Stepper Motors

I'd kind of like to clear up a few things about the motors, so that we can tell whether a particular motor is a good alternative to the official, slightly expensive choice.

So here are some questions:

1) What are the suggested specs on the motors? Things like shaft diameter, holding torque, bi/unipolar, voltage, resistance, etc. It would be nice to know what the minimum requirements for the motors are, as they are probably the most expensive parts. Unless you buy the plastic parts from a commercial rp company, that is.

2) Is the precision of the Darwin determined by the motor, or by the gears/rods/belts? In theory, we can increase the precision of the motors either by buying better ones, or micro stepping them. Would it even matter? If the gears/rods are too sloppy, then we shouldn't worry too much about the accuracy of the motors.

3) In that light, how hard is it to get our driver cards to micro step the motors? Is it just a change in the firmware, or should I use another design? It is theoretically possible to run the motors at 1/8 step or smaller, but I don't know if the current driver boards can do so. Again, does it matter? If accuracy is more dependent on the rest of the Cartesian assembly, it wouldn't. But it would still be nice to buy a regular 200 step/revolution motor and half step it or more.

Here is one driver card I found, in case you're interested.
[www.embeddedtronics.com]

Sorry for the long post,

Samuel

Here is a Stepper motor that is very close to the specs. of the Nanotec ST5709S1208-B Specified in the current BOM.

It appears to be available via web order for same day shipping.
More importantly it is available within the U.S.

If you are an electrical kind of person and "skilled in the art" as they say, please take a look and tell me what you think.

Best,
Dan

Edited 2 time(s). Last edit at 06/27/2007 12:09AM by Dan P..

wow, thats a pretty steep price at $84 /ea. i'm playing with a ~$25 one i got from jameco that mounts well in the existing spot for the nanotec motor. unfortunately it doesnt have two output shafts, so i may need to buy one of these to mount in its stead. i should be able to fire up the x and y axes here soon, so i'll let you guys know how it goes. hopefully the $25 motors will suffice =)

Agreed $84 is a little steep. If we knew what the actual required torques and speeds were we could right size the motors (price) no problem. I am guessing that the current motors were used because they were kicking around the lab, not because they are optimized for the application. For those of us in the U.S. that want to get going I think this is a good though pricey option.

Is the second output shaft used? I thought it was just there for decoration... I'm playing around with cheap-o motors too. I grabbed a few 200 step NEMA-17 motors and just got them running yesterday. It's only rated half as strong as the nanotec, but it's still damn strong. I was figuring the mounting brackets would need tweaking and the firmware would need to be recompiled to half-step, but if the second shaft is needed it could be a bigger problem.

It'll be nice if these motors work, it looks like there's a pretty good selection of 200 step NEMA-17 motors in the $10-30 range, and the motors are a pretty good hunk of darwin's price tag.

The $84 motor is the closest thing I've seen to the nanotec one. Don't see any reason it wouldn't work. If the rrrf can do a group buy of the nanotec ones, they should come out to be quite a bit cheaper, though.

I have studied the Darwin mechanicals and as best I can tell the second shaft is not currently being used. My assumption is the second shaft is a provision for adding shaft encoders if needed.

My goal was to provide an appropriate second source for the nanotec motors.

Core members please step in here and provide appropriate specs if you have them.

N.T. if there is the expectation to microstep these motors down stream, then we need to be a little careful as microsteps don't provide the same torque as full steps.

Edited 1 time(s). Last edit at 06/27/2007 01:31AM by Dan P..

@Dan,

The second shaft is there because farnell doesn't sell the single shaft motors.

Microsteps are also very inaccurate (without an encoder). They are mostly used to reduce vibrations and make the steppers more quiet. With an encoder microstepping is very nice, but with an encoder servo motors are even nicer (and cheaper, faster, stronger, simpler, more efficient, more scalable, need I go on?). Okay, the driver firmware is more complex, but that is software. Oh, and I say encoder here for a purpose. Shaft encoders aren't the only way to go. I saw a project where somebody used a modified optic mouse and a ruler as an encoder (on the slide, not on the shaft).

And losing torque with microstepping is not a problem here. The motors have a torque rating of 0.8 Nm or about 80 Ncm. with the tiny pulleys the X and Y assemblies use (I count 12 teeth at 2.5mm/tooth is 30mm around or about 5 mm radius) that would be 160N of raw force. That is more than enough to do light milling of aluminum! With microstepping you get only 70% or about 110N. Should be plenty still.

