The journey toward CNC

[wareagle]

The CNC option intrigues me. I see many instances where it can be an effective tool to aid in making things in the shop. In other words, I think the benefits outweigh the costs for my needs and purposes.

In addition, I am the type that is a hands on learner. Putting together a CNC system and converting a machine sounds challenging and fun. To that end, I will probably try to find a machine (likely a small to mid sized mill) to convert to CNC when I get down the road a little further. Unless I find a screaming hot deal on a ready to go CNC machine. For now, I am basically on a fact finding mission and am a long way from doing anything with hardware.


Computer controller & CNC software
The first step in my journey is to get a CNC program and start playing with it. I happened to have an old tower PC that I was about to rip apart and discard. It has the parallel port, runs Windows XP and meets the minimum requirements for Mach3. Yesterday, I downloaded and installed Mach3 on the machine and the has been a success.

Before I do any kind of interface with CNC hardware, I will reformat the hard drive (or buy a new one), reload the OS and then load Mach3 to avoid some of the issues I have read about with other programs eating resources and such. For now, Mach3 runs seemingly fine. I ran a wizard and was able to watch it simulate a run. So considering this step one, I guess that has been checked off of the list.


Screws
So, my first question on doing a CNC conversion is regarding the X, Y, and Z screws. For whatever reason, I haven't found much info (although I haven't spent a great deal of time searching, either) on ballscrews versus the regular screws for the axis movement. I see some of the sources for ballscrews, but am curious if ball screws are an absolute necessity, or if guys are getting along with the regular screws.

Could one feasibly fashion a second nut for the regular screw to take the backlash away, or is that more or less chasing your tail? In my online travels, I see where some are utilizing two ball nuts to reduce backlash, but it looks like (from what I can determine) those are machines set up for production.

Axis motors
Looks like there are lots of choices here. The sizing is no big deal. However, what are the pros and con between a stepper motor and a servo motor in regards to the CNC application? Again, I haven't spent a huge amount of time searching, but have so far come up empty.

Also regarding motors, are people oversizing their motors by a fairly wide margin for safety? Or is it a bad choice to do so?

Drives
The motor drives don't seem to be an overly complicated deal. The only caveat being that the drive must meet the requirement of the motor.

Breakout boards
So far, I don't see anything there that seems to be voodoo here. Just the interface between the computer and the drives.

Power supply
Again, not anything here that appears to be mystical. Are guys oversizing their power supplies by a fairly wide margin for safety?


Am I missing a key component to consider when looking at the system as a whole? Sorry if some of these things seem to have been have been answered to death. I admit that I haven't devoted enough time to seeking out information and could probably answer a lot of this stuff with some time looking around online.

Thanks in advance!

 

[jpeter]

I've done 2 conversions. The first one was done on a Grizzly mill at least 10 years ago. I used steppers. I was pretty green to cnc at the time. I didn't realize the advantages of climb cutting or how often cnc needs to climb cut. I tried preloading the acme screws with weights and pulley. Nothing worked. I converted the x and y to ball screws using McMaster Carr 1/2 or 5/8 ball screws. The z axis stepper was connected to the downfeed and worked ok. That mill did a lot of work. Several engines were built on it. The cheap ball screws worked only 100 times better than the acme screws. The backlash in the nuts was maybe 0.003. My next project was converting a Bridgeport tool room mill, 36 inch table. I bought a kit which contained preloaded nuts. They consisted of a double nut with a strong spring between the two, kinda like a belleville washer. That mill is the one I use today. For me its 10 times better than the Grizzly. Never-the-less the Grizzly did work. You'll find you won't be able to deal with backlash. You'll have to find a way to eliminate it.
Initally I wanted to be able to manual mill too. I've since found I never manual milled anything. When I manually mill I just jog feed or manually enter code in the midi screen of mach 3.
All my electronic conversion stuff came from Dan Mauch at WWW.camtonics-cnc.com. Dan helped me get the right stuff, its been working for a lot of years and he's still in business. Speaking of stuff, I think he knows his stuff. Good luck

 

[RonGinger]

Ive done several conversions and now also have a KX1 ready to run machine from Little Machine Shop. You seem to be onto the main issues.

