August 27, 2015

Mini Lathe Pulleys

I'm back from the left coast, and for once I'm not busy building giant wooden contraptions for orientation.  Instead I'm trying to finish the tiny lathe before classes start.

I made some 2L sized v belt pulleys for the motor and spindle.  I considered using timing belts, but the spindle's going to be pretty high speed (>7k rpm max) and figured v belts would be a bit quieter.  Not to mention, v belt pulleys are much easier to machine.

The pulleys were cut turned using a grooving tool ground from a HSS blank:

For the motor pulley, I had a lot of trouble getting a clean 9mm bore using the MITERS small boring bars, so I tried a technique seen here for making reamers.  I turned some 9.5mm semi-hardened steel shaft down to just under 9mm, and ground three facets in the end.  The faces were cleaned up on a whetstone, to debur the corners and get a sharp edge.  I drilled a slightly undersized hole, and expanded it to final diameter with the reamer.  The results was a pulley with a nice 8.99mm bore, which I tapped onto the motor shaft with a hammer.

Both pulleys are retained by set screws.  Normally I greatly prefer split collar style clamps to retain round things, but in this case it was hard to beat the compactness of set screws.  And it's dealing with fairly minimal torque compared to most things I build.  On the spindle side, the shaft diameter is fairly huge, which makes the force on the set screw smaller, and on the motor side, the pulley's press fit should deal with most of the torque.

On both the motor shaft and spindle I milled a little recess for the set screw to sit in:

Also notice the pair of lock rings for preloading the spindle bearings

And here's the spindle and motor pair:

Next up: Building a tiny quick change tool post.

June 7, 2015

Spindle and Headstock Machining

Machining the spindle turned out to be kind of a disaster, but I was mostly able to salvage things.

I started out with a chunk of crufted 1.625" steel round of an unknown alloy.  For whatever reason, I found it pretty much impossible to get a good surface finish on the stock with the MITERS carbide insert turning tools.  The only way I could get a finish resembling anything reasonable was to crank up the lathe's speed to the maximum and take really heavy cuts.  Unfortunately, this meant it was impossible to take finishing passes, so I screwed up the surfaces that mate with the inner spindle bearing races, and made them a little too small.  I did the final surface finishing using sand paper pressed against a precision ground steel block.  This seemed to fix the surface pretty nicely, but I was a little overzealous with sanding, and took off at least a thousandth too much off the front bearing interface, and a tiny bit too much off the back interface.

Since I didn't have an extra piece of steel to start over with, I attempted to salvage the spindle by using the newly acquired MITERS knurling tool to expand the shaft at the bearing interfaces.  It did work, although I lost a little concentricity in the process.  Fortunately, I can virtually eliminate the spindle's runout by taking a final pass off the critical surfaces on the mini-lathe itself.

Some design notes about the spindle:
- The front of the spindle is tapered and threaded to accept ER-32 collets.  I did the M40 x 1.5mm threading on one of the fancy Prototrak CNC lathes in the new N51 CNC shop.  This was a huge pain to set up.  All I wanted was a menu to input thread pitch/feed per revolution and a lever to engage/disengage the autofeed, but instead I had to actually program a job to do all the passes for me.  It took many hours poking through menus, reading the manual, and asking for help to finally figure it out.  Conclusion:  I should find MITERS a set of change gears for the Clausing lathe to approximate metric threading, so I can do my metric threading there.
-The spindle is extremely disproportionately large in diameter for a lathe of this size, and for good reason.  The bore through the spindle is .875" in diameter.  A stock taig spindle has a bore of just over .375", so this will be able to hold much longer large pieces of stock.

