Strutt Epicyclic Train Clock (maybe?)

[kvom]

The CAM program allows over-cutting of a profile so I can enlarge the gap as needed.  I considered scaling but this was easier.

[kvom]

I have managed to model a significant portion of the clock train and have begun working on some assemblies.  Because of the dense text and somewhat unclear B&W photos, I haven't progressed to modeling all of the mechanism as I'd like. 

My DMC2 Mini kit arrived this afternoon, so I'll be concentrating on its build for the next little while.

Smith's way of documenting his build means that the drawings often leave out holes for screws because he uses the mating part to locate them. 

One "problem" I've run into is in modeling the escape ratchet wheel.  It has 34 teeth meaning the angle for each tooth is not an even number of degrees.  Smith shows the form of the tool he uses to cut the teeth, but just says cut full depth in one path.  I derived the shape of the teeth from the tool profile, but the SW part doesn't look exactly like the drawings.  The matching "anchor"  also looks somewhat different than the photos.

[Jasonb]

Have you allowed for the fact that the cutters are often not positioned centrally to the blank but offset to one side.

[kvom]

The drawing for the ratchet shows the straight side of the tooth on the centerline, and gives the radius of the cutter (and hence the tooth). 

I assembled a model of the hour train.  The sun wheel, with 66 teeth is attached to the dial and is stationary.  The hour wheel with 72 teeth is concentric and can rotate.  It's shaft does through the dial and has the hour hand attached.  A common planet pinion mounted on an arm rotates about both wheels once per minute, and rotates the hour wheel 1/12.  The planet wheel rotates along with the pinion and meshes with the inside of the ring wheel (not shown).  Rather makes my head spin.  The central arbor, on which is mounted the minute hand, is not attached to any of these components.

I took a look at the mesh between the planet wheel and the ring.  This is critical, and if as designed they don't mesh properly either the ring or the arm would need to be remade.

Yesterday was day one in the shop assembling the CNC mill kit.  I got the frame done, and discovered why my legs aren't really up to standing 6 hours in a day.     Today's steps are to attach and align the linear rails and attach the ball screws.  I can add some photos if there is interest.

[Krypto]

I wouldn't mind seeing a mini-review DMC2 Mini.  Some of the reviews I've seen are from people with more experience with 3D printers or CNC routers and their expectations might be different than someone who is a modeler used to cutting metal.

[crueby]

I wouldn't mind seeing a mini-review DMC2 Mini.  Some of the reviews I've seen are from people with more experience with 3D printers or CNC routers and their expectations might be different than someone who is a modeler used to cutting metal.

Same here!

[kvom]

My old CNC mill was a bed mill with a pretty large work envelope.  3HP, 6K max RPM, R8 collets, no flood coolant but I mounted an air blast.  I think it weighed 1500 lbs.  Started with Mach3 but later converted to Tormach Path Pilot.  I paifd about $11K in 2010.  Sold to a friend for $5K in 2019 expecting to get a better mill (like tool changer and coolant), but then COVID happened.

The DMC2 Mini is about as opposite as you can get.   Kit cost $2700 shipped.  Mill will weight about 120 lbs. when complete.  3HP motor with 24K max RPM.  Probably never run below 8K as torque will drop off.  ER20 spindle so max tool diameter is 1/2".  Has an enclosure and flood coolant.  Work envelope is 7x10".  I'm not sure what the max spindle to table distance is.  Workholding is based on a 3" grinding vise, but the table is just a piece of aluminum that can be be replaced, so one can machine it for other needs.  Unfortunately only control is Mach3, and a license is extra.  It has a probe of sorts:  an electronic edge finder that will be modified to signal the control.  Hopefully will be easier to use than a mechanical edge finder.

The assembly instructions are very good,  190 pages in 50 steps.  They claim 20 hours to assemble, but that's not in the cards.  A machinist instructor shared his build on YouTube, and he took 50 hours.  But he was also filming.  I spend a lot of time double-checking the instructions on my laptop.  In addition, the kit comes in 5 boxes, so you can spend a lot of time rummaging for parts if you don't have the space to open them all and spread it all out.  I'm good for about 4 hours a day before my back gives out.  All the fasteners are metric, mainly socket screws and locking nuts.

[kvom]

Made a better assembly of the ring/planet gear mesh.  Should be good if made to the design.

[crueby]

Looking good! 

Though, we may need to up the level of difficulty a bit and go REALLY old school: cut the gears with a jewelers saw and hand files!      
Kvom? Hello? Oh well, he ran off...   

[kvom]

CNC machines use ball screws with ball nuts to move components accurately.  By default there is minimal backlash.  On my previous mill, the Z-axis (spindle) was mounted on a rigid column, while the X and Y axes moved the table.  On the DMC2 the X-axis is mounted above and across the work envelope.  While less rigid, the ball screws and nuts won't be exposed to coolant and swarf.  Only the Y-axis will have that exposure, and it will be covered by a rubber expanding tube.  The nuts will be under the table.

Each axis screw has two ball nuts fastened together by two screws and a stack of flat and spring washers.  By tightening the nuts against the washer stack, preload is applied between the ball nuts to reduce the minimal backlash further.  The target would be 0.001".  If tightened too much the torque required to turn the screws could overpower the axis' stepper motor.  Setting the preload will be done once assembly is complete.

Linear rails improve accuracy over extrusions, which are typically used for router tables.

[Krypto]

Thanks for going through with the extra effort of describing and documenting your journey with the DMC2 mini!  Your past experience with a home CNC mill for modeling work makes your posts much more illuminating for my interests.

[kvom]

Not much done today as I had lunch and then dentist appointments.

The Y-axis ballscrew assembly with the bellows covers is attached to the frame.  The ends of the bellows will cover the bearings at the ends so that no swarf can get into them.  The plastic tube attached to a ball nut will channel oil.  The bearing nearest the motor mount is screwed down tightly, and the near bearing is left a little loose to allow aligning the ballscrew closely with the linear rails.

[kvom]

Returning to the clock model, the attached rendering shows the ring gear interacting with the planet gear.  Combined with the components shown previously, this is the train that drives the hour hand.  The outer part of the ring gear is driven a pinion that is connected to the escapement ratchet and hence the pendulum.  Evidently the arm needs to circle the stationary sun wheel once per hour, and the hour wheel then rotates at a 1 to 12 ratio. 

The minute hand is attached to the central arbor which passes through all of the above but is attached to none of it.  Rather it has the center pinion at its rear end that meshes with the great wheel that is attached to the fusee.

The great wheel has 120 teeth and must rotate at 1 RPM, so clearly the fusee must turn at 2 teeth per minute.  The ratchet engages the great wheel, so provides the connection to release the fusee. The math involved here is not too clear to me.

I did find a drawing for the "stopwork snail", a part that matches one Kim made for his clock.

[Kim]

Kvom,
This is a fascinating project! I'm looking forward to seeing the epicyclic hour hand in motion! 

Also enjoying your DMC2 mini build/review. 

Kim

[crueby]

Agreed, looks like it will be a fascinating  motion. Is that a counterweight on the one side, or will more gears go on that disc?