Steppers or servos are what I call "holy war #1" there are probably a zillion pages of discussion on it in the various forums. I believe the following points are true:

Steppers will always cost less. At the least a servo requires an encoder and more complicated electronics.

A stepper motor, when used within its load rating will NEVER loose a step. There are thousands of machines in use every day for many years that make perfect parts with steppers.

For machine smaller than a Bridgeport steppers are completely adequate for the speeds and loads needed. For bigger machines steppers wont do it, servos are the only way. Putting servos on a little mill is a foolish waste of money and needless complexity.

I have been working on a plan for a CNC conversion project that might be of interest. We are going to do a group buy of 10 machines, the new High torque mill from Little Machine Shop, a ballscrew kit from CNC fusion, steppers from Keiling and Gecko drives. We will hold a class at the CNC workshop in June to make these conversions.

This will be a chance for 10 guys to get a machine for less than the street price of the parts, to do the conversion under the help of several experts (including Brian Barker of Mach) and attend a week of seminars on CNC.

The details should be released early next week by George Bulliss of the HOME SHOP MACHINIST magazine. Ill post a topic here when I have the details.

 

[stevehuckss396]

Screws:

You can use lead screws if the backlash is minimum (1 or 2 thou at the most). Mach can compensate for lash but it's best to get out as much as you can mechanically. Lead screws are inefficiant so the size if the motors will need to be bigger than a machine with ball screws.

You can buy the cheap McMaster-Carr ball screw and nuts. Dump out all the balls in the nut and order some high quality balls from VXB or another bearing supply. I had less than 1 thou lash in all 3 axis.


Motors:

If you are building or retrofitting a small mill, you would want to use some steppers. Keiling has some guide lines on there site for motor selection.

Drives:

Can't beat Gecko but some of the hobby boards are good. Depends how fast you want the pieces done and what you are willing to spend. My mill has some 300oz motors and a hobby board and it works well. I have all the stuff now to upgrade to Geckos with 575 oz motors. I need some more speed.

Power supply:

If you use Gecko you will want to size the power supply and voltage by the motors inductance. I have some formulas for that so if you get motors selected I can get some info from you and supply some numbers. You can never excede the max voltage of your controller.

 

[jpeter]

I have 600 oz-in servos on my Bridgeport. Got 2:1 reduction on x and y and 5:1 on the knee for z. I use 300 pound nitrogen springs to help lift the knee. BTW, I added gas springs to a manual BP too. Sure saved a lot of effort when cranking up the knee. The drive is an old Camtronics servo drive. I'd use Geckos' if I was doing it today. I like the servos cuz they're quiet. Had steppers on the Grizzly. You could hear it running all over the shop. Otherwise, I don't think it matters on my small Bridgeport. My rapids are about 26 inches per minute on x and y and 9 inches per minute on z. Its not fast but I don't do production either. If I was to do it again I go with 1000 or 1200 oz-in servos that way I could wreck more stuff. I've got enough power to wreck cutters as big as the BP can take. I'd really like to get the z rapid up cuz sometimes I need large z moves like for example when I chuck up a reamer. I wouldn't plan an doing much manual milling on the mill after installing ball screws because the slides move to easy to stay in place. For example, if your cranking x, y will move on you if you don't lock it or hold it or something.

 

[wareagle]

Jim, thanks for the info. Great food for thought!

Ron, that program sounds like an awesome opportunity. It probably won't work out for me, but when it is announced I'll take a good close look at jumping in.

Thanks for the info on the differences between the stepper and servo motors. I hadn't gone into the hardware requirements (or differences) for the two, but that sheds much light on the subject.

Steve, I appreciate the insight on the screws and drives. The inefficiency of the regular screws makes perfect sense. As I get deeper into this, I will probably take you up on the offer for help in sizing the power supply. For now, I am just looking at the abyss... As soon as I can gather up some cash, look out!

 

[Tin Falcon]

amat victoria curam
Victory likes careful preparation. (Anonymous)

Wareagle:

I have several threads going about my cnc builds and some on general CNC info.
There are lots of choices. Do your homework.

