Stepper Motors

[Brian Lawrence]

This may be a dumb question, but how do I determine a stepper motor size for my bench top milling machine? What are the determining factors & why?
I may be novice , but gotta learn somewhere ........... and CNC is calling my name for some reason???
You can also email responses to [email protected]

Many thanks to those who respond,



Regards

Brian

 

[RM-MN]

Stepper motors for your milling machine are specified by size (NEMA17, NEMA23, etc.) which specifies the mounting hole spacing and their holding torque. Most often you see the steps per revolution (most often 200). They may have 4 wires or 6. Lots for you to learn about.

When I decided to make my mini mill CNC, I estimated the torque needed to turn the crank wheels while milling and then chose motors to accomodate that.

 

[wolf GFX]

Hello Brian.

I did a CNC conversion on a Sieg X2 clone (sold here in germany as Rothwerk EFM200).
The short form:
I use on all axis Nema23 Steppers with 3A coil current, 1.9Nm holding tork, about 72mm long from Stepperonline, toghether with a 36V / 10A power supply and Digital Stepper Drivers DM542T also from Stepperonline.
All motors are geared down 1:3 with GT2 10mm belts with 20/60 teeth puleys. The belts are held under (heavy) positive tension by idler pulleys.
Don't use motor couplers of any kind, they will break, especially the spiral slot ones. The claw type couplers will lead to excessive backlash.
The machine is then driven by SFU1204 leadscrews and ball nuts (4mm displacement per turn, 12mm thick).

This is a very strong setup, intended to mill steel parts - and it works very well.
In effect, the drives are stronger than the machine. Upon a crash, I managed to neatly break a 10mm drill and offset the inclinable z-axis pillar by quite a bit (it was positively locked, believe it).


Some remarks:
* Don't opt for TBB6600 motor drivers:
They are very noisy, heat up themselves and the motors. On top, they won't deliver the torque the motors are capable of. I had these at first and changed to a.m. mentioned digital drivers. They are quiet, don't heat up, motors stay significantly cooler. I managed to get the rapid speed on my machine from 600 mm/min to a safe 1500 mm/min (all axis).
* High voltage from the power supply is important! Best is going to the max voltage the driver are specified for! This lets you achieve higher rapid speeds (before motors stall) and more torque overall. The only drawback is a slightly increase in motor (and less) driver heating. With modern digital drivers you won't notice. This is physics, I won't go into details here.
* As for the control board - I'm using a GRBL breakout board for 3-axis. There are more recent developments going to 32bit controllers. Look for GrblHAL. I use this for a 4-axis machine (including bachlash compensation)
Original GRBL is Arduino 8bit @ 16MHz, program storage cramped up too the last byte! But - it works like a charm - rock solid.

Regards,

Wolf

 

[Toymaker]

First, go to CNC Zone forums where you'll find answers to all your questions. Not trying to chase you away from the expertise you'll find here, but the CNC Zone specializes in CNC of all variety.

Back in 2011, I designed and built a benchtop CNC mill from scratch; and I still use that mill today. The big difference between my mill and a converted mill is that I use linear bearings on all 3 axis instead of "V" grooves, resulting in extremely low sliding resistance, meaning my steppers are only working to push the work into the milling bit and not overcoming sliding resistance. Eliminating sliding resistance makes a huge difference in your mill's performance, for instance, I've been using the spiral slotted couplers Wolf warned you to not use, and I have never broken one during the past 12 years of use. Also, my motors are directly coupled to the leadscrews which are 5/8" diameter with 5 threads/inch. All 3 axis can move at 200 inches/min., which is scary-fast in the Y direction's 7" travel distance.

I too use Nema 23 steppers on all 3 axis and adding to Wolf's info on choosing the right stepper, choosing the right voltage is also very important, and somewhat counter-intuitive. I too use a 36 volt DC power supply to power the steppers, which are rated for 3.6VDC. No, that's not a type-O, I really did type 3.6 volts. It may seem a bit crazy to power a 3.6 volt motor with 36 volts, but don't worry, your motor controller will make sure the stepper motor never gets too much current and burns up or even gets too hot. The 10x voltage allows the magnetic field inside the motor to build very rapidly, resulting in very rapid shaft motion of the motor.

I've only just started using a single TB6600 to open and close a ball valve, and so far at least, I haven't run into the over heating issues Wolf mentioned, and set for 8 micro-steps, it runs silky smooth.

Hope some of this is helpful,...
cheers !

 

[wolf GFX]

Hi all.
Just an interesting link concerning steppers:


Regards,
Wolf

 

[timo_gross]

Hello,

My benchtop machine has NEMA34 motors, the drivers are 6 A for x and y and 8 A for z-Axis. The machine is a little bigger than Wolfs.
The Torque I do not know. (not more than 6Nm) Occasionally I had lost steps when pushing the machine to the limit. The machine has the motors directly coupled to the ball screw with a claw/rubber coupler, no gear reduction. (that makes it quite similar to Wolfs setup with 2Nm and 1:3 belt reduction)

If direct coupling or belt is better, is not so easy to say.

