Basic CNC

[DMS]

Hey Matt, those Mesa Anything I/O boards look pretty cool. I was not familiar with them. I especially like the ethernet version 7I80DB Ethernet Anything I/O card. Looks like it would be especially nice for controlling multiple machines or something with more sophisticated things like tool changers. Do you happen to know if they require the real time extensions?

Double check compatibility first. I don't think the Ethernet versions of the Mesa board are supported by LinuxCNC.

 

[blacksmith]

Hey Matt, those Mesa Anything I/O boards look pretty cool. I was not familiar with them. I especially like the ethernet version 7I80DB Ethernet Anything I/O card. Looks like it would be especially nice for controlling multiple machines or something with more sophisticated things like tool changers. Do you happen to know if they require the real time extensions?

Hi David,

Sorry, I can't answer your question.

The folks at Mesa have been pleasant to deal with.

Give them a call.

REgards,

Matt

 

[jeffm]

I'll try a little on linear slides, ballscrews and acme screws
To be honest I didnt do much reading on items I can't afford except to note the differences.
Most people think the main reason for a ballscrew is backlash, while this is a reason to use ball screws it is not a feature of a ball screw that most hobby machinists can afford. So lets talk about friction first. The motors used for cnc are small and required to move very heavy loads in exact amounts. Friction is the biggest enemy. If you have ever tried to turn the handle of a machine that has set for a while you will feel a snap before the axis begines to move. The torque has to overcome the friction or adhesion that is between the surfaces before moving. This is everywhere in most manual machines. The thread mating in the nuts the dovetails mating to each other and the slides. Once this adhesion is overcome the motion becomes easier. Its the theory of motion things that are still want to stay still, things that are in motion want to continue moving. This plays hell on accuracy and the size of motor to overcome the initial friction/adhesion and a change in direction multiplies this as the mass has to be stopped then started to change direction (this happens fast but at one point it does stop ).
We overcome this with ball bearings. Bushings are mated surfaces rubbing against each other and the lube holds them apart so they slide on the thin film of lube instead of each other. These bearing surfaces are everywhere in a mill or lathe the dovetails are bearing surfaces and require lube to keep the two surfaces from dragging on each other. Ball bearings work different they do not rub the other part they roll against it always in contact at two points. If i put a ball on a table and put a book on top and move the book the ball will roll as i move the book not slide. There is far less surface area touching each other so there is less friction. Now if all the balls are in a straight line and "exactly" the same size they will all roll together and not against each other. Here is where precision bearings come into this, if one ball is the smallest amount larger than the rest of the balls it will catch up with the others and eventually end up pushing the entire line of balls along. This has the balls now touching in 4 locations top bottom front and back. The front and back of each ball is turning the opposite direction of the ball it is mated against so the friction and heat doubles. So ball size is very important
The surfaces the balls roll on have to be perfect also, a high spot on the table and a book that cannot be moved up leaves a tight spot in the bearing surfaces. So the balls have to be no bigger than this point in their path. Everywhere else they will be a loose fit. This is the reason for precision grinding a ball screw and it's nuts internal route for the balls to roll in.
To make the ballscrew antibacklash all the parts must be in contact at all times. but not ball against ball and they continue to roll along the thread including a passage to return them back to where they started from in the thread.
Ball slides work exactly the same way and the balls roll along the two mated surfaces and are returned through a passage to start over.
The cost of real precision ballscrews and ball slides are to high for most hobby machinist but anything that reduces friction is an advantage for our machines so even less precision ballscrews are a big improvement leaving us to just deal with the friction of our slides. this requires constant lube.
I know there is a lot to add to this and probably corrections needed

Steve

I'm way late here, but I wasn't a member yet when this thread was born.
This is a good description, I think, but I still can't quite see it. Can anybody point me at a link to a diagram or video?
Thanks,
Jeff

 

[DMS]

Here is a video on ball screw operation

[video=youtube_share;kl6qNn9-nkk]http://youtu.be/kl6qNn9-nkk[/video]

Here is a video on linear slides

[video=youtube_share;rq4Pis6Zhf4]http://youtu.be/rq4Pis6Zhf4[/video]

 

[David Kirtley]

Jeff,

It is just ball bearings vs two items sliding against each other. The ball bearings reduce friction.

