Effer, a spare parts story.

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

  1. Al2O3

    Al2O3 Administrator Staff Member

    Maybe take a look at using a tolerance ring as a fix. Ever use them? Usually for mounting bearings or press fits in dissimilar materials with greatly varying thermal expansion rates but also allows looser tolerance fits.


  2. BattyZ

    BattyZ Silver Banner Member

    I have not! Didn't even know about them till I read this post. Good to know those exist. However, I wanted to change some other dimensions so I just re-machined the stock. I did get a nice .0015 press fit this time!


    Apparently, these are the bearings for an 80's Honda front wheel bearing. I just sized them per my application. If they hold up a small car they should hold up to the cutting forces.

    Here this the fixed mount next to the rotary:

    20220330_164121.jpg 20220330_164616.jpg

    I have 1 or 2 more pieces to design and machine then we are installing! There will be a nice Lovejoy coupler to take the place of that I-just-want-to-see-it-spin one. I am aware of the stick out to the back. I didn't think it was going to be this long but oh well. I have the cross plate notched out so the whole rotary can tuck right up in for max Z travels. Getting a little excited these days. Bolting all of these solid pieces together has a real this-will-work feel to it.

    On a side note, actually had to use the router to sign job. Was thinking about how to make a riser table with what I had on hand when I thought I would just try it at full z axis extension.


    No issues whatsoever. For once I felt confident lowering the spindle speed to make more of a chip then than nasty MDF dust. Big fan of this spindle on the first job. Quieter than a router to boot!
    Tops likes this.
  3. Jason

    Jason Gold

    Suspendaslurry coated foam works just fine when done right.... Or do I need to dig up the B Plug thread?:rolleyes:
    BattyZ likes this.
  4. BattyZ

    BattyZ Silver Banner Member

    Found some time to get back at this project.

    20220702_114411.jpg 20220702_114341.jpg 20220702_140014.jpg

    Successfully LF cast, T5 HT, and machined a trick little final adapter. I messed around with a few couplers before settling on this "rigid coupler."


    I say rigid because I am getting about .060 or so of play when I push with decent force up by the indicator, roughly 11 inches away. The play is only radially. The tapered small car wheel bearings are actually rigid as planned. Had this setup apart quite a few times trying to nail down where the play it is coming from. Either out of spec play from the eBay "new" parker precision gearhead or radially flex from the coupler. Feeling by hand it feels like worn gears, it is consistently there and just enough to really bother me.

    I have options that I could implement, including a redesign with a harmonic drive, but I really want to see how this affects cutting. The cutting forces of wood/foam/wax can't be too large. There are just a lot of variables at play and I would like to judge them by the end result. So basically, can we fix a hardware issue with software to be "good enough?"

    I have bushings/couplers coming in for the rotary table part of this build. In the meantime, going to focus on getting the controls back over to Dynamotion from LinuxCNC. As Effer currently stands, should be able to use the machine in an indexed spindle kind of operation. Place a piece of foam in, machine at 90 degrees head tilt then go 180 degrees the other way and machine the other half.
    Tops and MrCrankyface like this.
  5. Al2O3

    Al2O3 Administrator Staff Member


    Flex from a rigid coupler? I must not be understanding what you mean by radial flex.........wouldn't worn gears just produce backlash and hysteresis induced position error?

  6. Tops

    Tops Silver

    BattyZ, do you have any more impressions from the lever-action ER16 system on that AMB 1050 spindle?
    Is it 230VAC 50/60Hz or ?
    I am (still) using a repurposed Rototzip on my machine and being able to do a 4 second tool change with an offset table would be awesome.
  7. BattyZ

    BattyZ Silver Banner Member

    It is just an aluminum coupler, so it is allowing some twist under force.


    Funny, just noticing the "zero backlash." This is most likely true because I also drilled and tapped setscrews for the flats of the shafts. After sitting on it, I am pretty sure the play is coming from the gearbox. However, doubt I will ever have an 11-inch endmill hanging out. Could see a 6inch for reaching but most likely 4inch or shorter. This will reduce the error proportionally.

