An Orrary Help wanted

[Richard ParsonsParticipant @richardparsons61721]

A request for other peoples opinions

[Richard ParsonsParticipant @richardparsons61721]

 

I could do with a little bit of advice here. I am redesigning my Orrary which a want to write up either for publication or if the ‘Head Hitter’ (David) wants it as part of his ‘open project’

My problem is how far should I go with it? I have 3 ideas which are to represent the movements of the planets as follows:

• The 4 inner planets (Mercury, Venus, Earth (and its moon) and Mars.
• The Six naked eye planets which are the 4 inner planets plus Jupiter and Saturn
• All the Major planets which are the six naked eye planets and Uranus and Neptune I am not dealing with Pluto.

My question is which of those three choices should I make? After all the contraption is one that ‘SWAMBO’ could trot out to show here friends how cleaver her ‘HITS’ is. Oh yes HITS is “Him in the Shed’.

I know at the end of the day the choice will depend on the gears or as I prefer tp call them the ‘coggery’.

Rdgs

Dick

Edited By Richard Parsons on 15/05/2012 08:15:38

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Richard ,

I suggest the six visible planets plus moon – shows everything relevant to most people and not too complicated .

Purely aside I always thought that it would be a fascinating project to design and build a cycloidal Orrary with the Earth in the centre . I don't know if one of these has ever been built but before Copernicus early astronomers might have considered this a logical thing to do . There are certainly early astronomical calculating devices which effectively use Earth as the centre .

In order to please everyone in our hobby fairly you might also like to design and build an Orrary with a flat Earth in the centre and nothing moving at all .

Regards ,

Michael Williams

[BazyleParticipant @bazyle]

Dick,

One of the comments under the 'open project' thread was the comparison to open source software. A charateristic of which is that is starts small and is added to over the years and this could be a great candidate for that.

Therefore I suggest you consider a minimal system that has a good foundation that can be added to by others in a number of different ways over years. This would give lots of people a challenge to work it out and there would be no final end or 'right' way to do it.

It might however be a bt beyond beginners because of the gears, though an interesting way to overcome that might be for clubs to make them where someone has the kit.

Finally don't get too hung up about the exact ratios of the gears as in the end who actually cares – that part is entirely for the satisfaction of the mathematician who likes ot know he has worked it out..

[Richard ParsonsParticipant @richardparsons61721]

I would like to thank both Michael abd Bazyle for their advice. At the moment I am poking about with the problem of accuracy. I have decided to use 1 Earth year as the basis for the machine. The problem is how long a year should I use? (365 days, 365.25 days or 365.26 days.) My basic criteria is that I feel that the machine will be hand driven (you could have it steam driven if you want) but I feel that the manual drive should be about a 4 to 1 reduction. That is one turn of the handle should move the ‘earth’ through 90° (1/4 of its orbit). To show what would happen in say 10 years one would have to turn the handle 40 times. Is this too much? Ok Jupiter will have made almost ½ of its orbit and Saturn 1/3^rd of its. Uranus will have moved under 1/8^th of its orbit and Neptune just less than1/16^th of its.

How much error would be tolerable? One thing I am trying to avoid is the need for compound gear trains (if I can).

I also intend to leave enough space in the gear box(es) for someone to be able to add more planets without having to do a rebuild. The problem with mechanical things unlike software is that to add new facilities you sometimes have to totally rebuild previously made components. If the original designer has made allowances for this the person upgrading the design has to stay within the original design philosophy. So there is with mechanical things a great difficulty for purely ‘free flight’ development.

Michael

I was attracted to the pre Copernicus model of the universe but I feel that this would be an anachronism and would only be of interest to a historian. There was at least one of these machines. Have a look at the ‘clock of Richard of Wallingford’. This was at St Albans Cathedral it is a modern replacement. There is also the ‘Great Clock of Olomouc’ in the Czech Republic.

Regds

Dick

[Eric CoxParticipant @ericcox50497]

Would it be a cardinal sin to ask for gearing that was commercially available as I have neither the knowledge, experience nor equipment to produce said gears.

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Richard ,

On hand turning :

Borrow some ideas from dividing equipment and milling machine quill feeds . Have one handle which gives 'coarse feed' – say one turn equals 100 years and then a worm drive which can be dropped in to give one turn equals 10 years . Each feed handle could also have meaningful division markings if you want . I'd have to think about it more but division in astronomical time units may be more meaningful than Earth years .

On gears :

There is a trick gear train which is useful in building calculating machines and may have application in your Orrary . Gears with larger numbers of teeth have quite a bit of latitude as regards the exact pitch circle diameter – especially when gears are made by a generating method such as hobbing . So it is possible to make two gears same size but each with either right number of teeth or one more or one less . If two such mismatched gears are mounted on a common axis and with a single long pinion mounted on a radius arm engaging both gears a somewhat odd epicyclic gear train is formed . This has the usual three possible ratios but different set ups of fixed and moving gears and teeth numbers can give a huge reduction ratio and a very fine precessional rate of output turning speed relative to a reasonable input turning speed .

