Q: What size motor do I need?
Q: Is a higher resolution encoder better?
Q: Where can I buy ballscrews from?
Q: I'm trying to decide what machine to convert..
Q: What is a realistic goal for a CNC conversion?
Q: How much will it cost?
Q: What is the difference between servos and steppers?
Q: What drive should I use for 'x' motor?
Q: What size power supply should I use for 'x' servo motor?
Q: What do you recommend for drives and controls?
Q: Do you have any recommendations for surplus electronic parts?
Q: Where can I buy pulleys and timing belts?
A: Without getting into a lot of calculations, the basic answer is, for a Sherline/Taig or mini-mill, I would recommend a 320oz/in peak motor, for a medium size mill/drill - something in the range of an RF30, I would recommend a 400-600oz/in peak output motor. For a large mill/drill, such as an RF45 up to a full-size Bridgeport I would recommend 750oz/in or better (you can go 1:1 on the RF45 - I'd go 2:1 with the Bridgeport).
Keep in mind that gearing will effect your motor calculation. For example, let's say you have a 100oz/in motor with a nominal speed rating of 1,000rpm. At 1:1 gearing with a 0.200" leadscrew, you will have an effective rapid speed of 200ipm. If you gear it down 2:1, you will increase your torque at the leadscrew to 200oz/in, but your rapid speed will 'only' be 100ipm. The two areas where motor torque will make the most difference are in acceleration and rapids.
A: Maybe. There are several factors in choosing an encoder: pulse rate and resolution. With the currently available PC-based CNC controllers, you are limited to a clock rate of 25khz to 45khz (Mach2). With a 250cpr encoder, you will need to generate 1,000 pulses (with quadrature) to make one full revolution. Therefore, with a 25khz clock rate, your maximum rotational speed would be 1,500rpm. With a 45khz clock rate, your maximum rotational speed would be 2,700rpm. If you go to a 500cpr encoder, your resolution will be better, but your maximum rpm will be 1/2 what you'd get with a 250cpr encoder.
For example: If you have a motor geared 2:1 with a 0.200" leadscrew, assuming your motor has enough rpm, with a 250cpr encoder and a 25khz clock rate, your resolution would be 0.0001" and your maximum rapid speed would be 150ipm. That same motor, with a 500cpr encoder would get you 0.00005" resolution, but your rapid speed would drop to 75ipm. At 45khz clock rate, you'd have 270ipm and 135ipm respectively.
Ideally you want to spec your parts so they all reach their limits at the same time. For example, if you have a motor with 1,000rpm maximum speed, and a 25khz clock rate, your ideal encoder would have 375 cpr (1,500ppr with quadrature). However, in the real world you'd probably want to use a 250cpr encoder in this example.
Also, keep in mind that just because you have great resolution doesn't mean you can actually use it. By the time you add up all the inconsistencies and variables in your setup, your encoder resolution is probably the least of your problems!
A: I recommend Rockford Ballscrew, although there are plenty of other places also. Rockford Ballscrew sells rolled ballscrews which are generally used for power transmission but which are probably fine for most hobby-level CNC work and are significantly cheaper than ground ballscrews. They also sell conversion kits for full-size knee mills starting at $650 including X and Y ballscrews, replacement Yoke and pre-loaded ballnuts. For other uses, they will machine ends of raw ballscrew stock for $60 per end to your specs which is very nice if you don't have a lathe. Call them directly for other pricing.
A: I would recommend either an RF45 mill/drill or a full-size knee mill. Out of the RF45's I'd recommend WTTool's version. I have personally seen/used the RF45 from Enco, Grizzly and WTTool. WTTool's wins hands down in fit/finish. If you're going to go to the trouble of converting a mill/drill to CNC, I wouldn't bother with a round column mill as you lose your zero when moving the head. I would also CNC the head, rather than the quill of an RF45.
For a full size machine, Bridgeports, like Southbend lathes, have, in my opinion, an undeserved mystique. Worn out Bridgeports often sell for more than a good condition clone. Additionally, a current model clone with square ways and Meehanite casting is going to be a much better machine than even a rebuilt Bridgeport. Bottom line - condition of the machine is much more important than the brand.
If possible, I would also get a machine with a 3 phase motor and get a VFD/Inverter which will give you greater rpm range. One of the problems with the gear-head mill/drills is the maximum spindle speed is only 1,500rpm. With the VFD you can easily get 2,200rpm or better, depending on how comfortable you are with over-driving the motor. On my Comet, I run the VFD between 30hz and 90hz, which translates to 50% to 150% of nameplate rpm.
