Power Supply for Big Stepper
Bill Chennault
Posts: 1,198
All—
·
I have completed a project and placed it in “The Sandbox.” It is called ”Stepper Motor Driven Rotary Table.“ The point of the project was to determine the repeatability of a such a setup with the idea of building another unit and mounting it on the first at 90 degrees, thereby potentially creating a very repeatable and accurate robot shoulder.
·
I did not spend a lot of money doing this and all the descriptions, costs, and links are included in The Sandbox thread mentioned above. My experiment is on-going in that I am now stressing the little Jameco NEMA 17 stepper up to (what I think is) about 75% to 80% of its capacity. The Jameco stepper is connected to the shaft of the rotary table via a timing belt arrangement that provides 3:1 gear reduction.
·
However, obviously the Jameco 155459 stepper does not provide enough torque for the loads one might imagine a robot shoulder to encounter. Right now, all that is mounted on the rotary table is a magnetic clamp with a dial indicator that provides visual feedback regarding repeatability. (In The Sandbox thread there is a picture, a link to another picture, and a link to a tremendously exciting movie—for which I expect an Oscar—depicting the setup.)
·
So, I am drawing conclusions based on my continued testing, which I will report in The Sandbox. However, I have a question or two.
·
The Jameco stepper must be from a company that designed a little printer and ordered a million little steppers for them, that is why they are so cheap. (With cheapness comes weakness.)
·
I need a stepper that offers enough torque to drive the rotary table directly, eliminating the timing belt arrangement, as interesting as it is. I think I have found a 740 oz-in stepper that will do the trick. Additionally, the recommended Gecko controller is fine, as well. The problem, as I see it, is the power supply. The vendor offers a toroidal transformer, a big cap, and a rectifier and you get to build the power supply.
·
I don’t want to do that. I want to buy a bench-top power supply sufficient to do the job. What am I missing here? Are 68 volt, 12 amp (approximately), DC power supplies unavailable in a bench-top format? Are they ULTRA expensive?
·
I would certainly appreciate your input and guidance.
·
--Bill·
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
·
I have completed a project and placed it in “The Sandbox.” It is called ”Stepper Motor Driven Rotary Table.“ The point of the project was to determine the repeatability of a such a setup with the idea of building another unit and mounting it on the first at 90 degrees, thereby potentially creating a very repeatable and accurate robot shoulder.
·
I did not spend a lot of money doing this and all the descriptions, costs, and links are included in The Sandbox thread mentioned above. My experiment is on-going in that I am now stressing the little Jameco NEMA 17 stepper up to (what I think is) about 75% to 80% of its capacity. The Jameco stepper is connected to the shaft of the rotary table via a timing belt arrangement that provides 3:1 gear reduction.
·
However, obviously the Jameco 155459 stepper does not provide enough torque for the loads one might imagine a robot shoulder to encounter. Right now, all that is mounted on the rotary table is a magnetic clamp with a dial indicator that provides visual feedback regarding repeatability. (In The Sandbox thread there is a picture, a link to another picture, and a link to a tremendously exciting movie—for which I expect an Oscar—depicting the setup.)
·
So, I am drawing conclusions based on my continued testing, which I will report in The Sandbox. However, I have a question or two.
·
The Jameco stepper must be from a company that designed a little printer and ordered a million little steppers for them, that is why they are so cheap. (With cheapness comes weakness.)
·
I need a stepper that offers enough torque to drive the rotary table directly, eliminating the timing belt arrangement, as interesting as it is. I think I have found a 740 oz-in stepper that will do the trick. Additionally, the recommended Gecko controller is fine, as well. The problem, as I see it, is the power supply. The vendor offers a toroidal transformer, a big cap, and a rectifier and you get to build the power supply.
·
I don’t want to do that. I want to buy a bench-top power supply sufficient to do the job. What am I missing here? Are 68 volt, 12 amp (approximately), DC power supplies unavailable in a bench-top format? Are they ULTRA expensive?
·
I would certainly appreciate your input and guidance.
·
--Bill·
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Comments
In terms of bench power supplies, that voltage and current definitely limits what's available off the shelf.
Price-wise, you're not going to get much better than that what's being offered by your vendor unless you can stumble onto a surplus transformer somewhere.
Its going to have substantial ripple at full load, but the steppers probably won't mind. Needless to say, the electronics will need better filtering.
Cheers,
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Tom Sisk
http://www.siskconsult.com
·
An unregulated 24V supply should be more than adequate for the 2V motors you've spec'd. I doubt that increasing the voltage beyond that will affect the torque characteristics much, since you'll be shaving very little additional from the time it takes to get the windings near their rated 2 amps. Here is a company that offers unregulated supplies for a reasonable price. These kinds of supplies appear regularly in the surplus listings and on eBay, too.
-Phil
Post Edited (Phil Pilgrim (PhiPi)) : 6/7/2007 4:37:05 PM GMT
I have aCNC mill I have been working on and I use the toroidal transformer, Bridge rectifier and a large stiffin Cap. Miine puts out 20 amps and about 60 volts. I can give you the part #s and they are cheaper than from Gekco I made mine inside of a old Computer case and it works very good but is not regulated on the DC side like a benchtop supply would be but it does not vary much on my setup. A benchtop supply that can put out that much power would be expensive and not really needed.
Also if you would like me to I can put a Oscope on the output and see how much ripple their is.
Post Edited (bennettdan) : 6/7/2007 4:30:55 PM GMT
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Think Inside the box first and if that doesn't work..
Re-arrange what's inside the box then...
Think outside the BOX!
I have been looking at that site for a while now because I do want to CNC my Grizzly Mill. I feel building the power supply from the components he offers would give you a LOT more supply than what you would buy - and it is most likely the route I will take. I say go for it - you certainly seem capable!
I think I am seeing the light concerning the high voltage supplies. Of course, I built them every day when I was a kid, just like everyone else. And, used basically the same components. Let's see if I can remember: A transformer out of a junk TV, a U4 half wave rectifier (big tube, I THINK it was·called a U4) and a big capacitor. I can't remember the DC output these things produced or even why I built them. [noparse]:)[/noparse]
Paul Sr., I am about to buy that Grizzly G0619. I will doubtless play with a rotary table on it, too! I am looking forward to its vertical DRO and RPM indicator. (Thank you for the compliment. I have learned a lot on this forum. And, I could have figured out the power supply, but it seems so .·. . so what? Know what I mean?
metron9, this machine--the robot shoulder using two stepper-driven rotary tables--will be table-top based. I will use it to see if I can still do a little genetic programming. That is why repeatability is so important to me. I also wanted to see if it were possible to build such a shoulder at hobbyist prices. Apparently, it is not going to be. Maybe.
bennettdan, I would love to see your circuit diagram! Please!
Phil, could there be something about stepper motors that make the voltage specifications (I read them, as well) inapplicable without some kind of calculations that produce the REAL voltage? I am driving the windings of the little bitty Jameco stepper--which is so inadequate as a "working" stepper that they do not even show the holding torque--at the Jameco-recommended 12vdc . . . WHILE I WAS TYPING THIS, homeshopcnc e-mailed me in response to my power supply query and said they have a 63V, 12A power supply that is fully built for·$135.00, which is not advertised on their web site, yet. I e-mailed them back and asked them if that was the appropriate supply to drive the windings of the 740 oz-in stepper. I'll keep everyone informed. (Phil, further . . . I read the controller "data sheet" and it makes it clear that the stepper controller and the stepper expect a lot of voltage.)
Tom, I now understand that the power supply is not going to be a nice, variable, linear supply with multiple outputs as I had dreamed! [noparse]:)[/noparse]
Well, while I was writing this book, "Rick" at homeshopcnc responded to my e-mail and said the power supply they are about to offer on their web site would be perfect for the 740 oz-in stepper I have in mind, or even up to four of them. So, I need to add things back up and see how far we are from hobbyist pricing again. Maybe not far at all, given the fact that the power supply will handle up to four steppers. (Too bad the stepper controller won't.)
I imagine I will build this thing using those parts with the POSSIBLE exception of the power supply, which I might build if I can do it cheaper than homeshopcnc sells it for.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
The main specs you have to keep in mind with a stepper are the maximum current per phase (6A for your motor) and the inductance (2.7mH for your motor). The voltage rating is the voltage that will be present across a fully-saturated winding when the current is 6A (2V for your motor). This works out to a winding resistance of 0.33 ohms. The reason for the higher voltage supply is to overcome the inductance of the windings, forcing more current into them more quickly. Your motor controller will do this with a PWM output, while keeping an eye on the current so it doesn't exceed the maximum rating. When the winding is first energized, the PWM output will be nearly 100%. As the winding becomes saturated, that percentage will diminish quite rapidly. These measures ensure that you get the maximum available running torque, but they do nothing for the holding torque. You'd get the same holding torque with a 2V supply.
The spec sheet for the controller requires an input voltage between 24V and 80V. It suggests a voltage between 4x and 20x the motor's rated voltage, which works out to an 8V to 40V range. But with 24V being the minimum, it's really 24V to 40V. At 24V, the ratio is 12:1, which should be more than adequate.
BTW, and without knowing what the gear reduction on the rotary table is or the kind of loads you'll be subjecting it too, are you sure you need such a honkin' motor? 740 oz-in. is equivalent to applying 23 pounds of force to a 4" diameter handwheel. Will the rotary table actuator really that hard to turn under load?
-Phil
Thank you for the stepper education. I am sorely in need of it.
No, I am not sure about the big-torque necessity. Sherline has a CNC rotary table (top of page on the right: "POWER OF THE MOTOR") of the size I am running downstairs right·now: four inches. However, while my Jameco stepper doesn't even have a published torque rating, the Sherline CNC rotary table has a 100 oz-in stepper.
The mill I am buying is a bit larger than the Sherline mill for which their CNC rotary table was designed. And, I am not a machinist and my last mill experience was in 1969. But, that experience was on a GIANT machine. So, no, I really do not know.
Here is what I would like to do: Build a dual table setup, one table mounted perpendicularly on the other, both driven by stepper motors. I would use this setup as a table top robot shoulder. From the table of the second rotary table, I would construct a sturdy arm using aluminum and probably an AX-12 (scroll down to see the AX-12) or one of its bigger brethren (the RX-28 and RX-64 are at the top of the page) at the elbow.
740 oz-in would certainly do what I want in this situation. Like you said, more than enough especially since I am really building it to move something like children's wooden blocks around.
I have another purpose in mind for one of the stepper-driven rotary tables: I want to stick it on a soon-to-be bought mill (G0619). That mill is somewhat larger than the Sherline. I am also sure that the 740 oz-in stepper motor would be more than enough to drive the rotary table during milling operations. Could it be too much? Could I get in trouble having too much torque on the rotary table shaft? I do not have enough experience to answer these questions. Paul Sr. wants to do a CNC work-over on his Grizzly mill. Perhaps he has some input--or you have some input--that would help me decide. I am certainly not in love with the 740 oz-in stepper motor. There is a nice 425 oz-in stepper. Or a 260 oz-in stepper. What I would like is one motor size that will perform both tasks: The dual rotary table robot shoulder and the stepper-driven rotary table on the mill.
Got clues?
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
I will use a misalignment shaft coupler to couple the shaft of the stepper with that of the rotary table. The shaft of my Grizzly rotary table is 10mm. Rather healthy. No where does Grizzly mention the gear ratio of my 4" table. I imagine it approaches 90:1. When I am finished testing it, I will stick the handle back on and find out.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
I can offer one data point, so I don't know how useful it will be. I drove a Sherline mill for years using 30 oz-in steppers with a 50:16 gear reduction. This would be equivalent to 94 oz-in if directly driven — less if you consider the loss of efficiency in the gearing. The motors suffered a loss of 10 oz-in at resonance, which rendered them barely adequate at that speed.
Perhaps you can extrapolate from that...
-Phil
The 425 would be more than enought to mill just abut anything on that little mill. My mill has a 10" by 54" table and is very large and it has 1200 oz/in motors for what I have been told by other guys that know my Mill. their is no name plate on the motor so I am not sure but I know they are about 10" long and about 8"*8" square. Most of the guys recommended me to run 8 amps to each motor so thats why I have the large power supply.
I have to agree with bennettdan that 425 is MORE than enough. In fact, looking at your picture of the unit you built, it looks like you are running at least a 3:1 pulley ratio - you could probably do fine with the 260. I do know that most of those rotary tables need lots of deburring and cleaning to make them operate smoothly so if you haven't done that (I haven't done mine yet either and it NEEDS it!), you will eliminate lots of drag. From everything I have read, I would look to the 740 as a Z axis motor, 460's for X and Y, and the 260 for the rotary table. For the "shoulder" - not sure on that as it has lots to do with the length/weight of the "arm" hanging off of it...
That G0619 looks like a nice mill. I have the smaller G8689 - also nice and hasn't let me down, but if I were to buy it again, I would go with the one you have selected as the 8689 is a bit cramped at times. I'm not a heavy user, so it is fine for what I use it for (so far!).
Tubes, huh - yep I was weaned on them too. My last experience with them was in '71 - biggest was the final stage amplifier for long range radar - 2" water water cooling lines, filaments 3.5V@6,800Amps...... We had cranes on both floors (second floor was a standby channel) to lift them out of the "chambers" to do the monthly PM.
Anyway, BUILD that power supply- you'll be much more satisfied!
Paul
Thanks for the input on the steppers. I am getting a lot of good advice.
Think about this in light of your size recomendations: I will do away with the timing belt arrangement . . . and, yes, it gives me EXACTLY a 3:1 reduction! (Did you use the Mark I Eyeball?) I plan to connect the stepper directly to the 10mm shaft of the rotary table through a misalignment shaft coupler.
So, does that affect your recommendations regarding a CNC rotary table for the mill I have in mind?
Debur the rotary table? Give me a clue how to do that.
(I would have loved to see those big vacuum tubes! Could the radar fry a bird?)
Thanks!
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Can a Power supply from a PC be used?·· I am building a CNC Router table and almost ready to put all the parts together.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Fernando Gomez
revinc.us
gomez-rivera.com
Never compare yourself with anyone else, there will always be someone bigger·or·smaller·than you.
I had hoped someone far more knowledgable than I would answer your question. I do not have enough experienced with steppers to really be confident of my answer.
Compare the output of the PC power supply with the minimum voltage and current requirements of the stepper motor you choose.·If the output of the PC power supply meets those minimum requirements then it should work.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
Yep - Mark I Eyeball was it. Not bad for an old fart, huh!
Getting rid of the 3:1 and going direct would most certainly be reason to re-evaluate the stepper/servo choice - you need more torque (especially initial) for a given load without the reduction. Whether or not you need to completely make up for the loss of the 3:1 (i.e. do you REALLY need 3X the torque capability?) is again dependent on the weight/friction attributes of what you are going to move.
I read an article that someone wrote on deburring the rotary table. Can't remember where I got it, but I will look and see if I saved it anywhere or can point you to a site (there were 3 or 4 sites that I was visiting regularly that covered the Mini Mill/Lathe). Basically, you completely disassemble the unit, remove any burrs or anomalies in the gears and moving surfaces, lube it up and reassemble. Some folks realized great improvements in smoothness and usability.
Something like (old brain cells at work here) 5-6 MW would fry ANYTHING. When we had to go up and work on the antenna, there was a key that you would turn and remove and NOTHING could make the radar fire while you had that key in your pocket and/or the hatch to the roof was open. Low man on the totem pole had to change the marker light bulbs - on top of the 50+ foot antenna, mounted on the roof of a 120' high building situated on a 3800 foot mountain. It was truely a rush!! Did it many times - today, I am actually afraid (horrified to be truthful) of heights and can't even get up on the roof of my own house!
We should probably consider moving this to the Sandbox if we continue......
Ha! Loved the story!
If you can find anything on deburring one of the little beasties, I would appreciate it. In the meantime, I will Google around. If all else fails, I'll just tear it apart and see what I see!
The only reason I used the 3:1 reduction was to get enough torque to turn the rotary table with the little Jameco stepper. In my last experiment, which amounts to using different winding combinations and a few other settable parameters, it has been running for 24 hours. For the first several hours, I was getting +/- 0.002" repeatibility. Now it is hitting almost right on the money. The only thing I can figure is that my winding combination is better . . . or the table is getting "broke in" or both.
I switched to from a BS2 to a BS2px thinking it would work better at 9600 baud, especially the messages I receive from the stepper controller and display·using DEBUG. It turns out that either Parallax Stamp is plenty fast enough.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.
This is the stepper (scroll to the bottom) SherlineIPD uses on their CNC rotary table. I have a Peter Norberg Consulting BS0710-USB stepper board that will handle two, one amp stepper motors or one TWO AMP·stepper motor, which is what SherlineIPD puts on their rotary table. (136 oz-in.) By the way, I discovered the little Jameco 155459 stepper motor I am currently using is rated at 29 oz-in. At 3:1, I suppose that is more like 87 oz-in.
Still, if the BS0710-USB controller board·from Peter Norberg Consulting will handle the SherlineIPD stepper motor, I will save the cost of another controller board and, much more importantly, another power supply. The combination MIGHT put my robot shoulder price back in the range of the hobbyist.
I have e-mailed Peter Norberg Consulting to see if they think their board will handle the SherlineIPD (it's an NMB product) stepper motor.
The Parallax Stamp-powered rotary table is currently in one of its many tests. This one is in its 36th hour. I will let it go for about 48 hours. However, repeatibility with the current PBasic/BS0710-USB parameters seem to be giving me about +/- 0.002".
This is good. I think.
--Bill
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
You are what you write.