It is possible to play numerous tricks with the ratios and get things like 365 .25 to one .

This is the gear train which is used in some types of high class grinding machine to get (via a screw) very fine downfeed .

Regards ,

Michael Williams .

[David LittlewoodParticipant @davidlittlewood51847]

Richard,

Sounds like an excellent project; as I am just in the middle of an overhaul of the screwcutting gearbox of my M300 (theoretically 216 different positions) I have got complex gear systems dancing before my eyes already!

I'm fairly sure a search will throw up numerous resources – descriptions of antique ones, people who have made them recently etc. One suggestion though: do your search using "orrery", as this is the normal spelling. They are named after the 4th Early of Orrery, who had the first one made for him.

David

[Richard ParsonsParticipant @richardparsons61721]

Michael thanks. I was aware of some of the things you wrote about. I also came across another trick. This was to use gears of a different D.P./Module. If you wanted to make a say 6 to 1 reduction but keep the gears on the same shafts you set up say a 3 to 1 reduction whose output was input to a 2 to 1 pairing but the second set of gears were of a different D.P./Module to the 3 to 1. The second set of gears would have the same combined PCD as the first set. That can be quite fun to calculate. Actually I have done a lot of analysis and have come to a conclusion which simplifies everything. The basis of this is one rotation of the earth shaft.

Eric this is what I am going to do. Even if the gears were all of the same D.P./Module the cutters would cost an arm and a leg and you might well need the full set of 8. Hobbing would be different. I do not have a hobbing machine but I do not have one. I have challenged the electronic wizards to create a circuit which would allow me to drive a stepper motor at any programmable speed taken from my milling machine’s shaft. With this I could hob in my mill (or in the lathe). My gearbox is totally enclosed so that you can use the cheaper plastic cogs and no one will see them. If you want to show bright and shiny cogs then you can skeletonise the box.

David I afraid I have a dyslexic keyboard. The words Orrary and Orrery both show up as spolling mistooks. Actually you will find Google is very gentle on this and recourses are 10 a penny. In truth what I am doing is may be is ‘All me own work’ as I have turned things upside down.

[BazyleParticipant @bazyle]

Dick,

as you point out 40 turns of the handle and some planets have hardly moved. This is where I would suggest a little artistic license and speed them up a bit. Frankly in an age where nobody even uses the moon dial on a longcase clock for actually finding the state of the moon who is going to care where the planets really are.

In the end these devices are only going to be turned by little children who want to see the cogs whizzing round.

For large reductions do you need smooth continuous motion? A single tooth pinion or eccentric can advance a ratchet wheel only one tooth per revolution. A high count ratchet wheel can be made from a length of bandsaw blade. Perhaps a bit crude though and more suitable for a non machinist woodworker.

[Richard ParsonsParticipant @richardparsons61721]

I now have the main parts of the Orrery designed. I have taken an executive decision to display only the Naked Eye planets, but I have left enough space to add Uranus and Neptune should the constructor want to.

At this point I need to describe some parts of the design. Firstly the machine is bases on the Earth Year which is 235.26 Earth Days. This is used to calculate the ‘Compound gear trains’ needed to give the planets their proper relative motion. Each planet is mounted on its own shaft of which there are 7. One is for each planet and one stationary shaft for the Earth/Moon combination. The shafts which are hollow fit and run inside each other with Saturn shaft being the Shortest and the planet nearest to the Sun – Mercury is the longest (it is also solid).

I have four problems which I hope someone can help solve. These are

• In the normal Orrery the earth is represented by an oversized globe. Here the problem is should the earth rotate. If it does then it has to rotate 365.26 times for one rotation of the earth around the sun. As you can see if the Orrery were rotating at 10 RPM the earth would rotate at 3652.5 RPM. Should I make the Earth model rotate and would it matter if it did not? A second thought is that the Earth Must rotate once per year to demonstrate the summer/winter.
• The planet shafts have a bearing at their top ends and rest on each other. To limit the height of the display these tops are to be ¼” (6.35mm) high. This would give a total height of about 45mm (just under 2&rdquo. I want to be able to reset the planet’s position. So I need to have some form of ‘friction clutch’. The planet carrying arms need to be robust so that when SWAMBO does her ‘dust removal’ operations things do not get bent. The bottom ends of the shafts have the final drive gear attached to it and in many cases there is not a lot of room there. Any one any ideas how to do it?
• Should the distances between the planets be similar or should they be representative of reality. If you define the Earth orbit as 1 then Mercury is 0.39, Venus is 0.72, Mars is 1.524, Jupiter is 5.203, Saturn is 9.537. So Jupiter is over 5 times as far away from the sun as is the earth.
• I have a problem with the representation of Saturn’s rings these are at 29 degrees to the ecliptic . Anyone any ideas how to make the thing?

