Minamil: a minimal CNC mill. And friends.

Each axis: $̶5̶$8 motor+lead screw, 3x LM6UU, 3x 6mm x 100mm rod, 1/8in hardboard, PC case screws

Similar projects worth following
Developing a cheap small 3-axis CNC mill for $̶1̶0̶$15/axis.

Powered by:
$̶5̶$8 motor+lead screw CH-SM1545 (low price seller raised price - hello supply risk)
Linear bearing cost minimum at 6mm (smaller costs more)
1/8in hardboard is practically free per small area
#6-32 x 1/4in computer case screws are practically free
3x 6mm x 100mm rods per axis can be $1 each or 6mm rod is common in printers etc.

It's working. Log entries say more than the "Details" at the moment.

Not much here yet. See logs for what might end up here.

XY test
XY table test under fixed spindle

First test of minimal build of axis concept (already teased at #CDCNC):

First there was #CDCNC, a highly-constrained just-barely-functional one-off toy built by improvisation with found junk with simple tools.

Here I'm developing an idea that came from thinking about whether or not there is any space between a dumb stunt like CDCNC and a commercially (i.e. efficiently) produced entry level CNC mill. Encouraged by early results -- see project logs -- I think maybe so.

In contrast to CDCNC, this is about reproducibility from a simple BoM and economical access to laser cutting. And somewhat greater capability.

Next there may be #"Desk Accessory" CNC Milling Machine

Parts prices

tracking because they've been a bit volatile and small changes to small prices can be significant.

  • CH-SM1545 -- motor + leadscrew
    • Banggood 15 Oct 20:
      • spend: 3x $6.48 + $2.90 = $22.34 "10-15 business days" to US
      • unit cost: $22.34 / 3 =  $7.45
      • ... but IME BG shipping options at checkout vary between item view and checkout and over time
    • Aliexpress "TO-HO store" 15 Oct 20
      • was $4.78 shipped in Aug and some longish time before that, which made a stronger argument for designing around this part!
      • spend: 3x $6.97 + $3.47 = $24.38
      • unit cost: $24.38 / 3 = $8.13 (+70% alas)
      • Still says "free shipping" on search hit, but not
    • Aliexpress "Shenzhen Chihai Motor Store" 15 Oct 20
      • spend: 3x $6.50 + $8.45 = $27.95 ePacket "12-20 Days"
      • unit cost: $27.95 / 3 = $9.32
      • possibly more consistent quality
  • LM6UU -- 6mm linear bearing
    • Aliexpress "CNC Factory Supply" 15 Oct 20
    • spend: 1x 10pcs $5.15 shipped "19-39 Days"
      • unit cost: $5.15 / 9 = $0.57
      • extra: 1
      • current minimum $4.39 @8mm
    • Aliexpress "DZPR Official Store" 15 Oct 20
      • spend: 1x 10pcs $7.54 shipped "12-20 Days"
      • unit cost: $7.54 / 9 = $0.84
      • extra: 1
  • 6x100mm rod
    • Aliexpress "CNC Factory Store" 15 Oct 20
      • spend: 5x 2pcs $2.64 = $13.20
      • unit cost: $13.20 / 9 = $1.47
      • extra: 1
    • Aliexpress "SNWNYN Store" 14 Oct 20
      • spend: 1x 4pcs 6x300mm $4.07 + $2.24 = $6.31
      • unit cost: $6.31 / 9 = $0.70 + means to cut w/small kerf
      • extra: 1x300mm
        • or
      • spend: 2x 4pcs 6x300mm $4.07 + $2.24 = $10.38
      • unit cost: $10.38 / 9 = $1.15 + means to cut
      • extra: 4x <100mm, 1x <200mm, 3x 300mm (free rods for next unit if later find means to cut with small kerf)
    • Scavenge
      • pretty common in printers, etc.
      • need means to cut (probably not a hacksaw)
  • Screws #6-32x1/4in -- general assembly
    • Grainger
      • spend: 1x 100pcs pan head Phillips $1.79
    • Scavenge
      • Common PC case screw
        • but need a pile, so less likely by casual scavenging
        • unless building less, like just one axis
  • Screws #6-32xSmaller -- mounting motor/leadscrew units
    • roll your own
      • spend: $0.00 if you drop no more than 2 from the 100pc bag (current rev)
      • unit cost: $0.00 + file
      • 12x file heads to <2mm thickness
      • 4x file length to material thickness
    • Supermicro 3.5in hdd caddy screws (via Amazon seller)
      • spend: 1x MCP-410-00005-0N $5.49 shipped
      • unit cost: $5.49 / 12 = $0.46
      • extra: 88x screws, 24x caddy labels
      • but probably not if you don't already have some
    • other
      • not really a standard screw (that I know of - comment plz?)
      • #6-32x4mm screws exist
        • some of which have thin heads
      • thin head
      • thin/shallow countersunk flathead ideal for constraining position
  • Material
    • 1/8 in hardboard
      • finish one side
      • ~15"x15" for current design
    • Home Depot
      • spend: 1pcs 2'x4' $3.37
      • fits 3x current cut layout + ~9"x48" extra
      • unit cost by potential cut count: $1.12
      • unit cost by area: $0.66
      • extra: 2x complete units + ~9"x48"
        • or
      • spend: 1pcs 4'x8' $6.74
      • fits 18x current cut layout + ~9"x48" extra
      • unit cost by potential cut count: $0.37
      • unit cost by area: $0.33
      • extra: lots

