New Machine Build Changing X-axis envelope during machining, automatically. For machining long planks


Page 1 of 2 12 LastLast
Results 1 to 12 of 15

Thread: Changing X-axis envelope during machining, automatically. For machining long planks

  1. #1
    Registered
    Join Date
    Feb 2012
    Location
    Norway
    Posts
    7
    Downloads
    0
    Uploads
    0

    Default Changing X-axis envelope during machining, automatically. For machining long planks

    Hi everyone,

    I want to be able to machine and work on 2x4, 2x6 and similar boards with a relatively small CNC machine. Something around 15 cm Z-axis travel, 20 cm Y-axis travel and "endless" X-axis travel. I want to merge a typical CNC with a "Robot" to achieve this.

    My background is mainly in programming and electronics (B.Sc EE, nearly complete M.Sc EE). I have about 300 hours of experience with additive manufacturing, and some in subtractive manufacturing.

    Some goals:
    Extending X-axis infinitely by combining robot with CNC
    Able to do aggressive cuts in wood (and in time metal). "Driving" the material directly is out of the question.
    Able to work with common G-Code.

    Basically the needed action is this:
    The machine needs to move "work X-axis position" to increase X-axis travel.

    Planned first steps.

    Software:
    1. Learn more about G-code
    2. Learn more about motion generation.
    3. Figure out motion path generation that steadily increases x-travel, instead of using the "full" allowed travel often.
    4. Write software routines for that will be run on the CNC in "robot mode" - activatable by M-codes.
    5. Write a G-code interpreter/reader that manipulates normal G-code so that the new hardware can use it.

    Hardware:
    1. Build an actual 3 axis small CNC machine and make it work
    2. Figure out a suitable active work-holding solution
    3. Figure out a suitable "move work" - actuator/routine

    Own comments to first steps.
    Software 1, 2, 3: Basically fresh up on everything.
    Software 4: Probably the whole subroutine will be programmed in C or similar, and simply activated by a single M-code.
    Software 5: The modification of the G-Code can be tricky. Basically, once X-travel is nearly spent in one direction - insert M-code for robot move, change G92 temporary offset and maybe edit machine registers too. Long X-axis moves must also be split up into smaller parts, to make it easier for a script to modify the code.

    Hardware 1. Maybe I will just by a cheap china router to spare some time.
    Hardware 2. Very open for suggestion. Guessing pneumatics is a good start?
    Hardware 3. Also open for suggestion. Repeatabillity is super important. I think a pneumatic actuator, locking the work to the Z or Y axis (ZY plane) (of course releasing the normal clamp), - and then using the X axis for the move is the best option. I might try to use a repurposed computer mouse to somewhat verify the distance moved, but that is low on the priority list. (I have tested this before, and a normal computer mouse does not give a accurate and repeatable distance measurement - but it is more than good enough to verify that the work is moving and not slipping)

    This is a hobby project and the time needed to complete it is not easy to tell, wish me luck - I'm going to need it.

    Similar Threads:


  2. #2
    Gold Member Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    2828
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    First look at G54 - G59, work offsets

    Pneumatics is a really bad idea for positioning. Air is compressible and not good for positioning. Hydraulics would work fine, but certainly would not be my first choice.

    Some type of electro/mechanical drive system would be the best, a lot of options available, depends on the rest of the system design. Read the position with an encoder attached at the appropriate spot.

    Overall, what you want to accomplish seems very doable and not terribly difficult.

    Jim Dawson
    Sandy, Oregon, USA


  3. #3
    Gold Member
    Join Date
    Apr 2004
    Location
    Oakland CA USA
    Posts
    4630
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Take a look at the Carvewright machine - it does what you're talking about by feeding the board through its working area like a planer (although it doesn't work as well as commercial routers with enough room on the bed for long boards): CarveWright.com - 3D Carving CNC Router Systems

    Andrew Werby
    Website


  4. #4
    Registered
    Join Date
    Feb 2012
    Location
    Norway
    Posts
    7
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Quote Originally Posted by awerby View Post
    Take a look at the Carvewright machine ...snip... CarveWright.com - 3D Carving CNC Router Systems
    Interesting. Reminds me that I should plan for a vacuum system from the get go. Also like the adjustable Z height.