To put the hex at the front of the spindle, I grabbed a hexagonal 5C collet holder and fastened some scrap stock in it.  I then put the matching ER-32 collet and holder on the spindle, and tightened the two collets down face to face.  The vice stop made this a quick mill-and-rotate operation:

The headstock was machined from the wonderful 4"x4"x12" billet I scavenged last year.  I bandsawed off a 4.5" chunk, roughed it to size, and fly cut all the faces:

The quickest and most repeatable way I found to measure it was using a granite block scavenged from and old atomic force microscope, and those measuring surfaces on calipers that everyone seems to forget about.  By tramming the Bridgeport (a relatively rare occurrence at MITERS) and tapping down the block with a hammer every time I tightened the vice was able to get a thousandth or less variation between the corners on each face.

The headstock grips onto the dovetail bed.  I cut the slit for the clamping mechanism with the Epic Slitting Saw.  It made the poor Bridgeport's spindle sound extremely unhappy, but did a nice job.  Two counterbored M6 screws tighten down the dovetail to lock the headstock in place.

Bearing bores completed.  The tapped M4 holes are for a bearing shield on the front, and a motor mount on the back:

And here it is assembled with bearings.  Both the inner and outer races of the bearing are tightly pressed at the front.  At the back of the spindle, the outer race is a sliding fit held with retaining compound, and the inner race is a light press.  It should possible to disassemble without trashing the bearings by heating to loosen the retaining compound.

I have yet to machine the nuts that will thread onto the back of the spindle and preload the bearings with a disc spring washer.

And here the lathe updates come to an end, as I'm currently ~3,400 miles away from MITERS.

May 24, 2015

Mini Lathe Handwheels

The making of the handwheels involved a few revisions and many mistakes before producing something acceptable.

My first idea was to have a system vaguely like Bridgeport mill dials.  There would be a ring with scribed markings on it which could be rotated and locked down by another threaded ring.  Here's what the parts look like:  A simple aluminum handwheel, with a solid stainless steel handle press fit into it.  I really don't like the type with freely rotating handles - if there's any slop they feel cheap.

The ring which would later get scribe marks slips over the threads and rests against a lip at the back of the handwheel:

The threaded ring screws down to lock both rings in place, allowing you to set the zero on the dial:

This worked, but showed a major flaw when I actually attached it to the cross slide.  When I turn the dial, I tend to turn it by the cylindrical part of the handwheel, and only use the handle for extra support.  This means that when turning the dial counterclockwise, the threaded ring unscrews itself.

So that idea was binned.

I moved on to simpler solid handwheels.

These were turned from aluminum round, and like before, had stainless steel handles press fit in.  Each as 100 marks scribed around the perimeter, with every 10th mark longer.  With a 1mm pitch ballscrew, this gives 20 microns on the diameter per tick mark.

The handwheels are threaded onto the end of the ballscrews.  To to this, I cut a small section off one of the ballscrews, and used an angle grinder to cut slots into it, making a "ball tap" of sorts.

On the cross slide handwheel, I spaced out when scribing the marks, and made them on the wrong side of the handwheel.  I'll probably temporarily fix this by making a pointer that extends out to the tick marks, but eventually I'll just make a fresh handwheel:

I got the markings right on the compound though:

Ticks on the dial were scribed on the mill, using a single point cutter and indexing head to cut the marks.  The three jaw chuck on the MTIERS indexing head doesn't hold parts very centered, so I had to shim the jaws with paper to get the dial centered properly:

The 10 long marks were cut every 36 degrees, and short marks every 3.6 degrees:

The pinion that moves the carriage along the rack gear was crufted from an old stepper motor, and pressed onto an 8mm steel shaft with a shoulder turned in it:

Stock Taig lathes adjust pinion-rack clearance using an eccentric bronze bushing, but I wanted a pair of ball bearings supporting the pinion and carriage handle.  I machined a bearing holder and mating surface on the carriage casting:

The M6 screw that fixes down the bearing holder sits in a short slot, so that the pinion clearance can still be adjusted.

I made a carriage handwheel of similar style to the others.  It attaches using my go-to shaft collar style clamp.

