Shop OBEX P1 Docs P2 Docs Learn Events
Linear drive Screw questions. — Parallax Forums

Linear drive Screw questions.

chaosgkchaosgk Posts: 322
edited 2009-04-29 15:13 in General Discussion
I am currently working on a CNC machine for several applications, one is to do the pcb drilling on my relay board project.·
The CNC frame i'm building is all aluminum to save weight and prevent rust.·· I have the X and Y axis slides done, and I'm working on the Z axis now.
I am using threaded rod to drive each axis. X and Y are fine, the Z axis is going to weigh about 15 lbs with the router and all of the bearing blocks and the Y axis slide.·
Does anyone know how to figure out how many pounds my motors will be able to lift.
There will be a pair of motors for the Z axis, one on each side of the CNC frame.


Here is what I know.

Diameter:· 5/16"
TPI: 18
Motor torque each:· 160 oz/in.

I will be at the point of attaching my motors this weekend and I don't want to have to buy new motors if I don't need to.
This CNC will only be used to route wood, drill PCB's (separate attachment), and the Y axis rail will be detachable so I can use this for foam wing core cutting for RC Planes.

The reason for the motors on each side to lift the Y rail is so that when it is in Hot Wire mode, each axis can move on it's own to allow my wings to have different roots and tips.


·
«1

Comments

  • Carl HayesCarl Hayes Posts: 841
    edited 2009-03-12 20:01
    I guess I could just answer this, but I'd rather step you through the calculation so you'll be able to do it again with different numbers.

    First, though, a note on dimensional analysis.· Moments (torques) can be measured in oz in, but not in oz/in because you calculate it by multiplying, not by·dividing force by distance.· There are quantities that can be measured in oz/in -- spring constants, for example -- but torque is not among them.

    Now then:

    Let's assume your torque comes not from a motor but from your finger pressing on a moment arm.· Let's make this arm one inch long.· To get 160 oz in of torque, you must press this one-inch lever with a force of 160 ounces.

    Each time you turn through 360 degrees (one full turn), you will have moved the end of the lever through a distance equal to the circumference of a 1-inch-radius circle.· That's 2 pi inches, or about 6.28 inches.

    How far did the load move?· You went one full turn on an 18-thread-per-inch jackscrew, so the load moved 1/18 of an inch.

    Your applied force went 6.28 inches, and your load went 1/18 inch, so the mechanical advantage is:

    ···· 6.28 / (1/18)···· =····· 6.28 * 18·····=····· about 113.

    The resultant force is your 160 ounces x the mechanical advantage of 113, so it can lift 160 * 113 ounces, or 18080 ounces, which is about 1130 pounds.

    That's in the absence of friction.· In reality there will be friction, so I'd reduce it considerably to be safe.· I wouldn't load it with more than, say, 1/3 of the maximum no-friction load.· That's just more than 360 pounds.

    Since you say your load is actually about 15 pounds, I'd say you're home free.· It ought to work.


    (Edited to correct my own arithmetic errors)

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net

    Post Edited (Carl Hayes) : 3/12/2009 8:26:30 PM GMT
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2009-03-12 20:30
    chaosgk,

    Are you sure you want to use threaded rod for your leadscrews? Granted, it's cheap, but I think you're asking for trouble as the threads wear. I know this, because I suffered for years with a CNC-conversion Sherline mill that used leadscrews having 1/4-20 machine threads. You may want to consider either Acme- or square-threaded leadscrews or (best of all) ballscrews. Square threads are more robust than machine threads for a given size and have a flat drive surface that wears more slowly and requires less torque for a given load. Acme threads are somewhat cheaper and less efficient, but their truncated threads provide nearly the same robustness as square threads. Ballscrews are the most efficient and can be adjusted for zero backlash, a fact you'll come to appreciate when you start machining.

    You can often obtain good-quality leadscrews for pennies on the dollar by scouring the surplus houses. It will be worth the trouble, believe me.

