Ed Nisley's Blog: Shop notes, electronics, firmware, machinery, 3D printing, laser cuttery, and curiosities. Contents: 100% human thinking, 0% AI slop.
The Y-axis leadscrew on a Sherline milling machine sits exposed to all the crap blowing off the cutter; maybe it doesn’t matter, but it seems nasty.
Throw them out when they're this dirty!
So I made a set of way covers from the template available here or here, except I used plain old printer paper, stuck in place with double-sided tape. The picture shows what one looks like after surviving the rigors of a trip to Cabin Fever Expo.
The key feature is that, when they get too schmutzig, you just throw them away and fold up a new set. It’s easier than dismounting and cleaning something more substantial that you can’t just discard because you’ve developed a serious, deep, long-term emotional attachment.
Everybody at Cabin Fever Expo liked them and wanted the template. If those links have rotted out, I have a copy of the file in the Useful Stuff section: here.
Bellows Folding
Update: Here’s a closeup of a new set. Start with the printed lines up, then fold the end tabs up: the printed side will be down (as in the bellows on the right) and nobody will know how poorly you followed the lines. Click to get a big pic with decent resolution if you need more detail.
So I hauled my Sherline CNC milling machine gadgetry, an assortment of trivial projects, a stack of handouts with pix & G-code, and a pile o’ EMC2 doc to Cabin Fever Expo for two days of Performance Art…
Ed Nisley Demo-ing CNC at Cabin Fever. Picture courtesy of Brian Glackin.
The key is to have the knobs turning: an inactive machine is just background clutter that everyone walks right past. It’s not nearly as interesting as miniature tools or a chuffing steam engine.
There’s something to be said for being on the crowd side of the table, as that lets both of you see the monitor. A bigger display might be more helpful; I duct-taped a 14-inch 1024×768 LCD panel to the top of the desktop PC box.
Although I brought some blank stock along, it quickly became obvious that live-fire milling under show conditions is a Bad Idea: far too many distractions and far too many things can go wrong. So I contented myself with cutting air; nobody really minded and I could switch programs in mid-stride to show folks the G-code program they really wanted to see.
Plenty of folks stopped by, many of whom either have CNC running or are in the throes of getting started. A surprising number of conversations started with “I have this old Bridgeport …” and went on from there.
There’s a crying need for a comprehensive machine design tutorial that explains how all the pieces fit together, with sort of a flowchart outlining the choices (I know it’s more complex than that, but a diagram would be a starting point for discussion). I don’t know enough of the servo end of the biz, but someone should show how the machine’s size determines the motor size and, thus, the motor driver size, with plenty of examples. There’s a misconception that you can run a big machine on little steppers or puny servos, with the controller making up the difference.
Many people do not understand the difference between CAD, CAM, and what EMC2 provides. I described the process as three layers: CAD makes the pretty pictures, CAM digests those pictures and emits G-code, EMC2 converts G-code into motion. That seemed to help.
The single most attention-getting part of the exhibit was, to my astonishment, my Orc Engineering counterweight (described here and here) supporting the Sherline’s milling head. I had to explain just exactly why you need a counterweight in the first place (heavy offset motor, short Z-axis ways) and how much it weighs (13 pounds, a bit too much). Some folks commented that they put similar counterweights on their much larger machines and after a while I stopped feeling inadequate.
EMC Penguin Mascot
At least a dozen people picked up my EMC doc and asked if I was selling it; took me a while to realize they wanted to buy the booklets. I don’t know if you could make any money at it, but there’s a definite market for ink-on-paper books with no plot and weak character development. Now, if Chips were was way more shapely, we could have a real bodice-ripper cover. Somebody get on that, OK?
I make booklets using Adobe Reader’s print-as-booklet feature, a printer with continuous-flow inking, and an Ibico comb binding machine, but there’s enough fiddling that doing much of it for anybody else just doesn’t make sense. Something like Lulu might work, but there’s a stiff (to me, anyway) up-front charge and the EMC doc changes often enough that you’d have to run plenty fast to stay in the same place.