--Blerik

Dan P. Wrote:
-------------------------------------------------------
> N.T. if there is the expectation to microstep
> these motors down stream, then we need to be a
> little careful as microsteps don't provide the
> same torque as full steps.

No and neither does half stepping so anybody using 200 step motors needs to bare that in mind.

---

[www.hydraraptor.blogspot.com]

Actually, no... Currently the motors run at exactly the same torque as microstepping would. The code is the wrong way around:

177 void motor_click()
178 {
179 byte cp;
180 #ifdef HALFSTEP
181 cp = coilPosition;
182 #else
183 cp = coilPosition << 1;
184 #endif
185 switch(cp) {
186 case 0:
187 PORTB5 = 1;
188 PORTB4 = 0;
189 PORTA2 = 0;
190 PORTA0 = 0;
191 break;
192 case 1:
193 PORTB5 = 1;
194 PORTB4 = 0;
195 PORTA2 = 1;
196 PORTA0 = 0;
197 break;

As you can see it only activates two coils on the half steps. If you use the half steps instead of the full steps, the motors have 40% more torque...

Changing line 183 to:

183 cp = coilPosition << 1 + 1; // increase torque by 40%

will increase the torque a lot. But there isn't a shortage of torque anyway... This will double the amperage on the L298N as well, check if the heatsink can handle it first!!!

--Blerik

(This is why I love microcontrollers. Fixing the hardware by fixing the software)

Edited 2 time(s). Last edit at 06/27/2007 08:13AM by blerik.

Ah yes. I hadn't looked at the code as I am not using it so I had assumed full stepping used both coils as that is the convention. If torque is not an issue then using one coil will keep the motors cooler.

Of course torque falls off with speed so using one coil instead of two limits the top speed. Again that is probably not an issue with belt drive as the extruder speed is probably the limiting factor there.

---

[www.hydraraptor.blogspot.com]

actually the design does use the 2nd output shaft for the Z motor drive. infact, it depends on it. i think this is rather unfortunate, so hopefully it can be changed or we can find a cheap equivalent motor with two outputs. i believe that the mountings are all NEMA 23 motors. i'm going to be testing out the cheapo motor i have tonight on the X axis. i have a good feeling about it. it slides really nicely anyway, so I think it should work well.

Ah... I just took a closer look at the pictures. It looks like the X and Y axes use the usual forward-facing shaft and leaves the shaft that extends out the back unconnected. The Z axis motor seems to be mounted upside down, so it's using the back-facing shaft to move the platform. I can't find a picture of the other side of the Z-axis motor. Is there anything important connected, or will you just have to come up with a way of mounting the cheap-o motor upside-down?

Mounting the Z motor upside down is easy, just use longer bolts. All NEMA 23 motors I can find online have indentations behind the mounting flange, so that shouldn't be a problem.

--Blerik

Edited 1 time(s). Last edit at 06/27/2007 10:41AM by blerik.

blerik - great idea!!! i'll try that instead of trying to find a new motor. i love the internet and the idea that someone thousands of miles from me can suggest a cool solution to a problem i'm encountering!

Zach,

I had (accidentally) ordered the single shaft stepper motors from Nanotec. And this has not given any problems, but for the Z mounting bracket. I think you are now using the RP'ed parts for Darwin (?), where the mounting bracket for Z has been made so that the Z stepper "fits" (a little indentation for a circular part protruding from the motor flange) so you might need to modify that. I've also used longer M5 studs to mount the motor the normal way (i.e. "upside down" from the original design); works like a charm.

Joost

yeah, that 'flip the motor upside down' trick will work nicely. i also fully tested the motors i have and they seem to be pretty powerful. i'm going to order more of them today, and i should have them hooked up in a few days. i'll hook the one back up to the x axis and hopefully i can get a video of it in action when i get back to the lab tonight.

Nice motor flippin.

Fraid we can't spec the required torques for Darwin, as we're still in the research stage, so the nanotech motor ST5709S1208-B (spec'd here: [www.nanotec.de]) is something 'chunky enough'.

eD

The odd thing is that I found this same motor at [www.newark.com] for $52.51 each. A distributor in the United States, Canada, Mexico and Brazil.



Sure beats the $80+ for the other motor

would you mind supplying a part number? we dont have a US distributor for it yet. i'm working with cheaper motors that seem to work, but if for some reason they dont it would be nice to have that fallback.