My last comment on this - when I got my first milling machine (an RF40 mill/drill) I had never used a milling machine before. So I was pretty happy with it. Yes it had various drawbacks and I had to take fairly light cuts to get decent surface finishes but I didn't know any better! Rigidity is king - the bigger the machine the more rigid it will be. For comparison, with my RF40 if I pulled on the head with an indicator chucked and touching the table, I could get a 0.015" reading without pulling very hard. With the Comet, pulling extremely hard, I can get a 0.004" reading. That is not to say that you can't get very good output with even a small mill/drill - you can, but it will be much easier to do so with a larger machine. My recommendation is to get as large a machine as you can afford/have space for.
On the other hand, there are probably a zillion lathes and milling machines sitting in someone's basement because they discovered, after they got it, that they really weren't as interested as they thought they were and it is too much hassle to get the machine out.. 'course, if you're seriously considering converting a machine to CNC, it is probably too late for you anyway - your wife already thinks you're nuts, you already took the second mortgage on the house and you have permanent metal splinters in all of your fingers...
Your CNC machine will not be able to exceed the machining rates of the same machine under manual control. So, if you can't take huge cuts in steel with your manual machine, don't expect that converting it to CNC will make it possible. If you have a baby mill/drill that is capable of cutting delrin and not much else, after the CNC conversion you will still have a baby mill/drill.
Additionally, since I'm not using my CNC machine for production, I very rarely exceed 60ipm rapids. While it is nice knowing I have the capability for higher rapids, I just don't use it very much. Of course, most of the parts that I work with are less than 12" long, so I don't have very far to rapid.
In general, servos have lower continuous torque but have the same amount of torque throughout their rpm range. Whereas steppers only have their rated torque at very low rpm, and will run out of torque fairly quickly, which means that you'll need a stepper with much higher static torque range in comparison to a servo. Steppers can hold their rated torque all day while servos are generally rated in continuous and peak - peak torque will be anywhere from 2 to 5 times continuous torque.
It used to be that servos were much more complicated to setup than steppers. With step-and-direction controls, setting up a servo is comparable to settings up a stepper. You do need to have an encoder (which means sizing the encoder) but you don't have to worry so much about torque curves.
In theory, a step-and-direction servo system should be more accurate than a stepper system since there is a closed loop between the drive and the motor. In practice, a properly sized stepper system will operate very similarly to a servo system. Please note - I am not referring to true closed loop systems here, with feedback to the controller. With a stepper lost steps are invisible, whereas with a servo system if there are enough lost steps the drive should fault.
In event of drive/control failure, a stepper will simply stop moving. A servo will probably go running at full speed until it hits something hard which can be pretty unpleasant especially with the forces involved in a typical CNC setup.
A: For a typical Bridgeport size knee-mill, you're looking at:
Ballscrews for 3 axis @ $900
Angular Contact Bearings for 3 axis @ $150
Timing Belts/Pulleys for 3 axis @ $185
Gecko G320's for 3 axis @ $360
Power Supply @ $250
Servos with Encoders @ $500
Controls and Wiring @ $200 (could easily be $400 or more depending on how many extras you put in - such as Alan Rothenbush's GERC and/or Robert Campbell's breakout board. Also, you could easily spend $200/axis just on motion rated cabling/wires but you don't *have* to.)
Misc Hardware, Material, Screws..etc @ $300
Costs will be a few hundred dollars less for a Mill/Drill.
A: For a brushed D.C. motor, matching the drive to the motor is basically as simple as finding a drive that can supply sufficient voltage/amperage to satisfy the motor. Keep in mind that you do not necessarily need to supply the max voltage/current to the motor - you just need enough to satisfy YOUR requirements. For brushless or A/C motors things are a bit more complicated and beyond this scope of this FAQ.
A: Since most unregulated power supplies can provide greater than rated amperage for short periods of time and since you will probably very rarely use peak amperage on your motors, and you rarely have more than one axis running at full power, I would look for a power supply that can supply 2x full, continuous power for axis. So, if you have a motor that can draw 10 amps continuous, I would look for a 20 amp power supply (assuming a 3-4 axis system). Even this will probably be larger than you need. It is relatively easy to upgrade the power supply at a later date if you have to.
A: I've been very happy with the sales and support from Geckodrive and use their G320's in my Comet Conversion. I've also heard lots of nice things about Rutex but I've never used their products myself. For PC controls, I really like Mach2 from Art Fenerty which runs under Windoze. For Linux, there's EMC.
A: I've been very happy with Mendelson Electronics and All Electronics. For new parts I generally use Mouser.
A: My favorite places for hardware is McMaster-Carr and Sterling Drive Products.
Return to the Virtual Smidgescape Return to NC Gear Return to the Machineshop Return to the 928 Pages