Strutt Epicyclic Train Clock (maybe?)

[kvom]

Since model shows are far away and I no longer like to make long drives, I've thought about trying to make a clock.  In browsing various websites I came across this one, originally designed about 1829 by William Strutt.  The plan book by W. R. Smith details the build process.

To actually attempt this build I'm going to need a CNC mill, because chain drilling and filing are not something I'd enjoy.  The mill I'm considering is the DMC2 Mini from a company in Toronto.  It can be had in kit form or assembled, but both are back-ordered.  This mill has a 20K RPM spindle, and I have endmills small enough to cut the gears in profile. 

The front and back plates have overall dimensions 9/13".  Rather than investing in engraving brass to the tune of $400+ to begin with, I plan to cut them from 3/16" clear acrylic sheet using the CO2 laser my daughter and I just purchased.   The dial will also be engraved on acrylic sheet.  For the rest I have a decent stock of 260 brass.

Before getting in too deep, I thought I'd try to make the fusee.  This part requires manual machining on the lathe in an unorthodox manner.  I showed the blank on Kim's clock build as he was also trying to make the fusee for his clock.  But rather than barging in on his build I'll continue here.

The blank (pic below) was turned from a piece of hex brass rod that is 1-7/8" across the flats.  Since this is the desired diameter of the blank it just needed to be turned down.  Before that I faced to length, drilled a through hole 9/32" and reamed it for a press fit.  Then cut a pocket in one end where a ratchet will live.  The shaft is a piece of 5/16" diameter drill rod that was center drilled on both ends and then pressed through the blank.  I used the vise on the Bridgeport as a press.

The profile of the fusee is circular with a diameter of 2-7/16" and centered at the left side.  The cross slide on my lathe (Monarch 10ee)  rotates easily.  I moved the toolpost to the right so as to center of the tool holder over the axis of rotation and then adjusted the cutter in the holder to achieve the needed radius.  This all required some finagling and was as close as I could make it.  Then set the cross slide perpendicular to the blank with the tool tip at the left end of the blank and lock the carriage.

I took a shallow .02" passes.  The tool I used might not have been the best choice as the surface wasn't smooth and my head was continually hit with tiny brass bullets like being in a sandstorm.  Eventually I got the end and smoothed the surface somewhat with sandpaper and some small files.  Since the surface will be cut with spiral grooves its current surface isn't a real concern.

[crueby]

Very interesting  way to turn in that arc.      Did you have a long handle on the tool holder to control the rotation across the arc?  Looking forward to seeing the rest!   

[Kim]

That's a great start to your fusee! 

I don't know that the finish on this initial shape needs to be that good since you'll lose nearly the whole surface by the time you have the spiral groove cut   (that's what I told myself, at least!)

I'm looking forward to seeing how you cut the spiral groove.  I'm still playing around with mine, and I'm not sure the fancy rotating tool is going to do it.  I may end up trying the other method of disconnecting the cross slide and moving it, and see if that works any better.  But I'm stuck now till I get my lathe back together.

Kim

[kvom]

Since my lathe has a taper attachment the cross slide can be detached from the screw by loosening a bolt.  So I will try pushing a threading bit manually against the work. 

When cutting the arc I held onto the top slide, which is quite long.

This will have to wait while I do the assembly and setup of the laser.  The cart to hold it should arrive today, and it needs to be assembled.  Then the riser base for the laser needs to be assembled, and finally the laser needs to be unpacked and adjusted before the first test.

[Jasonb]

Can you not replace the straight taper attachment guide with a template of the curve. Then you just use that to have the tool follow the curve as the leadscrew moves the carriage along at the required pitch. Repeat as needed putting on a deeper cut each pass.

[kvom]

While waiting for Fedex, I wanted to set up the lathe for threading and take a scratch cut on another bar to verify 12 TPI.  I haven't single-point threaded for years, so I needed to reeducate myself.

My Monarch 10EE has a multitude of knobs needed.  1) Rotary dial to select the thread;  each setting has 5 different threads in rows marked ABCBC.  2. Turn another knob to select row C.  3. Turn a knob from Feed to Thread. 4. Select Back Gear to allow slow speed with torque. 5. Pull out a knob that unlocks the half nuts and interlocks the carried feed lever. 6. Turn a knob from Feed to Right Hand.

When I  did the test I got the second C row threads, which is 48 TPI.  After some head scratching, I seemed to remember that I need to swap two of the gears in the headstock, but which two?  I posted the question on the Monarch forum at Practical Machinist, where 10EE gurus hang out.  Awaiting a reply.  Pic of headstock attached.

[Jasonb]

I've just been testing an lathe with an Electronic Leadscrew. I won't want to go back to change wheels and levers.

I'd try one of teh other gear lever position as 48 is a multiple of 12. Looking at charts AE gives 12 while BD gives 48

[kvom]

You might be looking at the 10EE square dial model which has A-E.  The round dial has only one setting for 12 TPI (or most other threads as well).  In 1942 they had to relay on gears.  The A-C setting essentially double the thread count: 3-6-12-24-48.

