Any CNC Builders in Austin Texas

[Mr.Chips]

I am a newbie and have enjoyed this site very much, is there a CNC builder/users group in Austin Texas?
I have several questions, which lead to other questions about building my CNC.
Think it would be too time consuming for everyone if they were posted.
Thanks

[Laff Riot]

Post away - New users come in weekly who need to start at the begining. Its a bit intimidating pre-reading a couple thousand posts to make sure your question hasnt already been answered.

Fortunately the post areas are well defined. I know the groups grown so much since I joined that I would feel a bit overwhelmed by the volume of information to absorb.

It occurs to me however that maybe a "Read me first" posting area would be helpfull to ramp newcomers up to speed before they jump into the more specialized post areas.

All the real rough questions n answers n such.

[High Seas]

WOW a "Read Me First" is a great idea. (This note is from another fairly new user).

It has occurred to me, that a "Quick Start" Checklist/Discussion in the Technical Articles would be a great place to begin. For me getting into this it seemed like I was at at the bottom of a very steep learning curve looking a LONG way uphill. This is the best source I've found on the 'net. Thankfully there is a lot of help/encouragement/experience on this site!


Hitching on an earlier discussion (thread) that involved free/low cost plans and "pay as you go support" -- access to the Technical Articles" could help fund the "Zone." A homebrew quick start article, plans, etc would be real helpful to newbies. It could also help refresh those little "lost great ideas" we all have misplaced from time to time. A series of articles that could lead a newbie from "thinking about it" to "doing it".

I started putting together an outline of what I'd be looking for - (was and still am looking for -- but have been reluctant to ask a hundred questions and tie up the great exchange of info here).

Mind you right now I'm staring at a pile of 80/20, a bunch of linear rails, screws and a near empty martini glass! Got a way to go 'til I can truely SWEEP CHIPS!


I've been very encouraged by level of knowledge on this BB -- its good old fashioned machine knowledge, high tech electronic wizzardry and a whole heap of programming skill here. If some could be captured, organized, and put in a set of articles wouldn't that be great! As a newbie I'd say yes!

As a member - I say how do we get it started?

JIM

PS I know this is a slippery slope!

[chuckknigh]

Look at the top of the subject listings -- you'll see "Sticky" mentioned for 2 threads, at the top.

This means that, rather than scroll off over time, these 2 threads will "stick" around at the top.

This would be a good way to make a "Readme.1st" listing.

Since you're obviously not intimidated by the task, I'm guessing that the mechanical part of the router is not a big deal for you. The electronics may be a bit "iffy" though -- I know that was my greatest worry.

Let me suggest a thread that I started, a while back. It's called "I think I figured it out..." and it can be reached by clicking on this link.

http://www.cnczone.com/showthread.php?threadid=438

Basically it's my "Eureka" moment, when I finally understood how the mechanicals and the electronics finally fit together. It might help you.

-- Chuck Knight

[Mr.Chips]

Thanks for the link, the haze is rizing slightly, but still no light.
Here is a segment from Balsamans comment shown below my questions.

He mentions a "Gcode interpreter". Is that part of the Machine control software such as TurboCNC, CNCpro, Mach 1 etc. Or is it another software.

Let me parafrase (SP?) what I understand:
Connect everything up
Have a program that drills four holes making the largest rectangle you can on your machine. Then measure the four holes to see it they are properly spaced and square. If they are not you can edit something in the software and make the corrections. Then run the routine again and recheck it to see if the setup adjuctments made all the correct corrections.
Is this kinda what happens or am I lost again.

What controlls the cutting speed?
Are there recommended cutting speeds for different materials/bit sizes/HP of router?

How does the CNC know where the item to be cut is located?

I have seen some people cut pieces that are longer than their work area. How do they reposition the work so the finished piece is the correct length?

Have red some horror stories that because of an error in programming the router was trying to drive through the work at a too high a rate. Is there some sort of feedback sensor that would sense the higher load and shut down?

