# Understanding clock gear trains

Discussion in 'Clock Construction' started by Raynerd, May 13, 2014.

1. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#1
Dear all, I have been a member of NAWCC for many years and have now built a few clocks from scratch but always from plans. Over the years I have tried to understand the complexity of gear trains, but I am despirate to now make my own clock from my own plans and so gear trains are something I will need to have crystal clear in my mind.

Please read through my previous threads before asking me to do my own research – I think it will be clear that I read an aweful lot, but I still do not understand how to calculate teeth numbers and ratios! I am quite happy calculating the wheel size and tooth shape given the number of teeth required, but not calculating the number of teeth required for gear ratios.

I could explain what I understand and what I don`t, but I think an example is best.

I have here a very simple clock.

It has:
Escape wheel – 30 teeth, pinion 8 leaves
Intermediate – 60 teeth, pinion 8 leaves
Minute Wheel – 64 teeth, pinion 18 teeth
Drive wheel 60 teeth

Clearly this gear ration works as it is a running clock! What I can`t see is why and in all honesty, what is trying to be achieved other than a full rotation of the minute wheel once per minute!

I have done calculations based on the ratio of gearing and can calculate that the drive arbour turns one rev for 195 revs of the escape wheel….but this doesn`t show me anything!

I guess I`m missing the period of the pendulum (double pendulum on this clock) controlling the escape wheel and I should be working back that way?

I`d really appreciate any help on this. When deciding on the teeth counts, what exactly are you looking for and how would you work backwards from the clock I have described to show “it works”

Any help appreciated.

Chris

2. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#2
Hi Chris
There are several places you want to think about it.
If you want to fit it into a case, you need to think about the pendulum
length.
Do you want a seconds hand?
What source do you want to power it? How many turns will it
take to run it for the time required?
It is as you think, matching the pendulum to the hands is
how you select the total ratio. The size of the pinions helps to
optimize the efficiency. More teeth better efficiency.
Six is the minimum you could use. 8 or 10 is better.
If you have a 1 second beat pendulum and you want a second
hand on the escapement wheel, you'll need a 30 tooth escapement.
You'll need a ratio of 60:1 to the minute hand. You can choose to do
that with one big wheel or more.
You have to decide if you want the hands to rotate the same
direction ( cw ) or it you don't care.
Going from the minute to the hour is a pain because the pitch
of the gears will need to be different to make the ratios work.
Typically a 4:1 and a 3:1. If you want to make separate hands
you can use the same teeth cutters. If on the same shaft you could
do the 12:1 as a single gear and then a 1:1 back to the hour.
Since it isn't much load, a 5 tooth pinion could be used with a
60 tooth wheel.
Tinker Dwight

3. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#3

I`m not really concerned with the size and number of leaves on the pinion, how long the pendulum is or if it will fit in a case – quite simply the calculations, how the gear ratio is worked out to calculate that it WILL make a working clock.
Even if the example is not practical for a working clock – I want to know how the gearing will “work” and more to the point, what the calculations involved are in making it work.

Chris

4. ### Jim DuBois Registered User NAWCC MemberSponsor

Jun 14, 2008
2,867
594
113
Male
Magnolia, TX
Country Flag:
Region Flag:
#4
5. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#5
You need to talk about energy lost in the train. This is not
something that is easy to calculate up front. It can be estimated.
The quality of your pinions and wheels will determine what you'll
need.
There are tables and rules of thumb. Other than that, there isn't much to
it that you don't already know. You know what the ratio have to be
between the minute shaft and the pendulum. How you arrange the
gears to achieve that is really up to you.
That is the best I can do for you.
Jim has some pointers to some stuff that might help.
Tinker Dwight

6. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#6
Jim, I'm a little confused by the links, he ppthornton link is calculations for working out wheel measurements once the required wheel is known. The other link is a calculator which I can't see the calculations it is actually doing.,, which is what I want to understand.

7. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#7
Let me ask another question then or in a different way if anyone can help... And I do appreciate all comments.

In my original post I listed the wheel in a working clock, I said for example, the intermediate wheel had 60 teeth. If I changed the design to 64 teeth - obviously planting have train differently to fit the new wheel size, other than adjusting the pendulum period, will this still work??

8. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#8
Which calculator do you want to know?
Tinker Dwight

9. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#9
Beats per minute from the full train wheel teeth numbers and pinion numbers and how from this, it decided what pendulum length is needed.

For the wheels I listed in my original post, I apparently get a 60 beat per min clock with a one second pendulum, how??

10. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#10
Last edited: May 13, 2014
Pendulum length in inches = L
beat/sec = B

L = 39.1 SrRt( 1 / B )

Each beat on the escapement wheel counts for 2 beats ( once each way )
To calculate the ratio, if it is towards the pendulum, you multiply. If it is towards
the minute shaft you divide.
In your example 2*(30/8)*(60/8)*64 is the ratio to the minute shaft.
3600 is the end number. This gives a 1 second beat.
To know what ratios might work, you need to calculate the
various options.
If you know the desired total ratio, you can look at the possible
selections of wheels and gears.
For this, it would be best to setup a spread sheet. You need to consider
that you can't make a wheel hit an arbor if the gears are all at the same
pitch. As an example: if you had a 8 tooth and a 60 tooth, the radial
distance between arbors would be proportional to 8+60. A gear
wheel of 68 teeth would be too large to work because the tips of the
teeth would be on the middle of the other arbor. In you case, 8+64 is
enough greater than 60 to clear.
So, ratio and clearance are two factors that you can use in your spread
spread sheet to determine what ratios you might want.
In the spread sheet, you might want to show different combinations of
ratios and what end beat you get with each.
Only put down practical ratios. An arbor with a wheel of 68 teeth
and a pinion of 8 teeth is not practical to be used with a wheel of 60
teeth ( at least not the same pitch ).
In general, it is best to make smaller wheels towards the escapement
to ensure you don't have to saw through and arbor. It is possible,
though, to have a secondary plate or cock to allow the 68 toothed
wheel to work with the 8 toothed pinion and 60 toothed wheel.
Just keep in mind that some ratios just can't be made with an integer
number of teeth.
Practical pinions need to be 6 or 8 to carry loads. 5 could be
used at the last stage of a fly where friction is a good thing.
10 is fine.
If you put the minute wheel off to the side of the drive train,
just realize that this causes an extra reversal of the wheels.
A second hand on the escapement would go CCW unless
you added another set of gears or idler to reverse it back.
Does this help a little?
Tinker Dwight

11. ### harold bain Registered User NAWCC MemberDeceased

Nov 4, 2002
40,850
162
63
Male
deceased
Country Flag:
Region Flag:
#11
Chris, you should update your profile to show your NAWCC membership number.

12. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#12

Wow! Thank you...

I need to digest the information and I will be back shortly!

chris

13. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#13
Last edited: May 13, 2014
Here is another way to look at tooth selection:

If we factor what you had it would look like this:

2*(5*3*2)*(5*3*2*2)*(2*2*2*2*2*2)/((2*2*2)*(2*2*2))=3600

We can't play with the first two because the escapement
to pendulum is always 2. So we have:

(5*3*2)*(5*3*2*2)*(2*2*2*2*2*2)/((2*2*2)*(2*2*2))= 1800

To clean things up a little, we can get rid of a lot of 2's:

(5*3*2)*(5*3*2*2) = 1800

Now we might say we want one of the pinions to be a 10
tooth. How might we rearrange the numbers to go with one
10 tooth and one 8 tooth.

1: 5*3 = 15, 2*5*3 = 30, 2*2 = 4
2: 5*2 = 10, 3*5 = 15, 3*2*2 = 12
3: 5*3*3 = 45, 5*2 = 10, 2*2 = 4
4: 5*5 = 25, 3*3 = 9, 2*2*2 = 8
5. 5*3*5 = 75, 3*2 = 6, 2*2 = 4
A few more that don't look practical

Now we need to select a size for the escapement
wheel. I would think 25,30 and 45 are our best
options.
This leave combo 1,3 and 4.
Now lets multiply our pinions to the selections.

For # 1 we have 30 tooth escapement and
A. 15*10 = 150, 8*4 = 32
or
B. 15*8 = 120, 10*4 = 40

For # 3 we have a 45 tooth escapement
A. 10*10 = 100, 8*4 = 32
or
B. 8*10 = 80, 10*4 = 40

For # 4 we have a 25 tooth escapement
A. 10*9 = 90, 8*8 = 64
or
B. 8*9 = 72, 10*8 = 80

Now which ones making the wheel size smaller
towards the escapement we could do for:
center, wheel/pinion, escapement/pinion,
#1A 150, 32/10, 30/8
#1B 120, 40/8, 30/10 ( not as good as 1A because pinion getting larger )
#3A 100, 32/10, 45/8 ( won't work 32+8 < 45 unless the pitch of the 45
was smaller.
#3B 80, 40/8, 45/10 ( 40+10 > 45 so OK )
#4A 90, 64/10, 25/8
#4B 72, 80/8, 25/10 ( not as good because pinion getting larger )