Sorry, I just did a search for the part: ST5709S1208-B

[www.newark.com]
Newark Part Number is: 42M2604
Newark is a Farnell partner. There is a $15 charge per order because they ship directly from Farnell. The only other charge I could see was the tax charge because I don't own a manufacturing business in my state. So it is 9.75% in Tennessee. So the totals as far as I could see was $187.89 to ship 3 motors to my front door step.

If you're planning on using these motors any time soon, I would recommend ordering them now, as it says on the site that they have 0 in stock and a lead time of 29 days. I don't know if that time is accurate, but it suggests that getting the motors to your doorstep in a timely fashion is not likely.

The Newark page at [www.newark.com]
seems to say they have 0 in stock and a lead time of 29 days. So you'll need a fair amount of patience as well as your US$187.89 :-)

Jonathan

P.S. Eric beat me to this info, but his post was on the second page and I didn't notice it when I posted! Oh well.

Edited 1 time(s). Last edit at 07/06/2007 09:41PM by Jonathan Marsden.

Sparkfun Electronics has a $15 stepper motor:
[www.sparkfun.com]
They're happy to sell one at a time, or hundreds. They're very good about publishing specs.

I don't know much about this kind of stuff but isn't that a 200 step motor? Don't we want a 400 step model? Just a question.

Cheers!

Yes but you can make a one line change to the firmware to use half stepping which will get you the same resolution with a 200 step motor.

Edit:

But with about 1.4 times less torque.

Edited 1 time(s). Last edit at 08/15/2007 07:31AM by nophead.

---

[www.hydraraptor.blogspot.com]

The down-side to that motor is it's rated for 15.4V. The Reprap hardware uses 12V, minus around 2V losses in the driver chip, so the motor would only be getting around 10V, so that'll make the motor even weaker. It might still have enough torque to get the job done, but since the hardware specs are still evolving it's probably smart to err on the side of stronger motors closer to the specs of the official motor.

This is also going to be a bit smaller than the official motor, so you'd need to make custom mounts for it. This one looks like it's NEMA-17 size, the official one is NEMA-23 (I think).

There are quite a few places you can get cheap surplus NEMA-17 200 step motors fairly cheap. I picked up a handful of the $5 (used) motors from allelectronics.com, so far I'm very pleased. alltronics.com also has a good selection, and I've seen good deals on ebay too.

Half-stepping shouldn't cost you any torque over the default configuration since the standard firmware only drives one phase at a time.

> the official one is NEMA-23 (I think).

correct.

It seems to me that RepRap would benefit a lot by switching to cheaper motors. As it stands, they are a huge (and, given the alternatives, one could argue unacceptable) portion of the total project cost. It seems to me that the easiest and most obvious solution would be to tweak the mechanical design to allow for NEMA-17 motors, then tweak the firmware to allow for half-stepping and 200 steps-per-rev unipolar motors. This simple change could allow the cost of the project to be cut by $100 or more with only a few hours of effort by the developers.

The biggest obstacle to this is that no one has actually spec'd out exactly what the torque requirements of RepRap ARE. If we knew that, choosing and integrating a replacement motor would be simple, there are literally hundreds of potential replacements out there. Is there anyone who has a RepRap prototype in their possession and is willing to give this a try?

If we do find a suitable replacement, how would the core developers feel about tweaking the mechanical design at this late stage to allow for the inclusion of this cheaper alternative? Keep in mind that this change would probably literally only affect the design of 3 printed parts, plus perhaps a couple others such as the length of the Z-rod coupled to the motor. It would also be a very simple change that would be extremely unlikely to "break anything," especially since it would be moving from a larger- to a smaller-framed motor.

Kyle

I'm sure it will happen eventually. Once these things start multiplying, I don't think it will take long before you'll be able to download parts files that have been modified to work with NEMA 17 steppers, standard (non-metric) hardware, you name it. If nobody has done it by the time I get there, I'll be doing it myself.

The tolerances on repraps that have been built by repraps will be different than those built by other commercial FDM machines. That might change the torque requirements, so measuring the minimum torque on existing machines may not match up with what the final reprap needs.

There are also other cheaper options in the NEMA 23 size. I think Zach is testing out some 200-step motors that were in the $30 range new from a major supplier. That's more than surplus motors cost, but it's seems to be a decent price for a new motor of any size.

i think that its much more likely that we'll switch to DC motors and encoders than change the type of stepper motor. i'm using the 1.8deg (200 step?) motors from automation direct, and they seem to be working just fine. i havent gotten to the point where i'm building anything yet, so we'll see. i'll definitely let you guys know how it goes.