I have seen several controllers for sale, both fully assembeled and bare boards, how can a newbie know which one is better? Is there a independant comparison of the various boards? I am retired and on a limited budget so I want to go with the most bang for the buck. Board assembly is not a problem for me.

I'll stop asking questions don't want to over burden.
Thanks

[chuckknigh]

OK, do you know anything about programming? G Code is a generic programming code that describes the part being made. It says things like: move the Y axis 1 inch.

The G-Code interpreter reads these generic lines of program code, and turns them into instructions that make *your* machine move the right distance, in the right way. On mine, the Y axis is defined as pins 2 and 3 on my parallel port, so it'd vary the signals on pins 2 and 3, till it had told the motor to move the right number of steps, and in the right direction.

Does that make any sense?

Here's the way it works. You have 3 pieces of hardware...your CNC machine, your PC, and the black box in the middle, that connects them. That black box is called a driver, because it turns computer signals into powerful pulses that drive the motors. Make sense so far?

OK...the CNC machine, itself, is dumb. It doesn't know how to move. All it has going for it is just that each time each motor gets a pulse, it moves some pre-defined amount. Mine are 200 steps per revolution, which means each pulse moves it 1.8 degrees around. 200 pulses, and my screwdrive goes through 1 whole revolution -- 18 threads per inch.

The box in the middle is where the computer signals go to, and where they're turned into pulses that control the motors. This gets complicated, and is unnecessary for this discussion. (Ask about this stuff, next...)

And finally, the PC is obvious. It's your desktop or laptop PC computer. It has a parallel port, and you hook the parallel port to the driver box with a 25 pin cable.

On the PC, you run some software...something like TurboCNC. TurboCNC is a favorite program because it's shareware...free to try, and only $20 to buy. That's cheap enough to play with... It also is very flexible, and can be configured to run all sorts of machines, with up to 4 axes (I think). And, to answer your next question, a 4th axis is a rotational axis, usually...imagine engraving a wedding ring...you'd want to turn the ring while you were doing the engraving.

BalsaMan (Eric) mentioned a G-Code interpreter. This is built in to most of the CNC control packages, and just turns an instruction to move the Y axis 1 inch, into the requisite number of pulses and the right direction signals...on mine it would be 3600 pulses, with a high signal on the direction pin.

Does that help you understand it?

------------------------------------------------

Now, on to your other questions.

The cutting speed is controlled primarily by the speed with which the computer sends out the pulses. I said mine takes 3600 pulses per inch -- those pulses could be outputted through the parallel port at 100 pulses per second, or at 100,000 pulses per second -- the software (TurboCNC) has a way to adjust the speed, and also the acceleration curves.

Yes, there are recommended cutting speeds for various materials, but that goes WAY beyond the scope of this particular message.

How does it know where to cut? That's a very good question. You define a "home" position, which you can liken to a zero-point. That's where it's X=0, Y=0, Z=0. All Z dimensions are negative...that's down. X and Y depend primarily on how you set it up in your software.

But, basically, you take your CAD program and draw a rectangle the size of your CNC machine's capacity. One of the corners is (0,0) Any dimension is measured from there...your starting point.

That's where backlash comes in -- backlash is when you turn the screw drive, and nothing moves. It's the "wiggle" in the system. You try to minimize backlash, since everything is indexed off of your home position.

Feedback sensors. What you described is called a closed loop system, and requires something along the lines of an optical encoder. Basically it's something that *guarantees* the position of the bit. Most homebuilt routers do not use this system, as it is expensive and difficult to implement -- however, there are at least 2 members of this forum who have servo driven (feedback controlled) machines. A more common approach is to use a stepper motor, which turns 1 step per pulse -- you can't guarantee that it's not losing steps, but it's a much simpler system, and much less costly for a first time builder to implement. I got my stepper motors out of old surplus HP LaserJet printers, so their cost was negligible.

Lost steps are not much of a problem for woodworking routers -- know your limitations, though. If the machine is flexing under load, you're taking off too much material in one pass. Just use some common sense, and it's not a big deal. BalsaMan's first router used the same steppers as I currently do, and he NEVER lost a step...and that included cutting aluminum.

