[JavaDog]

Ok, so the CNC machine is finally done! I just need to order my Pulleys and Belts from SDP/SI. Problem is, even after reading through the catalog and selection-guide - I'm not sure what ones I should go with. So, I'm hoping I can get some guidance.

Information:

Spindle Sizes:

Y-Axis and X-Axis Ballscrews = 1/4" O.D.

Z-Axis Ballscrew = 3/8" O.D.

Servos (3) = 17/64"

Ballscrew Pitch:

Z-Axis Ballscrew:



X-Axis and Y-Axis Ballscrew:



Servos Specs:





Plus, because of how the servos are, a pully with a longer hub would be ideal. But, the hub needs to be less than 3/4" O.D.

Servo Mounted:

[JavaDog]

Almost forgot to mention, when I bought the Servos it was recommended to run with a 2:1 ratio...so I am assuming that is correct.

[JavaDog]

Hmm...holiday weekend = slow weekend...

[ger21]

Almost forgot to mention, when I bought the Servos it was recommended to run with a 2:1 ratio...so I am assuming that is correct.

Depends on the screws. A 5 tpi screw would need twice the reduction of a 10tpi screw.

[Al_The_Man]

Almost forgot to mention, when I bought the Servos it was recommended to run with a 2:1 ratio...so I am assuming that is correct.

Almost every servo manufacturer has free sizing programs Like Kollmorgen for example, but nobody seems to bother using them.
All you need to do is plug in your systems design and required feed/speeds etc.
It seems most build their machine, buy the motors and then wonder if they are the correct size.
Al.

[ZipSnipe]

See if this helps ya http://www.marylandmetric.com/tech/beltpulleycalc.htm
Maryland Metrics: JavaScript Calculator of Belt and Pulley Dimensions

[JavaDog]

Depends on the screws. A 5 tpi screw would need twice the reduction of a 10tpi screw.

Unless I am doing it wrong, it looks to me that my Z-Axis Ballscrew is 5 tpi; my X/Y-Axis Ballscrews are 6 tpi?

Almost every servo manufacturer has free sizing programs Like Kollmorgen for example, but nobody seems to bother using them.
All you need to do is plug in your systems design and required feed/speeds etc.

It seems most build their machine, buy the motors and then wonder if they are the correct size.
Al.

Well, I bought my servos first (they are from Jeff of HomeCNC) - before I had even designed the machine. On the recommendation of quite a few long-time board members. From the other machines I see them used in - they look to be pretty powerful. They weren't cheap, so I hope I wasn't mis-informed...

So, it is more that I am trying to figure out what type of belt, size, material; as well as pulleys - I should use. SDP/SI's GT series looks great, but unless I go metric, they are only 3mm. Now, that may be enough - but I'm not sure. They have a 5mm GT (but it is under Metric only) system - and looking at the specs it would be more than strong enough.

But, there is also the HTD series (and they have a 5mm HTD system). It's rather confusing, even after reading all the information from the catalog...

See if this helps ya http://www.marylandmetric.com/tech/beltpulleycalc.htm
Maryland Metrics: JavaScript Calculator of Belt and Pulley Dimensions

That's a handy little tool, a lot like what SDP/SI has on their site - only with the inclusion of Input/Output RPM...

[JavaDog]

Ok, for example:

Assuming 2:1, for my Y-Axis - which has a 4" center-to-center - this is what I get:



I went with the 3mm HTD - since that seems to be the one for quiet and strong.

On the Center Distance - Why does it say ' Actual: 2.4746" '?
You can change the number of grooves for Pulley-A, Pulley-B and the Belt - how does one know what is right?
Now, I am assuming I want Pulleys with Flanges - but the choices for Pulley-B are all without flanges?

[ger21]

Accordingto what I've read at SDP-SI, GT2's are the best choice, with the least backlash.

You want to take the speed you want to move at, and take the top rpm of your servos, and gear accordingly, so that your servos are spinning at close to their max rpm when your machine is making rapid moves. There is no general rule, like use 2:1.

Also, I see about 6-1/2 turns per inch. Maybe a 4mm pitch? You can fine tune it with the software once evreything is up and running.

[JavaDog]

Accordingto what I've read at SDP-SI, GT2's are the best choice, with the least backlash.

