Weekend Letter

Some progress since the last post, albeit slow.  

The supports for the selector assemble have been installed complete with bearings upper and lower.  The trick was to get the angle right as the selector has to follow the slope of the magazine.  (Photo-1)

Because I’m using a jig to work on this assemble, it is important to duplicate the mounting hole on the main clock.

The selector arm has been extended to be in the upper track of the magazine.  This will allow it to be stopped by the balls in the upper rack when the cam drives it.  (Photo-2)

Once it contacts the balls in the upper rack, the selector plate will cam out enough to place the lower extension into the balls on the lower rack.

This will separate the same number of balls as there are in the upper rack.  Once I get a few more parts made, I’ll be able to make a movie of that action.

By separating the balls on the lower rack, I will be able to release that number into the hour striker section.

The release plate that kind of looks like a saw will be used to detect when the selector cams out and then release the balls on the lower track.  (Photo-3) The plan is to have another cam that operates the ball gate for the lower track..

When the selector cams out it allows a pawl to engage with the teeth on the release plate and be carried around by the cam. That action will only happen when the selector has moved.

Still needs to be made is the ball release gate and associated parts that will work with the release plate.

That’s for another day.

Now that I have the jig to hold the Hour Ball Magazine, I am able to see how things fit.

There are a number if parts that had/have to be made for the selector mechanism.  The first was the slides that will cam out to separate the proper number of balls to release to the striker assembly.

Next came the Selector Cam which will drive the selector until it is stopped by an hour ball, it then will cause the slide to cam out as it continues on around for one revolution.  I still have to work out how I’m going to accomplish the one revolution.  If I can’t make the selector work the one revolution will be a mute point.

The selector assembly has to do several things:

1. Provide drive for the selector arm with the cam
2. When selector arm is stopped the selector slide must cam out far enough to wedge between two balls in the lower magazine
3. The selector slide must lock into an index to keep it in position while the cam continues to turn
4. When the selector arm cams out, it has to trigger a release of the balls in the lower magazine so they can move into the striker mechanism
5. The plate that triggers the release of the balls has to reset before the selector cam reaches it’s full revolution, but after the balls for the striker have cleared the magazine.
6. The selector slide must reset and return to a home position before the cam completes a full cycle.

The photo shows the partial assemble as of today.  The Selector Cam, Selector slides and return spring, the spring wheel which will be used to return the selector to a home position towards the end of the cycle.

The selector slide will have an extension that will reach into the track area of the magazine.  I won’t add it until I get the selector mechanism mounted to know how long it has to be.

The next step will be to make the mounting for the selector assembly.  Once that is done I can align to other parts and figure out the drive gearing.

The hour balls appear to be moving through the striker mechanism smoothly now.

It is time to focus on the section that will select and release the proper number of balls to the striker.

Back in February I talked about the hour Magazine and its eventual function.  That was to act as a counter for the hour by the accumulation of a ball every hour.  Also to store balls that would feed into the striker mechanism when selectively released to power the striker.

Things a getting a bit crowded in that area and it became obvious that I would have to build and test much of the mechanism for the selection outside of the clock.

I built a jig the simulated the vertical columns of the clock where the magazine mounts.  This should allow me the freedom to work on all sides of the selector and manually test the operation before mounting it back into the clock proper.

The selector has many parts that work together and I suspect will be challenging.

I used aluminum for the jig because:

1. It is a soft metal and easy to cut and tap.
2. It is cheaper than brass.
3. It will be obvious which pieces are part of the clock.

The second photo shows the Hour Magazine mounted on the jig.  The horizontal brace is part of the clock.

Since the last posting I have been working out the design of the mechanism for the selectors.  I think I have enough to at least get started.  As in the past some of this will be worked out on the fly.

That is the exciting part.

Now that the balls are transitioning through the hour lever, the next things to do are provide an exit path for the balls and to harness the power from the lever.

Since there was already a track that went by the lower level of the travel of the hour lever,  (Ok that was  a weird sentence) I decided to divert the balls to it.

A bracket was made to provide a stop for the hour lever and also divert the balls to the back track.  (see photo) The photo shows the hour lever cup in the down position.

To get the power from the hour lever up to the top, I made a bevel gear assembly (see photo) and mounted it to one of the vertical columns.

Ball bearings were installed to insure a smooth operation of the gears. A rod then carried the movement from the lever assembly up to a hammer by the bell.

When the weight of the ball carries the hour lever down, the vertical rod will rotate cocking the hammer.

When the ball exits the cup, the lever returns up allowing the hammer to strike the bell.  That is the theory anyway.

I quickly learned that I needed more power to make the bell louder.  The only way to accomplish without starting all over was to add more weight to the hammer.  Since the hammer rotates horizontally, it didn’t affect the hour lever very much.

However, there was an erratic timing of the strike as the balls transitioned through the lever mechanism and what I felt was just too fast.  I needed a way of pacing the strike so it could easily be counted.

