Zeamons Orrery

[Alan CharlestonParticipant @alancharleston78882]

Hi,

I would like to have a go at making Zeamons orrery – http://zeamon.com/wordpress/?page_id=468

The only plan he offers is a pdf file which can be downloaded from the link above. There is a 3D rendering of the item which can be manipulated to show different views and the individual components can be viewed in isolation.

Unfortunately this isn’t of any help in actually making the orrery. Is there any way of extracting the information from the pdf file to produce orthographic views of the thing as a whole and the individual components? It wouldn’t necessarily need to contain the dimensions as I can scale these from the ODs of the gears. The problem with the 3D model is that it’s not clear that you are looking at a component absolutely square on after manipulating it.

Regards,

Alan C.

[John HinkleyParticipant @johnhinkley26699]

Alan,

Have you viewed the series of videos which start at the bottom of the page to which your link refers?  I haven’t, just the first one.  Life’s too short!  The screenshot below, taken from that first video, should give you all the information that you require to make the gears without any need to estimate anything.  The narrator states that he used a MOD 1 gear cutter to produce the gears, and simple maths or the use of a gear calculator will give the exact dimensions of the gears.  Depthing is covered elsewhere in the video series, and the length of the support arms for individual planets and the star can be either approximated from the photo, or chosen by you.

The rest is up to you.

John

Edit:  That’s a lot of gear teeth to cut.  It should build up your handle-twiddling muscles, though!

[Alan CharlestonParticipant @alancharleston78882]

Hi John,

Thanks for the reply. I was hoping to be able to convert the 3D render in the pdf to generate a set of drawings but it looks like I’m going to have to try and suss out how to proceed from the pdf file and the videos.

One thing I’m unsure about is the gear form. I have a set of m1.0 involute gear cutters but Zeamon bought a clockmakers m1.0 cycloidal gear cutter to minimise the friction between the large number of gears in the train. These are expensive from Thorntons and I’m wondering if it will make that much difference if I use my involute cutters.

Regards,

Alan

[Alan Charleston]

On Alan Charleston Said:

One thing I’m unsure about is the gear form.

Please read this and assess which and how strongly each of the three points apply:

https://en.wikipedia.org/wiki/Cycloid_gear#In_clock_and_watch_making

Make your choice accordingly.

Pay particular attention to the depthing requirements of cycloidal meshing pairs vs. the pre-calculated centre-distance approach used by involute gears.

[bernard towersParticipant @bernardtowers37738]

Yes they are expensive but looking at the post from John they will get a proper workout so maybe not expensive. On the other hand you could make them and Mod 1 is one of the easiest as you can use standard tooling. I made a tangential cutter to make the shape on one side of the cutter then flipped it over to do the same on the other side.

[SillyOldDufferModerator @sillyoldduffer]

I looked at this yesterday and came to the same conclusion as John.   Seems all the information needed is in the PDF, but stored in a way that forces the builder to do a lot of work.  Though impressive, the model available inside the PDF is more for admiring than making!  I don’t believe the PDF model has any facility for outputting 2D-plans.   As far as I can tell, getting engineering data out of the PDF has to be done manually.

There is another way.  The Orrery was designed in Gearotic, which can output 2D-drawings, with a good chance of producing a suitable set of workshop friendly plans.    As Gearotic only costs $120, and there’s a link to the Gearotic model here, I suggest buying and learning Gearotic is worthwhile for a project of this complexity.  In principle, save time by spending money.  In practice, I don’t know how difficult Gearotic is to learn, and not everyone takes to 3D-CAD!

Cycloidal gears are advantageous when the gear ratio goes up, and the most common example is a mechanical clock where a big gear at the bottom turning very slowly, drives the much faster moving hour, minute and second hands via a succession of gears.   Involute gears are best the other way round, as when a high-revving car engine often gears down to turn the road wheels slowly.

Which way round does an Orrery go?  I don’t know.  If the Orrery drive input is fast turning, and most of its gears go down, then involute will be OK.   But, if the drive input is slow, and most of the gears go up, there’s a risk friction will make an involute gear train too difficult to turn, making cycloidal the best choice.

Dave

[John HinkleyParticipant @johnhinkley26699]

Oh dear, I had hoped not to get drawn too deeply into this discussion, but, in for penny ….

I’ll just add a couple of links which a) may be of general interest and, b) might help.