Example cost

$22.34 - 3x motors

$5.15 - 10x bearings (1 extra)

$13.20 - 10x rods (1 extra)

$1.79 - 100x screws

<$1.00 - ~15"x15" 1/8" hardboard

total <$43.50 (+tax)

Missing the early $10/axis target, but still <$15. Motor price hike accounts for ~$2.50 of that. Scavenging 6mm rod would get to $10.

  • Naming the project/instance/class

    Paul McClay10/20/2020 at 05:14 0 comments

    Writing about things requires words to call them by.

    This project started with the explicitly provisional title:

    Cheap small CNC mill - "Formula 1551" for now

    This log entry corresponds with renaming the project to:

    Minamil: a minimal CNC mill. And friends.

    So far this project has mostly described a single line of development of a specific machine. I've had difficulty thinking of how to write about both the single design and the general ideas that make it go. Thinking of other variations on this theme that I hope to build aggravates that by disallowing lazy conflation of class and instance.

    So I'm thinking of

    • describing the class as "sub-mini CNC mill", and
    • calling the current instance "Minamil", it being a minimal CNC mill.

    Minamil as a play on minimal mill.

    Sub-mini because mini mills already exist as a category and this doesn't pretend to compare with any of that. Some mini mills already use "micro" to imply the smaller end of mininess so I could say "nano", but SI prefix escalation rarely ages well IMO. The "sub-" can connote both smaller-than and less-than.

    And friends because I hope to complete at least one variation that will be intentionally less minimal, and to provide enough information to enable replication of the specific machine and extension of the general ideas.

    But isn't this more like a CNC router than like a CNC mill?

    I can't refute that. I do think the generally taller-than-wide configuration with a relatively small work area on a relatively bulky XY cross slide arrangement under a relatively massive spindle attached via a Z-only axis that's not short to a bulky vertical frame evokes mill more than router. And I think the pretentiousness of calling it a "mill" hides in plain sight the pretentiousness of calling a toy-scale device a tool at all. Kind of like an Easy-Bake® "oven".

  • Go Faster? Yes.

    Paul McClay10/19/2020 at 21:48 0 comments

    tl;dr: 12x

    So far I've run the same "feeds and speeds" as the prior project  just to get things working. Lately I've tried pushing a little harder.