And here's the lathe with all axes attached:

The carriage handwheel, unlike most lathes, has very little backlash:

May 20, 2015

Cutting a Rack Gear on a Bridgeport

I considered buying a little metric rack gear for the lathe on SDP-SI, but it would have cost around $30.  Instead, I figured out how to machine one on the MITERS Bridgeport.  

The basic idea, which came from a thread on a machining forum which I can no longer seem to find, is this:  Put a single-point cutter in a fly cutter.  Angle the head of the mill such that the cutter points straight down.  Cut a tooth, then feed by the  pitch of the rack.  The forum-poster thought it would create a slightly incorrect tooth, because of the angle of the fly cutter.

Turns out, though, if the mill is angled between 90 degrees  and 90 minus the pressure angle of the gear, the teeth actually come out perfectly shaped.  Here's why:

This is what a single point cutter looks like when it has been swept around the axis of the mill spindle.  In this case, the mill is angled to 65 degrees, and the cutter is for a 20 degree pressure angle rack. 

The tooth profile that would be cut in this configuration is shown below.  Since the mill is angled to 25 degrees from horizontal, but the pressure angle is 20 degrees, the cutter removes an extra 5 degrees from one side of each tooth.  Simply rotating the head of the mill to between 70 degrees and 90 degrees means that the plane the tip of the cutter moves in doesn't actually interfere with the shape of the tooth being cut.  The result is a normal tooth shape, despite the funny arc the cutter moves in.

Here's what it looks like in practice:

I machined a .1" pitch rack out of some aluminum to make sure the process worked:

On to the actual rack, which is approximately a .5 module metric pitch.  The cutter was ground from a small brazed carbide lathe tool.  The rack was machined in three passes, to avoid loading the tool too much.  This took a very, very long time.

But I'm quite pleased with the results.  The rack feels like it meshes quite nicely with a (presumably) .5 module gear pulled off an old stepper motor.

Coming soon, making shiny hand wheels, and machining the spindle and headstock.

May 19, 2015

Compound Slide

School's over, so it's time to catch up on documentation.  The lathe isn't finished yet, but I've made a lot of progress.  Sadly, I won't have shop access this summer, so the lathe won't get finished until August or September.

The compound was made from two pieces, a dovetail slide which holds the ball nut, and a cone-shaped base.  The two pieces were silver-soldered together.

The base of the slide had a ring milled in it with a rotary table, so that the force clamping it down to the cross slide acts at a maximal radius:

The cone is clamped pulled down by a pair of set screws, the ends of which are turned to 45 degree cones:

The carriage, compound, and cross slide assembled on the bed of the lathe:

May 4, 2015

Cross Slide, Take 2

Man am I behind on documenting the mini lathe.

My original plan for the cross slide was to anodize it like Taig does, to prevent wear and galling on the aluminum-aluminum interface.  I decided instead of dealing with nasty chemicals and temperature control (required for hard anodizing), to just remake the part out of steel.  Steel machines much more slowly, but since I had already machined the part once, it went reasonably quickly.

Fortunately, Nick had a big chunk of steel which was very close to the size I needed:

The block was milled to size and top and bottom surfaces fly cut:

I forgot to take more pictures of it, but it looks pretty much identical to the original aluminum one.

I also attached one of the ballscrews to the carriage:

Bearing holders for the ballscrew were milled on the MITERS CNC mill.  The ballscrew is retained radially by a single ball bearing, and axially by a pair of needle thrust bearings (not shown here):

Prepare yourself for the making the compound slide and machining a rack gear on a bridgeport in upcoming episodes of Tiny Lathe.

April 19, 2015

The Longboard Has Gone Too Long Without a Motor

As previously indicated,  my longboard (can it be called that?  It's tiny) recently sprouted a battery pack and motor.

I started out with this board.  Nick crufted the deck out of a dumpster a while back, and I got an extremely cheap set of wheels and trucks for it.  The construction is interesting.  The deck is made from a 1/8" thick sheet of wood laminated on each side with a very thin layer of aluminum.  The resulting deck is extremely light and flexible, which makes the board really smooth out bumps (which are rather abundant on Cambridge roads and sidewalks).