Other people picked up the books and asked if I was selling the software. They seemed puzzled when I said it was free for the download and that not only was the software free, but the GPL meant that they were, too. I need to work on that part of the schtick… should’a had a few CDs to pass out, too.
I remembered to bring a bag of cough drops, ate ’em like candy, talked almost continuously, and wound up hoarse anyway. Probably convinced a few folks to try EMC, didn’t terrify many children, and a good time was had by all.
Although live-fire milling is scary, it’d be fun to make something like a finger ring (as in Dan Statman’s gorgeous designs, but plastic) as a hand-out freebie. The whole process should take no more than five minutes, tops, which might be tough. Running a rotary table and the mill would be a real crowd-pleaser; my 4th axis attracted some questions. Perhaps an EMC tag-team would suffice: one to mind the mill while the other works the crowd?
As always, Cabin Fever is stuffed with gorgeous examples of machine-shop work. Those guys actually know what they’re doing; I can write G-code, but it’ll take many more years of experience before that code actually makes passably pretty parts.
The first thing you do in any CNC milling setup is to locate the part’s origin to the spindle axis. Big mills use homing switches and carefully calibrated fixtures. I used to stick a pointy carbide scribe in a collet, push it finger-tight into the spindle, and align by eyeballometric guesstimation.
Sears laser level mounted on ceiling
That actually works pretty well for most of my projeects, as a few mils (heck, a few dozen mils) one way or the other doesn’t make much difference.
I kept lusting after an SDA Laser Center / Edge Finder (as advertised in Home Shop Machinist), which a friend says works really well. They have a 1/4″ (6 mm) shank for smaller machines, but it’s still nigh onto three inches long and headroom is a precious commodity in a Sherline mill.
I’ve seen projects using laser pointers as alignment tools, usually with an intricate six-axis gimbal hoodickie to aim the beam in exactly the right direction. If you’ve ever sighted along the body of a hand-held laser pointer, you’ll quickly see that the laser chips and optics project the beam in a generally forward direction, but any alignment is purely coincidental.
Then I had an insight: the spindle always stays aligned along one vertical axis. The mill is firmly bolted to the table, the table to the concrete-block wall, and the floor joists overhead rest on the wall, which means a laser mounted on a floor joist could shine right down the spindle bore. Do all the fiddly alignment once, then just it’ll Just Work forever after.
The top picture shows the result. A cheap-after-rebate (ten bucks, tops) Sears carpenter’s laser level provided the guts of the project; it’s no longer available, but you get the general idea. The fancy housing includes a cylindrical lens that converts the dot into a line, but stripping off the housing gets rid of that unwanted feature.
I carved out a plastic base plate and tapped it for 2-56 screws for fine adjustments. They fit directly in the holes that originally held the top of the housing, with springs to hold the level in position.
A plumb bob showed that the level had to live about halfway between two joists, so I screwed a scrap of 2×4″ in place and then screwed the level to that. The cardboard shims reveal the fact that the side of the board isn’t vertical, but I doubt cutting it would have helped much. After that, it’s a matter of sliding it this way & that, tweaking the screws, and fiddling around until the beam falls straight down the middle of the spindle.
The laser spot is far too large, but a small lens will do the trick. Long ago I got a sack of small lenses from a surplus outlet that included a teeny plastic lens with a 1-inch focal length. A bit of lathe work turned out a holder with a bore just a bit larger than the laser spot.
Lens holder rear viewLens holder front view
The hole trims off all the junk light around the spot itself and cuts it down to a few mm in diameter, well-centered on the spindle axis. You can’t get too skinny here, as diffraction gets in the way, but the holder bore presents a decent spot to the lens.
A dot or three of cyanoacrylate adhesive (I just hate the term “crazy glue”, don’t you?) holds the lens in position. As long as the lens is centered and reasonably perpendicular to the axis, it’ll work fine. The lens is much larger than the spot, so you could use one that’s even smaller than this with impunity.
