Machining CNC mill conversion project begins

Discussion in '3-D Printing and Related topics' started by caveduck, Feb 13, 2019.

Help Support The Rocketry Forum by donating:

  1. Feb 13, 2019 #1

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Time to actually start a thread. I've started a project to get a decent bench mill and convert it to CNC operation, and the base machine is on order. After a fair amount of research and YouTubing, here is the general plan:
    • Base machine: Precision Matthews PM-30MV
      • Better grade Chinese unit
      • 2HP motor, 2-range variable speed with belt drive, 220V 1-phase (can run on my dryer circuit)
      • table 8.25 x 33"
      • travel xyz 23 x 8.75 x 14" (CNC conversion actually increases X travel by a few inches!)
      • ~500 lb total, no single component over 100 lb so quite manageable
    • CNC mechanical conversion kit: ArizonaVideo99 (Dave Clements)
      • Double nut ballscrews for zero backlash
      • Z screw size increased to 25mm
      • Mounts and couplers for NEMA 23/34 motors
    • Controller - 4-axis stepper box from Automation Technology
      • Leadshine 4 channel stepper driver
      • Ethernet SmoothStepper interface
      • 4th channel will be used for a rotary table
    • Software - Mach4 and/or LinuxCNC
    This photo is from the PM page and shows what it looks like before conversion.
    upload_2019-2-13_14-18-25.png
     
    eggplant, ttabbal and DRAGON64 like this.
  2. Feb 18, 2019 #2

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Update #1. After a lot more rummaging I've decided to go with DMM servo drives and motors. This machine is on the upper end of being reasonable to run open-loop steppers, which leaves servos and closed-loop steppers (stepper with a shaft encoder on the back). There are not many vendors of large-enough closed loop stepper rigs - basically just 2 - so it's gonna be servos. That overturns being able to use a ready-made control box though.

    The control electronics is going to be an Ethernet SmoothStepper from Warp 9 Technologies with a C62 breakout board from CNC4PC. The servo drives need noise filtering so I'm going to talk to CNC4PC about getting an enclosure with the power system set up that way. The C62 has all the I/O's we'll need, with 6 drive channels plus VFD speed and direction control outputs and plenty of inputs for home/limit switches.

    I made block diagrams of the control system over the weekend, but am going to wait to post since they are still in flux and you can't really update posts here anymore.

    Still waiting on a shipping notice for the PM base mill, hope it's pretty soon! Got the garage just about cleared out enough now.
     
  3. Feb 21, 2019 #3

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    The mill shipped today...should have it in about a week. The new revision ArizonaVideo99 mech conversion kits are about ready to ship as well. Working on the electronics orders to Digi-Key and CNC4PC.
     
  4. Feb 26, 2019 #4

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    I called Saia to check on shipment status and It's getting delivered tomorrow! Meanwhile I've ordered the bulk of the electronics from cnc4pc, though I realized yesterday that I forgot to order the induction limit switch hydra. They were very helpful on a couple of pre-order email exchanges and confirmed that the C62 BOB (break out board) will do what I need. Still waiting on the mech ballscrew kit to be ready. Dave Clements (ArizonaVideo99) has also been really good with suggestions and info. Also started looking into 6" rotary tables for the A axis. Most of the ones I've seen so far look like putting on a NEMA 23 servo will be pretty simple.

    Photos begin in earnest when the crate shows up.
     
  5. Feb 27, 2019 #5

    Mendal

    Mendal

    Mendal

    Well-Known Member

    Joined:
    Feb 5, 2016
    Messages:
    209
    Likes Received:
    8
    Curious what your overall investment in the machine will be and how rigid it will be. You may find that tolerances might be hard to hold if the machine has flex in it. I don’t know anything about this exact machine and I wish you luck and will be watching this thread.
     
