Strutt Epicyclic Train Clock (maybe?)

[crueby]

Have you tried inserting a scan of the drawing into fusion as a canvas? I've  done that many times to trace off it as a sketch. You need to calibrate the size after importing.


 

[Kim]

That's what I did to make a CAD model of the clock I'm making.  Just like Chris said, I scanned the frame drawing and imported that into Fusion, then placed circles & splines superimposed over the image.  And like you did, I only created half the image, then mirrored it so that both sides were identical.

And like you said, the exact shape isn't important, it just has to look good.  The important thing is the location of the holes!

Kim

[Jasonb]

You are going to need to find a way to produce DXFS otherwise the laser will be gathering dust.

Assuming you are using free F360 and that is why you can't get a DXF then I just exported a part from F360 as a Step file and imported into FreeCAD and was able to open that and create a DXF. Was able to open the resulting DXF in a couple of other CAD Programs OK.  Alibre will also do it but you need pay for Alibre.

As for getting the image I use similar in Alibre and it is called "Trace"

Scale and rotate the image a sneeded. Put in the critical holes dimensioned from the main axis and then use circles, ellipses and splines to just draw over the image. F360 is much the same.

[kvom]

I'm not sure the trace method will be of any benefit now that I have the set of points the hard way.  But it's good to know that the technique is available for the future. 

I just reapplied for the veterans' SolidWorks program, and if I get it I'll switch over to it as I'm much more comfortable vs. Fusion.  And with SW I feel capable of modeling the clock.  I still need to decide on a CNC mill before getting beyond the plates and the dial, which are likely the only parts that will be laser cut.

And now I'm going to try to cut the fusee grooves using the technique Kim did.

[kvom]

Tried to start cutting fusee grooves using carbide insert threading tool but found that the right side of the tool was hitting the work.  But did 3 turns to verify it appears to be possible.  The tool allows moving the insert to the other side, so I'll do that and fiddle with the setup to match the scratch already made.  I needed to switch out of back gear to be able to hand-turn the spindle when cutting.

 

[Kim]

Good luck, Kvom!  You'll get there! 

I found this to be a rather intense operation.  I got there eventually, but as you saw, it took me a few tries.  Hopefully you'll get there sooner!

Kim

[kvom]

I was able to obtain the veterans' edition of SW 2025, so I'll redo the point list spline curves and hope to have better luck with DXF output (no need to generate STEP).

After reversing the thread insert I wasn't able to resync the scratch marks on the fusee, so I'm going to try a different tool.  In addition to the plates, the dial will be cut from acrylic.  I should be able to apply powder coating to the hour markers while engraving them.  The laser has "air assist" meaning it blows an air stream over the trace, and my first trial meant most of the powder got blown off.  I can reset it to 30% (minimum), but I could also disconnect the airline from the head.  Lots of things to try.

[Dave Otto]

So you are going to use the laser to fuse the powder? Acrylic also engraves with a nice bright frosty look, you can engrave on the back side so the front remans flat and smooth.

Dave
 

[crueby]

I didn't know that powder coat worked on acrylic? Are you applying it as a second pass, or engraving through it?   

[kvom]

I'm going to experiment with the powder vs. leaving the frosted look.  The reverse side mirrored engraving is interesting;  I hadn't thought of that. I have a lot of 12x12 acrylic sheets to experiment with.

I got SW installed and redid the point list, and this time I got a good DXF that the laser software could load.  After adding holes that are common to the front and back plates, I'll need to keep two separate parts for front and back.   For the front there are 4 holes for stand-off rods that attach the dial.  The location of these holes on the plate would normally be located by the rods themselves once the dial assembly is made, so I'll need to do the same in software.  There are numerous holes in the back plate to be located and then transferred to the front, but Smith locates these with the depthing tool.  So I suspect these will need to be drilled though both plates rather than being laser cut. 

The reason the threading insert didn't work after being relocated to the right side of the holder is that it was upside down, so there's no way to get the proper orientation.  I'm going to start over using a 1/16" parting blade.

My daughter came over to use the laser for some of her craft projects, and she discovered some things that I didn't know on its use.  Nice to have more than one brain helping the learning curve.  Her projects involves cutting paper, wood, and foam.  She'll be back for the July 4th weekend for more trial and error.