Don't use motor couplers of any kind, they will break, especially the spiral slot ones. The claw type couplers will lead to excessive backlash.

I would not agree. A lot of industrial machines have the motors directly coupled to the ball screw. Of course the design (accurate alignment) and materials used, should be fit for the job. ( couplers can look similar, but there are differences in runout backlash, price, hardness of the rubber inserts etc. )

Some remarks:
* Don't opt for TBB6600 motor drivers:

I agree with Wolf TB6600 drivers are quite horrible, they were maybe O.K. years ago. Today you get much better drivers for 10-20 USD, so no reason to use those.

What I suggest

• get a matching set of driver and motor from same supplier. Motors with integrated drivers are also available.
• avoid weird voltages (my 2nd hand machine came with 80 V system, which was a pain when one driver was damaged and I wanted to find a replacement, not many choices)
• consider upgrading to closed loop steppers, they just are nicer to use (my opinion); specially when you can feed the fault alarm back into the CNC controller
• look around what other machines that are similar to the one you want to drive are using (and ask owners, if they would do the same again, beware of confirmation bias)

Determining motor requirements in the correct way can become quite involved.

 

[Toymaker]

I agree with Wolf TB6600 drivers are quite horrible, they were maybe O.K. years ago. Today you get much better drivers for 10-20 USD, so no reason to use those.

Which stepper driver(s) for $10 to $20 do you recommend ?

 

[timo_gross]

Which stepper driver(s) for $10 to $20 do you recommend ?

Oh oh, may be you got me here, was I comparing apples with peas?
I think to aswer this seriously I need to put some more "brain" in this. Hope I can get away withe the bold claim . I will let you decide.

 

[timo_gross]

Spent the last 90 min to open lots of tabs on my browser copying the information.
I had luck as the TB6600 is (luck for me ) weaker than I feared (when the question came), indeed I found some alternatives.

You can have a look. Those tiny 3d prinder drivers can put out more power than some of the older bigger boxes. They must have gotten the power electrics really efficient compared to 30 year old drivers. Amazing, what the tiny drivers can already do. So if 1-2 A are sufficient, then you get away with really cheap printer hardware.


This is maybe not as interesting for the average CNC hobby, but it is giving some explanation how things work.




p.s.
@Toymake, if you are seriously searching, I can send you the links (do not want to post them here, do not like adverts and links will stop working in two weeks)


Greetings Timo

 

[Toymaker]

Spent the last 90 min to open lots of tabs on my browser copying the information.
I had luck as the TB6600 is (luck for me ) weaker than I feared (when the question came), indeed I found some alternatives.

View attachment 151843
You can have a look. Those tiny 3d prinder drivers can put out more power than some of the older bigger boxes. They must have gotten the power electrics really efficient compared to 30 year old drivers. Amazing, what the tiny drivers can already do. So if 1-2 A are sufficient, then you get away with really cheap printer hardware.


p.s.
@Toymake, if you are seriously searching, I can send you the links (do not want to post them here, do not like adverts and links will stop working in two weeks)


Greetings Timo

I did not expect you to go through so much effort to create the chart,...but I suspect the OP will appreciate the results. However, I did notice one little mistake: The TB6600 is rated for 40v (not 24).

Comparing all the voltages and currents, there's really not a huge difference between them, though the DM860H does seem to provide the most power.

 

[timo_gross]

I did not expect you to go through so much effort to create the chart,...but I suspect the OP will appreciate the results. However, I did notice one little mistake: The TB6600 is rated for 40v (not 24).

Comparing all the voltages and currents, there's really not a huge difference between them, though the DM860H does seem to provide the most power.

The voltage for the TB6600 board I copied from one of the sellers spec. Chances are that parts on different boards like capacitors are rated different.
That would also explain the range of shapes looks and prices of different TB6600 boards.

I think the list is on the very low end for driving the main axis of a medium milling machine. Probably better to get something more substantial, but I tried to find things below 20 $.
I personnally would consider spending somewhere between 120-180 per motor with driver (probably closed loop steppers).
AC servos are the next up, but I am not convinced that is the best choice for absolute beginner.

This cheap TB driver says 24V.


This boxed one I just grabbed from an electronic store out of curiosity. I was surprised how good it was considering its 10$ price.


The issue here ist, it does not give any information what you get. No maker, no hint concerning the internals and I had to crack it open to see it is a S109AFTG (whatever that means)

edit: surprise surprise the S109AFTG is also a Toshiba chip, obviously a little newer and it is tiny for beeing able to handle four A.