If I were to replace the 10EE I'd likely go for a Tormach CNC lathe, especially if I were to restrict myself to maximum 8" diameter parts.

[Jasonb]

looking at one of the round dials it looks like the Stud and box gears need swapping over so find a mating pair that are 24T and 48T and swap them.

that would take that pair of the gear train from 0.5 to 2 so you will get 4 times as much leadscrew rotation which will be your 12tpi

[Vixen]

Hello kvom,

Having a lathe with either an Electronic Leadscrew, or better still, a lathe modified for CNC is highly desirable. I modified my EMCO Compact 5 to run under LinuxCNC, which as you will already know, is the basis of the Tormach operating system.

The screen shot below shows how I have available at my finger tip a number of different canned cycles for the thread types I most commonly use, both internal and external. Each thread canned cycle is pre-programmed with the correct feed rate (pitch) and corresponding depth of cut. All I need to do is to alter the length of the required thread, everything else is pre-programmed.



The G76 Threading Cycle is very powerful. The single G76 code is followed by a set of parameters which define pitch, length of thread, final thread depth, number of passes, number of spring passes, retract position and even taper angle for tapered threads.

A typical G76 Threading Cycle command can look something like this:    G76 P 1.5 Z-10 I -0.1 J 0.1 R1.0 K 0.92 Q0 H2 L1

You may have noticed I have also added several canned cycles for the more common lathe operations: turn, bore, face, taper, chamfer, step etc.

If you ever decide to get a Tormach CNC lathe, you will never regret it.

Mike

[kvom]

Answer on Practical Machinist

[kvom]

It took me a while to remember how to exchange the two gears.  The other two larger gears are idlers with the same tooth count, so I just needed to loosen the center nuts.  The upper idler can slide up and down enough to accommodate the two different sizes of the topmost gear.  The lower idler can move left to right, but by loosening the hotdog-shaped clamp on the right, the casting holding the gear can rotate down to make room.  Then it's a matter of holding the gears meshed while tightening the nuts on all gears and then tightening the clamp.

Other than this I spent the remainder of the afternoon unboxing the laser and doing the mechanical parts of the setup.  I had assembled the riser base yesterday, and needed a friend to lift the laser out of its box and onto the riser.  The next step will be testing if the laser beam is aimed accurately, and adjusting the two mirrors that do the aiming.  In theory they do adjust them at the factory, but most of the YT videos dealing with setup seem to show adjustments needed.

Not sure when I'll get back to the fusee.  Since the threading bit won't be perpendicular to the surface.  I think it may be best to just start the grooves and use a triangular file to open them further.

[kvom]

What I now realize is that with the original gears, the center row of pitches could be selected with the A setting.  The rows could be labeled A-B-C/A-B-C.   

[kvom]

I determined that I can't free the cross slide from the screw on my lathe, so I'll need to fabricate a tool holder to cut the spiral.  Smith shows one in this book, so I won't reinvent the wheel trying to design my own.

I got the laser set up yesterday, but wasn't able to do the first simple test engraving because we lost power during a thunderstorm.

I also heard back that the small CNC mill I've been considering will be back in stock in a couple of weeks, but only in kit form.  There's a YT video of one being assembled, and he said it took him 50 hours.  The difference between the kit and the assembled version is $2000, so that's a $40/hour labor rate.  Back in my working days as a consultant I made a lot more than that, but at least I won't have to pay income tax.  I  might still look at a Tormach.

The profile for the plates is provided on paper to be glued onto the plates, similar to Kim's clock.  I need to translate that into a DXF file somehow.  One idea I have is to measure the drawing and come up with a list of points.  Then the Spline tool in Fusion or SolidWorks should do the job.  The reality is that the curves are decorative so there's no need to be too concerned with any small variations.

[kvom]

To layout the plates I madea copy of the right half of the drawing and marked dots along the curves.  Then I copied the page with dots.  One copy I fastened to some aluminum sheet and put in the Bridgeport vise.  I put a center drill in the chuck and used it to locate each dot, with the DRO giving me the coordinates.  I wrote the coordinates on the copy tediously, and then even more tediously created points in a Fusion sketch.  The only way I could figure to assign coordinates to the points was to move/copy a point at 0,0.  Once they were all entered, I used a combination of 3-point circles and spline curves to draw the outline.  I then mirrored the outline.  The profile looks fine, and the sketch can be extruded, so there aren't any gaps.

I then exported to a STEP file and then used an online STEP to DXF converter.  Unfortunately, none of my software that deals with DXF could read the result.   

Up until now the only Fusion parts I've dealt with were for printing, and the STL files generated were fine.  So I don't know where the problem lies.

I did get the laser to cut some practice holes in some acrylic, and also tried engraving it.  That looks good.  The laser is pretty accurate;  I cut a 1/2" diameter hole in 1/4" thick material, and with an inside micrometer it measured .504".