Effer, a spare parts story.

Discussion in 'CNC machining projects' started by BattyZ, Jun 21, 2021.

  1. BattyZ

    BattyZ Silver Banner Member

    Coming up on a decade ago I was doing my first metal melting and casting. Quickly found out that the actual metal part is the last 5% of the process. "Oh, I will just build a CNC to cut this foam perfectly and cast it...I will call it FR (Being short for Foam Router)" Next morning the name changed quickly to Effer; because wordplay. Feel free to laugh or snicker at Effer's VERY humble beginnings. I was ultra-broke trying to pay for a luxury service known as higher education. 20170412_122046.jpg 20170416_212710.jpg
    Cobbled from surplus scanners/printers and the cheapest drivers/boards from across the pond; this first interation would barely cut foam but it did! Gave me my first real taste of CNC and the LF process. Made parts to upgrade the 3d printed ones.
    Snapchat-2005143640.jpg
    Then years later when I found a pile of cold rolled flat steel and the local junkyard I got the itch to strengthen up my cnc game. A friend gave me a clapped-out old table saw that I turned into a table grinder (don't ask lol). Iteration #2 was born.
    Snapchat-1377990036.jpg Snapchat-2013831658.jpg Snapchat-518652720.jpg
    Shortly after this, I upgraded to a 1.25 Makita router which was a colossal upgrade from the WEN die grinder I was using as a spindle, if we can even call it that. Soon realized that this bad boy was way sturdier; but not square...like at all. This is when I got real deep into LinuxCNC's skew correction kinematics. Steep learning curve for me but was able to totally fix a hardware problem with software. Made some of these inserts for a friend who builds actually square cabinets and they popped right in!
    Screenshot_20180512-002116.png
    Effer stayed in this form for a couple of years. Did lots of money-making woodwork and some mold work with it. Made 8 foot long concrete molds with it even.

    Well, my lunch break is way over so this will have to be continued...
     
    Last edited: Jun 21, 2021
    Bldr J, Billy Elmore and Tobho Mott like this.
  2. BattyZ

    BattyZ Silver Banner Member

    Part 2:
    With a handle on CNC/CAD/CAM and a Clearpath Servo converted Van Norman No. 16 milling machine I set out for a thicker, more perpendicular gantry setup with linear rails this time. 3d printed the positive mold in 4 pieces and used sodium silicate sand to cast. It has a few inclusions but a very usable casting.
    Snapchat-535966738.jpg Snapchat-1052048983.jpg Snapchat-1740725273.jpg
    This 3rd iteration became my workhorse in anything wood or aluminum. Wood do 160 ipm all day in most tree carcus alloys. At this time most of my castings were 3d printed ceramic shell, greensand or semi-permanent die cast so not much foam work at all for Effer.
    20200210_150839.jpg Snapchat-1721916981.jpg Screenshot_20180523-062028.png Snapchat-1889642959.jpg
    I know we are all most likely cringing at my uncovered ways and lack of wire organization but this setup was really reliable for me. Gone were the days of missed steps. The pile of chips is from a 5-hour tool path for a die I was making. over 300 minutes of surfacing moves and not a single missed step. I was pretty stoked. I know I have not mentioned it yet but the early setups suffered from binding and lost steps. Which might be pretty obvious to some.

    I have been running this iteration of Effer for over 2 years...contently...that is until I joined this forum (I had been forever lurking at AA) AND Kelly was posting some of his masterful LF process work.

    SO..I have dusted off some old designs, ordered some bigger linear rails, and found some software to learn that can handle a couple more axis than 3...Effer is getting back to his roots!
     
    Mark's castings likes this.
  3. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    BZ, what's Effer's the cutting envelop?

    Best,
    Kelly
     
  4. BattyZ

    BattyZ Silver Banner Member

    It was 15x35x8ish <XYZ>. Depending on tool.
     
  5. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    I need to raise my rails so I have a little more than 8" under spindle. I have 4" z stroke so that would allow me to use all of that with 4" cutting length......in foam. XY are both 35".