    Quote Originally Posted by Jim Dawson View Post
    Pneumatics is a really bad idea for positioning. Air is compressible and not good for positioning. Hydraulics would work fine, but certainly would not be my first choice.

    Some type of electro/mechanical drive system would be the best, a lot of options available, depends on the rest of the system design. Read the position with an encoder attached at the appropriate spot.
    Sure, I meant pneumatics for the clamps. I plan to use the normal X axis motor during repositioning. Planned routine is something like this:

    (Machine has normal clamp that clamps the work to the table, and extra clamp that clamps the work to the gantry)

    • Stop program
    • Record Z position
    • Lift Z axis to home
    • Make sure X axis has enough travel //possible move
    • Activate gantry clamp
    • Deactivate normal clamp
    • Move X +10 cm
    • G92 X+10 //Probably wrong syntax
    • Activate normal clamp
    • Deactivate gantry clamp
    • Move X -10 cm
    • Move Z axis to previous position
    • continue program


    If the clamps are doing a good job I think this can be done open-loop for most woodworking. Clamping a 2x4 or similar should be easy. Oak trim and similar will need something different.

    Quote Originally Posted by Jim Dawson View Post
    First look at G54 - G59, work offsets

    Overall, what you want to accomplish seems very doable and not terribly difficult.
    On a long plank I think 5 work offsets will be too few. , He mentions G92 temporary offset at 8:00, hence the idea to try to use that.

    I look forward to getting this project underway. The hardest part I think is the G-code post processor/modifier and debugging the hardware and software. Part of the challenge is actually starting. Chopping it down into manageable bits and starting somewhere is the first hurdle. I have had this project in my head for a long time. Hopefully writing it down accelerates the process.

    Thanks for your replies.



  5. #5
    Gold Member Jim Dawson's Avatar
    Join Date
    Dec 2013
    Posts
    2828
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Sorry, I misunderstood the use the pneumatics.

    I like your sequence of operation, written like a programmer

    One of the problems with G code generated from a CAM program is that many times it is not linear. It will work on one area of a part then move to another area and do some work then move back to the first area. Getting the G code to cut where and when you want it to will be a bit of a challenge. The one exception to this that I am aware of is using CamBam and turning off the optimization, it can be forced into cutting the exactly the way that you want it to. It might also be possible to create separate operations for each X offset thus forcing it to work only in the active area, so in that case any CAM program would work.

    One possibility you might concider is using feed rollers to act as the X axis. I think that is the way the CarveWright might work. I have positioned a lot of wood using rollers, +/- 0.5mm/M @ 200 M/min (high speed optimizing chop saws) At router speeds, I would expect positioning to be very good.

    I'm looking forward to watching your project develop.

    Jim Dawson
    Sandy, Oregon, USA


  6. #6
    Gold Member
    Join Date
    May 2005
    Location
    USA
    Posts
    3689
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Interesting ideas and concept.

    First I’d nix using an industrial robot to position the work piece in the machine. Assuminga moving gantry the gantry can easily be designed to grab the part and index it along. Even a moving table machine can do this you just need to think different. I’m thinking about bar pullers on CNC lathes here and similar concepts. There are many ways to accomplish this though. I’d also would research walking beam implementations.

    You didn’t mention much beyond the use of construction lumber so the question of what you will be routing into the wood comes up. A repeating pattern is pretty much a piece of cake. A random array of features on a 20 foot board might be far more difficult maybe even impossible. At least impossible with a high degree of accuracy. The problem comes from construction lumber being highly variable in straightness and twist.

    In any event the approach I would look at first would be a moving table design that walks the lumber through the machine. This would be easy to relate in a movie but I don’t have one. So here we go. Imagine that you have a long infeed tabletable/convetoy upon which the lumber sits (it is on your right). The moving table is moved to its right hand position, the lumber is rolled into the machine across the table until it reaches a clamp. The clamp is engaged and the moving table takes control of the lumber. You machine X amount of length at whic point the table moves to the extreme left (or a well defined amount) where a clamp on the infeed side of the machines frame engages. The lumber is now clamped rigidly to the frame, the moving table clamp releases and the table slides under the lumber to the extreme right again. The cycle is then started all over. You effectively walk the beam through the machine a predefined amount each time.

    Most G-code interpreting solutions out there can already handle something like this (you only need a few digital outputs) so no G-Code interpreter to write. If the goal is to machine repeated patterns into the lumber, that work within the tables travel, you effectively have a loop of instructions to repeat until the end of the beam. More random patterns would require more effort.