“Clickspring” on YouTube has many videos on clockmaking, as well as his renown reconstruction of the Antikythera device.  This link will take you to a video describing the manufacture of a cycloidal gear cutter – or, at least, a close approximation thereto.  The notes in the video description link to the (free) spreadsheet calculator to which he refers in the video.  It downloads as an XLS file, which I loaded into LibreOffice Calc with no problem:  Cycloidal gear calculator

Let me just add that I have no experience in clocks or cycloidal gear making.  You must decide for yourself if the linked information is any use to you.

John

 

[John HaineParticipant @johnhaine32865]

The involute/cycloidal debate is based on the fact that clock going trains gear up from the drive (slow) to the seconds hand and cycloidal gears are reputed to be better (though this debateable).  I think an Orrery will (or can) work the other way and therefore involute should work.  Involute gear cutters are quite easy to make – see

https://mikesworkshop.weebly.com/designing-gear-cutters.html

Actually similar methods can make cycloidal cutters though the dimensioning is different.

Gearotic is not very user friendly and is oriented to generating gcode for CNC.  In my experience something which is just as good and easy to drive is

https://hessmer.org/gears/CycloidalGearBuilder.html

This generates dxf files which can be imported into a CAM program to generate the gcode – I have successfully used CamV}Bam for this for cycloidal gears.  On Hessmer’s site there is also a link to an equivalent that generates Involute gears.  I used cycloidal because the minimum practical size endmill in my CNC is 1mm and you can’t generate a 1 mod involute with that size cutter though involute gears would have been better.  The calculator tells you what the centre spacing needs to be and you can rely on it.

[DC31kParticipant @dc31k]

One thing that might point towards cycloidal gears as a good choice is the fourth row down, where it says ’11t lantern (pinion)’.

An 11t involute gear is technically without the range of a standard set of gear cutters. It is not a showstopper but would require some thought.

[John HaineParticipant @johnhaine32865]

This is where for example the Mike’s Workshop calculator comes in handy.  The lower limit on tooth numbers is set by the tooth getting undercut (or rather not, because that isn’t possible with a conventional cutter).  You can for a given pair of gears choose a pressure angle for which there is no undercut on the pinion – usually that will be rather more than the “standards”.  For example I found that for an 8 tooth pinion a 26 degree PA was needed.  For a low load application such as an orrery there would be no issue with this.  Exactly the same can be done with the methods I mentioned above.

[Alan CharlestonParticipant @alancharleston78882]

Thanks for all the useful information. I was intending to cut the 11 tooth pinion by hobbing it. With only 11 teeth it’s no trouble to have to use say 5 passes to cut it. I didn’t consider undercutting so I’ll look at that again. The sad thing is I used to have a m0.75 cycloidal gear cutter but gave it away years ago when I got out of clocks. Let that be a lesson to you all!

Regards,

Alan C.

[Alan CharlestonParticipant @alancharleston78882]

Hi,

Sorry chaps, but you’ve been replaced by a machine. I’ve just hat the following exchange with ChatGPT.

You said:

Is it possible to convert a 3D model in a pdf file to an orthographic projection.

ChatGPT said:

ChatGPT

Yes, it is possible to convert a 3D model in a PDF file to an
orthographic projection, though the process depends on the software you
use. Here’s a general outline of how to do it:

1. *Extract the 3D Model from the PDF*

Many PDFs that contain 3D models use formats like U3D or PRC to embed
the 3D content. Some PDF viewers allow basic interactions with the 3D
model, but you’ll often need to extract the model to work on it further:

* *Adobe Acrobat Pro:* Allows manipulation of 3D models, but may not
directly provide export to orthographic views. You might be able to
interact with the 3D model (rotating, zooming, etc.) and take
snapshots from desired angles.
* *Specialized Software:* Use software like *3D PDF Extractor* or
*PDF3D ReportGen* to extract the model into a common 3D format like
STL, OBJ, or STEP for further manipulation.

2. *Import into 3D Modeling Software*

Once you have the 3D model extracted, import it into software that
supports both 3D modeling and orthographic projection views:

* *AutoCAD*
* *SolidWorks*
* *Blender* (open-source, can generate orthographic views)
* *SketchUp*

3. *Generate Orthographic Projections*

* Open the 3D model in the software.
* Switch the view to *orthographic projection* mode (most 3D modeling
tools have this option in the view settings).
* Position the model in the desired view (e.g., front, top, side).
* Capture the orthographic projections by either taking snapshots or
exporting 2D views as drawings.