Ok, so GT2 belts it is.

Going with the 3mm GT system (largest in the list, no GT2 listed) - I get this:



I changed the Belt Grooves to get the Actual center distance as close as possible to my Desired - I'm guessing this is why you can change that value!

I also see now that you can change the Pulley Grooves to maintain the ratio you desire...

You want to take the speed you want to move at, and take the top rpm of your servos, and gear accordingly, so that your servos are spinning at close to their max rpm when your machine is making rapid moves. There is no general rule, like use 2:1.

Ok, so on the servo spec-sheet it says that the rated speed is 2150rpm. But, on the chart/graph - well I honestly can't figure out what speed is ideal. Looks to me like around 700rpm? But, if I am reading the chart right - 700rpm would be 85oz/in? That seems really really low...

Also, I see about 6-1/2 turns per inch. Maybe a 4mm pitch? You can fine tune it with the software once evreything is up and running.

On the X and Y Ballscrews, right? Yes, they are 4mm pitch. I'm not sure of the pitch on the Z-Axis ballscrew.

But, you saying I can work all that out in software - just gear them all the same (same pulley-pitch, belts, etc)?

[Ken_Shea]

JavaDog,
I do not see any mention of the encoder pulse used or the drive, it has been my (confused) understanding that all these components need to be optimized to one another including the controller, I presume it is Machx.
These include
1) Encoder Pulses
2) Servo no-load rpm
3) Drives (mine operate in quadrature(what ever that means, in my case it equates to 2000 pulses per second with 500 count encoders)
4) Screw TPI
5) Controller PPS capability, my reading finds that 25K in the MACH3 is best.

Also, that optimally you would design around using 80% of the servos no-load speed.

As ger21 pointed out, gear ratios should not be arbitrarily chosen, but selected with the total system in mind.

confusing, huh

Ken

[Al_The_Man]

Although many do not go through the mathematics of designing a system, it is however usefull to get an idea of the demands placed on the machine in order to at least get a feel for what the outcome will be.
The highest demand on the servo is when the motor accelerates or decelerates and the amplifier has to inject a high degree of energy in to the motor due to the Inertia of the load. And the current practice is to keep at least an inertia ratio of motor to load below 10:1
The advantage of reduction is the inertia is reduced by the square of the reduction.
This is a PDF that may at least give an idea of how a system is designed, even if you don't bother with the math portion.

3) Drives (mine operate in quadrature(what ever that means, in my case it equates to 2000 pulses per second with 500 count encoders)

This term is a bit mis-used as an incremental encoder is a quadrature encoder, even before you take it out of the box.
It refers to the phase difference between the pulses and not the count etc.
If you use the basic 500 counts, it is still a quadrature encoder.
If you multiply the pulses by four, then it is quadrature x4.
I can understand the confusion, as I have seen at least one web site that states the wrong definition.
Al.

[Ken_Shea]

Al,
So then it is not the drive but the encoder that operates in quadrature ?

Differential encoder the same ?

The highest demand on the servo is when the motor accelerates or decelerates
Is this the reason for the suggested 80% operating rpm.


Thanks
Ken

[Al_The_Man]

Differential quadrature encoder are not necessarily the same thing. There is a 90° phase difference between pulses (quadrature). The differential refers to the type of *drive electronics*, the alternative is single ended.
Differential has an inverted quadrature pulse for each of the pulse outputs, this is for noise reduction effects, single ended quadrature is only one out put for each A,B,Z, pulse etc.
I don't know where the 80% came in, I have operated servo's up to the max. rated rpm, as long as the load/inertia demands are not exceeded.
Al.
BTW by 'Drive Electronics' I mean the Driver I.C. in the encoder itself.

[Ken_Shea]

BTW by 'Drive Electronics' I mean the Driver I.C. in the encoder itself.



Thanks for the clarification, that did result in further confusion when the post was read.

[ger21]

I think what Ken was trying to say, is to also make sure that MAch3 can output pulses fast enough to accomodate the gearing you choose. This is dependant on the drives and encoders you're using. 320's or 340's? And I think Jeff was using 250count encoders?

