I've spent a couple of weeks digesting the vast amount of info. contained on this forum. Trying to find the answers I need is kind of tough to do considering how much info is here. So I'm gonna throw some questions out there.

I've decided that I want to build a home cnc router. I'd like to use it for building wood parts for acoustic guitars. I'm thinking that if the table could travel 24 in the x and 18 in the Y that I could use is for just about any part.
The router will be a craftsmen router for starters. I have a pc that is running windows 2000 that I would like to dedicate to the project. It is running an amd 1.1 gigahertz processor with 256m of ram. I want to start with a machine that I will adjustt the Z axis manually. So.... just a two axis machine.

I tried to download the emc brain dead install 4.6 a couple of times but the checksum was always wrong. When I started up off the cd it never worked right. I would always just end up at a \bash command prompt but not very much happening.

I tried to download mach2 on windows 2000 but that got tripped up too and gave me the dreaded blue screen!

I've heard many people say great things about a dos based g code interpreter but then it became evident that it was neccesary to buy an external board to use the software.

Then I started reading about emc and what was neccesary to make it work with controls. You either had to change the software or get some radio shack parallel port conversion kit. If possible I'd like to just plug the controller board into the parallel port and install the software in a windows environment. Is there anything out there like that? Is there anything like that for free?

Maybe I should back up a little. For what I'm trying to do I'm guessng that a step motor is the thing to do. What would be a good control to use? What type of tolerance will I get with this type of a control and a decent stepper motor?

That leads to the next question. I have read about people using every different kind of stepper. What kind of torque do I need to move this size table? What brand or type of motor do you suggest using and why?

Do ball screws work well on a machine that cuts wood? Can software compensate for backlash in acme screws? What is the best price to performance ratio for linear guiding a small cnc router?

Controllers are the next thing that I don't understand. I read much about gecko. They look like nice packages but it seems like they only run one axis and cost about $120. Xylotec looks like a pretty commonly used controller bu t how precise is it and does it work seamlessly with any of the gcode interpretters.?

Last question....Do major cam packages like surfcam, mastercam, or gibbs have output for controller like xylotec or gecko. I know they do for controls like fanuch or haas but do they for the home cnc controls??

Whatever info you could provide would be verry much appreciated. Thanks -Keith

When installing Mach2 in Win2K, it's very important that you download and read the docs before you install. Just read the firest part about installing it. You may also have much better luck with a clean install of windows, but there's no guarantee. Also, you should download the newest version of Mach3 instead of Mach2. A final release version should be available in the next week or so, but whatever is currently available on the website should work fine for you.

If you want to try DOS software, there is TurboCNC from http://www.dakeng.com or CNC Zeus from http://www.cnczeus.com
Neither will work under Win2K, you must use a DOS boot disk or install DOS to your hard drive on it's own partition. If you use a boot disk, you most likely won't be able to access your hard drive.

All 3 of these are actually the machine controllers. Xylotex and Gecko's are motor drives, or drivers. The drives are controlled by the software.

Neither Xylotex nor Gecko is any more accurate than the other. They both do basically the same thing. The main difference is that the Gecko can run bigger motors, and can run smaller motors faster (usually).

As formotor size required, I'd consider 200 oz-in the minimum. Bigger is always better.

The tolerance you get is not so dependant on the motors used, as it is on the quality of construction and components used.

Ballscrews work fine, most commercial routers use them. (if they don't use rack and pinion). The only problem is keeping dust out of them.

Software backlash compensation is available, but anti backlash nuts is a better option, as software compensation will not always work well. Ballscrews can have backlash as well as acme.

CAM packages output g-code files. All the software controllers previously mentioned use a fairly generic form of g-code. Most CAM programs should be able to create useable code using a generic post processor.

A router that size to cut wood should be pretty doable in mdf with steel pipe. It's usually 1/8" thick or less wood routing and most people buy the guitar neck already done. Especially if you only build a router, no neck making with this of course.

I'd recomment you make it a mill instead as you can cut the corner parts, neck with fretboards lines, inserts. Well, you may at least make provision for it in your design. Check the free cnc plans. JGRO's design is pretty all right as a starter.

Concerning the size you chose, don't forget that the guitar sides are much longer than the top as they curve and go around the butt. If you want to cut those as well you will have to make the table much longer. I would take a guitar plan and double check this... Kind of unfold the side part and check the lenght

It's up to you, but if you get into this, in not time you'll be wanting faster, biger, stronger... :) Take your time to build it strong and right, you'll be glad you did'nt skimp on the details...

I'd love to do guitars... But what I like most is to make em scream...