    Note the same general idea can be applied to a moving gantry machine. Most likely the walking beam would be more independent of the CNC.

    The big problem will be meeting feature accuracy requirements which you didn’t mention. The killer here is the rough nature of wood that makes precise positioning with robots or whatever problematic. You end up with slippage in the clamps or grippers and things move when the work clamps engage. I don’t ever see accuracy comparable to what you could achieve with a long CNC machine. Even a clunky cheapbut long CNC would work better.

    This brings up the question of why not just build a really long CNC machine? You would need to make sure the CNC controller can accurately track a 20 foot movement but that really isn’t a problem. The machine would actually be shorter as you don’t need guarded infeed and out feed areas (any method of beam walking requires this). This brings up an interesting point CNC controllers don’t have the capacity for infinite lengths. How far you can go depends upon the number of counts (16, 32 or larger registers) and the resolution of those counts. If you do 20 feet at a resolution of 0,001” , you need 240,000 counts. Actually you will likely need a lot more depending upon the drive method. In otherwords your controller needs to handle the length of the axis at the resolution you need plus the extra required by the drive method.

    In any event there are many possibilities for achieving what you want. A straight forward long CNC would offer the fewest possible complications. On the otherhand if your requirements are simple the walking beam approach is cool to watch run.



  7. #7
    Registered
    Join Date
    Feb 2012
    Location
    Norway
    Posts
    7
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Quote Originally Posted by Jim Dawson View Post

    One of the problems with G code generated from a CAM program is that many times it is not linear. It will work on one area of a part then move to another area and do some work then move back to the first area. Getting the G code to cut where and when you want it to will be a bit of a challenge. The one exception to this that I am aware of is using CamBam and turning off the optimization, it can be forced into cutting the exactly the way that you want it to. It might also be possible to create separate operations for each X offset thus forcing it to work only in the active area, so in that case any CAM program would work.
    Thanks for the tip about CamBam. For others checking in: It is free for use up to 1000 lines, with a licence running about 100 euros. Furthermore it can be expanded on with scripts.

    My summary is like this:

    Method_a: Engraving toolpath
    Pros: Easy to start with, will steadily climb in X-Axis
    Cons: Slow, poor surface finish

    Method_b: Common toolpaths
    Pros: Fast and good surface finish
    Cons: Not readily suited for my idea

    Method_a will work out of the box with a modifier, method_b will not.

    The preferred case is a script that will split the model into smaller models that will fit inside the work area.
    Example: Part length 60 cm. Non-recurring features. 25 cm X axis travel.
    Part is split into part_a, part_b and part_c.
    Separate g-code is generated for part_a - part_c
    A script makes a new master g-code file where the structure is: part_a, move X-axis, part_b, move X-axis, part_c.

    I need to learn more about the possibillities within CamBam, and this should be able to work.

    Quote Originally Posted by Jim Dawson View Post
    One possibility you might concider is using feed rollers to act as the X axis. I think that is the way the CarveWright might work. I have positioned a lot of wood using rollers, +/- 0.5mm/M @ 200 M/min (high speed optimizing chop saws) At router speeds, I would expect positioning to be very good.
    Yeah, feed rollers are a interesting option. I haven't talked about spindle power and cutting forces yet, but I plan to use a 3-5 kW spindle once the motion control is mature enough (My day job is in ebikes and emotorcycles, so I have access to relatively high power low weight motors). Routing is fun, but routing with high power is so much more satisfying. I believe feed rollers and high cutting forces are a no-go. Not too keen on making a pneumatic clamp, but more so than feed rollers.

    Quote Originally Posted by wizard View Post
    Interesting ideas and concept.

    Assuming a moving gantry the gantry can easily be designed to grab the part and index it along. Even a moving table machine can do this you just need to think different. I’m thinking about bar pullers on CNC lathes here and similar concepts. There are many ways to accomplish this though. I’d also would research walking beam implementations.

    You didn’t mention much beyond the use of construction lumber so the question of what you will be routing into the wood comes up. A repeating pattern is pretty much a piece of cake. A random array of features on a 20 foot board might be far more difficult maybe even impossible. At least impossible with a high degree of accuracy. The problem comes from construction lumber being highly variable in straightness and twist.