Controllers -- I assume you mean drivers. The boards that tell the motors how/when to turn? Most of them are built on the L297/298 chipset, and are roughly comparable. If you buy premade, I've heard nothing but good things about Mariss, from Gecko Drives.

I think that answers most of your questions. Speaking only for myself, I welcome questions -- don't worry about overburdening us. If you do...we just won't answer for a little while! ;-)

-- Chuck Knight

[Mr.Chips]

Thanks for the info Chuck.
Couldn't fool you, I haven't done any programming.

1. Did you drill a small hole at your zero-point to measure from when setting up work?

2. I understand X and Y zero positions now. As for Z, wouldn't that change if you use different length bits?

3. Would it be possible to download TurboCNC and try it out without being connected to anything, just for practice?

Thanks
Hager Hay

[abasir]

Hager,

1. There is no need to drill the hole. Once properly set-up, the software will 'measure' any move correctly, e.g, G0 X1 Y1 - will move the bit +1" in X and +1" in Y axis.

2. Yes. Manually (via software or handwheel) move the Z axis until Z is at your Z zero position then set Z axis to zero in the software.

3. You can download TurboCNC at www.dakeng.com and you can run it without connecting to anything.

Regards,

[Mr.Chips]

Sorry, Im not making myself clear on my location question.

1. If one time I place a piece to be cut in the corner of the work area and cut it. The next time I place it in the center of the work area how do I communicate the new location to the software so it will cut that piece in a different location. Seems to me I would have to measure from the original X/Y position to give it the new coordinates.

2. The way you explained the Z axis may be how you set the new X and Y locations, doing a Zero-point for both axis.

So you must have three buttons X-Y-Z. And you would establish a zero-point before any cutting. Is this correct?

Thanks
Hager Hay

I must be getting real smart, I ask so many stupid questions. ;=)

[chuckknigh]

Remember, and this sentence is very important, that you always start cutting from your zero point...and all locations are indexed off of that zero point. (Your zero point is not necessarily the corner of your machine, though it usually is)

There are 2 ways to do what you ask. One is to redraw the drawing, and generate new code. Not a bad approach, but not ideal. These days it's a matter of clicking and dragging to a new location.

The other is to "jog" your machine, or in other words, move your machine to an appropriate starting location, and then tell it to assume that the current position is "home." Jogging the machine is choice 4 on the TurboCNC menu, if memory serves.

A question, though..why would you position your workpieces all over the place, randomly?

-- Chuck Knight

P.S. There's no such thing as a stupid question... A favorite teacher once told me this...it's the difference between being stupid, and being ignorant. Ignorance is a treatable condition -- stupid is forever.

[Mr.Chips]

Chuck,
Thanks for seeing me through this.
I just gave the example of cutting the same piece. I could see for some reasons it would be necessary to change the placement of the material because of the size of the material.

Jogging the machine to the start point and declaring that the 0,0 point makes sense.

Thanks

[chuckknigh]

Not a problem -- and I should probably thank the other members of this group. I sounded a LOT like you, a few months ago. Yes, that recently.

I'm an anal retentive geek who wants to understand *everything* completely, before doing *anything.* The members of this group are great. I'm just glad I can give back a little...

Now, the best suggestion I can give you is to just build "something" and get it moving. Think of it as a proof of concept model, if you'd like.

If you choose reasonable components, you can reuse the expensive parts (motors, driver boards) in your second generation machine. But, regardless of how you do it, just get something moving under computer control. You'll learn more by doing, than you ever will by thinking about it. And, no matter how good you are, you *will* want to change something after you've built it...it's just human nature. That's why I'm suggesting a proof of concept model -- you'll learn, first hand, where the weak spots are, and how to correct for them.

Search for my old messages -- I built my machine from scrounged parts, and I think I set the record for cheapest machine ever built. Cost need not be an issue.

-- Chuck Knight