Para

I checked out jgro's design. Very nice I think. Thanks very much for guiding me in the right direction. I have to sit down and figure out how much all of this will cost to get going. I now have a pretty good idea what the setup will be. -Keith

Out of the experience of building my own machine, here is a quick evaluation of costs:
3 motors + controler board: 300$
Power Supply: 100$
MDF: 30$
Hardware: 20$
Aluminium angle, bearings, gas-pipe, ACME lead-screws: 200$
Dremel or Rotozip: between 30 and 100$
Spindle speed control: 50-100$
Cutters and bits: 50$ for a nice selection
Vacuum cleaner: 50$ for a big shop-vac

That's about it, so you should be good to build this under 1000$ or less if you can scrounge the MDF, bearings, motors (old laser printers) and already own a dremel. A good power supply may be done rewiring a microwave power supply or you may pick-up 28v AC (40v DC once rectified) transfos in printers as well. I made a power supply using 2 of those and have 40V DC 5.8 Amps for free!

Also, old printer shafts are sometimes big enough to use on the Z axis. This can save a few bucks there and the quality is better than gas-pipe. Minimum is 7/16 or 1/2 for a 10 inches shaft.

As for the assembly, tapping all those 1/4" holes is quite a job. Most people at some point revert to good quality flooring screws with a 1/8" guide hole as MDF would crack and split otherwise. At some places, I simply drilled thru and put a washer and nut in the back, MUCH more solid this way...

I'd recommend going with the JGRO plan as-is since I modified the original design and got into all kind of challenges to make it work...

Well, good luck, it seems all big and bloated but in the end it's not complicated, just complex a bit...


Para

Keith ,
As an instrument builder myself I would go a bit bigger than the 24x18 inch.
You could make tons of jigs and tooling for your projects ( dished out workboards , fox benders , etc etc ).
I still have to start building my own machine but It will have more or less the same purpose .(I'll put the photos on cnc zone once I start )
Good luck

Paul

I've spent a couple of weeks digesting the vast amount of info. contained on this forum. Trying to find the answers I need is kind of tough to do considering how much info is here. So I'm gonna throw some questions out there.

Digestion is good! Hopefully we can do somethig with the questions.


I've decided that I want to build a home cnc router. I'd like to use it for building wood parts for acoustic guitars. I'm thinking that if the table could travel 24 in the x and 18 in the Y that I could use is for just about any part.

Size is something you need to think long and hard about. I'm not sure what parts you intend to make with the router but you my want to seriously consider more than one machine. CNC-ing a small import mill might also be usefull. This of course depends on what sort of parts you intend to make.

In any event I think you are going to small for this machine. Do understand that you need room for clamping fixturing and other tooling.

The router will be a craftsmen router for starters. I have a pc that is running windows 2000 that I would like to dedicate to the project. It is running an amd 1.1 gigahertz processor with 256m of ram. I want to start with a machine that I will adjustt the Z axis manually. So.... just a two axis machine.

Sounds like a doable PC. I would like to suggest that you look at Linux/EMC/RTAI as an alternative. There seems to be mixed opinions about EMC but it is under constant develpment and people do use it for high speed machining of wood products. I would be concerned abotu the amount of RAM installed on the PC but that is not a big deal.

Manual Z is possible but I would certianly make a point to engineer in the possibility for a driven Z axis. For the interim you might want to consider a pneumatic driven axis that can clear the work piece under CNC control. Either way the incremental cost for adding a driven Z axis isn't all that great so do consider that in your plans. A driven Z will instantly make the machine more usefull to you.


I tried to download the emc brain dead install 4.6 a couple of times but the checksum was always wrong. When I started up off the cd it never worked right. I would always just end up at a \bash command prompt but not very much happening.

Interesting! BDI (Brain Dead Install) has just been released as a up date. Which is now version 4.23.

It sounds like you have trouble either with your internet connection or the software you are downloading it with. A program call "wget" is very useful for cdownloads on Linux and I believe it is available for windows. I find useing it over a dial up connection, that is questionable, to be very reliable. Use the -c option to restart failed connections.


I tried to download mach2 on windows 2000 but that got tripped up too and gave me the dreaded blue screen!

Sounds like you need to resolve your download problems before getting to serious about CNC software.


I've heard many people say great things about a dos based g code interpreter but then it became evident that it was neccesary to buy an external board to use the software.

There are a couple of these DOS based interpeters available but I'm not familiar with them. However I'm pretty sure that at least one of them can make use of the parallel port.