    I am really liking it. It takes me longer to notice it needs a tool change than to actually do it. I have the manual speed adjustment because it was simpler. Just tune it by chip load/material and forget about it.

    It is 230V yes, which may be an issue for some.

    Oh, it totally is. I know it was a bit of coin but it has really unlocked things. I have 7 tools on my table currently and had one part that used 5 of them. Really removes all the friction points.
    Tops likes this.
  8. BattyZ

    BattyZ Silver Banner Member

    Did some more deal hunting on fleaBay, and scored a crossed roller bearing and a harmonic drive! Made it through the CAD process in one sitting. Yesterday I spent the afternoon machining some parts out and this is what we have so far: 20220818_164303.jpg 20220818_161945.jpg 20220818_161932.jpg 20220818_130235.jpg

    Used the Superfly cutter w/ an insert that NYC CNC uses and am really happy with the surface finish on the backside of the big part. The yin-yang artifact on the front side is cool but given a 2nd chance, I would do a different tool path.

    This B-axis joint will basically be set up like a robot joint. Closed loop servo to harmonic drive. In my case a 120:1 reduction, the servo can do 3-4,000 rpm. The crossed roller bearing is P4 class and rated for like 1200 lbs static, hopefully, won't need near that! The whole thing centers itself with counter-sunk screws. I still have a lot of holes to tap! I may have to redo one piece with the holes way too close to the outer edge. Going to attempt a Helicoil first.
    Tops likes this.
  9. BattyZ

    BattyZ Silver Banner Member

    Pretty excited/concerned to see how everything was going to fit together after tapping tiny holes and whatnot.

    20220820_095749.jpg 20220820_095755.jpg

    Like I actually measured! The rotary axis came together quite nicely. A stought unit. All but one hole threaded nicely. Said hole is on the inside so no one will ever know! The output flange is RIGID. Makes my previous setups look like gymnasts.

    The axis isn't fully finished, I have to tap 2 holes on the back side to affix the end of the z axis ballscrew block. A new subplate may need to be made, we will see. Also, need to figure out a limit switch for homing. In the meantime, my angle meter seems to be precise when placed on the end of a long cutter.

    Now to bolt it up and give it a spin...and not hear binding or crunching.
  10. BattyZ

    BattyZ Silver Banner Member

    Well.... we have movement, but...we also have a very hot servo. Torques-out after a few minutes. I think I know where some misalignments are. Reperations underway.
    Tops likes this.
  11. BattyZ

    BattyZ Silver Banner Member

    Tops likes this.
  12. Tops

    Tops Silver

    Reperations? Is that like a snack you eat while you fix the CNC?

    The video short is awesome.

    I am 0 for 3 on running my CNC the last couple days. First strike was a dodgy clamp job and a bad program that failed almost instantly. The second a bad layout on stringy wood where the wooden pieces broke free before completion. The third I am trying to figure out, easier wood and better layout, but the z-axis 'dove' about a 1/4" while the program was running. Bit looks ok in the collet so I may loosing power to the Z-stepper.
  13. BattyZ

    BattyZ Silver Banner Member

    I wish! I remade a part and change how I assembled. Put it together "live" this time. Would snug up some fasteners
    I have been here many times! I always tried to skate the time it takes for good workholding. Most of the time it promptly bit me in the rear. Keep up the battle though! Good fight to fight.
    and then operate the harmonic, tighten down and move it some more. Think it found it's happy spot now.
    Tops likes this.
  14. BattyZ

    BattyZ Silver Banner Member

    ...I typed stuff but it got cut...snug up fasteners a little then operate the axis. Continue in this fashion till all fasteners in place.
  15. BattyZ

    BattyZ Silver Banner Member

    ...I typed stuff but it got cut...snug up fasteners a little then operate the axis. Continue in this fashion till all fasteners in place.
    Tops likes this.
  16. Tops

    Tops Silver

    I am going to do some 'reperations' on mine and whittle/sand by hand, the good fight via CNC will have to wait. Sort of bummed since I got a new power supply after the other quit and hoped it'd keep up. Can a person parallel DC power supplies to a stepper controller like a Gecko G540?