    -Phil
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-12 21:08
    I've used 1/4x20 all thread for this type of stuff before. You can address those concerns as follows.
    Check with your local plastics distributer you can get machinable plastic that's easy to work with. Use this as your moving piece, it has less play and will not damage the threads over time. You should also get a piece of stainless steel all thread. It has a less abrasive surface that what you might find at home depot.
    I've attached some pictures of my application of this in my IR-Ranger project.

    Now, all that being said, I got a sale flyer the other day from Enco and Amce threaded rods are on sale for very cheap. A 3' 1/2"x 10 was $4.99 and they have the nuts to go along with it.

    try: http://www.use-enco.com/CGI/INSRAR?PMSECT=0000000451

    Let us know how this turns out.
    BTW, seems like for your one-off relay project it might be simpler to just use a per-relay drilling template.
    But why do that when a CNC router hangs in the balance? I love this hobby!
    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    When your problem seems insurmountable, read the forum for a while.
    Solve someone else's problems and yours won't seem so bad.
    1200 x 1600 - 196K
    1200 x 1600 - 185K
    1200 x 1600 - 218K
  • chaosgkchaosgk Posts: 322
    edited 2009-03-13 03:00
    Carl,

    You know I sat with one of the guy I work with today that has a degree in math and physics and he could not think of the equation to make this work. I was at least on the right track, I was trying figure it with the circumference of the screw itself, not at 1in. I knew I had to factor in how far it had moved in the distance it had traveled, but I got stuck there.

    To answer the other questions about using threaded rod. For the first few projects, the threaded rod should be fine, but now that I'm seeing the ACME rod for $1 a foot, I'm going to order some.
    The mounts I'm making for the bearings and motors will be easily replaceable just for upgrades like this. I will be ordering some of the 6' pieces though right away.
    My other option was to use a piece of HDPE plastic and thread that to match the all thread. For what i'm doing, the backlash shouldn't be a big concern yet. I know HDPE may not seem like a good thing to use or it is a little soft to thread well, but recently we have had some -20(f) weather here and I left some outside to see if it would cut easier and wouldn't you know it, It cuts like wood at that temp and breaks pretty easily if you hit it with a hammer too.
  • Carl HayesCarl Hayes Posts: 841
    edited 2009-03-13 05:06
    chaosgk said...

    You know I sat with one of the guy I work with today that has a degree in math and physics and he could not think of the equation to make this work. I was at least on the right track, I was trying figure it with the circumference of the screw itself, not at 1in. I knew I had to factor in how far it had moved in the distance it had traveled, but I got stuck there.
    Well, you could use any nonzero·radius at all for the calculation, and any degree of rotation.· Just for fun, let's redo it at the median radius of a 5/16-18 screw thread, and rotate only one radian instead of a full turn.

    The 5/16 screw thread has (approximately)·a mean diameter of 0.29 inches.

    So the radius then is 0.145 inches.· For 160 in oz of torque, you would use a force of 160/0.145 ounces, or 1103 ounces.· That's about 70 pounds.

    How far do you push that 70-pound force around?· You push it through a 1-radian arc (that's 180/pi degrees, 57.2957795131 degrees approximately).· The length of a 1-radian arc is equal to the radius, so you push 70 pounds for 0.145 inches.

    How far did the load move?· You rotated the jackscrew 1 radian, or approximately 57.3 degrees.· That's about 0.16 turns.· In a full turn·the load would move 1/18 inch or 0.055555 inches.· You turned only·0.16 turns, so it moves 0.16 x .0555 inches, which is 0.00888888888 inches.·

    You pushed your 70 pounds through 0.145 inches, and the load moved 0.0088888 inches, so the mechanical advantage is now about 18.125.··Multiplying your applied force of 70 pounds by 18.125, we find that the jackscrew can lift 1268 pounds.

    Before, we calculated 1130 pounds.· Now, with different numbers, we calculate 1268 pounds.· That's a 12% difference, and·it is attributable entirely to the different rounding errors in the two calculations.· Of course the load that the jackscrew can actually lift does not·change merely because we calculate it a different way.

    Homework questions:

    (1) does the load that the jackscrew can lift depend on the diameter of the jackscrew?

    (2) does the load that the jackscrew can lift depend on the thread pitch of the jackscrew?