Then I started reading about emc and what was neccesary to make it work with controls. You either had to change the software or get some radio shack parallel port conversion kit.

You seem to have come into a bunch of disinformation here. EMC can be used with the parallel port also.

If possible I'd like to just plug the controller board into the parallel port and install the software in a windows environment. Is there anything out there like that? Is there anything like that for free?

I'm not sure what you are after here. I won't comment on Windows as I wouldn't reccomend that at all. What I can say is that many of the NC control software pacakges can drive hardware from the parallel port. You can also buy hardware that works with EMC that will enhance the capabilities of the software when using the parallel port.

As to hardware you do need to understand the nomenclature. The device that drives the motor is often refered to as a drive or amplifier. In the case of using the PC to generate step and direction signals the PC becomes the controller. I'd suggest looking at www.linuxcnc.org for more information about EMC specifically and PC CNC in general.


Maybe I should back up a little. For what I'm trying to do I'm guessng that a step motor is the thing to do. What would be a good control to use? What type of tolerance will I get with this type of a control and a decent stepper motor?

For motor "DRIVERS" look for Gecko Drives. Steppers would likely be fine for a CNC controlled wood router. As to tolerance many things come into play there. The very construction of your router will impact accuracy. In any event the stepper and the rest of the mechanics are variables. You could design a system that will resolve one ten thousands of an inch if you wanted to.


That leads to the next question. I have read about people using every different kind of stepper. What kind of torque do I need to move this size table? What brand or type of motor do you suggest using and why?

This is an engineering problem and depends on how fast you want to go and how strong you router will be. In general though a stepper will be successful on this size machine.


Do ball screws work well on a machine that cuts wood? Can software compensate for backlash in acme screws? What is the best price to performance ratio for linear guiding a small cnc router?

Ball screws will work if you can keep them clean. That is a big if!

Think about alternatives including rack and pinion or belt drives.

For linear guiding on a wood router you have problems similar to ball screw if you should use linear bearings with balls. That is you run the risk of things getting gunked up, though linear guides may work better if they have full seals. As for trade offs that again is an engineering issue.


Controllers are the next thing that I don't understand. I read much about gecko.

Gecko is a drive not a controller.

They look like nice packages but it seems like they only run one axis and cost about $120. Xylotec looks like a pretty commonly used controller bu t how precise is it and does it work seamlessly with any of the gcode interpretters.?

Again you are talking about a drive not a controller. Both of these work with step and direction signals put out by popular software. The software and PC being the controller. Obviously for the difference in price you are betting a lot more out of the Gecko as a drive.


Last question....Do major cam packages like surfcam, mastercam, or gibbs have output for controller like xylotec or gecko. I know they do for controls like fanuch or haas but do they for the home cnc controls??

If you can get standard G-code out of the packages EMC should be able to run the software. There is one consideration though they will expect your Z axis to work.


Whatever info you could provide would be verry much appreciated. Thanks -Keith

I hope that some of the above helps. You really should read up a bit on motion control. Sadly I could see you spending a lot of time and money and not being happy with the results if you don't understand what you are getting into. A few more weeks on this forum will help as will the Yahoo groups dealing with CNC.

As for you machine:

Before you start do nail down the size that you think you will need and design around that. Better yet nail the size down and then increase the machines size by 20 to 50%. Going to small is a path to unhappiness. To big isn't good either but should be less harmfull than a machine that can't do what you need done.
Do install all three axis as driven axises. That is the little extra it costs to build in a driven axis for Z is well worth it. It will become a requirement if you expect to use cam developed G-Code.
Take some time to learn about the electronics behind a CNC machine. Download and read all of the documentation on Geckos web site and their Yahoo support sight.
Likewise look into the EMC site and read all the info contianed there.

I've got one thing straight now. The drive is what actually sends power to the motors. The controller is the pc running emc. I will try to do the bdi thing again. I do have a lousy connection at work where I was attempting to download the bdi. Afterwards I checked the mdsum and it was biffed. So that was probably part of the blue screen deal. I have a pretty reliable connection at home (dsl) but I don't have a burner in the pc currently.
I would like to get emc working though. I have always thought that the concept of linux was great but it just never really jived once I loadeed and tried it. I've tried to use redhat linux a bunch of times over the past 5 years.

One thing that might make useing emc a little bit laborious will be making the models and g code on the windows side then bringing it into emc. Anyone familiar with doing that?? Do you need to restart the operating systems everytime you need to make a change to your model, generate the code, back out of windows, boot linux and somehow bring it into emc??

So it seems like the gecko drives are a good bet. How much will it cost to have gecko drives for all 3 axis if I'm using steppers at a modest torque for what I'm doing??