  6. Mar 1, 2019 #6

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    I promised some photos...here it is in my garage with the crate off, before starting to clean off the protective uber-viscous oily stuff. The little plastic box taped to the X table contains the drawbar, handwheel cranks and a few other bits of hardware. The Saia driver and I got the 649 lb crate it up my slightly sloping driveway with nothing more than a pallet jack; it was actually not much trouble. Mechanically it seems to have come through shipment PA to CA just fine. I took a bunch of video snippets to make an uncrating vid with iMovie, but I have no experience doing that and if it sucks too badly it will not see the light of day :)

    There is a lot to discuss about the project but I'm going out to Holtville Havoc in the morning to actually launch something.
    upload_2019-3-1_1-6-54.png
     
  7. Mar 17, 2019 #7

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    I have not been idle on this project, just not posting :) I've finished reverse engineering the factory spindle power/speed/direction controls and designing the mods for CNC conversion.

    Background: The CNC control electronics are an Ethernet SmoothStepper card driving a CNC4PC model C62 breakout board. The C62 has a large number of IO ports that are controlled by the two emulated parallel ports on the SmoothStepper. I have a drawing going of this but am waiting to post until it's closer to final. The C62 has outputs specifically designed to control VFD spindle speed and direction.

    Most of the factory spindle control is pretty standard, and there's a nice sticker with a schematic on the side of the PM-30 electrical control box. Things of note:
    • The power switch / E-Stop loop that controls the main power relay is 240V. So I need a 2nd pole on the E-Stop button to bring it out as a low voltage contact to the C62 external E-Stop input.
    • The motor speed control is 0-5VDC. This should be OK as the C62 breakout board VFD outputs will handle up to a 0-10V range. We're giving up 1 bit of precision but that will not matter.
    • The spindle direction inputs are 5VDC active low; these will be driven by the C62 breakout board.
    • The factory E-Stop button is actually a fairly nice IDEC style one (though I suspect it's a clone rather than the real thing), set up for single pole but with space for two more contact blocks to add extra poles. I was going to replace this with a good Omron 2-pole switch, but instead I just ordered a couple of YW-E01 contact blocks from Newark and will use the Omron in the servo box.
    • The DC grounds on the spindle speed and direction switches are tied together, so I need just 4 conductors on the CNC transfer switch, which is going to get added to the front of the PM-30 electrical panel.
    • The motor speed is a bit jumpy at the low end. Not clear whether this is due to a cheesy pot or the motor itself. It probably doesn't matter; at all useful speeds it's pretty smooth.
    • At full speed on the high pulley setting we get 2940 RPMs vs advertised 3000. Not bad. The last 10-15% of the pot travel seems to have no effect so the motor is probably maxed out at around 4.5V on the control input.
    • Circuit protection is with a fuse. Planning to replace with a resettable breaker (I HATE expendable fuses) but I haven't picked one out yet.
    • If you want to remove the motor (which I'm going to do before removing the head from the Z stage to reduce weight), you have to unplug the motor connectors from the control board and thread them back through the holes in the head. This in turn requires removing the wires from the 3-pin motor power connector, since the connector will not pass through one of the holes.
    • The unit comes without a power plug, so I wired it up with a Leviton L6-20 locking plug to match my power hydra. I wish the power cord was beefier (for mechanical rather than electrical reasons - the spindle only draws 7A or so) but the connectors inside the power panel aren't well set up for that.
    • The spindle tach runs on 5VDC from the motor control board, but the DRO on the quill downfeed is battery powered.
    servo_ctl_schematics_spindle.jpg
     
  8. Apr 22, 2019 #8

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Now I'm *way* behind on posting. A lot of the electrical mods are done and I've got the CNC transfer switch in place with the spindle powering up and operating fine in manual mode, the electronics tray is 2/3 built and wired, the mechanical conversion kit has arrived, the stand is on a mobile base, whew!

