. .

Thread: Sisyphus

  1. #1
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 1 Week Ago Has been a member for 0-1 years. Has a total post count of 19.
    Hi All

    I'm a new member, who recently joined the forum and am starting the journey. I've made some progress on designing my machine but have hit a few design issues and questions which I will appreciate help on as well as good general feedback on my overall design. I have experimented a little bit with a machine that I used as a proof of concept and have experienced some of the pains and joys of machining. I have little experience though and no point of reference for a build to know how strong/fast/accurate etc. builds, components or materials are.

    My objective
    I intend to build a CNC machine that will be used in product R&D (I'm building GPS tracking devices, and I will use the machine to build prototype cases and small production runs of a few thousand cases). It has to be fast and as accurate and precise (aiming for 0.01) as I can get for a budget of 3.5k or ideally less. I accept there may be tradeoffs and adjustment of my expectations! The machine will be used in my (small) office and will eventually be housed in an enclosure to reduce the noise. The machine will mostly cut hard engineering plastic sheets (good finishing quality is essential) and some occasional aluminium for custom moulds or jigs.

    I know accuracy and precision are relative in context and does depend on the collective results of all the components. Still, I do want to endeavour to get it as good as I can (while keeping sane and being somewhat realistic) as poorly machined parts can influence the performance of my devices significantly. I've decided to use two ballnuts on each screw as a trial but still need to work out the detail and to determine how to pre-load them. In the smaller runs I've done this far I found that poor machining and poor machine accuracy/repeatability adds excessive post-processing and clean-up time and leads to a higher percentage of rejects and I want to avoid this. To this extent, I may be "overbuilding" my machine but my view is that and overbuild is less expensive than a rebuild, and I need a robust machine I can trust so that I can get on with my business. I also want to keep the build as simple as possible.

    The target size is to be able to cut at least 500mm x 500mm and ideally 600mmx600mm ( so a machine size of 800mmx800mm). For the Z height, I don't expect to cut anything thicker than 50mm although I plan to have at least 125mm of travel.

    I "inherited" some starting components:

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

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

    I intend to use at least 4Nm Closed Loop motors or potentially 5.5Nm motors

    I've not worked on the electronic and electrical design yet and will add this once I have the mechanical design settled

    At first, I will use a Dewalt Router and later switch to a better spindle

    I have limited space and tools for fabrication so intend to use aluminium profile for the build as well as 20mm aluminium plates (preferably something like Ecocast that is ground flat)

    Frame
    Click image for larger version. 

Name:	Base_Frame.jpg 
Views:	80 
Size:	239.7 KB 
ID:	28544
    For the frame, I plan to use 80x80 and 80x160 profile. Please see the attached image for the frame.

    Questions:
    1. Are 4 cross-members on the X-axis too many - would 3 suffice?
    2. Are corner brackets and the connection with t-nuts sturdy enough to give a rigid frame? Or should I tap the ends and drill M8 holes and fasten them from the outside of the Y beams?
    3. Should I use the heavyweight or lightweight profile? (from KJN)

    Fixture Plate
    Click image for larger version. 

Name:	Base_Frame_Fixture_Plate.jpg 
Views:	79 
Size:	217.3 KB 
ID:	28545
    For the fixture plate, I plan to use 20mm aluminium drilled and tapped for M8. The plate will fix to the profile, and I will fit a 10mm - 20mm spoiler material to the top

    Questions:
    1. What grade aluminium would be best?


    Gantry design
    I've looked at different gantry designs and have looked at two different styles: Option 1 with the rails on top of the profile and Option 2 with the rails on the side of the profile. It is hard for me to gauge the difference between the two designs and which will be best or be "better" in the long run, and I will appreciate people's views on this. I attach images of the two designs that I have done. On both designs, the Y and X plates are all 200mm wide, and the Z plate is 180mm wide.

    Option 1: Rails on top of the profile
    Click image for larger version. 

Name:	Rails_On_Top_1.jpg 
Views:	103 
Size:	236.0 KB 
ID:	28548
    Click image for larger version. 