I am sort of turned off on the ball screw configuration. Belts sound like a good way since it would be more impervious to dust and chips. Where do I read more about this type of a system? Are there companies that sell kits for doing this?

That for all your help. I will continue to read.

Keith,
Geckos are $134 each. You don't need a breakout board. You can just solder the wires to a parallel port connector from radio shack. Seen some nice 280 oz/in nema 23's for $150 for 3 on ebay the other day. Guy said he had 8 sets I think. Do get shielded cabeling for the wires. Cat 6 shielded is good and cheap.
You mentioned belts are better for chips than screws. After tonight I have to disagree. I just got done cutting a 15in dia 25in long vase on my 4th axis out of foam. I forgot to cover the belt and pulleys and the belt got clogged with foam and stalled my 2700 oz/in motor. That foam gets everywhere. I had a rack on my X axis but just pulled it off and replaced it with a screw. Much better. The foam would clog the rack too and it faced down not up.
Have fun with your project. There is a ton of help here at the Zone.

Donny

www.whiterivermfg.com

I've got one thing straight now. The drive is what actually sends power to the motors. The controller is the pc running emc. I will try to do the bdi thing again. I do have a lousy connection at work where I was attempting to download the bdi. Afterwards I checked the mdsum and it was biffed. So that was probably part of the blue screen deal. I have a pretty reliable connection at home (dsl) but I don't have a burner in the pc currently.

Well there is a good way to look at this. At home I have a dial up connection and was able to download bdi-4.23 over a period of a couple of days. Not bad but before I got it burned to disk what should happen but bdi-4,24, comes out. That fixes a couple of issues so I just finished downloading 4.24. Now I have to go out and buy some blank CD's.

By the way a CD burner is a good investment and you can get by with a dirt cheap one.

I would like to get emc working though. I have always thought that the concept of linux was great but it just never really jived once I loadeed and tried it. I've tried to use redhat linux a bunch of times over the past 5 years.

Linux is very interesting and at this point I've given up on MS software on my machines. It does have a leanring curve but the college I went to introduced me to UNIX on sun systems like 15 years ago, so I'm biased.

If you find making the jump difficult you can look at EMC/Linux as a machine specific thing.


One thing that might make useing emc a little bit laborious will be making the models and g code on the windows side then bringing it into emc. Anyone familiar with doing that?? Do you need to restart the operating systems everytime you need to make a change to your model, generate the code, back out of windows, boot linux and somehow bring it into emc??

What one does is to set up a partition that is shared across both operating systems. Store your data there. You may get lucky though and find that you can run some of your software on Linux under WINE. Wine emulates (sort of) a Windows environment. The problem is results are highly mixed.


So it seems like the gecko drives are a good bet. How much will it cost to have gecko drives for all 3 axis if I'm using steppers at a modest torque for what I'm doing??

As far as I know there is only one size Gecko drive and you have to configure it to the current that the motor demands. One thing to point out here is that the drives require a DC power supply which may be an issue. So it is not just a matter of buying the drives, you also need a properly sized powersupply.

As for cost contact Gecko and see what they will quote you.


I am sort of turned off on the ball screw configuration. Belts sound like a good way since it would be more impervious to dust and chips. Where do I read more about this type of a system? Are there companies that sell kits for doing this?

I think it would be a mistake to call belts imperious to dust and chips. Just about all motion producing hardware will have trouble with that. Some may argue this but my point is that it is easier to guard and maintain a belt drivne system in a dust environment than it is a ball screw. Ballscrews may start out looking good but you will have gunk build up in them and the required lubrication just supplies a tacky surface for said gunk to attrack to. Note that it is likely that some ball screws are better that others with this respect and some provide excellent provisions for lubrication. Further ball screws can end up being very expensive relative to the rest of the machine.

There are not absolutes here either. At work (a very clean environment) we have a number of robots that use a combination of belts and lead screws to move hardware very vast. That sort of motion is not really appropriate for machining though. Do realize that belts have there own design requirements, you really can't ignore the engineering.

One thing I saw at a recent model engineering show is the use of chain on an axis. Something to consider but I'm not sure you would get the results you want. Others may want to comment on this technology if they have used it. One nice thing about chains is that they do hold up in harsh environments.


That for all your help. I will continue to read.

You will find all sorts of ideas in thes forums. Further there is much to be learned on the net. You may find that ballscrews may end up being the only thing that you are comfortable with due to the precision you want. Just realize that the ball screws are only as good as the surrounding hardware.

Dave