Most electric longboards out there are stiff, because you generally don't want your batteries and electronics bending.  Alternatively, some electric boards have a small battery at one end, motor and controller at the other end, and nothing in the middle, allowing the board to be flexible.  This requires a full-sized longboard and/or a small battery though.  I wanted to keep the flexibility of this board, even though to get reasonable range the entire underside would need to be packed with battery.

My solution was the bendypack, which is the only part of this build that's anything new.  The rest of the electrical system is typical diy-electric longboard stuff, with a HobbyKing motor, motor controller, and remote control system.

The key to the flexible battery pack is its corrugated enclosure.  The corrugations allow the pack to flex at the gaps between rows of cells.  The inter-cell connections run against the deck, which is very close to the neutral bending axis of the board and battery assembly.  This means that the wires barely have to stretch as the pack flexes.

To make the corrugated casing, I cnc milled a foam mold on the MITERS cnc mill.

The mold was milled in three parts, because of the size limitation of the mill:

Two layers of fiberglass were vacuum bagged over the mold:

The foam was dissolved out with acetone. It left some color behind though:

The case fits 18 26650 A123 26650 cells.  These are arranged in a 6s3p configuration, for a nominal 19.8 volt, 6.6 amp hour pack.

Cells were soldered together with grounding strap to minimize the thickness of the connections:

The pack was insulated from the deck with a sheet of rubber.  Around the edge is a much softer synthetic foam, to act like a gasket and seal off the pack against the bottom of the deck.

The pack was screwed to the deck with a bunch of 2-56 screws around its perimeter.  These screws were tapped directly into the bottom of the deck.  The material of the casing is very thin, so the screw heads alone would have pulled through the fiberglass.  I 3d printed a series of oddly-shaped washers which were glued to the casing to distribute the load.  The washers aren't shown here:

The motor-wheel-pulley system is nothing too special.  The 42 tooth HTD-5mm pulley on the back wheel was downloaded from this thread on endless-sphere and 3d printed.  Sounds a bit sketchy, but there are 6 M5 bolts going through it to hold the wheel hub, and the load is spread across a bunch of teeth.

The motor pulley was CNC milled out of some square 7075 bar stock:

I much prefer split collar style clamping mechanisms to set screws, so I manually machined some slots in the pulley, and drilled and tapped a hole for the clamping screw.

The super cheap trucks had no square faces or cylindrical surfaces, so I stuck if in the lathe and turned a small section down.  We don't have a steady rest for the MITERS lathe, so I used the old drill chuck, which has smooth jaws, with lots of oil:

I CNC milled a two-part motor mount, with about a centimeter of adjustment room for tensioning the drive belt.  The motor is a SK3 5065 236 Kv, left over from a project of Nancy's a couple summers ago:

The mount was temporarily fixed in place with a big set screw, and tack welded to the truck by Mike.  The 9mm wide HTD 5 timing belt was borrowed from Ava's motorized ripstik.

I shoved an RC airplane controller (160 Amp dlux HV) and rc receiver into a 3d printed case, and glued that to the battery pack.  I don't really want this to be the permanent solution, but it works well enough that I probably won't change it until something breaks

Right now the "switch" is just the battery's deans connector.  Also not ideal, and it will get replaced with a proper switch eventually.

Belly up:

The board is geared for a no-load top speed of just over 22 mph, so realistically can probably get to about 20.  I very rarely max it out though - with a board this short, it's just too easy to hit a rock and go flying.  I haven't measured the range exactly, but I'd estimate it at 7 or 8 miles, although this will depend on how hilly and windy it is.  Riding back and forth from East Campus to MITERS 3 times (.7 miles each way) drained about 3.5 amp hours from the battery pack.  In terms of power, I've watt-metered the board at a bit over a kilowatt peak out of the battery pack during acceleration.