    #CDCNC can manage extremely conservative cuts with a 1 mm end mill. Typically 0.15 mm step-down, 40% step-over, and nominally[1] 500 mm/min feed. That yields a material removal rate (MRR) like the material addition rate of grass growing. Numerically, 0.03 cc/min or two seconds per microliter. This machine should be able to do better, but I hadn't tried much beyond an early stab at cutting deeper. Until eventually I got around to trying.


    (summary below video)

    More briefly:

    • 3x feed rate: 500 -> 1500 mm/min
    • 2x depth of cut: 0.15 -> 0.3 mm
    • 2x stepover: 40% -> 80%

    Those gains multiply out to 12x MRR in a straight line.

    I think it's fair to call that "faster".


    The X & Y maximum "rapid" motion speed capped the feed rate. In other words, the machine cuts at g0 speed. In other words, "rapid" isn't very.

    On one hand, that limit has floated around pretty conservative numbers while sorting other faults and may hide some more available speed. I didn't try.

    On the other hand, the 0.5 mm pitch lead screws have to spin pretty rapidly to produce "rapid" motion and the top end of that may be not very tall. That's the trade for useful axial thrust from 15mm can-stack steppers. At 40 steps/mm and 8x microstepping (iirc what worked best when I poked at that a couple years ago), a diagonal 1500 mm/min "rapid" eats 16k step pulses/second. Doubling that would exceed Grbl's nominal max pulse rate (unless that's per axis?). All assuming sufficient torque at whatever speed. So if the motors have potential to make torque faster, the microstepping may need review.


    I've been running the dremeloid rotary tool at minimum speed to avoid rubbing at slow feed rates. Except that's not really how it worked. A major part of the prior project was pushing up the feed rate to escape from rubbing at the minimum spindle speed. Either way: minimum speed. About 10,000 rpm but spontaneously and seemingly randomly variable in some range. Increasing the feed rate made opportunity to try dialing up the spindle speed. About that:

    • MRR
      • For shallow cuts, the machine can do fast feeds (thicker chips) at minimum spindle speed. 
      • Faster feed does enable faster spindle speed. Up to maximum spindle speed at maximum feed. The spindle maxed out at 29,000 rpm. Tripling the spindle speed at triple the feed rate seemed like a reasonable thing to try and it did work agreeably well. I didn't look for a minimum feed for maximum speed
      • Faster speed enables more aggressive step-down and step-over at high feed rate.
    • Surface finish
      • tl;dr: better. Prior logs show a history of variably imperfect surface finish quality for flat face cuts, with evidence of vertical oscillation in Z. And the tool never sounds really right at min speed but smooths out at higher speeds. Apparently min speed allows something to rattle axially, apparently somewhat coupled with radial loading, which apparently shows up in the cutter path. That calms down at higher speeds (the internal fan makes enough thrust to preload the bearings?) and cut surfaces look and feel much better.
    • Noise noisenoisenoisenoise
      • At minimum speed the sound, while qualitatively rough, doesn't overwhelm. While the sound smooths out with increasing speed, it also picks up strong tonal peaks that make it unpleasant if not crazy loud. At max it's just #!%&$☆ loud. Earplugs (you were already wearing eyeball protection, right?), pre-emptive apologies and short runs. Apart from max MRR attempts, mid-range spindle speeds turn out decent results and there may be a sweet spot. tbd.

    Depth of cut / step-down

    Not a strength. Trying that first (see first log entry) didn't encourage me to try more of this earlier. Doubling the step-down helps MRR but it's still just 300 microns. Subjectively, it seems like the machine...

    Read more »

  • One screw

    Paul McClay10/07/2020 at 16:20 0 comments

    I've been saying this is designed for assembly with 6-32 x 1/4 -- common PC case screws -- only[*].