    So let's start catching up. Here's the original PM-30MV schematic sticker they so handily put on the side of the electrical box. This plus a bunch of (careful!) DVM probing let me come up with the design drawing above.
    upload_2019-4-22_13-32-53.png

    Now on to the first problem - a problem of POWER. I had a nice 30A 240V dryer socket in the garage, fully available because we have had a gas dryer for a long time. However, it looked like this:
    upload_2019-4-22_13-39-30.png
    This is an old style 3-wire plug from when the house was built in 1981 that only has the two hot poles plus neutral; there is no separate ground. The code was changed in 1995 to require a discrete ground conductor. Neutral basically IS ground - it's tied to earth ground at the panel - but having a separate ground helps reduce the chance of accidentally making the outer case of your dryer live if somebody swaps hot/neutral when wiring up the appliance. Soooo...a look inside the J-box:
    upload_2019-4-22_13-45-44.png
    shows that we do have an actual ground lead. Hooray, everything is there to install a modern correct 4-wire 240V dryer socket.
    upload_2019-4-22_13-49-14.png
    Nice n' shiny! 7.2 kVA of available power, just what the doctor ordered.
     
  9. Apr 22, 2019 #9

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Moar POWER! OK just having a righteous outlet is not enough. I need to get power from that outlet out to the 2-3 machines that are going to need it. Following a lead from some home shop video (sorry forget the exact source) I built a power hydra designed to be capable of carrying 25A @240V. Was hoping for the full 30A but the ampacity of the 10/4 cable limits it to 25. I started to construct it with 10/3 cable that is rated for 30A but had to go to the 4-conductor 10/4 in order to make 120V available. The hydra was a decent project in itself.

    Here's the parts layout. I managed to get almost everything at Home Depot except the eventual 10/4 cable.
    upload_2019-4-22_14-7-38.png
    Cutting the plastic conduit that joins the nonconductive J-boxes on the lathe:
    upload_2019-4-22_14-9-35.png
    Wiring underway. In this revision it still has 3 L6-20 240V 3-wire outlets, before I switched one to be a 4-wire L14-30 for the servo control box. The interconnect wire is all 12 gauge THHN, stranded but still stiff as all perdition.
    upload_2019-4-22_14-13-52.png
    Things are tight in the central J-box; I had to be careful with wire length and curling to fit. A cable tie around each group of conductors helped a LOT with getting the wire nuts on securely.
    upload_2019-4-22_14-17-58.png
    Buttoned up, all leads checked with the DMM, and labeled.
    upload_2019-4-22_14-19-23.png
    And finally, here's the L6-20 plug I installed on the mill's power leads, which are unterminated from the factory. Yes the 20A plug is somewhat above the needed capacity, but I wanted to standardize the outlets so I would not have to re-scramble the hydra depending on what kind of lathe and bandsaw are eventually used.
    upload_2019-4-22_14-23-49.png
    Proof of the pudding: spindle has POWER!!
    upload_2019-4-22_14-26-14.png
     
    Last edited: Apr 22, 2019
  10. Apr 23, 2019 #10

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    One thing I glossed over in the power hydra build is that the wall plate covers are customized. I looked through everything Leviton makes, and it turns out they don't have anything for L6-20 and L14-30 outlets in the dual J-box format. So for the dual L6-20 box I got a blank wallplate and bored it using hole saws. I did a prototype with a Dremel, then got fed up with the tedious work and found you can use a hole saw in a drill press just fine; my concerns about cracking the plates were unfounded.

    Bonus fun - the 1 5/8" hole saw you need for the 1.6" diameter plugs is one of a smattering of sizes that are *not* stocked by Home Depot. You'll need to order this online. If you do that, be sure to also get a 2 1/4" saw because that's the other major gap in the smaller sizes.

    The plate on the left with the 240V socket and two 120V sockets is made from a "modular" wallplate, where the dual 120V panel is stock and the other one is a drilled out lightswitch (old style flip type) panel.