If so, that gives you 1000 steps per rev of the motor. Depending on your PC, Mach can output a MAX of 45000 steps per second. 45*60 = 2700 rpm. Faster than your motor. If you run your motor at 1800 rpm, that's (1800 * 1000)/ 60 seconds = 30,000 steps per second. Run Mach3 in 35Khz mode and you should be fine with either drive. (Provided they are 250 count encoders)

Say, for example, you want to spin your motor at 1800 rpm. And you want your Z axis to rapid at 100 ipm. 5 turns per inch * 100 = 500 rpm at the screw.

1800rpm / 500 rpm = 3.6:1

Cut that in half to 1.8:1 and you should be able to get 200ipm rapids on your Z-axis. But, cutting the gearing in half cuts your torque in half.

Like Al says, there are formular for finding out the required torque for your machine to move at a certain speed. Decide how fast you want to go, find out if your servos are capable of that, and gear accordingly. A little math involved, but it may give you a much better understanding of what you're doing.

You're not reading that chart right, (Not a torque curve chart) the torque curve on a servo should be relatively flat throughout the rpm range.

[JavaDog]

Ok, tons of good info there to digest, thanks guys.

I am using Gecko G320 drivers, and the encoders are US Digital 250 CPR units.

Using Mach3 on a P4 1.6Ghz machine.

[JavaDog]

Say, for example, you want to spin your motor at 1800 rpm. And you want your Z axis to rapid at 100 ipm. 5 turns per inch * 100 = 500 rpm at the screw.

Why wouldn't I want to run my Servos at their rated rpm (2150rpm)?

1800rpm / 500 rpm = 3.6:1

Cut that in half to 1.8:1 and you should be able to get 200ipm rapids on your Z-axis. But, cutting the gearing in half cuts your torque in half.

Ok, I'm not sure if 200ipm is even neccesary on the Z-Axis (I know it was just an example). Also, if cutting the gearing in half also cuts my torque - that would seem to me that I would just be wasting what power my servos have.

So, for the X-Axis and Y-Axis - since they are 6tpi (from what I see):

6tpi * 200 (for 200ipm) = 1200rpm at the screw

2150 (Servos rated RPM) / 1200 = 1.8:1

So, a 1.8:1 reduction with the belt drive would give me 200ipm rapid without cutting the torque my motors can put out, right? 200ipm rapids would, imho, be pretty darn good...and at slower cutting speeds there should be plenty of power on tap, correct?

Like Al says, there are formular for finding out the required torque for your machine to move at a certain speed. Decide how fast you want to go, find out if your servos are capable of that, and gear accordingly. A little math involved, but it may give you a much better understanding of what you're doing.

Yeah, I've read the PDF Al linked to through a few times. I get what it is saying and I see on the chart the rotor inertia listed - but how do I get the other numbers to plug into the (OMG Math) formula? I mean, it's not like I can call the machine's manufacturer!

By the way, Al - I'm seeing now why using a "one size fits all" 2:1 isn't the right way to do this, so thanks. It seems like in homebuilt machines and on this forum, I've yet to see all this come up - it makes my wonder how many people have mis-matched systems and aren't getting all they could from their set-up.

You're not reading that chart right, (Not a torque curve chart) the torque curve on a servo should be relatively flat throughout the rpm range.

Ah, ok...good to know.

[Al_The_Man]

Yeah, I've read the PDF Al linked to through a few times. I get what it is saying and I see on the chart the rotor inertia listed - but how do I get the other numbers to plug into the (OMG Math) formula? I mean, it's not like I can call the machine's manufacturer!

This is why it is a good idea to play with one of the graphic sizing programs like The Kollmorgen one I mentioned, it gives a good sense of what you may expect, even if only to get the gist of what may happen.
You plug in your machines characteristics/dimensions etc, so you don't need a 'Machines Manufacturers' specs, you have them.
Al.

[JavaDog]

This is why it is a good idea to play with one of the graphic sizing programs like The Kollmorgen one I mentioned, it gives a good sense of what you may expect, even if only to get the gist of what may happen.
You plug in your machines characteristics/dimensions etc, so you don't need a 'Machines Manufacturers' specs, you have them.
Al.

I downloaded and ran the Kollmorgen software and plugged in what I could from my CNC Machine's specs. I filled in what I could and got this:



It is worth noting, however, that the software doesn't seem to allow for gear reduction/belts when configuring - just direct-drive.