Rotary 4th axis build componants & tech ?

[Cruiser]

I am ready to CNC my totary table and need help in selecting motor (servo or stepper) & sizing of motor to have appropriate torque for holding, & being able to achieve a useable feed rate that will keep up with, rather than slow down the "x" & "y". I have a 65V 20A power supply driving my IH servo'd mill. I will also need some advise on the selection of encoder, appropriate enough to break down the degrees of rotation fine enough to handle whatever need arrises. I don't know if i should concider gearing motor down or go direct drive, would direct be too fast for worm ? The rotary table i have is an 8" import. but seems plenty solid. I plan on obtaining gecko appropriate for motor selected. Now what am i not concidering, I don't know. Any assistance I can achieve here will of course be greatly appreciated. I would like to get this project underway soon. It must have been done numerous times already but i sure couldn't find any links. I don't want to end up under powered and won't mind being maybe a little over powered. I would like something that will do for a few yrs or more ..... Thanks !

[Kipper]

I think the 90-1 reduction is enough? If you have servos already then get another one. And a rotary will only use X+Z+A afaik. Mine is a mere 1.8NM and has no problems. hth

[Cruiser]

I must admit that i favor servo's, I don't quite know what 1.8NM (newton meters?) would be for torque but it sounds small. must be deceptive then, and applied to a 90/1 it could be quite appropriate, but that is why i'm asking as i am completely out of my line here and don't understand many of the factors involved with motor selecting and understanding the lingo describing the science involved. I would like to read any shortcomings of your system tho and if you did it over again what would you do differant ? I really would like to build this thing smartly, rather than wish i'd done it differantly later on .... I'm sure that there are others looking on wanting to do the same thing without paying for a ready to go system. My primary concern is not having enough torque or resolution, for whatever may end up requiring this machine's capability's. Can you provide some more details of what your rotory will do for you ? and thanks for your reply ! I do appreciate it !

[Geof]

Originally Posted by Cruiser .... I don't quite know what 1.8NM (newton meters?) would be for torque but it sounds small. must be deceptive then,...

It is deceptive.

The force is 1 Newton at a distance of 1 meter from the axis of rotation.

A crude approximation is that a Newton is about 1/4 lb force; actually it is 0.23 or something like that, 1/4lb gives you 4 ozs.

Another crude approximation is that a meter is 40"; again the correct figure is 39.370" but 4 and 40 are much easier to do mental arithmetic with.

So you have a force of 4 ozs at a distance of 40 inches. I am pretty sure you can see this is a reasonable torque when expressed in oz-ins. And please express it correctly; force first then length of lever arm as in oz-in, NM.

[Al_The_Man]

The two common methods of operating a rotary table is the axis always active, if you want to interpolate the rotary with the linear axis, the other is to disable the servo at position and use a brake, (non-interpolated machining).
The former requires more torque due to machining pressure.
If you figure out the maximum rotation speed you require and then calculate the maximum reduction you can use, based on the max RPM of your motor type, DC servo's usually max out at 3000rpm 4000tops.
Remember reduction multiplies the torque by the ratio of the reduction.
The other thing to consider, is, if you have decided on a reduction and the encoder is on the motor, you will need to select a resolution that is within the reading (speed) capability of the system you are using.
Al.

[Cruiser]

Thank you Geof, So if i reason this out correctly then moving the fulcrum in to 1" would be 40 times the force to hold, so 40 x 4 would be 160 oz in ? and at 1/90 then it could be 14,400 for use ? I sure i blew it here but it sounds good. I think (bewildered) How about the encoder, would a 1000 line be too much ? 100 to little ? I haven't broken down the units into minuits or seconds of rotation yet, but applying it to the motor ?

[Cruiser]

Al, I have a vague understanding of what you say here. I would like this system to be capable of dynamic interpolation, so then i assume a factor should be used for the "what if" and i don't have a what if to factor, hense a bit of over engeneering might be in order here ! as to the encoder selection, I will be tuning it into Mach 3, and the speed ? is this in pc / mach speed ? or ? This is getting just about as deep as i suspected ! Now i'm sure it is going to get much deeper ! Thanks lots guys and keep it coming.

[Al_The_Man]

I have not used Mach, but I believe there is also a maximum encoder rate limit that Your drives can operate at, as the encoder goes back to the drives.
The basic resultant quadrature count of the encoder is multiplied by 4 to increase the resolution. i.e. a 100pulses/rev = resolution of 400p/rev.
Al.

[Geof]

Originally Posted by Cruiser .....This is getting just about as deep as i suspected ! Now i'm sure it is going to get much deeper ! Thanks lots guys and keep it coming.

Yes, so you take it in bite sized chunks. And don't be afraid to do the ball park calculations I did just to get a handle on things. Also when things start 'not making sense' break off and do something entirely different: Letting the subconcious processor inside your brain chug over understanding things in the background can work wonders.

[Cruiser]

With some basic number crunching i came up with 129,600 units per revolution (@rotary) would equal 10 seconds or .00019" of 25.133" circumferance which would be also equal to the resolution that i understand my IH mill is computed to have and would be very acceptable ! So with 129,600/90=1440 for units @ 1 turn of motor/encoder .... now i'm lost ! ..... my system uses step & direction altho mach has the encoder quadrature capability and i'm sure that this would be useable, i just am not sure how to figure it from here for a proper encoder count selection. But, would a 500 count @ quad 2000 be a good choice then to better my 10sec resolution target ?
edited in: or 2000 count encoder using simple step & direction hookup, is there even such a thing ?

[Al_The_Man]

Originally Posted by Cruiser But, would a 500 count @ quad 2000 be a good choice then to better my 10sec resolution target ? edited in: or 2000 count encoder using simple step & direction hookup, is there even such a thing ?

I assumed you were using servo's, if using steppers then you probabally do not need encoders, as the resolution is decided by stepper.
BTW the correct term is 500 quad count x 4 = 2000. the term (quad)rature does not refer to the count in this context.
AFAIK Mach does not read encoder input. The feed back is to the drive.
I make it .002 per deg. for a (500/rev) 2000 counts encoder at 90:1 red.
Al.

[CarbideBob]

Cruiser,
A couple of questions.
Where did the table come from and what brand name is on the box? The tables from Harbor Freight are not as good as the ones from Enco or J&L and take a lot more torque to turn.
What are you planning on using for amps? Due to their design the Geckos don't like really high resolution feedback.
Remember that this is a worm gear mechanism so you you must run it with at least 2 to 3 thou backlash (slop) at the gear diameter. Worm gears can't be run at zero backlash or they will lockup.
Max RPM of the worm must be held to under 1500 RPM and at this speed they don't last long.
I've run 8 inchers with 100 watt AC servos belted 3:1 with 2000 line (8000 count) encoders. This will just turn the table when new and will occasionally lock up once things start to get sticky. I'm building one with a 200 watt motor now.
Bob