Best "Affordable" 4th Axis
- Thread starter dunkelcustomcue
- Start date
- Status
That also looks like this , a gear drive not harmonic drive. Just made different housing
http://www.aliexpress.com/item/Gear...3-Geared-Stepper-Motor-23HS22/1794201550.html
RBC
Deceased
George
Well, I don't think this one is a cross roller, but I haven't taken that end apart as of yet. By the size and shape, I'd guess it's just a deep groove roller bearing.
Just so you know, this unit is a Harmonic Drive. I have removed the motor, and found the input side of a Harmonic Drive unit. It's actually larger than I'd expected as well.
As to steppers versus servo's, I have to disagree with your statement. There are a few things to consider as well as a few common misconceptions about both motors.
Steppers are axtremely accurate in full steps, meaning 1.8 degrees of rotation per step. Most drives today use what's called Micro Stepping in which they divide the normal 200 steps per revolution (1.8 degrees) into many more "micro steps. They do this by "balancing" the motor armature magnetically between the poles. Of course, accuracy falls down somewhat in this scenario, but not very much. Certainly not enough for us to worry about with cue work. Gecko drives use 10 micro steps, turning a 200 step motor into a 2000 step motor. Based on the tolerances and over lapping, anything more than 10 is pretty much empty resolution.
Servo's are true feedback type devices. The drive is constantly working to bring the encoder feedback to wherever it believes it's supposed to be. Obviously, the number of encoder counts per revolution will dictate how accurately the motor will be positioned. Encoders are typically read in "quadrature" which means it's not uncommon in small hobby style systems to see encoder counts at 1000 to 4000 counts per revolution. The "enhanced" version of the new clearpath motors offers real resolution to 6400 counts per revolution. Many commercial machine have much much higher counts than that.
So, given the above information, even allowing for micro steps (2000 steps per rev), most servo systems have higher real resolution than steppers, and can hold tighter positioning.
Now, if we go back to my statement about using a servo (clearpath) on the Harmonic drive, I was referring to backlash during reversing. This is where the micro stepping really comes in to play. When you stop somewhere in between 2 full step positions and then reverse the direction, something unique happens. Quite often you will need to execute several small microsteps before you get any motor movement. I've done this with just an arm attached directly to the stepper motor shaft. I found it very strange that the control could be commanding movement, but the motor did not move. My belief is that it has something to do with micro steps. It may be that when you back up it goes immediately to the previous full step position, and then builds the micro steps from there. Whatever the reason, I believe that the tiny amount of reversing delay I experienced would be gone with a clearpath motor.
All this is really just crazy stuff anyway because the differences we are talking here are extremely tiny. Much smaller than our physical machines are capable of anyway. But it does make for some interesting conversation for gear heads like me.
Royce
Hi Royce, harmonic drive will have cross roller as main bearing. if you get a chance if you could send me pictures of inside and if wave generator .
Servo only better if there is encoder pick up to controller card. Closed loop just to drivers is just a waste of money. But no better at accuracy . But you right , not much difference if both are set up right .
Easy servos are probably the best to use because low torque need for the routers.
its a stepper with encoder .
high speed servos good for mill with 5 to 1 pulley set up
RBC
Deceased
George
I'm sorry but you're wrong on pretty much all accounts. I don't mean to be contradictory, but the truth is the truth.
I've seen several Harmonic Drive systems without cross roller bearings. As a matter of fact, Harmonic Drive Systems, the manufacturer of Harmonic systems in the USA, sells most of their systems as components which don't have any bearings at all. Actually, I really can't figure out where your statement about cross roller bearings comes from.
The servo versus stepper comparisons were based solely on the motors and what they are capable of. These are facts, and how they are driven doesn't change the comparisons at all. Many major manufacturers of very large and very high accuracy machines don't bring the encoder signals back to the main control. It's very simple. Servo's with the higher encoder counts, can hold tighter positioning than can steppers, even including the "Micro" steps which are questionable as far as accuracy goes.