So, select the wheels you'd like to make.
You could do a similar reasoning for one of the pinions
being 6 teeth and the other being 8 or 10.
This of course all assumed the pitch of your wheels and
the escapement was the same. In general, the escapement
has a wider pitch so the 45 tooth escapement isn't too
practical.
Tinker Dwight

14. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#14
I left out one combo that isn't too bad:
center to escapement
120, 50/10 24/8 with the escapement being 24 teeth.
You can back track the math.
Tinker Dwight

15. ### Jim DuBois Registered User NAWCC MemberSponsor

Jun 14, 2008
2,867
594
113
Male
Magnolia, TX
Country Flag:
Region Flag:
#15
(#teeth on center wheel x #teeth on third wheel x #teeth on escape wheel x 2)
÷(# teeth on third wheel PINION x # teeth on escape wheel PINION)=BPH (beats per hour) and there are tables to convert BPH to pendulum length. Generally speaking you do not want to use less than 7 or 8 leaves in a pinion, yes I know less are used in some clocks, the ratio of leaves in a wheel to leaves in a pinion should be about 8:1, I tend to favor escape wheels of 30 teeth, the math and the geometry are easier to deal with than escape wheels with a lot more or less than 30 teeth. It is desirable to make a train geared so as to only require 1 size wheel cutter and one size pinion cutter, unless you have a large selection or are willing to spend a lot of time to make multiples or a lot of \$\$\$\$ to buy them.

16. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#16
All good rules of thumb but I think Chris was
also looking for how one even knew what ratios
would end up working right.
Besides rules are sometimes more fun to break.
All the one I said would work didn't require more than
2 cutters at most for the 8 and 10 tooth.
No reason, one couldn't use all 10 tooth pinions.
Tinker Dwight

17. ### Jim DuBois Registered User NAWCC MemberSponsor

Jun 14, 2008
2,867
594
113
Male
Magnolia, TX
Country Flag:
Region Flag:
#17
Generally speaking the more used ratios of clock gearing have been published many times over the last 400 years. The attached scans should provide a starting point for anyone desiring to build their own timepiece. On the whole, pinions with larger tooth counts are preferred over low tooth counts as they have better geometry as to rolling versus sliding friction, but of course that then requires either more arbors in the train and or more teeth on the wheels, and if using one module of cutter then tooth counts in the wheels need to descend while moving up the train, otherwise you may have a bit of an issue with wheels meeting other arbors .....mechanical clocks have been built using only one wheel and at least one design exists using no wheels as we generally speak. Then there is the work of Aaron Dodd Crane who designed, patented, and built some very interesting "gear trains" for all sorts of mechanical devices...

File size:
224 KB
Views:
616
File size:
168.7 KB
Views:
661
File size:
158.4 KB
Views:
567
18. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#18
Hi Jim, yes I know all too well about the gearless clock as I built one the other year...
http://www.raynerd.co.uk/?page_id=1540

thanks a lot for the attachments. So basically with limited options, there is only realistically the common ratios to choose from!

19. ### Raynerd Registered User

Apr 11, 2004
339
9
18
Teacher of Chemistry
Manchester - UK
Country Flag:
#19
Tinker:
Please can you just explain the issue that you seem to be identifying with 45 ? 40+10 > 45 OK and 32+ 8 <45 not good .... why?? I don`t see where this comes from. I follow it all up to there.

20. ### Jim DuBois Registered User NAWCC MemberSponsor

Jun 14, 2008
2,867
594
113
Male
Magnolia, TX
Country Flag:
Region Flag:
#20
Raynerd, that is some really nice work on the gearless clock, and even better documentation/photos/blog. Commendable on all fronts!

The attached pdf is the gear train of something I am working on right (building) at the moment. The are a number of reasons that the gear train consists of 2 different modules as well as using an "extra" 1st wheel in the train...obviously to provide more run time.....the gearing options I selected are also a bit strange but necessary in this mechanism as I am building it. My point in posting this train is to respond to some of your questions of conventional versus non conventional gearing....it can be pretty much what ever you want it to be, as long as ultimately it fits in the overall design you are creating and the train ultimately delivers the BPM or BPH you want.

I recall many years ago making a great wheel replacement for a French skeleton clock that was 577 teeth on a diameter of about 8 inches. Why 577 teeth? I have no idea as it would not have been all that easy to index in about 1810 when the clock was created.....point being gearing can be what you want it to be....all the rest is just recommendations based on work done over the centuries where convention generally prevailed over experimentation for most clockmakers.