Name:	Rails_On_Top_2.jpg 
Views:	88 
Size:	125.2 KB 
ID:	28549

    Option 2: Rails on the side of the profile
    Click image for larger version. 

Name:	Rails_On_Side_1.jpg 
Views:	98 
Size:	208.1 KB 
ID:	28546
    Click image for larger version. 

Name:	Rails_On_Side_2.jpg 
Views:	87 
Size:	118.9 KB 
ID:	28547

    Overhang
    The router is set forward quite a bit and extend 40mm forward of the front face of the gantry plates. I've considered setting the X beam back by 40mm to offset the overhang but I will then have more issues with the gantry plate then interfering with the back motor plate. I know an alternative is to place the rails on the top/bottom of the X-beam which will reduce the overhang but I initially steered away from this it then involves more plates and a more complicated design (I may still have to do this).
    Click image for larger version. 

Name:	Rails_On_Top_2.jpg 
Views:	82 
Size:	126.6 KB 
ID:	28550
    Click image for larger version. 

Name:	Rails_On_Top_2_Setback.jpg 
Views:	81 
Size:	142.4 KB 
ID:	28551


    Adjustability
    I know it is important to have easy access to all the bolts/screw/nuts nd also to have adjustability so that the machine can be squared etc. However, I am unsure on which critical bits I need it and whether I should then just add oversized hole or slots in these areas? Any input here will also be appreciated.

    Tapping
    Where I connect motor mounts or bearing supports, I planned to tap the aluminium plate and not use a separate bolt - is this a good idea and how well does Ecocast tap?
    Last edited by ian; 07-07-2020 at 08:05 AM.

  2. #2
    Hi Ian, Welcome

    Quote Originally Posted by ian View Post
    a budget of 3.5k or ideally less.

    The machine will be used in my (small) office and will eventually be housed in an enclosure to reduce the noise.

    The machine will mostly cut hard engineering plastic sheets (good finishing quality is essential) and some occasional aluminium for custom moulds or jigs.

    I "inherited" some starting components

    3. Should I use the heavyweight or lightweight profile? (from KJN)

    1. What grade aluminium would be best?

    Adjustability
    I know it is important to have easy access to all the bolts/screw/nuts nd also to have adjustability so that the machine can be squared etc. However, I am unsure on which critical bits I need it and whether I should then just add oversized hole or slots in these areas? Any input here will also be appreciated.

    Tapping
    Where I connect motor mounts or bearing supports, I planned to tap the aluminium plate and not use a separate bolt - is this a good idea and how well does Ecocast tap?
    Budget seems achievable. I don't envy you having this in your office; even with a watercooled spindle rather than a router don't expect to get much other work done without wearing ear protection. An enclosure will help... but...

    The inherited components are never a good sign to be honest with you. They tend tie one hand around your back before you start. You've said you're cutting plastics primarily which tend to want higher feed rates, so 5mm pitch screws might not be ideal as they dictate your speeds somewhat. You may find you'll have to introduce some gearing to get the speeds you want. Having said that plenty of people on here build with 5mm pitch direct drive and it works for them.

    I'd go with the beefyer profile every time. More rigidity, less vibrations.

    As you've mentioned, the 'best' would be ground tooling plate.

    Adjustability... This is a tough question.
    • Think about it from the spindle tramming perspective - What would you move to tilt the spindle left/right and forwards/backwards?
    • Think about it from an axis squaring perspective - How do I tilt Z to make it go vertical with respect to X, then wrt Y? How do I adjust the angle between X and Y?
    • Where you have components meeting that may have some positional variance from your CAD due to building errors or tolerances, how will you ensure they can mate appropriately? E.g. Ballnut connections - can you adjust in all three directions using slots/shimming?
    • Motor positioning - either direct drive (how can I line up the shafts precisely?) or pulleys (how do I tighten that belt?)
    • How do I adjust the ballscrew angle without modifying the linear motion components?


    The list goes on an on. Oversized holes are one method, sure, but don't put them everywhere just because. Slots, shims, grub screws...

    Tapping into ecocast is perfectly fine and works well.