    [*] except for different screws[**] to mount the motor/screw assemblies

    [**] which are something of a specialty item for mounting 3.5in HDDs in compact sleds

    Here's confirming that the qualifier can be reduced to:

    [*] and a file

    Well, one kind of screw. x100. Of which a dozen need the head thinned and four of those need to be a couple mm shorter. Or spend a few $ of the special screws.

  • Back for more -- but first review

    Paul McClay10/03/2020 at 23:36 0 comments

    Ok, it looks like I can take another swing at this.

    New motors were waiting in the mailbox when I got back here. After swapping out the crippled motor the revised machine cut a test pocket like it was supposed to a month ago. Yay. Onward.

    But first:

    (or just skip to the next entry)

    That Day of Fail

    tl;dr: I got get-there-itus and tried to grab the ring when I should have taken a day off.

    Retrospectively sorting out a timeline of the unrequited sprint of early September reminded me how poorly (my) memory works for some things, and thus the value of notes and pics/vids/files with timestamps. One day in particular I wanted to review to have a history of many loose ends that feel like the sort of thing that I'll wish I had a better history of when they turn up again later.

    Just about a month ago I logged that I had:

    • CADed, cut, built & tested another rev of the axes, that
      • [good things]
      • roughs out my test pocket encouragingly well, but with a new imperfection that i don't understand
      • has died, in three different ways, at about the same point in the first finishing pass of same test pocket

    What that looks like:

    Three Strikes
    three different ways to die in the same place. click or see gallery for full res

    After redesigning the XY table as described under the Direction header in the same log entry and adding a Z axis of similar design, I tried cutting the usual test pocket. In addition to validating the revised design, because of course it would just work, this would test conventional milling with no extra mass on the X & Y leadscrews for the first time since climb milling lost _the_chatter_wars_.

    For all three tries the roughing operation completed without difficulty. I had inserted a pause in the gcode at this point to get a look. Here's the first try below. It looked ok but not great:

    Because this version of the XY table uses less material and may be less stiff, the merely "ok" output didn't throw me too hard. But some of it was weird. Ungreat features included:

    • Variation in quality of the four vertical walls of the pocket. Although something of a regression, that's nothing new. I didn't try very hard to interpret that at the time -- and I still haven't worked out how the differences around the four sides correlate with various mechanical imperfections. Mystery.
      • 2nd try: worse but not awful
      • 3rd try: pretty good
    • The outside surface of the round feature has -- this is hard to describe and barely visible above but shows better in the next photo below -- a patch near the pocket wall in each quadrant that looks depressed or offset inward from the intended surface. I still have no idea what caused that. It's unlike some earlier tries where excessive runout punched out the pocket wall instead of punching in the side of the ring. Mystery.
      • 2nd & 3rd about the same
    • A conspicuous ring of too much horizontal offset around the outside of the round part between the 3rd and 4th slices. I think it looks like the 3rd slice was cut too small. No idea how so. Mystery.
      • 2nd: same but less so
      • 3rd: none

    Once resumed, the first attempt continued with the constant-X parallel finish profiles. About 3/4 of that appeared to run ok. Then the X axis stopped feeding and the machine cut progressively smaller profiles in the same spot, cutting into the part. Bah. Stop job. That's the first/left fail shown in the triptych above. Right after that the X axis didn't respond to movement commands. Then in a short time it fixed itself. Ok. But why did it die and how to not repeat that?

    I guessed something like this:

    • Steppers heat up more when holding position than when running. At least in my experience (this will come up again...)
    • The X motor mostly holds position(s) through the X finish operation
    • So the X motor temperature rose through that operation
    • Rising temperature increased copper winding resistance
    • Constant-current stepper driver increased voltage (pulse width) to maintain current
    • Crappy wall-wart power supply -- grabbed for first tests and...
    Read more »

  • Did I say frustration?

    Paul McClay09/14/2020 at 19:42 0 comments

    The crippled motor crippled work on this last week while opportunity remained open. Now I expect to disappear for at least a week and don't know what to expect after that. Really bummed to not do more functional testing because ... I think it would have been fun and exciting. And because I was sprinting for the 'ibles CNC contest and missing that means loss of reward for the sprint. Whine. Pout.


