View Full Version : Stepper Motors
02-25-2005, 02:00 PM
Any one out there know of a resonably priced source for stepper motors. Specificaly looking at a NEMA 42 with at least 2100 oz/in torque. I've finally decided to go the DIY route from the HSM. Any help on locating new reasonable priced motors would be greatly appreciated. I'm not yet 2 years into this project and I'm getting anxious to make some chips!!
Roger from Montana
02-25-2005, 02:07 PM
First surplus sites off the top of my head.
New steppers can be had here. Good company, fast service.
02-26-2005, 04:12 AM
2100 oz/in... hmmm... kinda large. Are you sure you don't want to use servos? At this end of the spectrum, steppers get quite expensive, power hungry, and (due to the typical design requirements) tend to be a bit slow - although it is possible to get a stepper motor with high rpm capibility.
Have a look at:
I am currently in the final stages of modernising my Shizouka AN-S. When I got it, it was equipped with a "Retro-Tek CNC-88" (no, I'm not kidding, that was it's name!) from the late 1970s - complete with a genuine 8080 processor and a punched paper tape reader! I have stripped the machine down, removed all the electronics and wiring, and built a new control consisting of a mini-itx motherboard running TurboCNC (and soon, my own control software), Rutex servo drives, a Hitachi SJ-100 vector drive for the spindle motor, several power supplys, and a relay card for extra M-Codes. The old control took up two cabnents (one 14x22 box on the mill, and one free-standing, about 6-foot heigh by 3 by 3 feet), and the new control fits into a single 16x20 box on the side of the mill (in retrospect, however, I probbably could have put it all inside the column).
Other surplus sources:
As you asked for new, I have read that 90 volt Leeson DC motors work OK for bridgeport conversions, but I have not actually tried this myself.
I might be able to be more helpfull if I knew what , exactly, you are retrofitting, and what components you already have?
02-26-2005, 11:40 AM
Im am fitting a 1985 Lagun 310 and when I bought the unit the only electronics it had was the spindle motor. I talked with an Ah ha rep and the system he quoted me was $7000+. That spurred the DIY route. Is there any thing else you need to know?
Roger from Montana
02-26-2005, 06:52 PM
The only info I found on that machine with a quick google search indicates that it's a cnc knee mill with a quill Z axis. As you say it has only the spindle motor, I presume that it was stripped of it's control. So:
What kind of feed rates and table load are you looking for?
A ballpark guess on my part might be 120ipm rapids, and 1300lb combined table & workpiece weight, giving a requirement of about 2/5hp - the closest standard motor size being 1/2hp. The quill motor can probbably be 1/2 that, so 1/4hp. Enco has new Leeson DC motors for $200 for 1/4hp, and $220 for 1/2hp (these need an encoder added to them; for about $50 each). If you are willing to try surplus, you can probbably come up with suitable motors for much less.
What HP and voltage spindle motor do you have? Do you have 3-phase power available (if so, what voltage), or do you need a phase convertor or VFD? Or do you want to put a VFD on anyway for speed control? What is the transmission setup (step-pulley, gear, or variable-speed-belt) and how is it actuated?
Did it have an automatic tool changer? If so, how is it actuated?
Are the ballscrews still in good shape? How do they connect to the motors (shaft coupler, timing belt, etc)?
02-26-2005, 10:30 PM
A breif history of what has been done to the machine since I bought it. Stripped it completely down general clean up, all new bearings where needed,new turcite installed on the wear surfaces(quite a challenge to obtain .0001's better than the factory specs listed in the machine file). At his point i'm ready for a control system.
the spindle motor is 5hp 220 3ph which comes
down the alley and to my shop. No problem getting the correct power requirements. The tranny is high /low with a varible speed belt
having a manual control. It was originally equiped with a tool changer and power drawbar,of which both were gone when I bought the machine.
The z & y axis ball screws are in perfect shape, x axis has minimal backlash.I may end up replacing the x ball screw but I will try it before making that decision. All the screws are belt driven.
If I understand you correctly I would only need two large torque motors x&y and a smaller motor for z ? You are suggesting a servo system rather than a stepper. Why? I am willing to spend the $$ for stepper just not $7000 for a complete system. At present I'm looking at less than 1/2 that amount. I can live with that.
Feed rates you mentioned will be fine as this will not be a production machine. Weights? Well I'm not sure but would lean towards the max rather than light.
Thanks for the response !!
Roger from Montana
02-27-2005, 09:05 AM
$7,000(?) wanna buy a complete mill with servos and centroid control? It is a INdex 3hp model much better then the Bridgeport I have at home.
They are out there, you jus gotta go fishin.