    (3) does the load that the jackscrew can lift depend on the torque we apply?

    (4) does the load that the jackscrew can lift depend on the thread profile, assuming that the threads are strong enough?

    (5) does the torque that we must apply depend on the diameter of the jackscrew?

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    · -- Carl, nn5i@arrl.net

    Post Edited (Carl Hayes) : 3/13/2009 5:22:39 AM GMT
  • chaosgkchaosgk Posts: 322
    edited 2009-03-13 05:18
    Oh yeah, cannibal, to answer some of your comment on the drilling of the relay board, you should really go read my thread on the Relay Board in The Sandbox. it's pretty interesting. There will be around 500 relays in this system, with about 15 drill points for each one spread out in clusters of 56 or however many I decide to put on each of the slave boards.

    Total 7500 drill points that I would rather spend my time making the machine to do it for me then do myself. [noparse]:)[/noparse]
    and it's also a good reason to finally build my CNC machine like I wanted.


    I did do a simple test with my HDPE plastic.
    It probably sounds dumb, but what I did was take some standard plastic cutting board from walmart, cut it into 2"x3" sections, melted the surfaces with my heat gun on high.
    Put them in the vise and let cool in the snow for a while.

    Brought them back in and quickly drilled and tapped the hole.
    It is about 1 1/4" thick and smoother then the coupling nut I was going to use. I couldn't feel any play or backlash in it and I put as much weight as I could on the bolt I was using and it still turned freely by hand. The weight I put on it is far more then will ever be put on it during use with the minimal 15-20 lbs that will be lifted.

    So I think I will be using this instead of going with the ACME thread. My CNC is modular enough to allow swapping out parts with upgrades without too much trouble.

    Coming to my next question.
    The way I have this CNC setup, for the Z axis (up and down) I am planning on using two stepper motors, one on each side of the outside rails. Can stepper motors be run in parallel off a single driver to sync them and have them each lift a side? I have plenty of matched stepper motors and would prefer to do it this way so in the future when I cut foam with the hot wire, each or the motors can work independantly for X and Y on each side, then when it is time for routing and drilling, link them in parallel to work together.

    Post Edited (chaosgk) : 3/13/2009 2:18:56 PM GMT
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-13 15:15
    You can drive the steppers in parallel but you can't put two motors on one high current driver. One of the problems with this approach is that you've got to be very sure you will never miss a step which means pretty beefy motors. I'd look real hard at a shaft encoder and feed back loop. Can you design it with a single motor, you'll be better off? Maybe a belt driven cluster of lead screws?
    Here's the problem, let's assume one of the motors misses just a single step on each up down motion and you are using a 1/2"x10 threaded rod. With a typical 1.8 deg stepper that's a complete lead screw revolution every 200 movements or 0.1 inches off. Depending on the distance from your lead screw mount to your cutting tip you will rotate the tip off of vertical and push it in at a diagonal. If your using a drill bit it will tend to float off center. If your using a mill bit it will start drilling ovals. Without feedback, you won't know about it until you've drilled an entire board.

    I'd avoid the two motor approach if possible.

    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    When your problem seems insurmountable, read the forum for a while.
    Solve someone else's problems and yours won't seem so bad.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-13 19:12
    After seeing the calculations from carl, I should be able to use just one side of the z axis to lift the entire y assembly. I did some testing with it last night and it seemed to be fine lifting just one side, or like you said, work out a system of pulleys to turn the lift screws.

    Post Edited (chaosgk) : 3/13/2009 9:51:00 PM GMT
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-16 13:27
    I saw a picture of a CNC router design yesterday thumbing trough an old Servo mag. It used an 'A' typr arrangement and I thought of your project. Here is a very bad but quick idea of the design. The nice part about it is that it hangs the weight of the router very evenly on the stepper motor.
    (I'm not on a computer with a real CAD tool so sorry for the sloppy MS.Paint rendition.)