    It turns out you don't want to start with a blank dual plate; their mounting screws go where the outlets need to mount in a populated plate! The right hand J-box above is actually made that way but there are extra nuts on the mounting bolts behind the plate so that the same bolt can both retain the outlet and hold down the plate. If I ever need to open up that side again I'll go get a dual flipswitch plate and drill it out for the L6-20's; then all the mounting hardware will be conventional.
     
  11. Apr 24, 2019 #11

    Bikechain

    Bikechain

    Bikechain

    New Member

    Joined:
    Mar 13, 2019
    Messages:
    3
    Likes Received:
    2
    Gender:
    Male
    Location:
    N. Fl
    Great job! I'm curious, are you upgrading from a previous mill? It's impressive the work you've completed so far. I'm fascinated by CNC milling and 3D printing, it's an endless rabbit hole. I have zero experience and thought I'd find some software to learn how to design and then possibly get a printer. Do you have projects ready to begin when your conversion is complete?
     
  12. Apr 24, 2019 #12

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    @Bikechain This is my first go at a real mill; been improvising with a drill press for far too long. Doing a CNC upgrade to a manual mill is a great project in itself...I came to it by way of spending 3 years using and improving a mini-lathe, which has been the handiest thing ever for making rocket parts. For software I'd check out either Fusion360 or OnShape - I like OnShape but Fusion360 seems to have the most traction. The mill is going to get used for making a lot of electronics bay and launcher parts in aluminum, molds for nose cones and glider pods, fiberglass fins (slower than a router but will work ok), pretty much anything that needs profiling, slotting, or repeatable hole patterns. Also might try making some guitar necks.
     
  13. Apr 29, 2019 #13

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Short but significant update - I got spindle speed and direction control fully working over the weekend, along with the limit switch array. I was able to issue MDI G-code spindle commands from Mach4 on the computer and have the spindle turn on and go to the commanded speed in either direction. Much happy!!

    There is a specific way you have to configure the ESS+breakout board to support a DC variable speed motor with analog speed control as found in the PM-30MV. The key is you must set up the spindle in ESS as a PWM output with the PWM and DIR signals on two specific pins P1_14 and P1_16. These pins are decoded automatically by the C62 BOB, and the BOB drives relays for the FWD and REV direction signals to the motor controller, and decodes the PWM signal to drive the dedicated 0-10V analog speed output. You then need to adjust a pot on the C62 board to cut the output range down to 0-5V. All this behavior on the C62 is pretty much hard-coded - anytime it sees the PWM signals on the required pins, it drives the relays and analog output as needed.

    For the limit switch array (cnc4pc model A61), there is a nice RJ-45 connector for the inputs. The X, Y, Z, and Z limits come in on ports 1_11, 1_12, 1_13 and 1_15 respectively. They are PNP type, so you need to set the COM JUMPER to the 12/24V or 5V position depending on what external power you are feeding into the C62 (I'm using 24V, the sensors will work on anything from 5-24). Then you jut configure those pins in the Mach4 ESS plugin as active high. I did have to restart Mach 4 to get them to get recognized afterwards. You can test the inductive sensors by holding a bolt head up in near contact with the sensor head; they each have an LED that will go out when the signal is activated, and you can check them in real time on the "machine diagnostics" tab on the Mach4 main screen.
     
    Last edited: Apr 29, 2019
  14. Apr 30, 2019 #14

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Time for some more backfill with photos. I'm going to start with the design for the AC power section of the electronics box, since it's mostly all built and working.
    cnc_power_section.png
    And here is the interconnect diagram for the digital brains of the system, the Ethernet SmoothStepper (ESS) and its C62 breakout board (BOB):
    ESS_C62_interconnect_diag.png
     
  15. Apr 30, 2019 #15

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    After a couple of iterations on the power section design, I felt comfortable ordering some of the parts to start doing a layout in the electronics enclosure.
    upload_2019-4-29_22-22-56.png
    Above are the Schaffner noise filters (lower L), Meanwell triple output 5/12/24VDC power supply (lower right), circuit breaker and contactor relay (center L), the Omron E-stop switch (upper L) and the Weidmuller terminal blocks (upper R). The terminal blocks are great; they have 8 total contacts that will take 10-12 gauge AC wire, lock together with side dovetails that let you make a "brick", and they slide on a DIN rail for mounting. Just about perfect for this application.