    I did cut the revision mentioned a couple logs ago and assembled that for fit testing and taking notes for assembly instructions and further revisions.

    That rev eliminated the remaining blind screws. Now the spindle clamp -- which should be easy to modify/replace -- can be separated from the Z axis without disassembling the slide. Experimentally, it uses some intentional interference to bias which way an otherwise loose fit tightens up. Assuming the material will deform a bit. It seems to work in hardboard - to the extent that it feels solid but no real testing until I don't know when.

    Also: I was able to tile the parts within Ponoko P2 dimensions. Likewise for the next rev ready to cut, but NO PROMISE re future revisions.

  • Dense

    Paul McClay09/11/2020 at 22:54 2 comments

    To illustrate to density of this design, here's an XY table next to a compactly arranged stack of what's in it.

    XY table vs compact stack of parts
    "compact" stack is slightly more compact

    (for completeness: the assembled example has none of the running gear actually "in" it, which would make it ~1.5mm taller)

  • Progress but frustration

    Paul McClay09/07/2020 at 09:43 0 comments

    Eye candy

    next rev at work: cool continuous curls of "chip" from this blue stuff


    My metaproject here, the real project in some sense, is to get better at writing about stuff like this. My silence lately demonstrates a weakness: reporting unsatisfying progress between tidy milestones. I've often found the work to simplify a complicated thing easier than trying to describe the complications, so instead of writing a lot I've kept working until the writing can focus on the point briefly rather than blur a pile of provisos. I think that trade has often worked well. In this context the habit works against the concept of logging progress and process. So what can I say without either writing a book, or waiting until I get to a point that doesn't require a book to describe, or "simplifying" to an extent that approaches dishonesty, or otherwise offending myself?



    • CADed, cut, built & tested another rev of the axes, that
      • simplifies assembly & accessibility
      • simplifies backlash adjustment
      • includes Dremel-clone spindle mount from laser-cut material and string
      • roughs out my test pocket encouragingly well, but with a new imperfection that i don't understand
      • has died, in three different ways, at about the same point in the first finishing pass of same test pocket
    • Injured, apparently, and maybe crippled, but not entirely killed one of my motors
      • it's complicated
      • how much time to spend on figuring that out?
      • ordered more
        • gonna take a while
        • which illustrates another element of supply risk
          • already noted price hike in "Details"
    • Mostly CADed the next rev including
      • minor design tweaks
      • major CAD refactoring - for eventual sharing
    • Deferred writing logs

    [...writing down a rabbit hole to discover the point & discard the detail -- a process that works but absorbs time...]

    I've been working this project with some urgency because I expect life to interfere "soon", but i don't know when or for how long, and i hope to get a useful increment of progress before this gets shut down for a while. And Instructables has a relevant contest open for what has now diminished to a few more days.

    So maybe-losing a motor and almost-but-not getting timely replacement combines poorly with my expectations for the future. Aggravated by several simultaneous uncertainties. So, it's complicated.

    Anyhow... after handling the second design some, I've been working on a next revision that I aspire to cut today. But right now I'm writing this instead of working on that, so, see the meta? While design changes are small, I've put time into refactoring the CAD to better suit eventually sharing that - but still leaving "opportunity" to put more time into learning better ways to do things in Onshape.


    Lately in this project I've concentrated on simplification for reproducibility from laser cut material & a simple BoM. At lowest cost, assuming "free" laser cutting.

    The first rev was tricky & tedious to assemble and required lots of dis/reassembly to work on. Some tricky/tediousity is baked into the design, but at least exposing all the screws will allow random access vs sequential access into the layers.