[kvom]

Some progress and some frustration with SolidWorks.  The progress was in figuring how to trace an image to generate the plates.  The answer is to scan the drawing in the plan book to a BMP file, and then use the Insert Sketch Image into the open sketch.  Then I was able to do the points/splines and circles on one half, then mirror it to get the entire plate.  The result was over 5 times the needed size, but I could measure the paper sketch to obtain the scale factor, and then scale the SW sketch.  Other than the hole diameters, none of the dimensions are particularly critical.

For the dial, I need to generate two separate parts.  One has the digits alone;  this is needed to engrave them onto the acrylic sheet.  The second is the rest of the dial structure and is needed to cut out around the digits.  I got this to work the first time but then it was pointed out that the bottom numbers (4-8) need to be reversed.  SW text is finicky.  Any text needs to specify a line or arc, and the text is then "glued" to it.  The the single text entity has to be dissolved into constituent lines/arcs/splines.  Now the individual numbers can be rotated to the proper positions.  I had no problems with the first part as each digit is its only outline, and I can then export a DXF.  The DXF is read into the xTool laser software for engraving.

The problems I get with the second part seems to be when the cut lines intersect with splines in the letter outline.  Everything looks OK but SW says there is a problem with the intersection, and I can't figure out how to repair it.     It happens only on one of the digits.

I pulled the trigger on a desktop CNC mill.  The one I was looking at earlier is the DMC2 Mini, but it has been backordered until this week, and it only currently available as a kit.  But I noticed that the larger DMC2 mill was in stock assembled with 220V power, and only 1 available.  So I put in the order and will need to wait for shipment from Toronto.  It has a 24K RPM spindle, so I'll be able to use small endmills to cut the gear profiles.  No torque at low RPMs. 

Now that I'll have a mill to use, I'll start modeling the rest of the clock parts.  Like Kim, I'll need to make a winder and depthing tool too.

[kvom]

My original order for the DMC2 mill was cancelled because it had been purchased offline, and the website wasn't updated.  So I ordered the DMC2 Mini kit instead.  Supposedly takes 20 hours to assemble, but I'm counting on more. 

I solved my SW issue on the dial (was a brain fart), and did a test run on acrylic on the laser.  This had some problems that appear to be caused by parallax in the laser's camera.  It needs two separate runs to first engrave the numerals and second to do the cutouts around the digits.  Positioning the second outlines over the first using the camera looked correct but ended up being shifted.  If I can't get it to work I'll make it on the mill.

Meanwhile I've started modeling the parts following the order that's in the book.  Mostly straightforward.  For the gears, I'll use the Gearotic Motion software I purchased back in 2021 from curiosity.  For the first one, called the "great gear", I specified 64 teeth modulo 0.8.  The generated outer diameter matches that in the book.  The output is a DXF file which can be imported into a SW sketch.  In this case it's just a spline curve, and the body and spokes are added normally. 

When I load the DXF into CamBam, I find that a 1/16" endmill is too large to mill the teeth, and 3/64 won't quite get to the bottom.  The only 3/64" endmills I can find so far have 1/8" shanks and flute length of 7/64.  So for a 3/16" thick gear I'd need to cut half the depth, and turn the blank over to cut the other half.

I'll have to see if this is feasible when I get the mill running.   

[Jasonb]

Time to go metric and try a 1mm cutter

That gear has a lot more than 64teeth? more like 144

Long neck may do it in one setting

https://www.aliexpress.com/item/1005003234398346.html?pdp_npi=4%40dis%21GBP%21%EF%BF%A13.80%21%EF%BF%A13.80%21%21%2135.84%2135.84%21%40210391a017522322318217484e1589%2112000024774951884%21sh%21UK%21855056161%21X&spm=a2g0o.store_pc_allItems_or_groupList.new_all_items_2007586245776.1005003234398346

[kvom]

I mistyped about the gear above.  It has 120 teeth.  The clock plans are all in inches, so metric (other than the gear modulus) isn't in the cards.

The other gears (wheels) do not mesh with each other and are quite thin.  Three of them are .052" thick  So these can be milled in a single pass.   The meshing pinions use music wire held in a body shaped like a thread spool.  A picture of one of the pinions is attached.

The first of these thin wheels that I produced in Gearotic has 66 teeth, and shows a diameter of 2.141".  However Smith shows a blank whose diameter is 2.227".  To get the teeth uniform he widens each by .002" on each side by taking two extra cuts.  After inserting the Gearotic DXF into Solidworks, I scaled it to Smith's size and measured the increase in tooth width as .003".

My mill kit is to be shipped on Monday,

[crueby]

Kvom, even if the plans/parts are in inches, couldn't the CNC be told to use a metric cutter?      I have no experience with CNC, just a thought.