    Here is the factory E-stop switch inside the mill head control box, showing what kind of contact I needed to get to add a pole. Details of the printing on this pole unit don't exactly match the photos on Newark, so I suspect that the factory parts are Chinese knock-offs, but they look serviceable, and the genuine IDEC pole switch I got plugged right in.
    upload_2019-4-29_22-29-53.png
     
  16. Apr 30, 2019 #16

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Next I started working out the mechanical layout inside the electronics enclosure. My general method was to slap the devices on my scanner and print actual size paper templates that I could shuffle around on the raised bottom panel of the enclosure.

    upload_2019-4-29_22-48-40.png

    One of the main concerns was positioning and clearance for the 240V inlet plug. The above photo shows the 3-pin version before I went to the L14-30. There was a hole for a 120V inlet plug in the I/O panel (at bottom) but there was no way to enlarge it enough for either of the 240 plug sizes.

    Here's another preliminary layout after I unbolted the panel and made some more templates. Things are getting close to their final positions:
    upload_2019-4-29_22-49-34.png

    From here the main changes would be to rotate the smaller noise filter 90 degrees to provide more clearance for the disconnect terminals, and to add a ground bus bar behind the terminal blocks.

    Finally here is the full layout of drilling templates as I'm getting ready to start center-punching the hole locations. I should mention that the quad DYN4 motor drive template was done in Gimp to replicate the base pattern with the specified spacing between the drives. I also found out after the fact that it's pretty hard to remove Sharpie markings from the orange paint they use, even with various solvents.

    upload_2019-4-29_22-50-15.png
     
  17. Apr 30, 2019 #17

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    About the same time, I modified the case to add the inlet plug. This was done by center-punching, drilling a ~3/16" pilot hole, and then using a 3" hole saw, which worked surprisingly well on the sheet metal. The inlet plug itself was then used as the drill template for three #8 mounting bolts. The larger precut hole below it is for a standard 80mm case fan, which will get populated.

    upload_2019-4-29_22-59-2.png

    Continuing with the case modding, I enlarged the factory-cut square hole for the circuit breaker to hold the larger 2-pole breaker needed for 240V systems. The aperture was cut with a Dremel cutoff wheel and finished up with a flat file.
    upload_2019-4-29_22-58-11.png

    The new and beefier breaker:

    upload_2019-4-29_23-1-43.png
     
  18. Apr 30, 2019 #18

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Here is everything but the DYN4 servo drives bolted down. You can see some position adjustments from the previous layout, and the ground bar has been added.

    upload_2019-4-29_23-12-13.png

    And now the power section has been wired up. I'm feeding it with an umbilical from the breaker in the enclosure...all the AC disconnect terminals have been fully covered with heat-shrink for protection. You can also see that the screws in the terminal blocks are recessed; you'd have to make some effort to zap yourself. In the picture here you can see LEDs on the ESS/C62 combo are lit, and the case fan is spinning.

    upload_2019-4-29_23-12-45.png
     
  19. May 1, 2019 #19

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    And now for something completely different!