    Last edited: Aug 22, 2022
  17. BattyZ

    BattyZ Silver Banner Member

    I once cut out the ground and put 2 dc power supplies in series to get 24volts but never attempted a parallel.
    Tops likes this.
  18. Tops

    Tops Silver

    Yeah, I should probably RTM or contact Gecko to see what they think.
    Carry on, sorry for the interruption, more about the Effer!
  19. BattyZ

    BattyZ Silver Banner Member

    No worries mam, good to hear about other's computer-numerical-control pursuits!

    Carrying on! Took me a while to force myself to start the rewiring journey because we all know how much I totally adore wiring and my haphazard approaches have never turned around to bite me....BUT I used the opportunity for some personal growth. I have even planned ahead and threw in some extra 3-wire and 4-wire cables for the future! Near future; will be for limit switches. The wiring is not totally finished so no pictures. I do have video. Up and running again on Dynomotion's KFLOP board.



    Currently, the heavy focus is on the C programming side for initialization and kinematics. Super happy to have a friend whose job title contains the words "programming-systems-architect-level 5." He has seriously kept me from some long-winded multi-week rabbit holes. "oh just press these buttons, use this tool to do all variables at once, debug with this not that because.. etc"

    The "precision" right angle planetary gearbox for the current C axis has 5 to 8 thou at 3 inches out. Doesn't sound like much but would be .010-.050 out at 10-15 inches radially and that is going to make for a lot of sanding and extra work when assembling halves of molds. Lo and behold, there is yet another fleabay-special harmonic drive coming in. Got a downright steal on it because it is missing a crucial part, the input wave generator. Harmonic Drive, "Couldn't possibly quote just an input/wave generator, because everything that leaves their factory is a perfectly matched set." Soooo, found someone overseas to ship me flexible bearing and I will be machining the input hub in-house. As long as we are close on the profile (+/-.003) it should work well enough for my needs.
    Tops likes this.
  20. BattyZ

    BattyZ Silver Banner Member


    I rewired and got all 5-axes up and moving in a calibrated fashion:


    After spending a couple of days(and late nights) on getting the post processor where it needs to be to continue on my sample part; I came to a frustrating realization. Lack of speed/power! Individually the servos seem to run great and at a good speed. In coordinated motion, though, it is a totally different story. The Z-axis being the worst. Tried 2 different belt setups and even tried a new 10:1 planetary. I can get non-faulting-out motion...but have to force all movements below 25 ipm. The touter needs to be machining foam above 100 imp for it to make any toolpath doable in a sitting. Would prefer ability to get close to 200.

    So in short, the is57 servos may work great for smaller cnc routers, Effer has just quickly outgrown them. SO if anybody is looking for some let me know! I will be keeping the rotary ones, with the gigantic harmonic reduction drives they seem to be doing fine. For the moment anyways.

    This is where things start to get Fishy...Fishbonzy that is!

    Connected with FishbonzWV to get these big boy Servos! Thanks again!

    Now on to driving them. The matching drives are very limited, quite spendy, and have no warranty. Dynomotion makes a very nice board for driving servos. Unfortunately, it only goes up to 80V, and these call out for 230V. Mesa's 8i20 is currently out of stock. An all-the-way-down-the-rabbit-hole would be STMBL, but that project stalled due to a main onboard chip going out of production. So mix n match Fleabay to the rescue:

    Compax3 Servo Drives. With the F12 encoder feedback option so will have to convert the resolvers over. Will have more to post about these after a successful power-up. I am cautiously optimistic though!

    Getting back to Effer, these servos have a continuous rating of 1.1KW. More than plenty for my needs! Actually, ballscrew whip and frame bending under acceleration are now problems I will most likely have to fix. Once done, Effer should be a smokin' speedy unit. I would like to say confidently that smokin' will only metaphorically describe the speed but a literal definition is definitely possible. LOL

    Oh, and my flex bearing came in for making the wave generator I need for the 2nd harmonic drive.

Share This Page