    In any event the approach I would look at first would be a moving table design that walks the lumber through the machine. This would be easy to relate in a movie but I don’t have one. So here we go. Imagine that you have a long infeed tabletable/convetoy upon which the lumber sits (it is on your right). The moving table is moved to its right hand position, the lumber is rolled into the machine across the table until it reaches a clamp. The clamp is engaged and the moving table takes control of the lumber. You machine X amount of length at whic point the table moves to the extreme left (or a well defined amount) where a clamp on the infeed side of the machines frame engages. The lumber is now clamped rigidly to the frame, the moving table clamp releases and the table slides under the lumber to the extreme right again. The cycle is then started all over. You effectively walk the beam through the machine a predefined amount each time.

    Most G-code interpreting solutions out there can already handle something like this (you only need a few digital outputs) so no G-Code interpreter to write. ..snip. More random patterns would require more effort.

    Note the same general idea can be applied to a moving gantry machine. Most likely the walking beam would be more independent of the CNC.

    The big problem will be meeting feature accuracy requirements which you didn’t mention. The killer here is the rough nature of wood that makes precise positioning with robots or whatever problematic. You end up with slippage in the clamps or grippers and things move when the work clamps engage. I don’t ever see accuracy comparable to what you could achieve with a long CNC machine. Even a clunky cheapbut long CNC would work better.

    This brings up the question of why not just build a really long CNC machine? ..snip..

    In any event there are many possibilities for achieving what you want. A straight forward long CNC would offer the fewest possible complications. On the otherhand if your requirements are simple the walking beam approach is cool to watch run.
    Thank you for the long and interesting post. Some of your questions reveal more of my motivations. CNC Lathe Bar pullers was part of my inspiration actually!

    Why not build a long CNC? I do not have the room for it. This machine will be small, probably about 20-40 kg and therefore mobile (which opens up cool possibillities aswell).

    I definitely see the problems with construction timber and accuracy. Step 1 is to select good material. Step 2 is to realize that this machine will be precise, but not accurate.

    Use, down the line, include cutting steel. This is a whole different game than wood. The accuracy can be expected to be better than wood, but then again I expect steel parts to be finished to a higher accuracy than wood. (Magnetics can also be used for manipulating the work)

    You mention some g-code interpreting solution already have the needed features for a moving gantry designs. My Google-fu did not find any relevant resources (just a lot of common gantry design discussion). Can you point me to some resources?



  8. #8
    Registered
    Join Date
    Jun 2011
    Location
    US
    Posts
    606
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Take a look at Tormach's new band saw. I think you might find it's work holding/feeding system interesting. If I was doing this, I think I'd implement it as a dual-vise hold-down system. One vise fixed, the other on a 4th (linear) axis. When cutting, both vises would be closed. When moving the workpiece, one vise or the other would always be closed. Hydraulic vises would do, but I'm guessing they're out of the budget, so something custom.
    Biggest issue I'd see is holding the workpiece down securely and consistently, possibly something on the gantry to push down between the vises before locking down a vise?



  9. #9
    Community Moderator ger21's Avatar
    Join Date
    Mar 2003
    Location
    United States
    Posts
    32938
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    but I plan to use a 3-5 kW spindle
    ..............
    ..............
    This machine will be small, probably about 20-40 kg
    A 5Kw spindle can weigh close to 20Kg. I'd expect the machine to weigh a fair bit more, if it's going to have any amount of rigidity.

    There's no point in putting a 5Kw spindle on such a small machine, as you'd only be able to use a fraction of it's power.

    Gerry

    UCCNC 2017 Screenset
    http://www.thecncwoodworker.com/2017.html

    Mach3 2010 Screenset
    http://www.thecncwoodworker.com/2010.html

    JointCAM - CNC Dovetails & Box Joints
    http://www.g-forcecnc.com/jointcam.html

    (Note: The opinions expressed in this post are my own and are not necessarily those of CNCzone and its management)


  10. #10
    Registered
    Join Date
    Aug 2005
    Location
    United States
    Posts
    12
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Sounds a bit like you’re re-inventing the wheel.

    You didn’t mention your budget, but a “hybrid” CNC can fairly easily be assembled with the use of an industrial robot as the heart of the system (see attached photo). This was a system I engineered and assembled largely from Ebay sourced components for doing sculptural work in wood. Add a rail as a compliment to the robot, and you’ve extended your work envelope considerably. For those that would cite accuracy issues; work in metals with this set-up is not a reasonable expectation. But the joinery that I generated on wood parts were on the order of + or - .002” or .003”. We’re talking about a quality of fit issue and not absolute accuracy.

    Regarding coding issues; I used a commercially available and supported bit of software to convert standard G-code to the machine specific Rapid code required by the robot. Little to no editing was required prior to running parts. Those times when the work piece needed to be re-indexed because of its size, I employed an inexpensive laser center/edge finder. With this set-up my only constraint was the size of the room the machine was in.

    My point here is that, if you have a need for milling large wooden (or similar density) parts, this approach will get you down the road fairly quickly. On the other hand, if this is only intended to be a mental exercise, then have at it. No doubt, there might be some interesting outcomes. Changing X-axis envelope during machining, automatically. For machining long planks-robot-cnc-station-jpg



  11. #11
    Registered
    Join Date
    Feb 2012
    Location
    Norway
    Posts
    7
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Quote Originally Posted by skrubol View Post
    Take a look at Tormach's new band saw. I think you might find it's work holding/feeding system interesting. If I was doing this, I think I'd implement it as a dual-vise hold-down system. One vise fixed, the other on a 4th (linear) axis. When cutting, both vises would be closed. When moving the workpiece, one vise or the other would always be closed. Hydraulic vises would do, but I'm guessing they're out of the budget, so something custom.
    Biggest issue I'd see is holding the workpiece down securely and consistently, possibly something on the gantry to push down between the vises before locking down a vise?
    Found the Tormach new band saw.


    I think pneumatics will give me the force needed to start. As for setup and a possible dual vise setup - I'll just have to try and check the performance. What clamping force would be advisable to shoot for? will 100 kg be sufficient for a first try?

    Quote Originally Posted by ger21 View Post
    A 5Kw spindle can weigh close to 20Kg. I'd expect the machine to weigh a fair bit more, if it's going to have any amount of rigidity.
    There's no point in putting a 5Kw spindle on such a small machine, as you'd only be able to use a fraction of it's power.
    You are probably correct in those assumptions. For reference the 5kW motor is 7 kg, but motorcontroller (2 kg), spindle (xx kg) and accessory hardware will quickly add up. The motion control would need to match the spindle to load it sufficiently, and the rigidity of the machine would also need to be suitable.

    Quote Originally Posted by MARV View Post
    Sounds a bit like you’re re-inventing the wheel.
    Yes, I agree. I'd rather someone set me straight now, than several months down the line.

    Then again, CNC is continually developing and while this machine will not make sense in a professional environment - it can be a fun addition to a hackerspace or hobby-shop arsenal.

    Quote Originally Posted by MARV View Post
    ..snip...

    My point here is that, if you have a need for milling large wooden (or similar density) parts, this approach will get you down the road fairly quickly. On the other hand, if this is only intended to be a mental exercise, then have at it. No doubt, there might be some interesting outcomes.
    Thank you for sharing your experience. This is meant to be a project to do complex stuff with simple means. I'm not quitting my day job or ordering a robot arm. This project might take 3 months, it might take 2 years - we'll see.

    Happy new year!



  12. #12
    Registered
    Join Date
    Jun 2011
    Location
    US
    Posts
    606
    Downloads
    0
    Uploads
    0

    Default Re: Changing X-axis envelope during machining, automatically. For machining long plan

    Quote Originally Posted by Storken View Post
    Found the Tormach new band saw.


    I think pneumatics will give me the force needed to start. As for setup and a possible dual vise setup - I'll just have to try and check the performance. What clamping force would be advisable to shoot for? will 100 kg be sufficient for a first try?
    Not sure for wood, might have an issue on heavy cuts. On metal it would be way too weak. A small Kurt style vise (even a Chinese knockoff) will have a couple thousand kg max clamping (for example, Kurt 4" vise has around 3500 kg max clamping force.)



Page 1 of 2 12 LastLast

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  


About CNCzone.com

    We are the largest and most active discussion forum for manufacturing industry. The site is 100% free to join and use, so join today!

Follow us on


Our Brands

Changing X-axis envelope during machining, automatically. For machining long planks

Changing X-axis envelope during machining, automatically. For machining long planks