02-27-2005, 09:27 AM
If you need that much torque your into servo drives. You can get 2000 ln lb easy enough on a stepper drive but unless you get into 170 volt high performance drives the necessary reduction makes the axis slow and sluggish. 2000+ in lb works out to about 1/4 HP when reduced to 500 lead screw RPM.
Servos don't need to be expensive and their high speed performance is hard to beat. Rutex, Gecko, and a few other make excellent servo amps capable of driving a 3/4 HP motor for less than $150 an axis. The work from step and direction signals like step motor drives do.
You can make a servo motor out of any permanent mag DC motor by mounting an encoder from US digital on it. The motors are where you find them and good scroungers can find them for $20 - $100 at industrial salvage yards, eBay, etc. The encoder cost about $40 to $80 depending on resolution, whistles, and bells.
There only remains the DC supply for the axis drive amps. Line voltage bucked down to 75 Volts, a full wave bridge rectifier and 2000 MFD per amp is about all the complexity you need.
02-27-2005, 09:33 AM
Bin fishin - caught one, need to finish the project I'm on 1st. Got too much time and effort into this one to abandon now. Besides I only need one good machine. The rest boils down to using the old imagination, right!!
Thanks for the insight though, Dave and BTW I get a kick out of reading your responses. A good source of humor for me. Just gotta love this site very informative and relaxing!
Roger from Montana
02-27-2005, 04:17 PM
Some answers to some of your questions, albeit somewhat out of order (also, I think you may know some of this info, but I have posted it anyway for the benefit of others)
>> Thanks for the response !!
No problem! http://bbs.homeshopmachinist.net//smile.gif
>> You are suggesting a servo system rather than a stepper. Why?
Steppers and Servos both have advantages and disadvantages, but not all of them apply to us.
No PID tuning required.
Low cost drive electronics possible in simple cases (such as unipolar, non-current-limited systems).
High torque for a given size at the small end of the spectrum, without gearing.
Torque falls off dramatically as RPM rises; limiting the maximum RPM that the motor can go to, more so than a similar servo (this effect can be combated somewhat with bipolar chopper drives, but the servo still usually wins speed-wise.)
If the motor is asked to provide more torque than it can, it will stall. Most stepper systems have no position
feedback, and when the motor stalls, positioning is lost.
Typical stepper motors setups are running at close to full power even at a stand-still, often resulting in large power requirements.
Stepper systems can exhibit resonance bands where the torque available drops off (sometimes to zero), causing the motor to stall at specific RPMs.
Often more expensive than comparable servo systems, at the large end of the power spectrum.
Extremely fine resolutions are not as available in steppers as they are in servos (typical stepper resolutions are 400 to 800 steps per rev, with 3200 steps commonly available through micro-stepping drives. A 2500cpr encoder for a servo will give 10000 steps per rev.)
The feedback system ensures that if the servo is momentarily overloaded (this usually occurs during acceleration), after the overload condition dissipates the servo will return to the commanded position; thus preventing a local blip in the torque requirement from throwing off the remaining moves, or completely stalling the system.
Often has a flatter torque-speed curve than a comparable stepper, allowing higher usable RPMs.
Higher typical resolution then stepper systems. (4000 to 10000 steps per rev is common, and I have a couple of 4000cpr encoders around here somewhere - those would give 16000 steps per rev.)
Typically smaller and lighter than comparable stepper systems.
Idling power consumption can approach zero with a modern switching servo drive (as the only thing that draws power when it's unloaded and not moving is the microprocessor, and encoder)
Requires PID tuning.
Improperly tuned servo systems can oscillate, sometimes violently. There are, however, tuning procedures that can just about eliminate this.
At the small end of the power spectrum, it's more complicated and expensive to implement than a stepper system.
Braking requires current flow through the motor, and correction by the servo drive, even though it's not moving (again, complexity for the person designing the drive).
Current usually flows through the rotor (DC brushed motors) and can be a cooling issue if the motor is having to dissipate significant power while moving slowly.
As I said above, not all of this applies to the current situation. We are not designing a drive system; but rather are getting one off-the-shelf. The motors that I am suggesting to use are designed to be run under load at slow speeds. So it boils down to power-per-dollar, speed, resolution, and what I feel is the most important, feedback -- and at this end of the spectrum, these factors usually come in on the side of the servo. (Oh, and a consideration that might be important for some applications is that servos are usually much quieter than steppers - although you said that it was in your shop, so you probably don't have other housemates yelling at you to "turn that doggone noisy thing off!!" )
>> I am willing to spend the $$ for stepper
Based on both experience and collected readings I feel that for what you are doing, you will get better performance out of a servo system, even though the initial setup is a little bit more involved.