 

[Toymaker]

The voltage for the TB6600 board I copied from one of the sellers spec. Chances are that parts on different boards like capacitors are rated different.
That would also explain the range of shapes looks and prices of different TB6600 boards.

I think the list is on the very low end for driving the main axis of a medium milling machine. Probably better to get something more substantial, but I tried to find things below 20 $.
I personnally would consider spending somewhere between 120-180 per motor with driver (probably closed loop steppers).
AC servos are the next up, but I am not convinced that is the best choice for absolute beginner.

Interesting side note concerning the TB6600. According to the DIY robotics crowd, most TB6600 controllers sold today and ALL controllers labeled "Upgraded", don't actually use a TB6600 chip, but instead use the slightly cheaper and much smaller (physical size) TB67S109AFTG, also made by Toshiba. The performance and specifications of these two chips are similar, but the TB6600HG does have a higher peak current rating (up to 5 A), and the TB6600HG only supports up to 1/16 micro stepping, while the TB67S109AFTG supports 1/32.

I opened the metal box of the TB6600HG I purchased from an online company here in Thailand and yep, it uses the TB67S109AFTG.

I suppose closed-loop steppers are nice if you intend to really push the limits of how fast your machine can cut through a piece of stock, where you're making such large cuts, and so fast, that your steppers are right on the edge of missing the occasional step (or micro-step). And that's great if you're in a rush to get a job done or you need to make 100 widgets as fast as your machine can go,....BUT, I'm not sure most hobbyists fall into that category. Even on the few occasions where I'm making, say 60, identical parts, no one is paying me to make them as fast as possible,...so, I can, and do, take my time. Also, my experience with my mill when I have pushed cutting feed too hard is that either my milling motor bogs down with large bits, or I break smaller bits,...but never do my motors miss a step.

 

[timo_gross]

Interesting side note concerning the TB6600. According to the DIY robotics crowd, most TB6600 controllers sold today and ALL controllers labeled "Upgraded", don't actually use a TB6600 chip, but instead use the slightly cheaper and much smaller (physical size) TB67S109AFTG, also made by Toshiba. The performance and specifications of these two chips are similar, but the TB6600HG does have a higher peak current rating (up to 5 A), and the TB6600HG only supports up to 1/16 micro stepping, while the TB67S109AFTG supports 1/32.

I opened the metal box of the TB6600HG I purchased from an online company here in Thailand and yep, it uses the TB67S109AFTG.

I suppose closed-loop steppers are nice if you intend to really push the limits of how fast your machine can cut through a piece of stock, where you're making such large cuts, and so fast, that your steppers are right on the edge of missing the occasional step (or micro-step). And that's great if you're in a rush to get a job done or you need to make 100 widgets as fast as your machine can go,....BUT, I'm not sure most hobbyists fall into that category. Even on the few occasions where I'm making, say 60, identical parts, no one is paying me to make them as fast as possible,...so, I can, and do, take my time. Also, my experience with my mill when I have pushed cutting feed too hard is that either my milling motor bogs down with large bits, or I break smaller bits,...but never do my motors miss a step.

Haha, that explains why you were happy with it. I find the S109AFTG much nicer running than the TB6560.

Closed loop steppers just make things a little more reliable and they seem to have more torque at higher speed.
Given the relatively moderate price penalty I think it is just a good investment. I think most of my problems were "overloading" the poor thing. Unreliable chip evacuation, causing increased cutting loads does not make things better. Dovetail machine much slower than 5m/min.

Would be interesting how heavy your mill is and what torque rating your motors have.

Greetings Timo

 

[Nikhil Bhale]

Hi,
I have a question in mind. Since this is a CNC related topic I will ask it here.
Why do CNC setup requires stepper motors and ball screws with minimal backlash?
My thinking is that if we can get position feedback from DRO to motor driver the setup will be backlash free.
And you could use normal acme feed screws and DC motors. Maybe DC motors with electromagnetic brakes for holding position.

My question is, Is this possible or I my idea is totally absurd?

Regards
Nikhil

 

[Toymaker]

This may be a dumb question, but how do I determine a stepper motor size for my bench top milling machine? What are the determining factors & why?
I may be novice , but gotta learn somewhere ........... and CNC is calling my name for some reason???
You can also email responses to [email protected]

Many thanks to those who respond,



Regards

Brian

You can significantly reduce the size of the Z motor needed by countering the weight of the milling head. The Z axis motor wastes a good amount of power just moving the mass of the milling head up. Most conversions I'm aware of use a larger motor on the Z axis to overcome the weight of the milling head. However, if you devise a method or means that counters the weight of the milling head, you can reduce the size of the stepper needed. There are several ways to accomplish this: consider how an elevator uses a counter weight on a pully with cables to offset the weight of the elevator car. If you have room, a simple pulley and cable connected to the milling head and a weight of equal mass of the milling head will result in eliminating the weight of the mill head. Another method uses "pneumatic lifters" (like those found on many cars) pushing up on the milling head, thereby reducing the weight of the mill head. There's no "one" best way,...just use the best method for your mill.