Royce
Hi Royce , cross roller are use to support the shaft on bigger harmonic drives.
Most mounted have cross roller on opposite end. Drive with wave and other end cross roller for maximum support. All the robots i have seen that way.
here you go . Again 2 cross rollers for support . if cheap you can use a radial bearing
http://www.orientalmotor.com/images/mainContent/hg-gear-internal.jpg
Last edited:
Servos are better. servos cost more. servos are more complicated to set up.
thats all it is really. But not more accurate if stepper has encoder .
depends on the encoder,
a stepper has 200 defined steps per turn. microsteps are not precisely positioned. they are repeatable though so its not randomly picking a spot either.
a typical cheap servo has 4096 to 10000 defined positions per turn. higher end servos like siemens use encoders with 128000 to 4096000 defined steps per turn.
so servos are more accurate. no question. the trick is how relevant is the difference? 5mm pitch screw divided by 200 steps is .001" precision on a stepper. thats really about as precise as a typical hobby machine can hope for. you need a ground c5 ball screw to achieve that accuracy.
so, then you just add an encoder to a stepper. if the drive is smart, its now every bit as accurate as a servo with the same encoder. from there, the stepper is only worse at high speed, and still costs a few hundred bucks less and is easier to set up.
For feedback to the control... With no feedback, you are letting the servo drive to all the work, which in simplistic cases, With feedback, you can let the control do some tasks, like compensating for various errors, and even just display the exact position and not the "assumed" position. basically a control with feedback can be made smarter. They should have feed back . the drive takes in the encoder, and then passes it back out to the controller (or to slave another drive).
Most proper industrial controls have feedback to the control. a few cheaper ones dont. The siemens , Centroid , Mitsubishi all do . Alot of the cheap chinese controls do not have feedback to the control or cheap cards .
On that 4th axis bit more info. they dont have a ton of holding torque. the restriction is on the gearbox itself its seems. only 8nm. so the lower reduction is only giving you precision, not torque.
on 50:1 you'll get a flat 8nm from 0 to 40rpm or so on a good nema23 stepper/driver. his is roughly the same as a 400w ac servo on a 5:1 belt drive bot for about half the cost. your steps will be 160000 steps per turn with 16 microsteps.
Again , Easy servos with E/C feed back to controller are better than a servo on a router.
George
how is closing the loop at he control relevant? Do you sell a drive that has feedback output? Easy servos will output digital signal to what kind of control? Mach? How?
I would agree that a high end stepper "not a cheap easy servo" would be most accurate if restricted to its full steps. I think that since micro steps are constantly morphing then there would be constant "while very small" discrepancies in linearity.
Do yourself a favor and bail on the easy servo kick, the clearpath people are referring to would be amongst the best business decisions you've ever made, maybe next to threading the ballscrews
how is closing the loop at he control relevant? Do you sell a drive that has feedback output? Easy servos will output digital signal to what kind of control? Mach? How?
I would agree that a high end stepper "not a cheap easy servo" would be most accurate if restricted to its full steps. I think that since micro steps are constantly morphing then there would be constant "while very small" discrepancies in linearity.
Do yourself a favor and bail on the easy servo kick, the clearpath people are referring to would be amongst the best business decisions you've ever made, maybe next to threading the ballscrews
I have the Centroid controller as in some video. Also for less money Planet- usb controller. I do have another low cost Ethernet controller i got sent a sample sent to me , but not hooked it up yet. But all have encoder feed back
I sell any controller your looking for if needed .
Clearpath is just a brushless DC motor with an encoder. Nothing wrong with them at all but would only use those on single motor unit .
There is a lot of integrated motors and drive out there. Do you know who makes the motors for the Clearpath ?
With out feed back to controller, none of the motors will actually now where each other is on the table . With closed loop to driver, motor will just go into stall mode if something happens.Machine then will have to be powered down to reset and go back home .
I have servo from Mitsubishi to Leadshine.