    Best,
    Kelly
     
  6. BattyZ

    BattyZ Silver Banner Member

    That sounds like a great upgrade. That 35 x 35 belt drive would be really nice at times. Also, take into account that they have 1.5 to 7 degree tapered ball end mills (2 flute). These can really cut down time on draft or, more of us LF guys, angled faces. Saunders machine works also makes mod vise plates.

    https://saundersmachineworks.com/co...-3d-shapeoko-3-aluminum-fixture-tooling-plate

    At 400 they are quite spendy, especially for foam work. This brings me to the thing I have to figure out yet. Work holding in foam. For flat pieces, a custom vacuum setup seems to be the way? When I get around to actually cutting on 5 sides I imagined a tall, tandem custom mod/split vise would work? I have had great results using the superglue-painters-tape though.
     
  7. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    It's rack and pinion. Z is acme screw.

    Cutting forces for foam are so low it doesn't take much holding force. Two-sided tape is my go-to. Sometimes just plain clear one-sided packing tape. You don't want the thick two-sided tape used you hang pictures or carpet tape. The thin stuff that is like clear packing tape only tacky on both sides. Even with that, I only use small pieces here and there or you'll never get the foam loose without damaging it.

    Vacuum is nice if you are going to do repetitive work or frequently mount/demount the work piece. Household central vacuum motors make nice vacuum sources.

    Best,
    Kelly
     
  8. BattyZ

    BattyZ Silver Banner Member

    I was going to leave everyone on that ferocious cliffhanger I typed up but changed my mind. So the original goal was to do casted legs and gantry but the gantry vibrated out of my big CNC vises during heavy cutting so that was a no-go. Big bummer having spent all that time...anywho I wanted to push forward so I welded one up. Every weld was clamped securely against my flat surface and cast iron welding squares. I do want to revisit the cast aluminum setup but design changes will have to be made. And the goal here is foam and wood only.
    20210616_190245.jpg Snapchat-1395979021.jpg
    I have 2 Z axis plates to machine and just a few more brackets to weld up then the mechanical side should be mostly complete. Then onto the wiring. In the long run, I would like to do an epoxy granite base with closed-loop glass scale feedback.
     
    Last edited: Jun 22, 2021
  9. BattyZ

    BattyZ Silver Banner Member

    Oh nice! What are your advertised travel and cuttins speeds? Do they give you anything for resolution specs?

    I really need to try this stuff. Seems way quicker than what I am currently doing.

    Any particular brand you like for the central vac motors or more just what the markeplace has?
     
  10. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    The manufacture's spec page. If I had done my homework better, I probably would have bought differently, but more so for software compatibility than hardware.

    https://millrightcnc.com/product/millright-cnc-mega-v-router-bundle/

    You'll be surprised how little it takes. On most pieces, I probably only apply it to 10% of the pattern area, but better small pieces in multiple locations. It always sticks to the waste board on separation, and sometimes you can reuse it, but usually not worth it.

    I like the Ametek/Lamb units with the cast aluminum housings, but there are other brands and they are all about the same performance. Some are plastic. If you search eBay for "3 stage vacuum motor" you'll get a bunch of hits. I posted specs here:

    http://forums.thehomefoundry.org/in...-spinning-cup-oil-burner.246/page-4#post-4468

    I used a pair of them to build this:

    http://forums.thehomefoundry.org/index.php?threads/vacuum-forced-air-cart.810/

    If you stage two of them creating a 6-stage turbine they cam pull 1/2 atm. They move a lot of air so can tolerate more leakage than other sources but tend to be a bit noisy and might give your router a run for it's money in that regard.

    Do you do anything for dust collection? Foam is a tough one because it becomes charged and sticks to everything.

    Best,
    Kelly
     
  11. BattyZ

    BattyZ Silver Banner Member

    It is 2021, you can always do brain surgery! They post 500 IPM max cutting and .001/12 in resolution, not too shabby at all. I know the precision will suffer from the flex of the machine. But even some 5-ton machines flex. Our cutting medium is really a saving grace. About the software, is it just not intuitive or somewhat clunky?

    All great info about the vacuum units. That cart of yours is really a pro build. I am sure I will need a rig like that in time.

    I individually pick up each piece and recycle it because I am a caring human. I would never open the garage door and set up 2 fans to blow it away from my machine... I have just occasionally used compressed air to clean off the whole machine while it was running in an open corner with decent results but yeah, it is sticky. I have a harbor freight dust collector I need to set up with a cage and maybe a timed periodic dump blast? Real open to ideas here.
     