The narrow shank fits into an end mill holder, of which you should have several anyway, and the wrap of tape makes it a snug slip fit. If I hadn’t tried to get clever, the brass would be the right diameter; I keep telling myself to make another one and some day I will.
Lens holder in end mill holder
Because the beam is essentially parallel, the lens focuses it to a brilliant spot about 1 inch below the lens.
Laser alignment spot in action
The maximum angular error (offset from the true spindle axis) is pretty small, no matter how bad a job you do, because the beam must pass through the middle of a 10-mm tube that’s 130 mm long. Assuming it’s slanted off-axis by 1 mm at the top and still makes it into the hole in the lens holder, that’s under half a degree. At the far end of the 1-inch focal length, the spot is off-position by 9 mils, call it 0.2 mm.
However, the lens focuses that beam down by a pretty good factor and reduces the error by the same factor. I don’t have the number, but in practice I think the spot size is larger than the alignment error.
You can do better than that by adding another aperture at the top of the spindle and getting really fussy with alignment.
In any event, it’s closer than I came with the carbide scriber!
Update: I mounted it on a new bracket attached to the new counterweight gantry: much better! Some tips on aligning the thing there.
The motor driver box on my Sherline mill started out as a stock unit, but I’ve tweaked the circuitry to improve the analog performance. Those adventures formed the basis of my Above the Ground Plane columns in Circuit Cellar magazine columns for August & October 2004.
Because the firmware for the PIC microcontrollers wasn’t available, I wrote a clean-room version so I could show how it all worked for the column. My code won’t run on a stock Sherline board, though, so it’s not a drop-in replacement for the stock firmware.
One of the reasons I attacked the controller was to reduce the audible noise coming from the motors. That’s an inescapable part of chopped-current stepper motor drive circuitry, but the noise was modulated by all manner of things that shouldn’t have affected it; just touching the box shouldn’t make any difference at all. The fact that it did meant the circuit board had some, well, infelicitous layouts.
Although the final result was much more stable, I decided to turn off each motor if it didn’t move for at least five seconds; that’s a simple firmware tweak when you write your own code. As a result, the shop was quiet when I wasn’t actively milling.
Solar Measurements Circuit Board – Top Side
Now, having a motor be completely turned off while milling is going on isn’t generally a good idea, because milling forces from the other axes can push the table against the leadscrew and, perhaps, turn the screw against the unresisting motor. I figured that on a Sherline mill, what with the sissy little cuts I take, that wouldn’t be a problem.
And I was right for the better part of four years!
A benefit of turning the driver circuitry off was that I could easily twist the knobs by hand to fine-tune the XYZ position during setup. That worked out really well.
However, drilling the seemingly simple circuit board pattern you see here (for my February 2009 CC column) produced exactly the right collection of forces (while drilling? Huh?), vibration (maybe) and motor pauses (for sure) to introduce an absolutely repeatable positioning error that Went Away when I tweaked the firmware to keep the motors enabled at all times.
I’ve since made another tweak that reduces the current to an idle level after five seconds. That both reduces the audible noise and drills the board correctly, so I’ll keep an eye on it for a while before declaring victory.
The PCB has a few unused (in my code, anyway) chip-to-chip connections that I could employ to let them all decide when nobody’s moving. I think turning the motors off 20 seconds after the last axis stops moving should work Just fine; my G-code doesn’t wait around that long except for manual tool changes.
This is the dingus that attaches the crossbeam to the central pipe rising up from the table for my counterweight gantry. I discarded a whole bunch of elaborate construction ideas in favor of just jamming a plug in the pipe and cranking down on a nut to tighten it.
Expanding plug overview
It’s pretty much self-explanatory; I cut everything to fit, cleaned up the cuts with a file, and added some lube to the tapers so as to make it nice & slippery.
The need for an O-ring to hold the halves together occurred to me after I’d bandsawed a 1 mm trench in the side of the plug. I chucked it up in the lathe again and used a round-nose tool to carve a groove around its belly. If you try this, do the groove first: an interrupted cut is murder on what’s basically a parting-off tool.