File size:
204.6 KB
Views:
290
21. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#21
Last edited: May 15, 2014
For simplicity, I'd made the assumption that the pitch or modulus of
the escapement was the same as for the other wheels.
If so, the diameter is proportional to the number of teeth.
the middle wheel would add to the escapements pinion to have a
diameter proportional to 38 + 8 or 40.
The spacing between the two arbor is 40*K. The escapement
wheel of 45 is 45*K. The rim of the escapement would be interfering
with the arbor of the middle wheel.
One could swap things, making the escapement 32 teeth and the middle
wheel 45, as in: 100, 45/10, 32/8
That would work.
If you have a wheel someplace that is two small, you can always increase
the size of the pinion to make things work out.
In the case of 100, 32/10, 45/8, you could have made a 20 tooth
pinion, giving: 100, 64/20, 45/8
That would have worked. Since the 32 and 10 both have factors
of 2, you could do 3 times 16/5 or 48/15, giving: 100, 48/15, 45/8
48 + 8 is greater than 45 so there is enough space for the escapement
wheel between arbors.
You only have two rules. The total ratio has to stay 1800 for the one second
pendulum and you can't make a wheel larger, higher in the train than will
fit between the arbors.
It is all about factoring to the prime factors and then arranging them
to get what you would like.
Of course, you could always use a table like Jim has posted.
Using different sized pinions means different addendums for efficient
gears ( different tooling ).
The one that always is a pain is the hour hand to the minute hand.
It is 12:1. That factors to 3 and 4. And the arbors need to have the same
centers. This mean two different pitches or do the 12:1 on one wheel,
using a second 1:1. To match the pitches, the diameters would need
to be the same totals as well. 24:2 and 13:13 would work and then
scale up to a useful pinion 72:6 and 39:39, just to use the same pitch.
With the one second pendulum, if you wanted the second hand on the
escapement wheel it needs to have a 30 tooth escapement.
If you used a shorter pendulum, you could move the second hand
to the middle wheel. The calculation of the middle would be the
same as was done for the minute hand to the pendulum. It has
to have prime factor because you can't have half teeth or fractional
teeth.
Tinker Dwight

22. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#22
The over all ratio worked out nicely with all the prime factor
in the numerator. It is possible that the denominator ends up
with a factor as well that doesn't easily cancel out. In such
cases, the least factored value might leave 7 on the denominator,
to get the desired pendulum length.
A total ratio of 2314.285714 is just such a total ratio. It would be
for a pendulum of 4628.571429 BPH.
Still, you have to end up with an integer number of teeth
and the diameters of the wheels need to fit between arbors.
To get integer teeth, the denominator and numerators need
to have prime factor.
Tinker Dwight

23. ### David Malphrus Registered User

Aug 27, 2014
9
0
0
#23
I'm a new member trying to follow along on this thread. That ratio: 2314.285714:1 would convert to: 575, 350, 395, 218 and 167 tooth wheels and: 50, 96, 191, 44 and 31 tooth pinions. Is this correct?

Best regards,

David Malphrus

24. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#24
Yes, that looks to be real close but originally it was a bunch of gears
with a ratio of 16200 and a pinion with 7 teeth.
You ratios are close enough.
Tinker Dwight

25. ### David Malphrus Registered User

Aug 27, 2014
9
0
0
#25
Thanks, Dwight.

Working with 16200/7 instead of 2314.285714 I get: 460, 540, 490, 288 and 150 and: 80, 92, 90, 98 and 35

--david

26. ### Tinker Dwight Registered User

Oct 11, 2010
13,666
70
0
Calif. USA
#26
You have a lot of values that can be reduced. You can factor out 5s and several 2s and a couple of 3s.
That could make the pinions a lot smaller.
Tinker Dwight

27. ### David Malphrus Registered User

Aug 27, 2014
9
0
0
#27
Thanks on reducing the fractions. Looking at their reduced size, it makes me wonder if Clockmakers strive for hunting tooth arrangements...

--david

28. ### John MacArthur Registered User NAWCC Member

Feb 13, 2007
352
29
28
#28
Occasionally, thought it's been a while, I'll grab the old slide-rule (what's that??). It makes short work of finding an integral solution to a fraction.

Johnny

29. ### David Malphrus Registered User

Aug 27, 2014
9
0
0
#29
"(what's that??)"

I'm old enough to remember them well. All the smart guys had them hanging from their belt in high school (around 1970).

30. ### mpampe Registered User

Jul 17, 2012
8
0
0
Aerospace manager
Washington State
Country Flag:
Region Flag:
#30
Chris,
It looks like you are having lots of fun configuring a clock of your own design. I understand. With all of the good advice from members, did you get one designed?

Just a thought if you don't have the resources for multiple wheel and pinion cutters: you can do a lot with standard gears just ordered off of the internet. I built this one without a single traditional clock part (wheel or pinion or anything else).