"Warning" vs. "Arming"

[Paul M]

I'm creating a tower clock maintenance manual for the E. Howard Model 0 striking clock that I help to maintain and need clarity regarding the following:

At 35 minutes after each hour, something releases within the clock mechanism causing the fly fan to rotate about one turn. I have always assumed that this is what is known as the "warning" event, i.e., "a short run of the strike train prior to the actual strike". But a 25 minute heads-up seems a tad early for a strike warning. A colleague believes that this event is what's known as an "arming" event. In his words: "The roller is spun over to the drop plate that's on the other end of the snail follower arm. It's intended to reduce drag on the snail so the effect on the clock regulation is minimized."

Neither of us are horologists so are terminology is probably a bit off. But I hope our questions are clear:

1. What's the purpose of the strike-train motion that occurs 35 minutes after each hour?
   
2. When should the warning event take place?
3. What actually happens within the clock mechanism during the warning event?

Thanks for any clarity you can provide. If it's easier to simply point me to a publication of some sort, that's fine too.
- Paul -

 

[glenhead]

Ok, really basic question here - does it strike correctly when it's supposed to?

In my view, the main benefit of a clock going into warn is that it lets the strike train run a bit to get the oil redistributed on the pivots. The pivots have been sitting for a half hour, and the oil has been squeezed out of the contact point. With modern synthetic lubricants it's not that big a deal, but dinosaur-based lubricants gum up over time. Getting the dinosaurs working again helps protect the pivot contact points better.

Another benefit is that the warn pin is stopped by a lever that drops away fairly easily. Between that and waking up the lubricants, the strike sequence can get rolling more readily and predictably.

As for what happens at the 35 point, if it's striking correctly I'll have to leave that to the tower clock experts. I dunno!

Glen

 

[Paul M]

does it strike correctly when it's supposed to?

Thanks for the reply, Glen. Yes, the clock is striking absolutely perfectly. It's performing as designed, I'm just not sure what the designers intended to happen at that 35 minute mark. I'll try and gather more detail, possibly attach a video ...

 

[FDelGreco]

I've worked on a couple of Howard round tops and they both warned at around 35 minutes after the hour.

Frank

 

[glenhead]

Very interesting. I didn't process that these don't do half-hour strikes (right?), so having a warn around half-way through the sit-and-wait cycle will stir up the dinosaurs.

You'd asked what happens when it goes into warn. I can give you a synopsis of what happens on an American time-and-strike mantel clock, and maybe that will translate to a tower clock. When the strike train is in full-stop mode, there's a pin on a wheel that gets stopped by a lever. When the train goes into warn, there's some sort of doodad (that's the technical term; exactly what it is depends on the clock, usually some sort of ramped-up lobe) that lifts one or more levers to free the stop pin. A projection on one of the levers stops the pin again after a partial revolution, and the strike train is now officially in warn. (That's why you hear sort of a "shshshlack!" when it goes into warn. Everything starts to run, then it's stopped abruptly.) When the time train gets to the hour, the doodad clears the lever (the ramp ends), the lever drops away from the stop pin, and the strike train spins up. I hope this translates!

Have fun with it, and please educate us with what you learn.

Glen

 

[Uhralt]

The time point for going into warning varies widely between clocks. Some of my clocks go into warning about 20 minutes before the strike, some ten minutes before und one of my clocks, a Black Forest clock, goes into warning just two minutes before the strike.

Uhralt

 

[Bruce Alexander]

Hi Paul,
In addition to what Glen details, I believe that the purpose of the "Warning" is that it allows a more precise timing of the start of the Strike (or Chime) run. The Time Train has to slowly lift activation levers/cams to take a Strike or Chime Train out of its locked state. When the activation lever precisely drops, the train in question will run until lock conditions occur. Just how that is accomplished varies, sometimes quite significantly. I recently worked on a Winterhalder Hoffmeier Chime Movement which is set up to move a weighted lever down against a spring force as the Time Train approaches the Quarter. At the Quarter, the lever is released and it pops up to knock the quarter rack hook up and off the rack. So the chime train itself goes directly from the Lock to Run and Back. I thought it was interesting. Please do let us know what you find out about the "Armed" vs. "Warning" States.
Regards,
Bruce

 

[Paul M]

All,

Thanks for the replies. Here, specifically, is the problem I’m facing and why I’m obsessed with knowing exactly when warning occurs. While putting together a Tower Clock Maintenance Manual, I came across the following statements in The Turret Clock Keeper’s Handbook by Chris McKay:

“It is best NOT to wind the chiming and hour striking trains when the clock has warned (is about to strike or chime). This is within the five minutes before quarter striking, and within the ten minutes before the hour strikes.”​

“Warning is the clock’s action just before it strikes. About ten minutes before the hour the going train partially releases the striking train; this is accompanied by a loud click.”​

If my E. Howard Clock warns at 35 minutes after the hour, this means that I should never wind the striking train during the 25 minute period between that warning and the strike. It’s hard to believe that anyone would design a tower clock with this type of restriction. Here is what I had planned to put into my Maintenance Manual:

“The clock warns when the lever drops the rack into position against the snail. After this happens, the striking train should not be wound until the strike has occurred. “​

Could this be what McKay was referring to?

