My new project

I have decided to extend the "Egyptian pyramid CNC" approach. I am writing a set of perl modules that would give simple instructions to human mill operators (move this handle, that handle, adjust backlash, etc).

Since I am not really limited by anything (I would operate the mill manually), I want these modules to take advantage of the fact that I have three linear axes, quill depth, and rotary table. If I am not mistaken, that means that I have a five axis motion.

The lowest level would be concerned with keeping track of X, Y, Z, depth of quill, and rotary table angle. (and computing cutter position relative to the part, based on these coordinates).

It would also output instructions (turn handle such and such etc). This is not simple for a variety of reasons (concerned with human nature of the operator since there is a need to account for possible distraction and resumption of work, I will explain later).

The higher level would compute the curve that the cutter should follow in order to mill a given surface (or line in 2d context). (the cutter has a radius).

This would basically let me create easy to follow milling instructions.

Good comments could be made that there is software that does something similar. I would like to be able to perhaps interface my modules to some cad systems like brlcad or some such.

I do think that even if my work is a duplication of something that exists, there is some value of it to me. So I will make steps in this direction, with particular interest given to proper use of rotary table. (milling circles is much more easily done, with higher quality, on a rotary table -- the circle I made was for illustration only)

As a bonus, I will try to write them so that I can plug in some sort of G code module for output rather than a human output.

i
Reply to
Ignoramus7101
Loading thread data ...

Reply to
E. Walter Le Roy

Sounds like a project and one full of fun and work.

Mart> I have decided to extend the "Egyptian pyramid CNC" approach. I am

Reply to
Martin H. Eastburn

Thanks... I can do much of it on the train, since it is mostly computer programming & some math...

i
Reply to
Ignoramus7101

As much as it breaks my heart, I will be brief and implicit:

You'll have a useful tool, and will not have wasted your time, by converting your Bridgeport to CNC instead.

Regards,

Robin

Reply to
Robin S.

I have nothing against it, my problem is that I do not quite understand the issues involved (what motors, drives, controls, feedback etc). I am totally in the dark about it and have hard times finding anything concise.

In fact, I would not mind spending money if there was a book available about CNC conversions.

Plus, a good set of scripts might help with some CNC projects, tests, etc.

i
Reply to
Ignoramus7101

Iggy

There is a magazine published by the same people (Village Press) who publish Home Shop Machinist devoted to, well, home shop CNC called Digital Machinist. In fact the last issue started a multipart series on converting a Bridgeport type mill to CNC. This is published quarterly so it could be awhile for the whole article.

I've been watching what's available on the subject as I'll be doing something similar eventually. At this time the controls such as axis drives, breakout boards, and PC software are relatively inexpensive, and more or less 'modular'. The expense, as I see it, involves the hardware. When I worked in machine shops I ran a fair amount of NC and CNC equipment so I probably have some preconceived notions about what's required.

First, go with ballscrews. More expensive but leadscrews are for cranks and ballscrews are for drives. I've seen some nice router tables done with allthread for lead screws and it worked quite well but for metalwork go with ballscrews. Servos instead of steppers and if you can afford it position feedback (as in digital scales) instead of encoders on the drive motors. I haven't really seen much on the last part but I'm a more or less casual observer of the state of the art at this time.

Anyway, I know you have the computer skills and I'm pretty sure you have the mechanical ability for such a project.

Regards Paul

Reply to
Paul

You won't find a standardized opinion about the process because you have quite a few options.

Have you visited the CAD_CAM_EDM_DRO group? You do have to do some digging, but everything basically scales. As your machine gets larger you need larger motors and larger drives. Software is the same, and the principles are all the same, regardless of machine size.

Roll-your-own CNC moves at more of a magazine speed, than a book speed. There are certain things that stay consistent (leadscrew options, motor options) but the real trick is the drives and the software (which evolve).

Gecko seems to make some good drives that are well supported, both from the manufacturer side, and the user community side. EMC seems to be an excellent option for software. The Mach software seems to be an excellent option as well, although I don't think it's free, and it does run under Windows (not sure if there are actual issues there though).

Once you have the iron, everything else is just plugging wires into the right places and configuring software. You don't need to write any software or reinvent the wheel.

I've built a small two axis CNC machine. I even soldered the board (purchased as a kit) and built the iron. That was in my last year of high school. There really isn't all that much to it.

Learn how to write G-code. Just learn it (!) Learn how to write the entire program, top to bottom. Saftey block, plane selection, etc., etc., all the way to rewind.

For work that can be done manually (as opposed to using a CNC machine) the MDI function, which allows you to enter and execute single lines of G-code, is much faster than CAM programming in certain circumstances. You need to be able to look at a g-code program to troubleshoot what your CAM system spits out anyway.

Regards,

Robin

Reply to
Robin S.

Sounds very interesting, I may even subscribe to it.

I have no such notions...

Paul, just how well can I compensate for backlash if I use positioning input in addition to servo motor feedback (which is implied by the term "servo". It seems to me that with good positioning input, much of backlash canbe properly compensated for.

I think that I will spend a long time learning about it.

What mostly surprises me is that most people mount large motors with relatively low gear reduction (1/4? or so), instead of using much smaller motors that are substantially geared down. I have a lot to learn, anyway.

i
Reply to
Ignoramus7101

I have not, but I am a regular reader of cnczone.com.

Where I am confused is what is needed. Let me make up a list. I would need

- A computer with EMC2 and Linux (no problem)

- A computer PCI board to connect to breakout board

- a breakout board to connect to the computer, quadrature inputs, and servo drives

- Servo drives.

Just how far off am I?

Very nice.

OK, I will read about it. Thanks a lot.

i
Reply to
Ignoramus7101

On Thu, 14 Jun 2007 23:18:53 -0600, with neither quill nor qualm, "Mike Henry" quickly quoth:

I wonder if that DVD will be available to the public...

- Metaphors Be With You -

Reply to
Larry Jaques

Try

formatting link
for several forums that are oriented toward just that sort of thing.

The CNC Workshop is being held in Galesburg, IL this week and is apparently a sort of informal gathering for those interested in building and running hobby-type CNC machines. I believe that it ends Sunday and you might find it productive, or at least interesting, to spend the day there on Saturday.

formatting link
Mike

Reply to
Mike Henry

In your case, I would stick dollies onto the handles. A frog onto X, a turtle onto Y and a bear onto Z. This way, instructions will be understandable even for you.

Computer: "Now Iggy, turn the turtle clockwise half a turn and you get a cookie".

Of course, you could *learn* how to use a mill. But you prefer making nonsense, or you wouldn't be Iggy.

Nick

Reply to
Nick Mueller

Iggy,

Spend some time turning the handles and making chips yourself before jumping into cnc.

Wes

Reply to
Wes

Well, most software does have some means of backlash compensation... but I tend to view this as a workaround and suspect there will still be issues. The servos I've seen do have encoder feedback, but that is on the wrong side of the nut to screw connection to deal with it. I believe the main issue is cutter forces acting on the table, as like climb milling. Seems finish, cut profile, and cutter breakage would be likely consequences.

I'm far from an expert on this, I'm sure somebody else here can better answer your questions. Remember though that a long carbide end mill profiling deep in a pocket is a different game than engraving plastic laminate signs:)

Personally I think I'd be disappointed in the results of a Bridgeport type conversion that didn't use ballscrews.

Paul

Reply to
Paul

I'd be happy to help you along with converting to CNC. Believe me when I say you will be extremely happy if you do so. I knew nothing about CNC when I converted my first mill, but I did get good advice on the web. If you like, send me an email and I will give you my phone number and we can talk about what's needed and how to do it. I'll be available for advice as you progress.

Reply to
Rich Goldner

You need the computer (1 ghz is fine). I would suggest using Mach3 for the control software (I have been using it for years). It's very complete, easy to install and use and has great conversational functionality. It's only $159, however you can use it forever for free as long as your GCode program doesn't exceed 1000 lines. The computer's parallel port connects to the breakout board which goes into the control box that you build. The breakout board connects to the stepper or servo drives (Gecko's are great) that are also in the control box. The power supply (also in the control box) also connects to the servo or stepper drives. The stepper or servo drives connect to the stepper or servo motors which are mounted on the axis leadscrews. This is the minimum configuration. There other niceties such as limit switches, safety circuits and front panel hardware switches which can be added later.

As far as feedback, steppers do not use feedback. Servos typically use encoders on the servo shaft. Encoders anywhere else (such as linear scales or encoders on the leadscrews) can lead to uncontrolled oscilation. It is a good idea to eliminate as much backlash as possible, however Mach 3 does a good job of backlash compensation. I have converted three mills to CNC, but because of expense have always used the existing leadscrews. I would prefer ballscrews, but they cost almost as much as the rest of the hardware. The mill will be very useful even without ballscrews.

Reply to
Rich Goldner

My suggestion would be to do one intermediate project first. Get yourself a couple of 220 oz-inch steppers, even surplus would be fine, buy or build a chopper driver board, and then build a simple little cnc router. Plywood, some all-thread, a few aluminum bits would build a good learning tool. You could mount a dremel, a small router, or even just a pen to start. I have a board which was a kit and no longer manufactured, but it has PCI input via the parallel port, stepper drivers, and all the electronics needed on one board. It drives a CNC-equipped Taig mill. Once you have done this project, go ahead and research the equipment necessary for servo drives, or you may find that steppers work just fine for your work and stick with them. The board I have can reach a 1/16th microstep, which translates to positioning tolerance of only a few tenths, which is pretty good. The limits are really in the Taig machine rather than the steppers or drivers themselves. Another thing that really streamlines the setup process is that EMC2,

formatting link
now comes as a live cd already configured with Ubuntu and the latest version of EMC2. Checking the configuration files should be a fairly simple process, and there is very good documentation available. Hope this gives some starting ideas. ww88

Reply to
woodworker88

My mill has a feature that you might find interesting. It uses digital readout scales for position, for positioning, this automatically compensates for backlash and lead screw error. The motors themselves have tachometers that connect to the drives, this closes the loop twice, once for motor velocity and again for position with the linear scales. I know you've been considering DRO's and interfacing with the computer.

I would recommend that you download the Live CD for EMC2 and play with / learn about it. You could interface DRO scales to parallel port(s) and display the position on the screen. There are many possible configurations and your experience with Linux and programming would give you a lot of possibilities. Steppers are generaly cheaper than servos but you get what you pay for (to over-simplify).

Reply to
Roger_N

Email sent.

thanks.

i
Reply to
Ignoramus27770

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.