    At this point I'm focusing on simplification of "user experience" and accepting some compromise to capability. In the current design there are

    • no screws buried under other parts
      • almost, and maybe zero soon
      • except for attaching the motor units, which are very interior
    • no counter-sinks
      • one less tool
      • one kind of screw
        • for structure - different screw for motors
        • 6-32x1/4 "PC case" screws
          • commonly scavengable
          • but need many so probably a purchase item anyhow
      • better for building with acrylic
    • greatly improved backlash adjusters - yay
    • fewer large-area parts
      • area (not shape) of latest cut job would fit within a Ponoko P2 sheet (384mm*384mm)
      • but that's not a feature claim yet
    • [is] mount for Dremel-clone spindle - reduces add'l BoM to: string

    (design feature list overlaps personal accomplishment list -- janky redundancy but I'll say that's ok for a progress log)


    Read more »

  • The X and Y that were

    Paul McClay09/05/2020 at 06:05 0 comments

    The next rev of the XY table improves, I hope, in some respects other than looking good.  So I took some beauty shots before gutting the first one. This seems like a good time to say a little about the mechanical design-in-progress.

    Telescoping axes:

    Or: "6mm rod starts at 100mm; what could I do with that?"

    The next rev continues this basic arrangement.

    Spoil board replaceable without messing with anything else. The initial cut job could include extras. The screws within the work area could be:

    • omitted
    • countersunk more deeply
    • swapped for shorter screws with smaller/flatter flat heads.
    • swapped and sunk deeper

    Integrated workholding:

    Including burried nuts, adjustable corner fence, and some clamps. The fence can slide in to (0,0) or out to align material with some margin around the work area. The little clamps work at the edge of the table. I made more of the larger clamps to use when not trying to hold material close to the table edge. The hole spacing is a little sparse, so the larger clamps have short- and long-reach positions for the back screw while the clamping screw runs in a slot. The fence and clamps fit within the tabletop area while holding 60mm x 60mm material -- but that's overconstrained because the footprint allows some overhang of the two "far" sides. The cap screws+toothy ferrules work for finger tightening, sometimes, depending on hand space & angle of access, fingertip cleanliness, etc.

    While I like this idea and hope to get back to it, the next rev omits all the tastey workholding goodness & reverts to double-stick tape. Other fish to fry.

    About telescoping axes

    I don't know of other CNC machines using a similar arrangement. If you do, please comment. I'd love to hear what textbook/reference already has all the answers that matter. For this particular combination of 89mm long actuator w/~60mm throw, 100mm rods & ~3mm thick material, it seems like a good way to get long travel in a compact footprint with stiffness varying in an acceptable range. I hope to work through a more careful comparison with the more usual two-bearings-on-a-rod+roll-constraint arrangement but it's time to post this and sleep.

  • Made a thing

    Paul McClay08/30/2020 at 03:33 0 comments

    Apart from an initial attempt and unsurprising failure to cut out the part at Q, I've just been trying to cut clean pockets without trying to cut out parts. The last log ended with the pocket at R which I'm willing to call "pretty good".

    So it's time to try cutting a part!

    It helps that exactly one month ago Stewart Allen checked in a "wide cutout" option for Kiri:Moto. That solves the cause of predictable failure at Q above (actually the failure is under W) without any other CAM gymnastics.

    This time I double-stick taped the material in place. Kiri wanted to cut the part outline as part of the roughing operation before finishing the curvy bit. Kiri does support tabs, but I didn't want big tabs and didn't know how unbig the tabs could get without messing up the finishing operation. So, tape.

    "Send File"...

    click or see gallery for full resolution

    ...and this time it worked.

    pocket has four beautiful straight flat smooth sides -- click or see gallery for full resolution
    20 contours/mm - done that once prolly don't need to do it again
    Ok, that's a cheat - I can squeeze the plastic more or less to dial-a-dimension

    Relative to my subjective hopes/expectations for this project, I'll say that turned out well.


    Next: try conventional milling without augmented lead screw inertia.