    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
    960 x 1280 - 29K
  • chaosgkchaosgk Posts: 322
    edited 2009-03-16 22:27
    I'll take some pictures of my CNC i'm working on tonight. The X and Y Axis rails and slides are done, the Z axis would have been done, but I had problems with my aluminum welder, then I solved those, but NEEDED some aluminum of the same size as the Z rails for another project for work,·so I had to sacrifice·one of the rails, I will be cutting a new one tonight.
    Can we post video here?
    ·
  • PhilldapillPhilldapill Posts: 1,283
    edited 2009-03-16 23:29
    I think you can post a video on YouTube, then link to it.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-17 01:50
    Ok, that should work, I'll try to get something up tonight. I'm just about done with my wire spool rack for my actual job. We've got around 100,000 feet of cable to spool off in the next week, so I really need to finish it.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-19 23:40
    Here are some quick pics of the CNC project, I will be working on either the router rails and slide tonight or the X axis drive screw or something else on it.
    Let me know what you all think.
    I decided to use standard 5/16th all thread for the screws with my threaded HDPE for the nuts to move it, mainly because of budget reasons and other projects that need the fun money more. (ie, the relay board project.)

    The current plan is to have steppers on each side of the Z axis vertical·rails which will allow them to act independently when running as a foam wing cutter.· The X Axis will also have dual steppers, one for each side of the rails, they will be detachable, along with a single stepper in the middle that will be used for running both X axis slides at the same time for routing.·

    I am also planning on one stepper to drive the Y axis that is connected to the Y rail itself so when the Y rail is removed, the X axis can operate on their own, forward/backward, up/down.
    I have to work on the attachment for the router assembly, it will be mounted to the Y axis bearing block· and have it's own motor for controlling up and down on the router.·
    When the dual X axis slides are in the routing position, I will use the drive screws to lock it in (Y rail)·the up position and allow the router stepper to conrol the up and down of the router or rotary tool, whatever is attached.

    Does this make sense?
    855 x 642 - 135K
    855 x 642 - 142K
    960 x 634 - 219K
    694 x 470 - 105K
    855 x 642 - 112K
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2009-03-20 00:12
    Your slide bearing arrangement is pretty clever! I wouldn't have thought of doing it that way. No critical machining required! smile.gif

    -Phil
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-20 01:10
    Coming together very nicely.

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2009-03-20 01:10
    chaosgk--

    Do your steppers have encoder feedback?

    Would you use ball screws if money was no object?

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-20 03:53
    Of course i'd use ball screws if money wasn't an issue.
    I got some of the ideas for the slide bearings from buildyourcnc.com he had some good ideas there, but was pretty obvious he didn't have access to any of the tools or machines I have in my garage. Aluminum welder (which I am just now learning to use nicely), drill press, digital measuring. Everything seems pretty sold on mine so far and I can adjust the bearings on all sides if anything gets loose from use.
    My steppers do not have feedback, I'm trying to do this on as little of budget as I can. The current purpose is simply to be able to drill thousands of holes in the same place on my pcb's i'm working on.

    I didn't get anything done on it tonight, I got caught up with a movie and by the time I got out there I had lost ambition for anything except grabbing a couple steppers to ponder over in the house.

    Just an oddball question, does anyone else take parts of what they are working on into the bathroom when you are going to be sitting around there for a while? I find it's a good place to think [noparse]:)[/noparse]
  • chaosgkchaosgk Posts: 322
    edited 2009-03-20 04:11
    Here's an updated picture with the descriptions of what the parts are.
  • Phil Pilgrim (PhiPi)Phil Pilgrim (PhiPi) Posts: 23,514
    edited 2009-03-20 04:23
    chaosgk said...
    Just an oddball question, does anyone else take parts of what they are working on into the bathroom when you are going to be sitting around there for a while? I find it's a good place to think [noparse]:)[/noparse]
    More commonly, in the BR, a crossword puzzle to get my mind off of what I'm working on. It helps the subconscious process things in the background!

    -Phil
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2009-03-20 14:25
    chaosgk--

    Gee. All of that is in the bathroom? [noparse]:)[/noparse]

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-20 14:39
    Bathroom?! How 2002, that's the "Library!"