You are off to a good start by exploring Mach 3

The gecko G 540 seems to be a reliable good value almost plug and play controller. and gecko has made in USA motors to go with it not too pricey at $79 each.

IIRC you have an electrical /electronics background and can probably build a good power supply.
others have discussed ball screws. I have yet to add them to my X or Y but likely will.
exploring CAM may also be helpful in your success later on.
Tin

 

[shred]

My little CNC Taig got by with the regular screws (not even acme, just the stock v-thread lead screws). Had to do some backlash comp in the SW, but the Taig itself is flexible enough that it didn't matter much. That machine had small servos on it. It's nice to have it stop when some bad things happen (not all) since the servo would fault, but otherwise there wasn't much difference. My Tormach has steppers (and ball-screws) and doesn't lose steps, so as Ron says, sized correctly it's not likely to matter enough to care about, though you can get into all kinds of holy wars around them.

Mach is good; almost ubiquitous in hobby-land. Also think about CAD-CAM. Lots of $ can be spent there.

 

[MachineTom]

Very interesting thread. Am I correct in assuming that these retro-fits do not use DRO scales, just an offset for backlash in each axis. Is the milling bi-directional for a given cut? or single directional. Seems like the positioning could grow into a large error, without scales.

My BP has ball screws as it was CNC at one time, but still has .001+backlash in spots. Help me understand the position control.

 

[jpeter]

You can manage 0.001 backlash but not much more. It's not all about position. Position can be dealt with by the backlask compensation in the driver software. It's more about being able to climb cut, when the cutter pulls the table along causing the cutter to "hog in." The result of that is a broken cutter and ruined work. Not only can't climb cutting be avoided in CNC but it's really the only good way to get long tool life and good finish. So basically you need to be able to climb cut and you can't climb cut if you have backlash.

 

[mu38&Bg#]

There are many CNC out there running with backlash. You can avoid climb cutting. Still, it's not the way I would do it. Preloaded ball screws are definitely the way to go.

Position feed back can be none as in a basic stepper setup. It relies on counting steps and setting things such that the torque required to move never exceeds the holding torque of the servo. Then you have steppers with encoders to make sure you haven't lost steps, but no real position feedback. It relies on setting acceleration and speeds so that commanded and actual position do not diverge much if at all. These can be stepper or servo motor driven. Finally you have real time feedback systems just like you see in industrial CNC. EMC2 with a motion control card and scales on the axis slides will do this. Even this is not optimal as the ball screw has error. You can add a lead screw compensation table to help with accuracy in some systems, EMC2 I know. Any system using scales on the axis will also have feedback on the motor in the form of tach or encoder.

I like servos. My setup is Gecko 320's, dc servo motors with encoders, and preloaded ball screws. Maybe one, day I'll look at scales for the axes, but for the work I do now this is very accurate. If you understand motor ratings its easy to chase down motors cheap on Ebay. Good ballscrews will be expensive.

 

[Tin Falcon]

Very interesting thread. Am I correct in assuming that these retro-fits do not use DRO scales, just an offset for backlash in each axis. Is the milling bi-directional for a given cut? or single directional. Seems like the positioning could grow into a large error, without scales.

My BP has ball screws as it was CNC at one time, but still has .001+backlash in spots. Help me understand the position control.

These retrofits can use a dro if you have one the dro can be set up ad separate inputs to mach 3.
Mach 3 has dros built into the system but the numbers are bases on output to the motors rather than real position.
If I had a BP with ball screws in good condition I would use them. I do not think a thou or two will matter for what we do.
You can always redo the bs later is really needed.
Tin

 

[don-tucker]

Just to give you guys a laugh this is my forray int CNC,needless to say that is as far as it got.

:-[
Don

 

[Tin Falcon]

Don no laughter here. CNC can be a long road when you factor in either saving up money or buying parts when you have enough for a piece or two. Time available for hobby stuff with home responsibilities even retired folks are busy with life these days. A nc project is sometimes like owning a hose and fixing up never finished, push through and make some chips my man.
Tin