  3. #3
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 1 Week Ago Has been a member for 0-1 years. Has a total post count of 19.
    Hi All

    After further considerations, planning and design, I have revised my previous design and made several chnages.

    I am now moving to the build stage and will likely start to build some assembly in the next week or two, so this is my last chance to take on board some feedback and fix any errors or oversights!

    This is the latest design...


    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.54.06.jpg 
Views:	54 
Size:	166.1 KB 
ID:	29057
    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.54.20.jpg 
Views:	49 
Size:	180.6 KB 
ID:	29058
    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.54.45.jpg 
Views:	54 
Size:	280.2 KB 
ID:	29059
    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.55.18.jpg 
Views:	49 
Size:	273.3 KB 
ID:	29060
    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.55.44.jpg 
Views:	59 
Size:	179.1 KB 
ID:	29061
    Click image for larger version. 

Name:	Screenshot 2020-10-23 at 18.55.55.jpg 
Views:	51 
Size:	193.5 KB 
ID:	29062

    Feedback and comments welcomed...

  4. #4
    ian's Avatar
    Lives in London, United Kingdom. Last Activity: 1 Week Ago Has been a member for 0-1 years. Has a total post count of 19.
    Giving this another try...

    Wiring 0.1.pdf

    I have very little experience in electrical wiring but I have drafted the first wiring diagram and will appreciate any feedback to make sure I don't make any serious electrical mistakes!

    Some key questions I have:

    Is it good practice to use the terminal blocks to distribute the power
    Should I connect the VFD as a separate feed from the mains or can I run it from a single main for the whole system
    Should I connect the VFD via the relays (not sure the relay can support the power draw)

    any help will be appreciated!

    Components I will use:

    5.5N Nema 24 Closed Loop Motors

    2.2KW HY Spindle

    Inductive sensors

    230V, 60V, 10A Toroidal transformer

    24V 15A Switching power supply

    Masso G3

    Masso Relay Module

  5. #5
    Hi Ian

    First thing that strikes me is you haven't fused things. That toroidal may also appreciate a D-Curve MCB.

    The limit switches will probably want powering. I'd suggest some power indicators to notify you when things are live.

    Just have a single earth bus rather than two.

    I used DIN rail terminal blocks with bus bars. Seemed to work absolutely fine. Nice and tidy.

    I connected my VFD through a single feed into the cabinet, but I used some noise suppression on the mains lines to reduce the noise output from the VFD. Just 3/unit from China. Commonly its suggested to take the VFD power feed from somewhere else in the room, but I can't tell you which is best. Some VFDs say in the manual not to turn them on and off via a relay, so I don't have mine on one.

    If it helps, I've posted my circuit diagram on my build log, post 89, around page 9.

  6. #6
    Design feedback wise, looks pretty reasonable given the components you've already got. Looks like you've placed the limit switches in an interesting position though - it they don't trigger they're going to get crushed :) I'd put them at a right angle to the direction of travel. You'll also want to be sure that they will trigger on the aluminium (they should - but with a significantly reduced sensing distance), or you'll want a ferrous metal target.

  7. #7
    JAZZCNC's Avatar
    Lives in wakefield, United Kingdom. Last Activity: 16 Hours Ago Forum Superstar, has done so much to help others, they deserve a medal. Has been a member for 9-10 years. Has a total post count of 7,687. Received thanks 1,326 times, giving thanks to others 86 times.
    Hi Ian,

    The schematic as a few issues I would address.
    First off you don't have a safe E-stop system because you are only sending the E-stop signal to the controller so are relying on software to stop the machine. The NVR switch isn't doing anything so may as well just be a ON/OFF switch. The only way to take advantage of the NVR switch would be to run it through the E-stop contacts so when E-stop is hit the NVR does it job and drops power and won't reset untill E-stop is cleared. However this means running mains power through the E-stop and probably around the machine which isn't ideal.

