# Basic Repeater Mechanics for the Unmechanical

Discussion in 'Complicated Watches' started by Ethan Lipsig, Apr 25, 2019.

1. ### Ethan Lipsig Registered User NAWCC Gold Member

Jan 8, 2006
1,950
991
113
Male
80% Retired Attorney
Country Flag:
Region Flag:
#1
Because I have no watchmaking skills, understanding the complexities of how repeaters operate is beyond me. Some years ago, there was an excellent multi-part article on repeater mechanics in the Bulletin. I read it, but with little comprehension.

With that mea culpa, I will sally forth (like Don Quijote) with an attempted layman's explanation. I would appreciate your comments on it. If my explanation is risible, be gentle with your comments.

I think every repeater has four basic mechanisms, a switch that triggers the repeat function, a power source that runs it, a mechanism that counts out the interval being repeated, e.g., hours, which I will call the "count mechanism," and a hammer mechanism that makes the repeat sound. I am sure that there are different versions of each of these four basic mechanisms.

What I want to focus on in this thread is what I am calling the count mechanism. I surmise that this mechanism would be substantially similar whether it is counting hours, minutes or any other interval. I further surmise that if the watch repeats more than one interval, it needs separate count mechanisms for each interval, plus a sequencing mechanism that makes it repeat time first for the larger intervals and then for smaller intervals.

If this is correct, then any two-interval repeater, e.g., a quarter repeater or a 5-minute repeater, of the same design would be substantially similar. The only difference would be minor changes in the second-interval count mechanism.

If so, then any three-interval repeater, such as a typical minute repeater, only would be more complicated than a two-interval repeater of the same design because it would have three count mechanisms and two sequencing mechanisms.

Is this analysis correct?

2. ### gmorse Registered User NAWCC Member

Jan 7, 2011
10,845
1,361
113
Male
Retired from Xerox
Breamore, Hampshire, UK
Country Flag:
Region Flag:
#2
Hi Ethan,

I strongly recommend that you read Richard Watkins' excellent treatise on repeaters, available here on his website. Even if you have no wish to actually work on these remarkable little mechanisms, his explanations of their working are the best I've come across; he clarifies some errors and omissions in some of the 'standard' works on the subject.

Regards,

Graham

3. ### Ethan Lipsig Registered User NAWCC Gold Member

Jan 8, 2006
1,950
991
113
Male
80% Retired Attorney
Country Flag:
Region Flag:
#3
Graham, I read Watkins' excellent articles when they came out: "Some years ago, there was an excellent multi-part article on repeater mechanics in the Bulletin. I read it, but with little comprehension." I was hoping that my message would trigger a simplified explanation.

4. ### gmorse Registered User NAWCC Member

Jan 7, 2011
10,845
1,361
113
Male
Retired from Xerox
Breamore, Hampshire, UK
Country Flag:
Region Flag:
#4
Last edited: Apr 26, 2019
Hi Ethan,

The count mechanism consists of one or more 'racks'; they all have sufficient teeth to make the maximum possible number of strikes, (12 for the hours, 4 for the quarters, and so on), and work on the principle that the operating lever/plunger/slide is only allowed to 'gather' enough teeth to correspond with the actual time shown on the dial. When the repeater is actuated, the repeat mainspring is only wound far enough to make the necessary number of strikes; you may notice that the movement is far less for 1 o'clock than 12 o'clock. It also only moves the rack far enough to make that number; it's prevented from moving any further by a tail hitting the appropriate step on the relevant snail. The rack works the hammer by its teeth contacting a pallet on the hammer arbor which is sprung so that the teeth can pass it without causing it to strike when the rack is moved into position but actuates it when the rack returns, propelled by the repeat mainspring. It actuates the hammer pallet just enough times to make the correct number of strikes.

The hour rack is typically mounted on the repeater mainspring arbor between the plates and is effectively a wheel with most of its teeth missing, just the 12 needed for the hours. The quarter rack is usually under the dial and has two sets of teeth, one for the hour hammer and one for the quarter hammer, since this type usually strikes a 'ting-tang' pattern for the quarters. Each hammer has the same spring loaded pallet mounted on its arbor.

Words don't tell this story very well, so here's an extract from a condition report I did on a quarter repeater a while ago, which I hope will clarify things somewhat. Have you watched Richard's videos of repeaters in action?

Regards,

Graham

File size:
994.2 KB
Views:
34
5. ### Ethan Lipsig Registered User NAWCC Gold Member

Jan 8, 2006
1,950
991
113
Male
80% Retired Attorney
Country Flag:
Region Flag:
#5
Graham, I promise to study your detailed posting and to reread Watkins, but that will take me some time. For now, I understand you to be confirming my surmise that there is a separate "count mechanism" for each interval that a watch "repeats." The proper name for this mechanism is a "rack", e.g., an hours rack or a quarter-hour rack. I further understand you to be confirming that the principal difference between a quarter-hour repeater and any other two-interval repeater of the same design is that the former has a quarter-hour rack and the latter has an, e.g., five-minute rack, and that the principal difference between either of those two-interval repeaters and a three-interval repeater of the same design, e.g., a minute repeater, is that the latter has an additional minutes rack. Correct?

