the controller as to how that is handled. There are a lot of options. There are servos that can be programmed to operate between two points, other options are more controller dependant. Since you are really going to have a multi-axis machine, some kind of a normal CNC controller would work well. But again, there are a lot of options.

For instance using any CNC controller a simple G code program would do exactly what you want. Let's call the ram axis X, then the G code would look something like this:
G1 X6.00 (move out 6 inches)
G0 X0.00 (rapid return to 0)

Then just loop the code, and add a Y move to move the table over for the next pass. There are several ways to do this. The proper code format could be generated in just about any CAM program.
 
Just to be clear, the thesis I wrote above assumed a retrofit of an existing machine, not a new build. I had also thought about building a small machine using a NEMA 34 stepper for the ram power. Using an encoder with or without a servo motor would require the use of more sophisticated control software, such as LinuxCNC or Mach 4 (which typically require more expensive processing and interface hardware). One other reason I chose LinuxCNC for my (possible) build is that it has a wait for input M code (M66). My plan was not to use any move commands except for G0. The moves would be programmed with a G0 move then an M66 command. The machine would wait for the signal that the ram was at the rear, then move the cutter to the cut position, then upon reviving the M66 command for the ram forward it would move the tool to the clear position, repeating this until the cut was completed. If you think about it, there are no other movement commands need for a shaper. Being able to change the position of the M66 commands so that it gets the command as soon as it clears the material on either end of the work would allow more time for the axis motors to move to either the cut or clear positions (as would more tool run from the edge of the material). I don't know if this is actually needed in practice, as a reasonably sized stepper motor may be able to make these moves in plenty of time for reasonable shaper run speeds. It's something I need to look into. You need to have some tool run from the edge of the material because the speed of the ram is not constant and slows down rapidly towards the ends of travel, so to keep a reasonable surface speed on the cut some is needed. I usually leave about 3/4" on the lead in and about 1/2" on the way out of the material, but that depends on the setup to some extent (and how much stroke I have left).
 
You must log in or register to reply here.
Back
Top