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-20 17:23
    OK, in all due seriousness. This goes back to·eliminating precise machining and eliminating trouble mantaining parallel on your X-Y table, here is a rig I used that works incredibly well. It's a throwback to the old drafting board days and was used to keep parallel bars parallel. I'm surpized it's not seen more as it's cheap and very reliable and simple. It uses a cable that snakes around the perimeter and crosses in the middle of the bar. It's ingenious in that the once it's anchored down and taught there is no way the bar can slip out of parallel without movement of the anchors.·I used it on an X-Y table I built. using nylon drawer pulleys (1/4" bolt hole inside of bearing) and steel 'brake' cable from a bicycle. I built the whole thing for less than $20 and it uses one stepper motor and has incredible accuracy.

    If you want some more photos I'll be glad to send.

    Keep us posted on your progress.
    Jim-


    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.

    Post Edited (CannibalRobotics) : 3/20/2009 6:13:49 PM GMT
    518 x 610 - 31K
  • Bill ChennaultBill Chennault Posts: 1,198
    edited 2009-03-20 22:20
    Jim--

    Well, that is pretty cool!

    --Bill

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    You are what you write.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-20 23:50
    Can you send some more pictures? I'll have to lay it out to see how it works, but basically, you drive one side and the tension of the cables forces the other side to run parallel to it? That seems like a good solution for when the CNC is in router mode, but switching to foam wing cutting looks like it would be very annoying to unstring all of the cables, maybe not, I'll have look at it more.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-21 01:43
    Ok, so after going out and looking at my CNC, I can see how it would work to have one drive screw on one side and the entire routing of the cable. Like Bill said, that is pretty cool. I think it would only easily apply to the X axis as the router may be doing 3D routing and the cables could interfere with the wood going up and down.
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-21 15:30
    OK,
    you happened to get me at the shop with nothing but a camera phone, my wife has the good camera. So, I dug out the table and took some pictures but the detail is not great. I can add some more probably Monday if you want. The project is kind of dirty too as it's been idle for a while.
    The pics were too many and too big to post here so I put them on my web site. I also have a detailed photo study of the entire frame construction as well. I'll try to dig that up too for reference. Go see.

    http://www.cannibalrobotics.com/x-y.htm

    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-21 17:38
    OK, to follow up.
    I did find all of the construction pics of my X-Y table. They show the lead screw version before I changed to the belt drives. This would be more applicable to what you are doing. I'll post those to my website as the photo quality is better and it really shows how it was all put together.
    Here are a few to give you the idea.
    I'll let you know about the rest.
    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
    2048 x 1536 - 1M
    2048 x 1536 - 1M
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-22 14:51
    chaosgk
    What is the plan for software? Are you planning on some type of G code interpreter or just hard coded steps.
    The reason I'm asking is that I'd like to get a prop driven G code interpreter working.
    j-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
  • chaosgkchaosgk Posts: 322
    edited 2009-03-22 15:45
    I haven't decided on software or stepper drivers yet so I'm not sure exactly what i'm doing there. I'm going to get the CNC itself finished with motors and everything then worry about how to make them move. Any suggestions here for a 4 channel driver, cheap?
    By the way, I checked out your site, those are some pretty ingenius setups on there. I started on the drive screws for mine last night, but I really like your belt drive system. where did you find the belts?
  • CannibalRoboticsCannibalRobotics Posts: 535
    edited 2009-03-22 17:45
    Well,

    I think I'm·actually heading back to lead screws on this one eventually. I put the belts on to increase the movement speed as the 1/4" x 20 lead screws were too slow. I've ordered some 3/8 x 12 Acme screws and nuts for this.

    The belts are faster for sure but lack accuracy and power so they slip·steps frequently. My application could work well on 1/16"·accuracy but the belts could not mantain that either. They also tend to be thermally sensitive. They would sag on hot days.

    I got them at McMaster Carr: http://www.mcmaster.com/#1840k3/=14nvdu

    they also sell the·cogs for the belts. If you've never seen their website, it's a wonderland. Small Parts, Grianger and Northern Tool all rolled into one.

    I'd discourage the belt approach for your application though.

    Jim-

    ▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
    Signature space for rent!
    Send $1 to CannibalRobotics.com.
Sign In or Register to comment.