Leadshine is far from cheap motor company lol , One of the largest motor company's in the world. Almost ever knock off is a copy of their drive systems and motors.
Again , i would take the easy servo on a router and i would without a doubt use like a clearpath on lathe headstock or something like that . Great for 5th axis spindle control
Funny you mention about threading on ballscrew, i do that now lol but only because now all parts are drop shipped , No press to press them on at place they are shipped to .
Do you know holding surface difference from M10 nut to 10MM clamp?
Sorry forgot, How do you thread motor shaft to ballscrew? As you say, clamp will not hold?
No slamming head on door is not good. What you say absurd. Use window is much better I say so must be true. Also, I have Acme controller but use Toshiba break board is better door is worst choice. Machine is by good window where much better to slamming head.
G
Here's hoping the new lathe and mill will have Clearpath and finer/slower lead screws.
Ball screws no like dust .
Ball screws no like dust .
Yes, hopefully there is a clearpath to it.
Must be the only person i hear saying servo with no feed back to controller is just as good as driver with feed back to controller .
George
Your a moron....i like you but your a moron. I actually didnt say that at all. What i meant to say is that your talking about stuff you've only heard about. I did say that mounting your ballscrew with a coupling is a half ass dis-service to your potentially fantastic business concept. I also said that your choice of electronics kinda blows and that there are much cheaper, much better products that would better suit you and your customers needs.
Its all good though, i still like you. Hell i check your thread on the zone almost everyday to see whats new.
RBC
Deceased
George,
He didn't say that, I did.
And, it's actually the truth. Especially if you're talking about motors like the leadshine stepper with encoder feedback.
In case you didn't know, Cleapath motors are built by Teknic. They are, and have been, a leader in servo motors and drives for automation for many years. Their Clearpath motors are not just a DC motor with an encoder on it. They are of extremely high quality and they come from an established solid company. They also don't cost very much more than comparable stepper systems, and they are much more reliable.
As for controls where the feedback goes back to the control
This thread is about 4th axis indexers. My comments originally were about 4th axis indexers. Your response to my comments were about which was more accurate, steppers or servos. I replied to your response, and attempted to redirect the conversation back to 4th axis indexers. So far, you've taken comments from experienced seasoned cuemakers and argued with them by referencing various unrelated information about things other than what the original poster is looking for. You made the claim that steppers were more accurate than servos and then said that servos are more accurate than steppers but you don't need it on a router. You flip flop like a politician.
Please, just let it go. Bring your belt driven rotary's to market and sell them. Some will buy them and some will not. The market is big enough for all kinds of different products to be successful. You don't have to make yours out to be better than everything else.
For what it's worth, I do have, and use everyday, a belt driven rotary axis. And yes, it's stepper driven. It works very well, but I wouldn't use it for any inlay work. We do use it to cut slots in our ring stock to make our stitch rings, but that's about the only index work we do on it. We do use it for turning, every day!
When it comes to inlay work, I currently use a worm gear rotary indexer. It works great, but it does have a tiny amount of backlash. Not enough to worry about, but it does wear and will need to be tightened up at some point. But, when I need another indexer, it will be a Harmonic Drive.
By the way, I don't find a cross roller bearing in the image you posted. Here's a cross roller bearing. And by the way, they tend to be very pricey!
http://timkenbearing.cn/UpFile/s2012116102050.jpg
Royce
RBC
Deceased
George
Here's something you don't seem to be aware of. Yes, on some controllers, the encoder feedback does go back to the controller, but do you know what the controller does with that information. On all of those systems, at least the ones that I'm a little familiar with like Fanuc, the motors and their positions are controlled by the drives or amplifiers. And yes, that's one drive for each axis and they operate independent of each other. The feedback that goes back to the controller is used in diagnostics. It's the only way the controller can provide "following error" which they all do. Even on my 1983 commercial CNC lathe. It can't know what the actual following error is unless it compare the commanded position with the actual position. It does not, however, use that information to alter the commanded position of that axis or any others.