  12. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    It's really just the postprocessor......GRBL. I could revamp the controller to whatever but I can live with it for now. I really just need to develop my CAD skills which doesn't have anything to do with the router.
    Didn't know it could be recycled.
    Well here's what I did for my pin router.

    http://forums.thehomefoundry.org/index.php?threads/machining-xps-foam-patterns.14/#post-1717

    You can see the table level duct opening connected to a 3HP bag collector via 5" flex. The hood is just film in back and top. The ends are acrylic and come off in a few seconds. So it is contained/enclosed on 5 of 6 sides. It still ejects a lot of foam, but to varying degrees depends on which side of the stock is being cut. As the operator, I'm usually covered in it. As mentioned, the foam is charged and clings to everything, but I have a compressed air wand I use to clean up inside the hood and chase everything into the dust collector opening and the rest easily vacuums up with the shop vac. It still represents a significant improvement over nothing. I don't think it would matter how much air you tried to move, the result would be more or less the same.

    For the CNC Router, I figure I'll put a collector duct across one entire end, a pusher fan opposite at the other end, and enclose the remaining sides with a clear film covered cage that either hinges or hoists from overhead. I don't expect it to be perfect either but similar use of chasing the residual dust down the chute with compressed air should finish the job, and I won't be covered head to toe. Cardboard and duct tape ducts work just as well but good for development.

    Best,
    Kelly
     
  13. BattyZ

    BattyZ Silver Banner Member

    haha the comment was mostly tongue-in-cheek. I guess if you put the foam bits in your attic that would be recycling?

    I was thinking something along these same exact lines

    One area of CNC which might help is no operator part...well sorta. Should still be able to put a lid on to close in the mess or at least most of it.

    With this dust collector has the cyclone worked well or at all for you? Just never tried and am interested.
     
  14. BattyZ

    BattyZ Silver Banner Member

    Had some days off and actual time to work on Effer, progress happened.
    20210624_131307.jpg Snapchat-2132571692.jpg 20210624_132354.jpg 20210624_132402.jpg
    Once the gantry linear rails were squared and fastened the rest went pretty quickly. Only had one plate where some holes did not line up from the cad model. I think I used different model rails than what I bought. The petg printed 5-axis head has a little play when pushed which I don't think will be a big deal in foam to get started. I want to cast a new head, naturally. The z-axis rails need to be like 10 inches longer, which I don't know how I did not see that from the get-go. Not a big deal, longer ones will bolt in just fine. Also plan on notching out the lower part of the cross plate so the Z can travel up higher as well. The goal was to be able to fully "hurdle" the part for maximum travel height. There will be a 2-6 in endmill in at all times so my lose some Z height on that. Also may be able to tell the cam to move around the part instead of over but I do not know about that. Movement is smooth on all axis, which is reaffirming. Have not moved over to Dynomotion yet, still on LinuxCNC. Waiting to finish the new control cabinet...
    Snapchat-1672987337.jpg
    This is the additional control/power box almost done. This will feed step/dir to the other 3 axis, control the additional 2 axis, and use PWM signal to control the spindle speed. All inputs will be fed back to this box as well. The green board in the middle is the actual kflop board. This box went together relatively easy because of the screw terminals and breakout boards for all the connections. I highly recommend them!

    Travels on this machine, when done, will be right about 35x37x22 which should cover all the castings I can really handle. Swapping in a longer set of rails on the Y-axis would not be too big of a chore if the need arises.

    Hoping to have a video up soon of real 5 axis motion and some cutting even maybe. May have some work to get the programming/post processing to work out smoothly from the CAM, we will see!
     
  15. BattyZ

    BattyZ Silver Banner Member

    Update: spent a bit of time chasing down a grounding issue but all five stepper drivers are moving smoothly on the Dynomotion KFLOP board! I tried to load up some short videos here but no go. (I bet pictures take up a lot of server space as is.)

    The next step is to calibrate each axis for counts/revolution. With the linear axis this will be pretty straightforward. However, with the rotary axis I might have to get a little creative. Any ideas here would be welcome. I do have a probe so I thought about making a valley with 2 blocks and sweeping in between and measuring the measured distance of the dro on the screen to what I know to be physically there. Do some unit circle math and load values. An encoder would be the best solution but little out of my programming reach for now. Maybe that's too high tech...couldn't I just put a thin bit in the spindle and print off a big degree wheel to visual calibrate?