I hadn't noticed that the effort needed to trigger repeating grows with the number of "repeats" the watch will be making.

6. ### gmorse Registered User NAWCC Member

Jan 7, 2011
10,845
1,361
113
Male
Retired from Xerox
Breamore, Hampshire, UK
Country Flag:
Region Flag:
#6
Hi Ethan,

Yes, quite correct. This is the minute snail attached to the cannon pinion, with its four quarter lobes each subdivided into fifteen minutes. You can see the quarter snail sitting underneath it.

The minute rack which engages with the minute snail sits on top of the quarter rack.

Just to clarify, these two pictures are of a late 19th / early 20th century Swiss repeater, whereas the English watch in my first post dates from the 1750s. The operating lever in the Swiss piece uses a rack to rotate the repeater mainspring arbor, whereas it's done by a chain and pulley arrangement in the English one, although that design is known generically as a 'Continental' style.

Regards,

Graham

7. ### Philip Poniz Moderator NAWCC Member

Feb 22, 2012
194
60
28
Male
Princeton, NJ
#7
An interesting point Ethan, I do not think that the basic theory of the repeater as an automaton or a semi-automaton has ever been raised. As a mathematician with mechanical inclinations, I find it quite interesting.

First, of course, a repeater has a power source, that is a coiled spring.

Activation mechanism: It can be activated automatically (trip repeater), or manually.

The next comes memory. These are cams of final number of stages. The output depends only on those stages (Moore machine).

Then comes an acceptor, a device accepting or rejecting the inputs from the memory. It is the all-or-nothing mechanism. Many early repeaters do not have it.

After that come transducers which generate output from the inputs. These are the racks. The output is generated by an algorithm written by the shape and placement of racks’ teeth.

Order-control mechanism. It controls which striking comes first, makes sure that the quarter striking “waits” until all the hours are finished and also controls the delay between the last quarter strike and beginning of the minute striking. To visualize this, imagine if the delays between the quarter striking and minute striking were constant, then the time between the quarter striking and the first minute strike, in the last quarter of an hour, would be short and quite long in the first quarter.

And finally the display mechanism which, usually is an audible output of hammers striking on gongs or bells.

Philip

Rick Hufnagel and Ethan Lipsig like this.
8. ### gmorse Registered User NAWCC Member

Jan 7, 2011
10,845
1,361
113
Male
Retired from Xerox
Breamore, Hampshire, UK
Country Flag:
Region Flag:
#8
Hi Ethan,

To expand a little on some of Philip's points:

A trip repeater has to have its spring wound independently of the activation process, like a clock-watch.

This is the device which prevents the strike unless the activating plunger/slide is fully depressed. The 1750s example I showed in post #4 does have it. The repeating work in that watch is practically indistinguishable from ones made 40 or 50 years later.

These are the snails, which are a species of cam I guess.

Regards,

Graham

9. ### dshumans Registered User NAWCC Member

Sep 17, 2009
413
44
28
#9
The rest of the excellent responses explain more than you asked for, but we love these machines and sometimes want to share that.

gmorse likes this.
10. ### Philip Poniz Moderator NAWCC Member

Feb 22, 2012
194
60
28
Male
Princeton, NJ
#10

The construction of quarter and 5-minute repeaters is not always the same. There are two types of 5-minute repeaters; one strikes a number of five minute intervals passed in an hour (2 events), the other strikes quarters and the number of five-minute intervals in a given quarter (3 events). I remember two Patek Philippes like that (one in Chile). The former has the same construction as a quarter repeater, the latter is more similar to a minute repeater but without the delay mechanism.

The same story is with 7½ minute repeaters. The vast majority of 7½-minute repeaters strike 7½-minute intervals in an hour (2 events). The construction is like quarter repeaters. But, for instance, Francois Czapek, the first partner of Antoine Patek made repeaters striking quarters and then the second half of the quarter was announced by a single stroke (3 events).

There are also 10-minute repeaters that work on the same principle as the quarter repeaters (2 events) (and minute repeaters based on 10-minute divisions (3 events). I can't resist not to mention an exotic date repeater (I do not remember if it was 1 or 2 event repeater).

Philip

11. ### Tom McIntyre Technical Admin Staff MemberNAWCC Star FellowNAWCC Ruby MemberSponsor

Aug 24, 2000
82,201
1,506
176
Male
retired SW dev
Boston
Country Flag:
Region Flag:
#11
In English repeaters, the racks would strike an extra blow for the quarters, but a flirt would prevent the action until the 7 1/2 minute point had passed. If you listen closely you can hear the action but not the blow. From ancient memory I think that was attributed to Stogden.

12. ### musicguy Moderator NAWCC MemberSponsor

Jan 12, 2017
5,166
2,395
113
New York State
Country Flag:
#12
cool

Rob