 

[glenhead]

Bob Tascione has a good series of clock class videos. Several of them are previewed on YouTube. Here's the one on the strike sequence of a count wheel clock:

While the counting mechanisms differ between a count wheel and a rack and snail, the warn sequence and how it leads into the strike sequence are quite similar. The video ends with the warn sequence, starting at about the 4:30 mark. There should be a similar sequence of events in your tower clock. (I don't know that I've ever before said "sequence" that many times in that short a paragraph...) If the clock is indeed going into warn at :35, then I agree it would be "best" to do the winding in the first 35 minutes of the hour. Maybe not an iron-clad requirement with that long a resting period in warn, but probably best-ish.

Hope this helps.

Glen

 

[Paul M]

Click here to see a short video of my clock and the warning(?) event that occurs at about 35 minutes after each hour. One thing worth noting: While the strike occurs almost precisely every 60 minutes, the warning event is somewhat irregular. The table below lists the time stamps from our camera feed today showing when the warning and the subsequent strike took place. Why on earth would the strike be so regular while the warning (if that's what it is) vary by as much as 17 seconds??

Warning....Strike
3:34:35......3:59:34
4:34:39......4:59:34
5:34:45...... 5:59:34
6:34:45 ......6:59:33
7:34:52.......7:59:33
8:34:44.......8:59:33
9:34:44.......9:59:33
10:34:46.....10:59:33
11:34:42.....11:59:33​

 

[Bruce Alexander]

I'm just guessing but that may be why there is a Warning state in the first place Paul.
It must be possible to engineer a more precise timing for the start of the strike when the train is in Warning and ready to run as soon as the activation lever drops.
Nice mechanism! It looks like the blades of the "Fan" have a wide range of adjustment to set the speed of the strike.

 

[flynwill]

I believe that the warning state is a mechanical necessity. Consider that the strike mechanism is triggered by a cam on the relatively slow turning minute hand shaft. If you implemented a mechanism where the hi-point of the cam means "start the strike" then the cam lobe has to be narrow enough to ensure it is back in the low state before the strike finishes otherwise the strike will start again immediately. Thus the need to have two states, one triggered by the cam rising above a trigger point and a second triggered by the cam dropping back below the trigger point on the back side of the lobe.

 

[Paul M]

When the strike train is in full-stop mode, there's a pin on a wheel that gets stopped by a lever. When the train goes into warn, there's some sort of doodad (that's the technical term; exactly what it is depends on the clock, usually some sort of ramped-up lobe) that lifts one or more levers to free the stop pin. A projection on one of the levers stops the pin again after a partial revolution, and the strike train is now officially in warn.

Glen, your description matches up fairly well with what I'm seeing on my clock. Click here to see a close-up video of what happens at approximately 35 minutes after each hour. The hourly cam is slowly raising the lever until it allows a bar that it is obstructing to pass beneath it. This then allows the front arbor of the striking mechanism to rotate about a quarter of a turn until it is caught. Based on all the feedback I've gotten, I'm fairly sure that this is what is known as a warning event.

It still bothers me, though, that this warning doesn't happen "like clockwork", i.e., at the exact same time every hour. A single cam, which rotates exactly once every hour, controls both the warning and the strike. Why would the warning vary by as much as 8 seconds over a 24-hour period while the strike remains constant? I've attached a photo of an even closer view of the lever obstructing the bar. Note the presence of something that looks like rust where the two objects make contact. As far as I'm aware, this area has never been lubricated. I applied a small dab of grease there this morning but it appears to have had no effect. Here are the exact times of the last three warning events:

11:34:30
12:34:25
1:34:40​

This makes absolutely no sense to me, especially since the clock is striking like a champion every hour on the hour. I know this isn't a big issue, but it's a bit worrisome when the clock doesn't behave in a predictable, repeatable manner.

- Paul -

 

[glenhead]

If you can get a folded piece of fine sandpaper (like 3000 grit) between the lever and the bar to remove the rust and polish the sliding surfaces, it's *possible* that would tighten up the intervals. It's possible that the rust makes it slip loose at different rates depending on <fill in the blank>. Once it's cleaned up and shiny a dab of light grease will be icing on the cake.

Either that or gremlins.

A dozen passes with sandpaper to polish the interface would at least make it look not-rusty. At worst it won't hurt anything.

Thus endeth my Guess of the Day.

Glen

 

[flynwill]

There is no engineering reason for the warning to have any sort of precision.

As you have observed is is a gentle slope on a slow-moving cam that at some point trips the mechanism to go to the "warned" stated (for lack of a better term).

What matters is that the strike happen on the hour, and that is governed by the sharp edge on the cam that the lever falls off of at the right moment.

If you do the maths you'll probably find that the movement of the lever over 15 seconds (the variation you observed) is ridiculously small such that even the tiniest imperfection or vibration will make a difference. I would say It's doing really will to happen on the same minute.

 

[Paul M]

What matters is that the strike happen on the hour

Yep, I firmly agree and have ceased to worry about the precision of this event. The important thing I learned is that this is definitely the clock going into the "warned" state and that we should refrain from winding the striking mechanism until after the hourly strike.