  • Thank you Orang

    Paul McClay08/29/2020 at 04:44 0 comments

    edit: more pix + result

    tl;dr: while writing this it occurred to me that the simple answer may be conventional instead of climb milling.

    Subject to further testing, it appears that Dr. Vahid has rescued me from not ever figuring this out on my own. His own thesis work is out of scope, but his literature review called out clues that I doubt I would have found apart from the context in which he framed them. Thanks, Orang.

    About wicked chatter when force is applied in the direction of motion...

    Yup, that can be a problem. Especially when the lead screw is/has:

    • self-locking [M3x0.5: yup]
    • low rotational inertia vs load inertia [wimpy screws/hefty XY table]
    • low axial compliance [low=fit for purpose; high=unsuitable]

    Well, that pretty much means "this project". And that

    does not encourage.

    About axial compliance:

    1. This chatter thing went from marginal to major after intentionally reducing axial compliance
    2. The CH-SM1545 motors (probably all these little can-stack stepper + lead screw assemblies) have asymmetric axial compliance because they have a spring on the motor end pushing the rotor/screw into a fixed bearing at the other end. Reviewing cases where similar motion in one direction or its opposite provoked more or less chatter: more chatter has correlated with loading against the unsprung end.

    This wasn't a thing that I really wanted to corroborate. Springing the unsprung end so that the screw floats between two springs might quell chatter in both directions, but that won't help with accurate or repeatable positioning. Instead I tried to do the opposite.

    About self-locking:

    The screw self-locks due to its shallow pitch at (d/6)/rev. The fine pitch multiplies the force & precision possible from the tiny motor relative to what something like #CDCNC could get from a similar motor driving a steeper pitch screw as in a CD/DVD or similar scavengeable device. So that's a given for this project. 

    [Hey - anybody know what sort of scavengeable stuff uses finer pitch stepper+screw assemblies?] 

    So how about rotational inertia of the screw?

    It would be nice if we could hang flywheel off an end of the screw, which is probably not often said around stepper motors, but both ends are captive. Instead I tried this:

    "flywheel" -ish

    I really don't like how this eats up useful length of the screw. 

    Other side-effects that I really don't like include the screw jamming hard when it carries augmented rotational inertia to either end of travel. At least that gives evidence that it makes a difference. How much difference is necessary to shift the ratio of screw/load inertias from too small to safely less small? Dunno.

    That randomly found piece is aluminum and smaller diameter than could potentially clear the ways. A couple of slightly larger diameter M3-threaded brass (or rhenium) disks jammed together would have the same or greater inertia in less length. But any metal part here adds to the BoM which I'd rather minimize. And I didn't find any really suitable standard part. Something like a pair of "DIN 467" nuts (that's a thing‽) but different. A best-fit part would be simple to turn on a lathe, but that departs further from the simple BoM intent.

    But anyhow...

    (edit: add pix + tweak text)

    Adding "flywheels" to the X & Y screws seemed to work for calming the chatter out of, for one test case, the four quadrants of the initial facing cut around my test pattern and the beginning of the four walls of the pocket. That was somewhat encouraging but less than definitive because at the same time the Z axis was flaking out -- then flaked out.

    (also shows runout and uneven cut depth of the two flutes)

    That test cut didn't get very far because the Z axis stopped lifting early in that job, so there isn't much to compare except the face cut and the first layers(?) of the square pocket. The sides of the pocket at least started straight. Some quadrants of the face cut had a slower/longer-period...

    Read more »

View all 18 project logs

Enjoy this project?



Dan Maloney wrote 09/14/2020 at 20:18 point

I like the idea that you're building on the ideas tested in the CDCNC. Looking forward to seeing what you come up with.

  Are you sure? yes | no

Paul McClay wrote 09/15/2020 at 04:43 point

Hi Dan - thanks for skulls! Sorry to keep you waiting for results. While slow shipping and other life block progress for a while, I think early results give [me] good cause for confidence.

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates