. .
  1. #1
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 2 Days Ago Has been a member for 0-1 years. Has a total post count of 11.
    Hi All

    I'm Ian. I've been reading various posts on the forum and find it very helpful - thanks to everyone who contributes to this body of knowledge.

    I've always been keen to build my own CNC Router and have now started this with more intent as I need to build a machine to produce to support my R&D projects and manufacture smalle batches of prototypes (GPS tracking devices for livestock). I started off with a small X-Carve machine and Fusion 360 6 weeks ago and quickly realised that I will need to / want to buil my own machine. I've since build a first iteration machine (800mm x 800mm using 40mmx80mm extrusion, lead screw, v-wheels and acetal sheets for plates) and made a lot of costly mistakes, but have also learned a significant amount about CAD, CAM and machine operation. I'm now moving to design and build a second iteration.

    The ultimate aim is to build a bench-top CNC router, able to cut mostly engineering plastics of up to 800mm x 800mm x 40mm with reliable precision and accuracy of 0.01mm at the most optimale speed (which seems to be 3500mm/min) and at 3200 resolution. Accuracy is important due to the products requirements and to achieve good surface finishes and potentially also cut PCB's. Ideally the machine will be housed in a sound/dust enclosure as it will be in my office and I will need easy access to it to build prototypes and keep an eye on small production batches. I may want to cut some aluminium plates for making prototype moulds. I'm not yet sure how achievable and realistic the above is as the budget is also somewhat limited, but I intend to build incrementally from the basics and add more automation and features later on (assuming I can define a good end-design in the beginning and build towards it).

    So far I have the following items for my next iteration:

    X & Y
    3 x 800mm 1605 ballscrews, C7
    3 x pairs 800mm Hiwin 15mm rails and 2 HGR15CA carriages per rail

    Z
    1 x 250mm 1605 ballscrews, C7
    1 x pairs 300mm Hiwin 15mm rails and 2 HGR15CA carriages per rail

    4 x Nema 23 motors, 4Nm (potentially need to change to closed loop steppers and drivers as step errors can ruin expensive components)

    40mmx80mm extrusions for the frame (need to change these as they are not rigid enough - intended for 3D printers)

    Blackbox controller (from Openbuilds), 24v power supply, limit switches (need to change to better controller and power supply)

    I will shortly be posting a few questions have!

    Ian

  2. #2
    Hi Ian,

    Welcome to the forum!

    As always, the first bit of advice for new builders is "DON'T BUY ANYTHING YET!" - the temptation is strong, but to get a good design you need that flexibility. Otherwise you'll hamper yourself immediately if you're intent on using pre-purchased components, leading to compromise after compromise.

    You've already done a reasonable job at the next step, which is setting out the requirements of what to cut and how accurately. Next, focus on the mechanical design. Try to CAD it out if possible, it will give everyone a much better idea of what you're planning.

    The first thing I've noticed is that you say you want to cut 800x800, and then you've specified linear rails and screws that are 800mm long. You'll need these to be longer than the actual cuttable area, because bearing blocks and ball nuts take some space. You also don't want to be hitting the bearings or endstops in your cutting area, so you need some margin. Plus, the gantry will have some width, which will reduce the cutting dimensions.

    Looking forward to seeing some designs and questions soon :)

  3. #3
    Kitwn's Avatar
    Lives in Exmouth, Australia. Last Activity: 8 Hours Ago Has been a member for 2-3 years. Has a total post count of 474. Received thanks 64 times, giving thanks to others 11 times.
    Just to add to what Andy said, the specified length of ballscrews is the total length including the machining at each end for bearings and pulleys. For 800mm travel you will be looking at screws around 1000mm long. If you go for direct drive your frame will neede to acomodare the length of the screw and the motor. As an example my machine base is 1300mm long, the ballscrews are 1100mm, the rails 1200mm and the total travel of the gantry is 940mm.
    Engineering is the art of doing for ten shillings what any fool can do for a pound.
    Wellington.

  4. #4
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 2 Days Ago Has been a member for 0-1 years. Has a total post count of 11.
    Hi AndyUK / Kitwn

    Thanks. I should have said 600mm x 600mm cutting area as I know that the cutting surface is less than the length of the rail.

    The rails I have are 800mm and the ballscrew are 800mm between the bearing ends, so I estimate this should be enough to get the 600mm travel. I'm still working on the design, but the carriage width is 61mm, together 122mm, but I still need to allow some space for the grease/oil nipples, some safety space and space for the homing switch (if it is not flush with the plate). It's a steep learning curve! I'm working on a Fusion360 CAD model for the design at the moment.

    What would you recommend for spacing of the two carriages on the rail? It seems most designs place them with no space next to each other (potentially giving more travel on the rail), but I've been thinking that adding space (perhaps the width of one carriage) will help distribute the load better on the rails (but given the strength of the rails/carriages, I'm not sure wether it will be such a big advantage).

  5. #5
    For a machine this size I,d have bearing plate 225-250mm. This gives a good balance between ridgidity and maximising the travels.

    Regards your first post and you mentioning 3200 resolution then let me just clairy something that you may not be aware of.? (Think I've said this before, if not I started writing then got disturbed.!!...Lol)

    Don't class micro steps as resolution because it's false, MS are mostly to help with smoothing the motor rotation and they do come at a cost in that you lose torque and require a much higher number of pulses from the controller for the same RPM causeing it to work much harder and with more chance for missed steps. Also much above 2000Ms is wasted on most steppers because they cannot resolve to those fine amounts. 1600 MS is more than good enough for most stepper systems.

    If you want higher resolution then either buy smaller pitch screws or use a gear ratio then this will be your true resolution that can be trusted 100%.

    Regards step errors and closed-loop then again don't rely on the closed-loop.? Step errors are often caused by over tuning the motors or binding axis etc and this is no different for a closed loop system.! The errors will still occur but the encoders will ensure the motors catch backup at some point. However the error will still have occured and will possibly show on your part, depending on what it is your cutting. The only way to ensure that you don't get step errors is to size and tune the motors correctly.
    When tuned correctly and provided everything on the machine is aligned and sticktion free then you will never get step errors on a correctly tuned machine and it will perform just as good any Closed loop system that use the same type of motors.
    Where Closed loop systems differ and are better is in the fact you can get 3 phase motors with a finer 0.9 or 1.2deg step angle, these run much smoother and provide more power at slightly higher Rpm's because of being 3 phase.

    But if you have Steppers already that are the correct Size then with good Digital drives and correct PSU they will match any 2 phase closed loop system on performance and accuracy provided they are tuned correctly.
    -use common sense, if you lack it, there is no software to help that.

  6. #6
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 2 Days Ago Has been a member for 0-1 years. Has a total post count of 11.
    Quote Originally Posted by JAZZCNC View Post

    Don't class micro steps as resolution because it's false, .....
    Thanks - noted on MS. When people refer to "resolution" in general, what does it refer to then?

    Quote Originally Posted by JAZZCNC View Post

    But if you have Steppers already that are the correct Size then with good Digital drives and correct PSU they will match any 2 phase closed loop system on performance and accuracy provided they are tuned correctly.
    Thanks. How do I tune it correctly?

  7. #7
    Quote Originally Posted by ian View Post
    Thanks - noted on MS. When people refer to "resolution" in general, what does it refer to then?
    It should refer to the true resolution of the ball screw pitch. But because MS does divide each step up into smaller steps then a slight amount of resolution is gained, however, it's not to the degree people believe and claim, or I should say it cannot hold that level of resolution accurately and repeatably. The keyword being repeatability, only the true pitch of the ball screw can be counted on to repeatably give the same resolution.


    Quote Originally Posted by ian View Post
    Thanks. How do I tune it correctly?
    Well every machine is different and motor tuning settings etc are different based on each machine. However, the method of tuning is pretty much the same in that you find the limit of where it stalls the motors and then back it off a certain percentage. How high that percentage is will depend on many variables and how you want to use the machine.

    For instance, If your mostly cutting 2D with the odd 2.5/3D work that doesn't have small tiny movements and speed isn't of great concern then you'll tune the machine with Velocity and acceleration in a 50/50 balance. However, if you are mostly cutting 3D or doing engraving etc that consist of 10,s of 1000's of tiny short moves then you'll tune biased more towards acceleration.
    The reason you would do this is that each move as got to accelerate up to commanded speed then travel at commanded velocity then slow down before the next move comes along. In reality with tiny moves you can never hit the commanded speed before the brakes have to be applied for the next move and the whole process starts again. So tuning for high acceleration gets you up to a higher velocity before the brakes come on and it can drastically shorten the cycle time on these types of Job.

    Another tuning scenario comes up for 2D profile type work where the moves are not so short so the commanded velocity is easily reached, however, if this velocity is high then it can cause issues with corner rounding when using G64 CV mode (constant velocity) because the opposite of 3D type work happens.!
    Now we are at full velocity and tramping along nicely then a corner appears, but CV says we must maintain the commanded velocity but the laws of physics say there's no way we are getting around that corner at this speed so the controller as to do something so it cuts across the grass. This is no good to us because we wanted a 5mm radius not 50mm.
    So we need to tune the motors to help with this, again acceleration comes into play, if we set a higher acceleration it means we can leave the point at which we need to cut across the grass until much further down the road and closer to what we wanted.
    Now in practice the controller provides a few tools to help with this, like specifying when CV can be turned off for corner angle degrees then turned back on again. But motor tuning plays a big part into how well the machine can be set up. A properly set up machine can shave hours off cycle times.

    This is where people go wrong and they try to tune the machine for all situations and it's just not possible. The other thing they do wrong is failing to realize that it's a simple equation in that you basically CANNOT have both high, it's either high Accel with lower velocity or vise versa or a balanced setup. The balanced setup will work for mainstream work and machine setups but if you want to do 3D type work or the high feed type work then a different tuning setup is required.
    When they get this wrong is when you get missed steps or stalling motors.

    Now the crazy thing is that most don't realize you can create profiles with different tuning setups and just load that setup when doing that type of work. It's easy really.

    But the basic rule of thumb for a general tuning setup is to tune velocity until the motors stall then back it off 25%, then adjust acceleration to a level which you are happy with. Often the machine will tell you when it's enough. If you want higher of either just know that the other as to be lowered and you shouldn't have any trouble.
    Obviously there are other parameters that come into play like Pulse rate, MS, Voltage etc that affect how the motors can be tuned and then there's the machine it's self and the Mass, sticktion etc which get thrown into the mix. Which is why every machine is tuned differently.!

    It's not rocket science but it's not an exact science either with fixed settings so a lot of feeling your way and tweaking comes into play. then machine also loosens up over time and will need tweaking as it settles down.
    -use common sense, if you lack it, there is no software to help that.

  8. #8
    Kitwn's Avatar
    Lives in Exmouth, Australia. Last Activity: 8 Hours Ago Has been a member for 2-3 years. Has a total post count of 474. Received thanks 64 times, giving thanks to others 11 times.
    deleted.
    Engineering is the art of doing for ten shillings what any fool can do for a pound.
    Wellington.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. My Journey To The Perfect Drag Knife
    By walter67 in forum General Discussion
    Replies: 8
    Last Post: 07-08-2018, 11:23 AM
  2. Starting to think about a new machine
    By charlieuk in forum Gantry/Router Machines & Building
    Replies: 0
    Last Post: 17-02-2017, 07:38 AM
  3. Ruffel's cnc cutting journey
    By DuffelBuffelWuffel in forum Metalwork Discussion
    Replies: 13
    Last Post: 14-09-2015, 05:12 PM
  4. NEW MEMBER: What to buy when starting out
    By smith99 in forum New Member Introductions
    Replies: 18
    Last Post: 11-04-2014, 12:13 AM
  5. The Journey begins.. new to me.. my first CNC - Black Cat
    By Fivetide in forum Milling Machines, Builds & Conversions
    Replies: 37
    Last Post: 22-10-2012, 04:48 PM

Tags for this Thread

Bookmarks

Posting Permissions

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