>> Weights? Well I'm not sure but would lean towards the max rather than light.
Weight affects power required; my earlier ballpark calculations were somewhat conservative (i.e, overbuilt), so 1/2hp X and Y, with a 1/4hp Z should be fine. However, if you really wanted to overbuild it, either of the basic Rutex or Gecko servo drives can put out up to 2 hp, and the Rutex drives can be had up to 10.7 hp max power.
>> If I understand you correctly I would only need two large torque motors x&y and a smaller motor for z ?
Close - not torque, but power; as you can gear them down to match the torque/rpm requirements -- or in this case, use different pulley ratios. However, for any given rotor moment of inertia, higher-torque/lower-speed motors will be more responsive than lower-torque/higher-speed motors. The Z axis is not nearly as heavy, and thus needs less power to move it (also, you usually don't have much in the way of side loads on the system when it's moving vertically).
>> quite a challenge to obtain .0001's better than the factory specs listed in the machine file
>> No problem getting the correct power requirements.
>> The tranny is high /low with a varible speed belt having a manual control
This simplifies matters somewhat - you don't have to implement a control system for that functionality.
If you have any surplus dealers near you, I would go look for some good DC brushed motors of about 60 to 100 volts nominal, 1/4hp (185 watts) and 1/2hp (370 watts). I would also check out the surplus dealers websites listed above. If you can find some suitable motors cheap, I would try them first; but for used I wouldn't spend more then $35 per motor, perhaps $40 if it's in really great shape (check out the brushes, and check that the rotor spins smoothly).
If that fails, I would go ahead and try the Leeson DC motors - Enco has them, but you may be able to get a better deal elsewhere (as I a: have several surplus dealers within 2 hours, and b: my mill came with some huge servos already attached, I haven't spent a lot of time looking.)
In either case, you will probably need new encoders (surplus store encoders that are attached to motors tend to be poorly documented, assuming they still work at all) so have a look at these:
These can be ordered with an adhesive mount and shaft pass-through so it will accommodate motors that don't have a mounting pad for an encoder on their back - you will just have to use some spacers or otherwise make clearance for the encoder.
I prefer Rutex servo drives, as they come as a nearly complete system, with a "motherboard" that plugs into the parallel port, and includes relays for control of things like the spindle, coolant pump, etc. Specifically, I would get 1ea R990MB, 1ea R990IO, and 3ea R990H. Also, I would get a DB15 breakout board. See:
The other popular servo drive is the Geckodrive. They have some disadvantages as compared to the Rutex drives:
They have a lower working voltage 80v, not 100v.
They have a lower servo lock range - 255 steps, not 20,000.
They don't have a motherboard that they plug into, and they don't have a relay card for that motherboard.
On the other hand, they are less expensive, and for many tasks that don't exceed their capabilities, do fine.
To switch the spindle motor on, you will need a heavy duty relay (i.e., contactor) - the R990IO can't handle 5HP worth of power going through it's little relays, let alone switch all three phases.
Then, you need a power supply for all this - a simple, unregulated one will do fine; and a 1 kva supply should be enough for this system.
For the bridge rectifier and capacitors, I would probably just order them from Digi-Key
The other component required is the control PC. For this, I personally am a fan of mini-itx boards - cheap, compact, doesn't require a card cage as just about everything is on the motherboard, and readily available. The other thing that I am a fan of is using CF cards as mass storage for the controller (with an adapter, you can plug a CF card directly into the IDE port on the mini-itx motherboard, eliminating the need to mount a hard disk - specifically, you want the "Female CF Card Adapter").
For software, I would pick up a copy of TurboCNC, and run it under FreeDOS:
Nice thing about these is you get the source code, so if you need to implement any oddball drivers (I've got a TPIC6595 hooked to a couple of spare pins on the parallel port as a relay driver - kinda' nonstandard), or do anything like that, you can (or, if you aren't up to programming it yourself, you might be able to get somebody else to).
Somewhere around here I've got a link to where I got my networking stuff for dos (it's freeware) - All my machines, including the CNC's, are on my ethernet. I'll have to look for it. However, I suppose if you wanted you could mount a floppy drive... I just find floppies, personally, inelegant. http://bbs.homeshopmachinist.net//tongue.gif
[This message has been edited by hammerhead74000 (edited 03-07-2005).]
02-27-2005, 04:31 PM
Forgot something - the Rutex drives need an AC 15 to 24 volt supply for the logic circuits - a doorbell transformer should work OK for that though, if you don't have one already.
02-27-2005, 05:06 PM
Brilliant post well though out and documented.