    I don't have a lot of space in the single garage slot that's available for larger equipment. The mill with its ~38" span along the X table with motor and around 14" of travel in each direction needs a fair amount of clearance. There was quite a list of constraints on placement:
    • Need to be able to remove dead washer/dryer units from the laundry alcove
    • Clearance in front of the mill to move around it with the Y motor sticking out in front
    • Not hitting anything at maximum X table travel in both directions
    • Need a place to set the largish electronics cabinet next to it (fortunately the X table height clears it)
    • Leave enough room for a 12x36 (or so) lathe behind it
    So I stole an idea from Adam Savage and made a foam board template with an Elmer's science project tri-fold board and some packing tape. It was *really* effective at letting me visualize clearances. The outcome was two-fold:
    • I became pretty sure would want to move the machine later; there are too many variables in the somewhat confined space
    • I couldn't find any documentation on how thick the concrete slab is, and got pretty nervous about putting even fairly short (2") masonry anchors into it. The builder of my house is known for taking some schlocky shortcuts and I did not want to risk puncturing a thin slab.
    At the same time I discovered mobile bases in the Grizzly catalog. These things are low-slung carts with casters and lever-actuated rubber feet that will let you reposition machinery, and also can increase the base footprint for machines with smaller bases (like the PM-30MV). They come in a range of sizes that will hold much larger and heavier machines than the PM-30...I ended up getting the smallest one:

    upload_2019-4-30_22-12-43.png

    The way you use these is to make a 1 1/2" plywood platform by laminating two pieces of 3/4" plywood, bolt the the plywood to the mobile base and bolt the machine to the plywood.
    I got two 2x2 squares of 3/4 ply at HD, trimmed to about 21"x21", and laminated with Titebond 2, throwing a few heavy things on top for pressure:

    upload_2019-4-30_22-16-57.png

    For assembly, I just turned the steel PM-30 base upside down and went at it. I drilled holes for 3/8" bolts through the steel and wood with a good twist drill in a cordless drill and assembled everything with nylon insert locknuts.

    upload_2019-4-30_22-19-50.png
     
  20. May 1, 2019 #20

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Here's the PM-30 steel base on the completed Grizzly mobile base. It seems pretty stable even when I get up and stand on top. You need to adjust the height of the deployable rubber feet carefully so that the base is close to level when they are deployed. It doesn't have to be perfect, the floor is not totally flat anyway and fine tuning will have to be done at the final location.

    upload_2019-4-30_22-23-26.png

    ...drumroll....here's the eagerly awaited PM-30MV mechanical conversion kit from ArizonaVideo99 (Dave Clements). Top to bottom are the Y, Z, and X axes. Top L is the 2nd support bearing for the long X axis. These are nice ballscrews with preloaded double ballnuts for zero backlash, and all-aluminum brackets for NEMA-34 motors. The ballnut travelers have Zerk fittings for lubrication, and the X axis has nice angular contact bearings at both ends (the stock PM-30 basically has nothing). Everything is very smooth and I can't feel any backlash at all. Before installation I'll probably change the Zerk grease fittings to regular oil line fittings so I'll have the option to add a one-shot oiling system later.

    The assembly plan is to dismantle the mill where it still sits on its pallet - none of the parts are over 100 lb, modify the Y saddle as required to clear the larger ballnut hardware, possibly do some other mods (oil distribution grooves in the ways, add oil lines and fittings) and reassemble the unit on the mobile base with the new screws installed.

    upload_2019-4-30_22-34-35.png
     
  21. May 3, 2019 #21

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    1052E3AF-3A37-48AA-B588-1954FDE484CA.jpeg 38D65C9A-19E5-4575-8380-FD2DF7945BE7.jpeg Wired up power to the DYN4 servo drives, hooked up a motor and got it to move with the DMM serial utility.
     
  22. May 10, 2019 #22

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Murderer's Row of DMM NEMA-34 750W servo motors hooked up for testing! Torque aplenty to move 200lb of table/saddle/vises. They have an encoder on the back for closed-loop positioning.
    upload_2019-5-10_15-0-5.png

    Here's an overview of the whole testing setup as it now sits. It's now 90% complete on wiring checkout, I think all the hard stuff is done in the electronics.
    upload_2019-5-10_15-5-23.png

    There is a little bit of a height issue with the DB25-to-RJ45 adapters on the servo drives: there is only about 1/8" of clearance between the cable bend in the orange cable and the cabinet door.
    upload_2019-5-10_15-8-10.png

    I've ordered some Tripp Lite right-angle ethernet cables that will give me about 1/2" more space. The other cables visible in the photo will not be a problem once dressed.

