CNC configuration

Other Versions
Spanish
Building a CNC router. May 1, 2003

Question
I'm in the process of building a CNC router. I'm having trouble selecting the proper lead screws. What is the average/standard threads per inch and does most CAM software have a way to configure this? I don't want the computer to think it's moving an inch when in reality it's moving two inches.

Forum Responses
(CNC Forum)
From contributor B:
Your CAM software will create G-code that your router will follow. Your control software for the router will read this G-code and move the router around your table. Control software is configurable to each individual router just as you will need.



From contributor T:
According to the machine data from the last point to point I worked on, the X-axis (rack and pinion) traveled 27.6mm per revolution of the servo, the Y-axis (ball-screw) 15.2mm per rev, and the Z-axis (ball screw) 5.0mm per revolution.

I think the more important question would be what diameter ball screw do you need? I would expect that the larger diameter ball screws would have a larger pitch, and that you must configure the software to account for that pitch.



Contributor B is right in that, as long as you know the pitch (and any gears, etc. attached), the CNC control software can accommodate it, i.e., if 2.54781 rotations make up an inch of travel, that's what you enter.

Contributor T's thought on the diameter of the lead screw is correct, but the size of the screw does not relate to the pitch, except in a general sense. A bigger screw is stronger, though, and hence can spin faster without deflection. And since the pitch determines accuracy but takes more RPM to move an inch, you need a bigger diameter screw to get more speed, and also to get more accuracy (and torque I think) for a given speed. Need both? Go big, or go ball screw, which doesn't rely on pitch as much for the accuracy.



What is it that motivates someone to tackle a project like this one? It seems to me that anything that the machine would save you in process and efficiency will be eaten by the amount of time it takes to make the machine. I am not criticizing your desire. I respect it. I just think it would be a huge version of the jig that takes a day to build in the shop and ends up saving about a couple hours on a project. I built a vacuum hold down table that spins and locks using a venturi device and I look back on it and wonder if it would have been cheaper to cough up the 1100 dollars to buy one. I guess I did learn a great deal about vacuum and air systems... Maybe I should try to build a CNC... hmm? Good luck.


From the original questioner:
Thanks everyone; that helps a lot to clear the fog.

I'm in the process of moving and will not have my shop back until February. I do have access to my tools, but in an extremely limited space. So, I've got time to kill. The reason I need the machine is because I've gotten a contract to do the milling for a company that sells aftermarket guitar bodies. I'm going from a dozen a year to 50 a month overnight.

At first I was just going to build it to make my templates, figuring that I could make the machine for what a machine shop would charge for one template (I need 12), but then I was able to trade a guitar for some stepper motors that would otherwise be way out of my price range. Now I've got the power, may as well build the full-blown ShopBot type machine.

Quick follow up question: is it the control software, then, that is told a) the size of cutter in the tool and b) where the tool is now (i.e. reset zero)?



What have you decided to use for motion controller? There's a bunch of stuff out there that plugs into a PC and can accept standard G-code. Also fully configurable for ball or lead screw pitch and a whole host of other parameters (limit switch sense, calibration markers, etc.). Most of these can be configured to drive a variety of motor types - stepper, DC servo, AC servo, vector control, etc. Check out some of the following:
Galil, Acroloop, Delta Tau, DSP Control Group.

Most can provide rudimentary operator interface software as well. These are just a couple that I know off the top of my head.



From the original questioner:
I talked to a guy at the manufacturer of these stepper motors and he indicated that the drivers connect straight to an LPT port and from there any CNC software should be able to control them. Is there a third piece to this that I'm missing?

I basically want to take my CAD drawings, pull them up on the machine's PC, tell it how much depth per pass and what size cutter, hit the GO button, and go make a sandwich.

Am I wrong that this is how it should work? If I do have to plot every step manually, I can live with that because I'm going to be making the same part for months once I get it going.



It's kind of funny, you know, the same motivation (make a sandwich) would make me just find a working, used, name brand CNC that someone has for sale cheap. The market is so packed with distress sale CNCs that are selling for much less than the cost of the components. Guitar bodies require a pretty heavy frame and horsepower, and long tool length and even a tool changer to be really effective.

I think that you are admirable for doing it yourself. I think the software questions are worth asking, and the more you know about your machine the more you can do with it in the long run.

The deals on turnkey systems are just too killer right now.



From the original questioner:
The ones I've seen just don't seem worth it to me. Maybe I just don't know where to look, but these five-figure machines don't have four-figures worth of parts in them to my eye. While I think that a lot of satisfaction will come from making it myself, a turnkey machine would meet my goals a lot more efficiently. Just can't see spending 10 times the money for a machine that runs on cables and has a rotozip in it.


It's definitely worth looking for used. The trouble to build one if you don't have one for a) reference and b) making the parts that make the machine accurate and square is not worth it if you find a good deal. The control issue is especially important, as I'm less good at electronics than mechanics.

Things to keep in mind. A) A Shopbot would probably be fine for your work. After 1-2 years or less you could always start making your machine, either upgrading the Shopbot or using what you've learned from it. B) Rack and Pinion is okay for the work you are doing, accuracy wise, and easier to find a good fire-used used unit. Trust me, if you can find a used unit with a servo drive, you'd probably be okay with a chain drive if you looked after it! The servo lets you run all day without worries about the steppers stalling and screwing something up. Doesn't happen often, but....



Check out Flashcut CNC. I used their signal generator along with Geckodrives for a retrofit to an old Thermwood router with real good results. The machine is slow by today's standards, but I also have less into it than a Shopbot and ten times the machine.


The comments below were added after this Forum discussion was archived as a Knowledge Base article (add your comment).