    So a better way is to drop the NVR switch for a normal ON/OFF contactor then create a latching Safety circuit using a relay and momentery button with 24V running through the E-stop circuit which is much safer.
    This creates you an hard wired safety circuit which you can then control many things using the relay contacts or several relays in a circuit so when the E-stop occurs you can leave other things turned on which you don't want turned off when E-stop happens, like Motion controller and VFD etc.
    You can also wire them to turn things on as well, like light towers or sirens warning you the E-stop happened or the Kettle so you can enjoy the Oh-Shite Hob knobs with a nice cuppa Yorkshire tea to console your self...

    One of these contacts you would use to send the signal to the controller informing it the e-stop happened so it can stop the G-code. The other contacts you would use to Kill power to the toroidal transformer or a Contactor that feeds the toroidal PSU if it's high current.
    Also one contact would send a signal to the VFD that an E-stop occured. Don't wire the VFD power through any relays, it should stay on when an E-stop occures but should be told to Stop and kill it's outputs.

    Regards the spindle and water pump, the VFD as a relay built into it, use this for the water pump so the pump only runs when the spindle is turning.

    The other thing is to only have ONE Ground point, STAR grounding, where every ground or shield wire runs too, if not you run the risk of creating Gnd loops.
    The only exception is the VFD, with the VFD you take the spindle wire shield to the Gnd point on the VFD, the VFD Gnd is then taken from the STAR point.
    Also only Gnd the shields at one end, usually in the control box.

    You need fuses on the PSU's and VFD, I would also think about putting fuses on the drives, also the water pump and dust extraction.

    The Terminal blocks are fine and a very easy way to create bus bars for pulling power from.
    -use common sense, if you lack it, there is no software to help that.

  8. #8
    Hi ian,

    I hope you havent started yet, because I would strongly reccommend to go with design option 2.
    - The reason is, that the arrangement of the x-rails of option 1 is allmost pointless. The idea to have more than one rail is all about having a broad base (by putting them apart as much as possible) to compensate larger moments (to make the design stiffer).
    In design option 2 you have a far better stiffness using the same components, hence you will get closer to your design goal of a 0.01mm precision (wich is tough to achieve)
    - Option 2 will also have less debris falling on these rails.

    More ideas:
    - Another improvement to option 2 in that same direction would be by placing the x-spindle also on the outside, having less debris on it and the nuts beeing in the middle of the 4 wagons (seen y-wise optimum for transferring forces) . The spindle and its nuts can even be put on the outside of the cheek, making it more accessible to adjust (you are going for a 2 nut spring loaded design, I take it), to grease and to clean.
    - Increase the cheek plates thickness to 4/3 of what they are now. Same goes for the plates of all the z-parts.
    - Consider side-plates for the moving z-plate to increase the second moment of inertia in x-direction (y-direction is never a problem there)
    - Also prepare the design to hold a "real" milling spindle later. Starting cheap with a handhold router is absolutely ok, but you will later want to get a spindle built for that CNC-purpose anyway, e.g. in the 2.5KW region.
    - If you then will go for a watercooled version, you will also have a quieter machine when milling plastics.
    Last edited by stormer; 3 Weeks Ago at 06:39 AM.

  9. #9
    Furthermore:
    - Rubber bumpers that stick out allmost as much as the end sensors on the x and y axes may save the thing in case of an end sensor failure. To find space for the bumpers on the z axis is more tricky, but there are doughnut shaped solutions, that mount right around the spindle end bearings or around the nuts.
    - Concerning the mentioned axial versus the tangential use of these type of approach-sensors: To my experience, the axial use you have chosen is far more precise (repettitive correctness). That would also improve such an end sensor as an absolute home sensor, if you intend to use them that way.
    - To avoid destruction of the sensors in case of collision, I use plastic (printed) adaptors around them, which will break under stress, so that the sensors will just be pushed through their larger mounting holes.
    Last edited by stormer; 3 Weeks Ago at 06:46 AM.

Thread Information

Users Browsing this Thread

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

Similar Threads

  1. BUILD LOG: New Build - For Your Amusement - MK-2 build
    By Karl in forum DIY Router Build Logs
    Replies: 12
    Last Post: 08-02-2017, 08:03 PM

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
  •