Ball screws are basically a way to reduce the force needed to turn a high pitch screw. This goes back to the problem with steppers. They have a lot of torque when moving slowly but less when moving fast. A regular screw has a lot of mechanical advantage You don't need a big motor to turn it. The down side is that you have to turn it many times to get a lot of movement which can mean that the motor is turning fast enough that it loses torque (and other problems). To counteract this you can increase the pitch of the screw which makes it move more for each revolution. Let's use a 1/2 in screw as an example. A standard 1/2 in acme screw has a pitch of 10. That means that you can turn it 10 times to move one inch along the thread. Then you can have a multi-start screw. A 2 start screw will turn 5 times to move one inch. The motor only has to turn half as fast to get the same movement. The problem is that it also has 1/2 the mechanical advantage. They also have 5 start screws. They have 1/5 the mechanical advantage. With less mechanical advantage, the friction of the threads in the nut are a larger part of the force needed to move the screw. To overcome this friction, they put ball bearings between the threads in the nut and the threads in the screw. They also put a pre-load on the ball bearings to reduce the backlash.

 

[plm]

First off, this thread is a great learning tool, especially for a beginner. I commend all of you who have contributed your time to help others. Now, my question: As I understand it, the basic procedure to get from an idea to a machined part is as follows. Design the part using some form of CAD. Upload that drawing to the Mach or other program which converts the design into G-Code and then sends that information to the controller which in turn, controls the motors and thus the machine. Is this about right or am I missing something? Assuming that I have this basic concept correct and I then correct in saying that without CAD you're up the creek without a paddle? In other words there is no other way around the need for CAD?

Thanks to all

Patrick

 

[Hawkeye]

Patrick,

You're really close. That's what I thought when I started down the CNC road. From Cad, you need to go through a second program which generates the G-code, then feed the G-code into Mach3, which tells the breakout board (controller) what to send to the stepper motors.

Some of us are using D2NC (Design to numerical control) software as the second step. It doesn't do everything (there's a list of shapes in CAD that it won't do directly, like ellipses), but is a good starting point for us beginners. There are some work-arounds you can make up, such as making up an ellipse out of sections of arcs, which it can handle.

D2NC can let you make up the shapes you want directly, without a CAD program. They have a shape library that you can draw from, and once you get used to G-code, there are some simple shapes you can write in from that knowledge.

 

[JohnAspinall]

Now, my question: As I understand it, the basic procedure to get from an idea to a machined part is as follows. Design the part using some form of CAD. Upload that drawing to the Mach or other program which converts the design into G-Code and then sends that information to the controller which in turn, controls the motors and thus the machine. Is this about right or am I missing something? Assuming that I have this basic concept correct and I then correct in saying that without CAD you're up the creek without a paddle? In other words there is no other way around the need for CAD?
Patrick

Real close, but just as a different perspective, here how I think about this.

There's an information flow; it starts in your head, and ends up with motors moving. You've mentioned some assorted players along the information flow (e.g. CAD program, Mach,...) but it's just as important to know what form the information takes along the way. A simple version might be:

1. A CAD program turns keystrokes and mouse clicks into a description of shape (e.g. dxf file, AutoCAD file).
2. A CAM program turns a description of shape into a description of tool path (nearly always some flavor of G-Code).
3. A G-Code interpreter (e.g. Mach) turns a description of tool path into motor motions (e.g. coordinated steps for stepper motors, position commands for servos,...).

That's the main highway. There are assorted shortcuts and detours. For example, if the shape is really really simple, you might know the best (or only) toolpath in your head. You could skip 1 and 2, and write the G-Code yourself, if you're so inclined. Sometimes people will take the G-Code written by a CAM program, and "tweak" it by hand. Hawkeye mentions D2NC which (my perspective) is roughly a CAM program filling slot number 2 above, but designed to take shape input directly from the human, instead of from a CAD program.

Just my perspective, but: shape >> toolpath >> motor motion is central for me.

 

[plm]

Hawkeye and John

Thanks for the description of the process. I believe I have a pretty firm grasp of the process now. I hope this thread continues as there is a host of information here for the person new to CNC.

Patrick

 

[jumps4]

I use d2nc to convert dxf files into g-code it is faster than using d2nc to draw. d2nc is really easy to use the import and convert functions and does a lot of really nice functions like tabs.
steve