    The only wiring remaining to do is for some RJ-45 plugs in the back panel (ethernet to the SmoothStepper and connector for the limit switch hydra), the 3 aux 120V outlets, and the ribbon connection for the manual control MPG pendant to the C62 board. Overall I think there is not a lot of technical risk left in the electronics box build.
     
  23. May 12, 2019 #23

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Here's an update to the spindle control schematic showing the wiring for the 6-pin CNC remote cable. This cable also carries the E-stop contact pair back from the E-stop button on the mill so that the CNC control can halt the servos as well as the spindle when you whack that button.

    There is some chance this will all get revised later if I eventually change to a 3-phase AC VFD spindle drive. That would give me a motor that could be driven up to around 5000 rpm - in conjunction with the existing 2:1 belt drive and a spindle bearing upgrade I would be able to get up around 10k rpm, which is helpful for using smaller tools on aluminum.

    servo_ctl_schematics_spindle.png
     
    warnerr likes this.
  24. May 12, 2019 #24

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    A little more on the motor upgrade possibilities - the ones of interest are the Marathon microMAX series of AC 3-phase motors from Automation Direct. There is a 1.5HP model Y284 and a 2HP version model Y285. They come with encoders and can be driven from 0 to 5400 RPM using a suitable inverting VFD that allows you to run the 3-phase motor from 1-phase 240V. They have constant torque from 0 to the base 1800 RPM and will yield 10k RPM on the spindle as long as you put in bearings that can take it.

    There is some information about using this motor on the russstuff Youtube channel here:

    The main page for the Y284 on Automation Direct is (sorry for ghastly URL, couldn't find a shorter one)

    https://www.automationdirect.com/ad...hon_max_plus_w-z-_encoder_(0.5hp_to_5hp)/y284

    and the tech specs for the whole line are here: https://cdn.automationdirect.com/static/specs/motorsmaxplus.pdf
    upload_2019-5-12_3-50-36.png

    A while back there was a batch of the 1.5HP Y284 motors being sold off for a bit over $200 from Dealers Industrial Equipment, and a bunch of people rushed to put them onto their Grizzly G0704 mills. Sadly the deal on that particular motor seems to be over and they are back to around $1K, so this enhancement is going to wait a bit.

    VFD: You can run these motors with a WEG CFW3000 series micro-drive VFD (check automation direct for these too). They're reasonably priced and come in up to 3HP sizes. Be sure and get the 1-phase input version. The DuraPulse GS3-22P0 should also work nicely and is a bit more nicely designed.
     
    Last edited: May 12, 2019
  25. May 13, 2019 #25

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Here's the backside of the on-mill spindle control panel after modification. You can see the 2nd pole on the E-stop switch at bottom center, and the 4PDT CNC transfer switch at middle right (small red rectangle). This switch had to be spliced into the middle of four leads behind the panel (see schematic). Heat-shrink tube is really your friend in this kind of work. The 6-conductor DIN plug is barely visible at the top left corner of the front panel - the twisted pair of yellow E-stop wires lead to it.