The method I chose for my mill was to use a beefy spring connected through a series of pulleys as shown in the pics below, which shows the back side of my mill. This "block and tackle" system allows for a large linear movement of the Z axis, while causing a very small movement of the spring; which is ideal since springs exert fairly linear force over a short pulling distance.

 

[Toymaker]

Haha, that explains why you were happy with it. I find the S109AFTG much nicer running than the TB6560.

Closed loop steppers just make things a little more reliable and they seem to have more torque at higher speed.
Given the relatively moderate price penalty I think it is just a good investment. I think most of my problems were "overloading" the poor thing. Unreliable chip evacuation, causing increased cutting loads does not make things better. Dovetail machine much slower than 5m/min.

Would be interesting how heavy your mill is and what torque rating your motors have.

Greetings Timo

How heavy is my mill ??? I have no idea,....only that its far too heavy for me to pick up by myself

That said, one of my goals as I was designing my mill was to make it as light as possible, while keeping it ridged, and to that end, with the exception of the purchased milling head, the entire frame and all other major parts, including the table, are made from 6061 Aluminum. If you've ever seen a mini-mill, or pictures of one, you may recognize the milling head in the pics below. I considered making my own mill head, and I hated to add the weight of all that heavy cast iron, but in the end, it was just too easy to buy instead of build. All the flat black surfaces you see are type III hard anodized, including the table.

In lieu of weight, I can give you dimensions.
The widest part is the aluminum frame holding the X axis rails and is 69cm across.
The table measures 19 x 49 cm.
Depth is 46cm and height is 73cm.
The majority of the frame is composed of 1/2" aluminum plates, screwed together with lots of M5 and M6 screws to form 3 boxes, all bolted together to hold the X, Y, and Z linear rails.
The 3 Stepper controllers are SLA7062M.
All 3 NEMA 23 stepper motors are:
P/N: 57BYGH405A
4 phase Unipolar 6 lead
Coil: 3.6v at 2.4mH
Holding torque: 21Kg-cm

 

[Toymaker]

Hi,
I have a question in mind. Since this is a CNC related topic I will ask it here.
Why do CNC setup requires stepper motors and ball screws with minimal backlash?
My thinking is that if we can get position feedback from DRO to motor driver the setup will be backlash free.
And you could use normal acme feed screws and DC motors. Maybe DC motors with electromagnetic brakes for holding position.

My question is, Is this possible or I my idea is totally absurd?

Regards
Nikhil

The simplest answer is; because steppers and ball screws are the best, easiest way to build a hobby CNC machine.
DROs are nice, but expensive. By tasking the computer to keep track of the number of steps each stepper motor makes, the computer always knows the X, Y, and Z positions without the need and expense of a DRO on each axis.
Ball screws and ball nuts are used mostly to reduce rotational friction, which the motor must work harder to overcome in order to move the axis, the fact that ball screws & nuts have essentially eliminated backlash is simply a side benefit as modern machine control software such as Mach3 can compensate for backlash.

So, yes, you could use DC or AC motors to drive the feed screws and put DROs on all the axis and include electromagnetic brakes on all 3 axis,...but all that would be much, much harder, and far more expensive to do than just using steppers.

BTW, high end industrial CNC machines do use DC motors to rotate the feed screws, and the motors have "encoders" which send position information back to the computer; DROs aren't needed. Steppers have become the motor-of-choice because they're the easiest to use, and plenty fast enough for most applications.

 

[RM-MN]

Why do CNC setup requires stepper motors and ball screws with minimal backlash?

The mills make the smoothest cuts when they are used in climb milling. My mill has about 3 thousands of backlash. If the cutter I am using requires a one thousandth of an inch per revolution and I try climb milling, at some point the cutter will pull the work toward it that 3 thousandths of an inch and that will overload the motor, make a step in the smooth cut, or break the end mill. Were I to replace the leadscrew with a zero backlash ball screw that would never happen.

 

[Toymaker]

Seems a few folks like the DM542T stepper controller, and it's priced at around $10 +/- a few bucks. So another controller to consider.
The guy in this video shows his router running on both the TH6600 and the DM542; clearly the DM542 is the better choice. DM542T vs TH6600

 

[Toymaker]

This is a simple no-load speed test showing all 3 axis of my 11 years old CNC mill; I'm manually pressing the direction keys on the keyboard. It's very hard to see, but, if you watch on a large screen, you can just make out the speed displayed on the Mach3 screen, showing 200 inches/minute for both X & Y, and between 197 & 198 for the Z axis. Ball bearing lead screws and linear bearings & rails make this possible.

I hope this serves to encourage others