Rdgs

Dick

[Jeff DaymanParticipant @jeffdayman43397]

For a clutch idea between the drive gears and the final planet moving arm shafts, a clock type bow spring might work. I have seen them in a bow or arc shape with two points of contact or a triangle formed into a spherical dome with three points of contact. A hub fixed above the bow holds tension to provide the drive friction. Or, a coil spring pressing on a fixed plate on the shaft could work.

JD

[Richard ParsonsParticipant @richardparsons61721]

Jeff

Thanks a lot for that idea. All I have to do now is to find somewhere and way of fitting it. The problem is I have not finished the ‘posh versions’ of the drawings yet. Most of the design consists of sketches and notes in my note book and ideas in my head.

For example the Planet Shafts each have a wall thicknesses of 1mm, (except for the Mercury shaft which is 4 mm solid) but as there are 7of them all told (one inside the other) at the point where they enter the gear box the overall diameter of the shafts is 16mm.

The best way seems to be in the intermediate gears as these can be made into individual modules pressed together using a Belden washer.

Rdgs

Dick

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Richard ,

A few random thoughts :

Most orrery drive mechanisms use a primary drive shaft and then effectively one separate gear train for each planet . There is an alternative approach wherein (eg) Mercury train drives Mercury but also acts as input for Venus which in turn acts as input for Earth etc . This is a sequential approach to designing the gear system rather than the normal parallel method . Since there is a quite simple mathematical relationship between the orbit rates of each successive planet it may be that the drive train becomes very simple .

Whatever you use for setting the relative positions of planets it would be a nice feature to be able to set up the Grand Alignment easily . I don't know how many billions of years will pass before it happens again in realty but it would be interesting to see on a model .

I have never liked friction locks . An alternative method would use clamp collars . For each planet the radius arm is mounted to a collar on appropriate spindle and trapped between two flanges – one on its own spindle and one on next nested spindle . Single split in collar and screw to clamp .

Regards ,

Michael Williams .

[Richard ParsonsParticipant @richardparsons61721]

Michael I was aware of the Mercury drive system, where Mercenary drives Venus, Venus drives Earth etc.. One big problem with that system is that the errors accumulate. There is a secondary problem which is with the compound trains. In order to minimise the gear plates you have to consider changing the pitch (Module/D.P.) of each pair of gears in the train. I will try to explain supposing the Driver(1) and Driven (1) have a certain centre distance. Since Driven(1) is on a common shaft with Driver(2) and you wanted Driven(2) to run on the same shaft as Driver(1) then you would have to search all combinations of pitches in the catalogue. That is tedious and the program to do it is not nice and would occupy Old TOM (my computer – well it is a Toughly Obedient Moron many a long hour. I have designed a gear box which is much simpler and is based on 1 rotation of the Earth to drive each planet individually. It uses Mod 0.5 as this brings the whole thing into a box some 100mm overall. By doing this I can easily achieve errors of between 1.38E-5 and 4.92E-7 of the relevant planet’s ratio achieved with that required. I do that with only 4 gear wheels for each planet.

The only problem with it is that the individual planet shafts are 2mm wider than the previous shafts. So with the Mercury of 4mm OD the Saturn shaft will be 16mm OD and If Uranus and Neptune shafts were added the Neptune shaft would be 20mm OD. The reason for this is that there is one false shaft which does not rotate. This is to provide a drive for Earth’s Moon and if you wanted to you could use the same static shaft for the two moons of Mars.

If you wanted to you could add a second static shaft between Jupiter and Saturn for the major moons of those two planets (I am not going to unless asked as this would require another cart load of gears.) as for Uranus (27 Moons some of which run retrograde) and Neptune (13 known Moons, some also run retrograde –oink!). The gear boxes for that lot would be ‘ginormous’ and would probably need several horse power to drive them.

Rdgs

Dick

Edited By Richard Parsons on 30/05/2012 08:11:40

[MICHAEL WILLIAMSParticipant @michaelwilliams41215]

Hi Richard ,

I envisaged for each stage :

Input and output gears on spindles , a layshaft carrying the other two main gears and floating idlers to match centres . All gears same mod . Hinged layshaft which can drop out for initial setting up of planet positions . This comes from mechanical calculator technology – something similar with pick off gears was often used for setting transfer ratios .

Nothing to do with above but have a look at the planet dial gear trains on Dondi's clock .

Regards ,

Michael Williams .

[Richard ParsonsParticipant @richardparsons61721]

Floating idlers are not needed to match the centres. if the input and output shafts are close enough the 'lay shaft' is put a different position on each plate. But you may have an idea about resetting the planet display. so i am thinking

Until you mentioned it I had not looked at Doni's clock., as it is before old "Copper-nickers", it is of only of passing interest. Any way i would need to employ 'Rent-a Goon' if I owned that much brass over here, I would also have to employ the local Rendoreseg to watch 'Rent-a-Goon' and Group 4 to watch the Rendorseg (the Bill).

Regards

Dick

[Author]

Posts