Expanding plug parts
While I know (thanks to Guy Lautard’s invaluable Machinist’s Bedside Reader books) that a self-releasing plug must have a taper angle with a tangent greater than the joint’s coefficient of friction, that really wasn’t much help here. I picked 40 degrees and, yup, it’s self-releasing, but not really slippery enough. Takes a bit of torque to expand the plug enough for a good grip.
Perhaps my grubby surface finish has something to do with it?
Memo to self: find out how to figure the taper angle correctly, then do better finishing.
It’s a simple fact of life that a CNC’ed Sherline mill requires a counterweight pulling the head upward, because, without some help, that poor little Z-axis motor has a hard time lifting the head’s nine-pound dead weight. The fact that it’s cantilevered way out from the Z-axis dovetails is another problem: there’s plenty of torque binding those ways.
For the last few years I’ve hung a random hunk of iron from a pair of pulleys attached to the floor joists overhead, but that’s not portable and I’m planning to bring the mill to the Cabin Fever Expo again this year. So I hacked out a sort of gantry that works reasonably well and, when I get back, it’ll replace my crude pulley lashup on the joists.
Mill counterweight gantry center support
The main beam is an aluminum extrusion that looks like it started life as a traffic sign post; it came with the house and Ol’ Gene was tight with the town DPW, so that seems reasonable. The center support is actually a pipe clamp that I’ll crunch on the edge of the table at the expo, with two plywood scraps to keep from embossing their furniture: the pipe sticks straight up from the table.
There’s an expanding plug inside the top of the pipe that I must go into more detail about later, but the general notion is that the beam becomes one with the pipe when I crank on the nut. I faced off the top of the pipe and cleaned out the graunched metal at the end of the threads.
The gray strap should hold electrical conduit to a wall, yet fits perfectly on iron pipe; who knew? Given that no plumbing size matches any physical property you can actually measure, specifying a match like that is impossible.
Counterweight pulley with Loctite
I made a pair of pulleys around 26 mm OD ball bearings, mostly because I couldn’t find anything else that would work. Yes, they’re open to shop crud, but I’ll add side shields before I screw it to the ceiling. I know I’ll be explaining how they work at Cabin Fever, so there’s no point in hiding the things.
The shaft is a steel rod, turned to fit and drilled out for a 10-32 screw. The greenish dab on the shaft is Loctite; I slid the pulley over the dab, aligned the cable with the hole, and let it cure in place. Loctite gets me out of making a quartet of fussy little spacers: better living through chemistry.
The outer ring is polycarbonate, chosen because the sheet was the right thickness. This ought to be a lathe project, but it was easier to clamp it on the mill, so I helix-milled the center hole to fit the bearing OD.
Spiral-milling the pulley OD
I hammered a piece of copper pipe into a mandrel / jaw pad, put a thin chunk of acrylic under the polycarb as a spacer, grabbed the whole affair in the three-jaw chuck, and helix-milled the OD. No drive dog to hold it in place, no special prep, no nothing: it Just Worked. Of course, I’m taking sissy little cuts on the thing, but you best do that on a Sherline anyway.
The copper pipe mandrel trick doesn’t give great precision, which, fortunately, isn’t needed in this application. I found a chunk of EMT in the heap with the right OD and walloped the copper pipe around it with a rubber mallet to beat it into shape.
The reason you need a mandrel is that thin ring of plastic deforms under the pressure of the jaws, producing a three-lobed effect: pleasing in an art project, yet strangely inappropriate in a nominally circular machined part. I discovered this exquisite little inconvenience a few years ago and haven’t forgotten the lesson yet.
I planned to use a slitting saw to cut a groove for the wire rope around the OD, but then I came to my senses: ‘way too much leverage on that poor little chuck and not enough traction from the jaws. A bit of rummaging came up with a 3/16-inch ball-end mill burr, which was just slightly larger than the wire rope.
[Update: What I’ve been calling a “ball-end mill” is actually a “ball-end burr”. A “ball-end” or “ball-nose” mill is basically an end mill with a hemispherical end. Sometimes “burr” is spelled “bur”.]