    Speaking of programming, have a friend (level 5, systems architect programmer) help me out with trying to wrap my mind around the C programming language and what is going on...good thing I can afford his payment method!
    Snapchat-1506270153.jpg
     
  16. BattyZ

    BattyZ Silver Banner Member

    With a lot of effort and tinkering...we are up and moving!



    The kinematics have been sorted. Tooke some reaching out to get it done. Tom at the Dynomotion forum was quick to respond. Now I just have to adjust the pivot length and calibrate each rotary axis. The first method I am going to try is with a digital level.

    After everything is dialed in I will work on wiring up the spindle power supply and tidying the wires. Then it should be first chips!

    I also have to sort out a homing routine. I do have a probe for this spindle so was thinking of building a probing routine for it. that would also set zero for the rotary axis. Like I said before, attaching quadrature encoders with a know home would be best but maybe a little above my paygrade right now.
     
    3Dcasting likes this.
  17. BattyZ

    BattyZ Silver Banner Member

    Been a minute since the last update. Had to get out some machining orders and get in some OT at work. In the meantime, the longer Z axis I ordered had shown up. Also picked up a 16mm pitch lead screw for the X-axis so I installed both of those.

    Took the main carriage down to do some machining on it. Made a gap so the Z-axis can tuck up just over 2 more inches which might come in handy getting the head over the top of a part. Needed to extend to the backside pocket to fit the longer nut housing.
    20210805_121903.jpg 20210805_122034.jpg 20210805_140559.jpg 20210805_140615.jpg
    The Z-axis does look mighty tall in that last picture. Not all of the linear rail length is fully used but I would rather have it that way than not enough like before. Down the road, there will be a plate or casting to cover the rails and add rigidity.

    Not really pictured- I swapped out the stepper motors on the head for ones with 5:1 ratio planetaries. Torque is greatly increased, like almost 5 times as much! lol I thought I would be giving up speed but not really at all.

    Have been working on cam strategies and the post-processor in the downtime so hopefully can run some gcode with foam mon/tues.
     
  18. amber foundry

    amber foundry Copper

    Ace posts
     
  19. BattyZ

    BattyZ Silver Banner Member

    Focused my efforts on Effer's 2 main areas for "growth":
    1) Working 5 axis CAM post processor for my custom setup,
    2) Metal body for the 5 axis head.


    1) This was a lot of banging my head against the wall. Was trying various posts; modifying them to see if they would work. I eventually had success in just doing some 3+2 machining but when it would flip to the other side the end mill would have shifted .08-.140 inches. I believe most of this is from my number #2 issue. I won't really explain the hours of tom-foolery but getting closer to a fully custom working processor. Main issues arising during the simultaneous 5-axis moves. Also, when I went to just 3 axis to machine some foam I had to deal with more issues that were very annoying, but I learned things. Also reverted to my ole' linuxcnc post and it ran without a hitch.

    2) The main C axis bearing mating in the printed PETG head has become too loose to be any kind of acceptable. Do earlier this week I designed and printed a mold for some PLA shell investment casting. This was all fine and dandy till the shell fell apart putting it in the sand. Pretty big bummer. Did find some wicked-looking inclusions going on. So I did not have a watertight print in the first place. Maybe it was a good thing it failed early?
    20210812_203814.jpg 20210814_133954.jpg 20210814_135832.jpg 20210814_194142.jpg

    So I set out to machine some foam blanks to go lost foam method...
    20210814_181520.jpg 20210815_144039.jpg
    One pour was with no coating, one was with a very quick coating of drywall. Have much to smooth out with my dip-coating process but it shows a lot of promise.



    The removable pouring basin was a huge win. Thanks for sharing Kelly!

    20210815_170525.jpg 20210815_172804.jpg 20210815_172816.jpg
     
  20. Al2O3

    Al2O3 Administrator Staff Member Banner Member

    After banging my head for a week with EMI and post processor issues you have both my admiration and sympathies in sorting a 5 axis post. I'm by all that now and tuning my hardware. Stalled waiting on a couple simple parts to arrive but have a half dozen programs running smoothly and should be at it in a few days.

    What shell material and how many coats? That is a bummer. That print looked nice. 3D Printing is next on my agenda but I have a fair ways to go and a lot of backlogged parts ideas for the CNC first.

    Soldier on!! You'll get there.

    Best,
    Kelly
     

Share This Page