Comment from contributor A:
I am a design and manufacturing engineer for one of the largest guitar manufacturers in the world. I have a few thoughts on your project, and some suggestions. I think it is admirable to try to build your own CNC (I'm doing it now, too), but I think you need to examine your needs and skills a little more before proceeding.

I agree with you - I don't think any of the small hobby type machines, and to me this includes several models that go up to around $50k, are worth the money. Especially for what you are trying to do. I also wouldn't suggest buying a used machine before learning a little more about them, and seeing how the business goes. I've seen a lot of people buy a machine before they needed it, and that's why there are so many used ones on the market. Besides, lots of the big boys make most of their bodies without CNCs, too.

What are your reasons for going to a CNC? What equipment do you currently use to make bodies? You said there were about 12 bodies you need to make. What about routing configurations for pickups, bridges, etc.? What type of neck? Glued in or bolt on? Are the bodies carved, like a LP, or flat like a strat? How are your electrical and mechanical skills? I'm not trying to be critical, I just think there is a better, more economical way for you to go about this.

I don't believe any machine you build, especially with stepper motors, is going to satisfy you from a quality and accuracy standpoint. I think you would be better served using your time and money to invest in a good, used pin router. Something with a heavy table, good spindle, and easy depth adjustment. I would get a good used two spindle shaper and bandsaw if you get into necks, too. With just 50 bodies a month, you can rough out the blanks with a jigsaw. Since there are just twelve different bodies, you could buy or make some standard templates to do the body shape, control cavities, neck pockets, pickup pockets, etc. If you do your own CAD, and can save to a dxf file, someone out there would make them, I am sure. Things are tight everywhere, and most people with a CNC would love to make a little extra cash for running a sheet or two of patterns. I would suggest Baltic birch for the templates.

If you still feel like you need to build a CNC, check out the plans at machinetoolcamp.com. You buy the plans and they guide you step by step through the process of building a machine like you are talking about. This machine would be great for making templates, jigs and fixtures for your pin router. I just think it would take too many passes and be too hard on the machine to do the whole process. And you have to consider fixturing the CNC, too. You would definitely need a vacuum system to hold the parts down, and that could cost as much as the machine. There's also the problem of not being able to change tools without stopping the machine.

A good source of info on guitar building is an ebook by Jim Donahue titled Guitars: Design, Prodution and Repair. He goes into how different sized operations, from home shops to large manufacturers, do it. And it's not me - I don't work for Ibanez. It's not the clearest written book in the world, but it does have some great info.
Also check out Building Electric Guitars by Martin Koch. He has an ebook and printed version.

Just remeber CNC is just a means, not an end. It doesn't guarantee success, but it has doomed many startups when done wrong. Stick with the basics, be innovative with your fixturing and jigs, and in the long run you'll be much better off.



Comment from contributor C:
If you waste the time to build a CNC controller out of motion cards you will also have cost that you don't need. There is software that will make a CNC controller out of the printer port of a computer and then you only need the drives, motors and the mechanical link.


Comment from contributor D:
You are correct in presuming there is a third piece that fits between the parallel port and the stepper motor. Parallel ports can, in fact, directly power and run tiny steppers, but anything that you are considering will need a driver for each axis. Additionally, in order to get from CAD drawing to G-code, you will need another software package that "generates" the appropriate XZY coordinants. Prices range quite a bit, from several hundred to several thousand. I would consider coding by hand, though, as your parts sound fairly simple and a one time task.


Comment from contributor E:
I too am building a CNC routing table. I suggest you find the book titled "CNC Robotics: Build your own workshop Bot", ISBN # 0071418288. Take the ISBN number to any bookstore and they can help you locate the book. I have built the frame and the gantry for mine. I just struck a deal with my company on two 6-foot Acme screws and the slides! Anyway, the book will walk you through eveything from building the frame to soldering the driver circuit board.


Comment from contributor V:
Cutting out guitar bodies or other contoured shapes, for that matter, is not difficult. You mention using Geckodrives as the stepper drivers and Deskam as the control software. Deskam claims 2 1/2 D. You can contour 3D as long as the output to DeskNC is STL. I made a machine using the Geckodrives with the Deskam software and it works very well.

I made a 5' X 12' CNC that is capable of cutting a myriad of materials. In parts, I spent approximately 4 grand. What kind of machine could I buy for 4 G's? Maybe Shopbot or Shop Sabre. However, you cannot buy these machines in these sizes for this price!

I started to build the Shopbot according to their procedures, but using Greenley struts as the x and y axis rails is nuts. I departed from them after I had built the table (I would differ from their table design in afterthought). I purchased stainless 1/2 diameter predrilled rod with the appropriate aluminum supports and Ina bearings from Mcmaster Carr for the linear motion guides for the x and y axis. The x and y axis on my CNC are driven by a gear and rack system using 1050 oz/in steppers with anti-backlash springs installed between the two gears. I have an accuracy of .003 with a speed of 400+ ipm. The traverse (no load) speed is much faster. It moves! The harmonic ossicilations in a non-tensioned ball screw system will limit the top speed (whether loaded or unloaded to approximately 150 to 200 ipm whether you rotate the screw of the nut to accomplish linear motion). And most machines that I have seen that cost upwards of a quarter mill use the gear and rack system anyway.



Comment from contributor W:
I am the owner of a small company, and I am presently developing a CNC Control package. The selection process of determining which lead screws to use is guided mainly in nature. The pitch and thread count will effect torque and speed. The diameter will affect maximum torque applicable to the lead screw. The controller must have a parameter in which to determine threads per inch. This is critical in assuring accurate movement interpretation by the controller.