Ball-end mill used the wrong way
Then I screwed up: I mounted the chuck on the rotary table, the table on a right-angle plate, the plate on the mill, and zeroed the ball at the top of the pulley.
The right way to use a ball-end mill burr is along its equator, so I should have zeroed it at either the left or right side of the pulley. The ball doesn’t have really great cutting edges around its South Pole.
Fortunately, it cut a 1.5-mm trench around the polycarb just fine. I suspect if I was using aluminum, this would not have had a happy outcome.
The parts heap yielded a pair of lead blocks, a sturdy eye screw, and some humongous heat-shrink tubing that made a tidy counterweight. I lashed everything to the countertop supporting the mill, added a 2×4 post and a machinist’s jack to support the countertop, and it passed the smoke test.
Lead counterweight
The Sherline seems happier with a counterweight that slightly outweighs its head. The lead blocks weigh 13 pounds, about 3 pounds more than the head, and Z-axis travel is nice & smooth.
It’s said that the disadvantage of removing all the weight from the head is that there’s less weight to press drill bits into the workpiece. I haven’t seen that problem yet; methinks a few pounds really don’t make that much difference compared to the forces generated by the motor through the leadscrew. We shall see when this thing is installed in its permanent home on the floor joists.
Counterweight hook on Sherline mill head
Real machinists make all sorts of lavish clamps to attach the counterweight rope to the Sherline mill head. I favor the Orc Engineering approach: a random hook from the heap, firmly attached with a big hose clamp, more-or-less over the head’s center of mass. What’s not to like?
Incidentally, one nice feature of a hook on the head is that you can hang the head out of the way under a nearby shelf. I have both the 3k rpm and 10k rpm heads and it’s really nice to have the other head conveniently located, yet out of the way.
Something that’s obvious in retrospect: that pipe’s gotta be very close to vertical, lest the cable drag on the side of the holes. I may need some shims on the Cabin Fever table to make the answer come out right.
Here’s the heart of the helical-milling rough-cut routine for the OD. We start at a safe Z-axis traverse level and the first pass is at Z=0 on the surface to reveal any gross alignment errors…
G0 X[0 - #<_Pulley_Radius>] Y0 (set up for cutter comp)
G41.1 D[#<_Tool_Dia> + #<_Cut_Finish>] (cutter comp on left)
G2 X[#<_Pulley_Radius>] I[#<_Pulley_Radius>] F#<_Traverse_Feed> (CW entry arc to right side)
G0 Z0
F[#<_Mill_Feed> / 2]
#<_Pass> = 0 (iteration counter)
#<_ZLevel> = 0 (current Z-axis level)
O100 DO
G2 X[#<_Pulley_Radius>] I[0 - #<_Pulley_Radius>] Z#<_ZLevel> (helix down)
#<_ZLevel> = [#<_ZLevel> - #<_Cut_Depth>]
#<_Pass> = [#<_Pass> + 1]
O100 WHILE [#<_Pass> LE #<_Num_Passes>]
G2 X[#<_Pulley_Radius>] I[0 - #<_Pulley_Radius>] (remove ramp to final level)
G40 (comp off)
G0 X#<_Pulley_Dia> Y0 (move away from part)
G0 Z[#5163 * 25.4] (get air)
Memo to self: use the ball mill’s equator next time.
Update: The thing fit perfectly on the floor joists over the mill; more info there.
The drawbar bolts on a Sherline mill (and, presumably, the lathe) have a fitting that adapts the bolt head to the top of the spindle. There’s nothing to keep it from sliding right off the bolt, which happens 100% of the time when you just pick up the drawbar.
The fix is easy: a short length of heatshrink tubing applied near the bolt head.
I happened to have some tubing with an internal hot-melt glue lining that stick like, well, glue to the bolt. Pretty nearly any heatshrink tubing will work, although you might need two layers to catch the fitting’s ID.
Sherline drawbar bolts with heatshrink tubing to capture the fitting
The Smell of Molten Projects in the Morning 