Just so you know, I can do the same thing with Mach3. You can wire encoders back to Mach3. It can monitor them and even compare them to the commanded position. You can even set out of position limits and have it enter fault if it gets out of position too far. But, it doesn't alter the commanded position based on any one axis getting out of position.
As for Jake's comments on your ball screw mounting, I'm not sure you understand what he's talking about. I have your Raptor machine, and it's a good machine. We use it everyday. I recently installed your ballscrews and used your mounting system to do so. They are not attached to the machine by either threading or anything that needed a press to do. What do you mean when you say they are threaded. In what I have, the coupler is pressed (by hand) against the bearing and then locked down with the clamping screw. While this does work, it's working on mine right now, most of us would prefer to see threads and a nut to pull the ballscrew up against the double row angular contact bearing. While the clamping coupler works, it may also move or loosen at some time and you could develop linear movement in the screw. It works, but it wouldn't cost anything additional to design it like all the other commercial machines I've had experience with. It seems that you've ventured away from the norm and don't gain anything from it.
Royce
Hi Royce, i am not arguing about anything , i am just stating facts.
It has nothing to do with selling 4th axis or anything. Still don't have time to finish design or anything lol . Been made 2 years ago and still only one. Waiting for someone to make a good one and sell it so i don't have to.
I never said steppers are more accurate? if i did type that it was a mistake. ill read post later.They have low end torque which is what you need on a router .
I only reply to what is said.
I deal in lots of motor kits. Centroid controller has Teknic motors for low cost and Mitsubishi for high end. i have sold a few of the teknic /centroid set ups They are a small company compared to leadshine if your stating leadshine small company. Leadshine is mostly AC servos. But now are getting into low cost high powered .
Stepper with encoder is just as accurate as BLDC motor with encoder .
Depending on what your rest of your machine is.On C7 ballscrew, how much more accurate will a servo be compared to a stepper? . Steppers will be just as accurate as servo. C3 ground ballscrew, its a big difference. Basically any motor with and encoder is a servo, stepper /ac or BDC or BLDC. What ever has highest encoder count will be the most accurate
That was just a link the had open view of of the wave generator and bearings at end.
In that view they are using Toroidal roller bearings. Cross rollers are used in others. Yes, your right cross rollers are very expensive .
On the 4th axis you have, i was interested in what is in it for more 5th and 6th axis spindle than 4th axis rotary . Pretty much same as this on these machine. But price to high right now to use same components
https://www.youtube.com/watch?v=acLeVwQTPeg
Teknics motors are very low tech bldc servos with high temperatue magnets. They have pretty low continuous torque, but higher intermittent torque. the encoders are cheap magnetic ones. not bad motors for the cost, but not remotely special and nothing compared to high end motors from siemens or the other big companies.
The drives.. "clearpath" is a pretty generic motion controller built into the drive. 1000 other companies have the same thing. The clearpath are relatively low cost. Thats about all they have going for them vs other brands. the clearpath feature set is USELESS for cnc applications. you're wasting money on features you don't need or want. But prefect for running a programed set up with single motor unit .Then you would get full bang for the buck and great deal.
Here is Centroid with Teknic 750w servos .That is one of the controllers i sell
https://www.youtube.com/watch?v=UILYonYb7r0
I know what Jake was saying. Yes the new ballscrew are threaded like i used ti have before. Ballscrews are now dropped shipped. So they are now slip fit and threaded.
On coupling, which is the same as collar clamp. Has 100x more surface holding area then nut. I have to get handbook out and recalculate it . It done to cut down on shipping cost now as to why they are dropped shipped . The company i deal with are getting bigger and bigger and order are now moving into 2 to 3 weeks to get machined .So it speeds things up
Ballscrew which are press fit are machined. 9.98 -0 +0.005 slip fit is 10 +0 - 0.018
On a steel mill or something heavy load i would use collar clamp and nut .
if i was using the 7000 series bearings i would have stayed with nut, but 5000 series you don't need nut to adjust bearing.
- Status