    upload_2019-5-12_20-9-8.png
     
  26. May 15, 2019 #26

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Got all 3 servomotors jogging along a couple of nights ago. Here's a quick outline of what I did (photos to come)
    • Set the 4 driver configuration DIP switches on the C62 BOB to OFF
    • Set the "driver fault" config jumpers on the C62 to the pins 1-2 position *only for the drives in use* - i.e. drives 0, 1, and 2. In the C62 manual it talks about this being a cable disconnect detect. In practice, if you leave them on the default 2-3 position, the motor will stop moving after a little bit but Mach4 will keep sending pulses and incrementing the position.
    • Use the provided USB-serial adapter and configure each DYN4 drive with the DMM-DYN4 utility
      • The utility doesn't install its own shortcut correctly - go fix it manually
      • Use "detect COM port" to find the correct port - should work fine
      • Use the "read" button to read out the current config of the drive
      • Change control mode from the default RS-232 to STEP/DIR
      • Set "gear num" to 8192 (you can probably get anything from 4K to 16K to work, interacts with Mach4 motor params)
      • Use "save" button to configure the drive
    • NOTE: we'll run the drive auto-tune once the mechanical loads are connected
    • NOTE: you can leave the DYN4 serial cable connected and the DMM-DYN4 utility running while you operate the motor via Mach4. This makes parameter tweaking quite a bit more efficient.
    • In Mach4 config on the Axes tab set motors 0, 1, and 2 to be step/dir mode (the default).
    • In Mach4 config, go to the "motors" tab and set for each motor
      • steps/unit: 40k (play with this value to get the distance indication right, varies with leadscrew pitch etc.)
      • units/min: 300 (I have on good authority you can get these motors up to 400ipm, being conservative for now)
      • accel in units/sec**2: 2.0 (also conservative, crank up later to find useful range)
      • The UI for this is flaky - after setting the values, be sure to switch to a different motor, switch back and check the values again. It appears that when you change the steps/unit, that the others auto-adjust to compensate, which I generally didn't want.
    • Now in Mach4 you can click "enable", go to the "jogging" tab and try to make the motors move. The Diagnostics | Logging screen is really useful when you get unexpected E-stop conditions and the like.
    • Adjust the steps/unit and gear_num parameters to get the correct number of turns per inch. in my case with 5 pitch ballscrews I'm setting things up to get the Mach 4 indicator to show 1.000 inches for 5 turns of the motor.
     
    Last edited: May 15, 2019
  27. May 16, 2019 #27

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Jog pendants! I had a look at these a couple of months ago; most of what I found then were some vaguely discouraging Chinese units. Somehow I missed VistaCNC, http://vistacnc.com/index.htm which is right in my backyard in San Diego. Here is their P4-S model, which has a USB interface to connect directly to the computer. You can get this with a 16' coiled USB cord and a magnetic sticker to put on the back which will let you hang it on the machine base or column. There is Mach4 specific firmware and a Mach4 plugin. In addition to 6-axis motion, it lets you control the spindle, cycle start/pause/stop, and constant feed rate for manual style milling. A version with an external hard E-stop relay is also available, though I don't really need that. I emailed them and will try to meet and pick up a unit in person if possible.
    upload_2019-5-15_16-56-53.png
     
  28. May 16, 2019 #28

    caveduck

    caveduck

    caveduck

    semi old rocketeer

    Joined:
    Jun 6, 2011
    Messages:
    1,172
    Likes Received:
    52
    Backing up a little bit, here's my brand new uncrating video!
     
  29. May 16, 2019 #29

    OverTheTop

    OverTheTop

    OverTheTop

    Forum Supporter TRF Supporter

    Joined:
    Jul 10, 2007
    Messages:
    2,662
    Likes Received:
    429
    Gender:
    Male
    Location:
    Melbourne Australia
    Nice find with that pendant!
     
  30. May 17, 2019 #30

    new2hpr

    new2hpr

    new2hpr

    Well-Known Member

    Joined:
    Jan 23, 2009
    Messages:
    918
    Likes Received:
    9
    This is a great read so far. Excellent detail and photos! I may be doing a CNC upgrade to my Dad's Enco mill and possibly his 13" lathe at some point.

    With all the effort and expense, did you consider linear encoders rather than the shaft encoders? Not sure what level of accuracy/precision you're looking for. I'm curious about the tradeoffs between the two. Shaft encoders are certainly easier to mount and keep clean!
    -Ken
     

Share This Page