4. *Export the Views*

Once the orthographic projections are set up, you can export them as:

* 2D drawings (usually in formats like DWG, DXF, or PDF).
* Images (JPEG, PNG, etc.).

By following these steps, you should be able to go from a 3D model
embedded in a PDF to a set of orthographic projections.

I haven’t tried it yet but don’t we live in interesting times.

Regards,

Alan C.

[Alan Charleston]

On Alan Charleston Said:

You said:

Is it possible to convert a 3D model in a pdf file to an orthographic projection.

There are some difficulties with your thought process.

In your first post you say “There is a 3D rendering of the item [in the pdf]”.

Yet you ask ChatGPT about a “3D model”. The answer it gives is correct for the question it has been asked.

The problem is that you have not asked it the same question you ask in your initial post.

A 3D rendering is not a 3D model.

The rendering is a screenshot or photograph of the model, not the model itself. Unless the model information is embedded into the pdf, it will not be possible to extract anything other than the picture shown above.

You can think of it as the distinction between attaching a screen capture of a spreadsheet in an email and attaching the actual spreadsheet file to the email. You cannot take the screenshot and do anything meaningful with it in Excel.

In order to embed the model, you need a pdf _writer_ with this capability. Unless the creator of the pdf has that software, knows how to drive it correctly and performs the embedding step, the model will not be in the pdf.

Give your ChatGPT a second problems: upload to it the pdf file and ask it to find the 3D model hiding within. Using my own natural stupidity, I tried it here (https://pdfcrowd.com/inspect-pdf/) and it tells me the pdf file contains two images (pictures).

[Alan CharlestonParticipant @alancharleston78882]

Hi DC31k,

It’s a 3D Model in the pdf file as the image can be rotated to show different views and individual elements can be viewed in isolation. I have extracted a U3D file from the pdf file but haven’t proceeded beyond that yet.

Regards,

Alan C.

[Alan Charleston]

On Alan Charleston Said:

Hi DC31k,

It’s a 3D Model in the pdf file as the image can be rotated to show different views and individual elements can be viewed in isolation…

Alan is correct in that what’s inside the PDF does much more than render.   Appears to emulate a sub-set of a 3D-CAD display.  The rest of the CAD tool is missing so you can’t edit the model or export technical data from it.   Part  dimensions can be viewed and extracted manually, but a huge amount of work.

“Universal-3D” feels like a good idea that never caught on.  An advanced viewer that allows PDF readers to rotate around 3D-parts and animate 3D-assemblies, probably using a stripped down version of the original CAD model.  Possibly too complicated a feature for management, and not complicated enough for CAD trained engineers!

Although  u3d was open-sourced, I can’t find it.  Might have come with an extract tool useful to Alan.  I don’t think it’s related to the active game platform U3D, or if it is U3D has moved a long way from the original concept.

Dave

[Michael GilliganParticipant @michaelgilligan61133]

My contribution to ‘the knowledge’ starts and ends with this:

Try a Google search using: u3d cad pdf

… it worked for me; but I was then immediately out of my depth.

MichaelG.

[Martin of WickParticipant @martinofwick]

Interesting project, having perused the orrery site and videos for a while and noted that Gearotic only offer a fully paid for version, have decided that the best route for producing gear drawings is via fusion 360 using their involute gear form generator. The other parts will have to be done by ‘hand and brain’. As the maker states, it is largely an artistic endeavour rather than a precise astronomical calculator.

Involute gears preferred for me because I don’t think friction build up is so critical as the mechanism seems to operate from high to low speed (possibly excluding the moon and Mercury). Mod 1 cutters are readily available, the pcd distancing is less critical and may be easier to set up. Also might be able to purchase small bevel gears rather than attempt to make.

I noticed the motion of the calendar wheel on the original version was rather erratic and thought it might be due to meshing issues between the small lantern pinion and its driver, so the performance of these two gears may be quite critical.

The query I have is the 11t gear undercutting in involute form. Not sure how to achieve this. Could it be done by carefully filling away the tooth flanks and checking clearances with the 147t driver? Would the driver teeth also need modifying or topping in some way for smooth action?

 

 

[ChrisLHParticipant @chrislh]

Martin, you have a PM

[Author]

Posts