It was also important not to tax the drive motor in doing this.

I ended up making a flanged wheel with a collar that had six socket head screws evenly spaced around one side.  Then mounted it on the horizontal shaft that turns at 5 rpm.

A block was made that held a small lever with a counter weight and would ride against the screw heads as the shaft turned.

Since the shaft turns at 5 rpm, that moves the lever 30 times a minute or every two seconds.

A plate was added to the hammer arm that when the hammer cocked, it could be latched by the lever until released as the shaft turned.

The balls were originally moving into the hour lever about every second or so.  Now when the first one arrives and operates the hammer, the second ball will cock and latch the hammer until released.  This creates a strike every two seconds.  (See Movie)

Now it is time to work on the selector for releasing the proper number of balls for the hour.  I’m pretty sure this will be the biggest challenge yet.

Note: In the last three months I have spent untold hours trying to get the main bell to be louder and finally replaced it with a larger one that works much better.  There were a couple of gremlins that popped up and demanded attention as well which required some rework of the track.  I also replaced some work I had done earlier that I wasn’t happy with.  It is pretty normal to see a better way after you have done something.  This post covers a fair amount of time and progress not really noticeable. 

It has been very hot here in Arizona the last month and that has made it difficult to work in the garage.  And then there were the four days I spent in the hospital and recoup after that didn’t help too much.

Anyway, things are back on track and we will see how things go from here.

Now that I have balls accumulating in the Hour Striker Magazine, I can work to route them to a mechanism for the hour strike.

The challenge here was to get the maximum power from the lever to eventually operate a hammer for the large bell on top.  A little CAD work helped figure out how much room I had to work with and ratios of the gears involved.  I case you are wondering, I have not documented the whole clock with drawings.  Much of it is made up as I go.

The principle was to have one ball at a time come into a cup and the weight of the ball would make the lever rotate down while cocking the hammer.

A the bottom of travel the ball would then exit and the lever to spring back up allowing the hammer to hit the bell and accept another ball if needed.  That was the theory anyway.

The first step was to build the lever mechanism.  Using flat bar, it was mounted to the vertical columns.  The cup (tube) on the end would hold the ball through the cycle.

There was a side entrance to the cup which was slightly tilted to allow the ball to roll out at the end of the cycle.  The pivot shaft of the lever was then geared to another shaft closer to center of the frame.

The way the gearing was set up, this would allow the hammer to cock when the ball forced the lever down.

A section of track was made to feed the balls from the magazine into the cup.  As in the quarter strike mechanism, I used a small flat piece of brass to block additional balls while the lever was moving down. This also stopped them from spilling all over the place during the cycle.

Originally, I had planned to have the balls exit the cup toward the outside of the clock.   Then I realized that I still had to deal with the balls from the hour counter when they dumped after the 12 o’clock strike.  Once I made that track, it became obvious that I could also use it for the balls from the lever to exit as well.  The track for the 12 o’clock dump is shown in the photo to the right.

With the counter weight in place, the balls would cycle through the mechanism smoothly.

The movie is of the balls going through the lever mechanism.  I used a plastic lid to catch them because the receiving track wasn’t built yet.

The next step will be to harness this power and drive the hammer up by the bell on top.

I’m having to make these titles up as I go.  Anyway there had to be a way to divert the excess balls for quarter strike to the rack that holds then for the hours strike when needed.  Making a fast moving ball take a 90 degree corner was problematic.  This took far longer to accomplish than I had anticipated.  I ended up remaking this section three times.  Ugh!

The first step was to establish a track between the exit path over to the magazine (lower track) that was just installed.

Considerations:

1. The balls had to have enough speed to make it to the magazine after being diverted.
2. There had to be a way of sensing that the magazine was full
3. The route had to consider the other tracks that would be involved with the hour strike

A section of track was made that connected the exit path to the lower magazine track.  There had to be enough slope for the balls to roll into the magazine.

A feeler lever was installed in the striker magazine to sense the presence of a ball.  This was attached to a shaft that moved a piece into the path of the diverter section when a ball was in the full position in the magazine.

The exit path was reworked and a new section made so that the weight of the ball coming through would make it drop to a lower position making the ball run into a solid diverter block and bounce into the track section leading to the magazine.  This happens very quickly.

If the magazine is full, the exit path section would be blocked from dropping down and the ball would continue into the spiral and return to the second lift.

I reworked this section several times before coming up with a valid solution that worked consistently.

The movie shows the ball being diverted to the magazine.  You will see the plate move that blocks the section from dropping when the ball crosses over the sensor wire on the way to the magazine.

So far I have gotten the quarter hour strikes working and routed the extra balls back to the second lift.

I struggled with how to keep track of the hours and also at the same time select the right number of balls to feed the striker mechanism.  That and fit it all into a rapidly diminishing real estate.

It quickly became obvious that I did not have enough room to use the straight line approach for selection and must figure out an arc solution.  Spending some time in CAD, a concept took shape.

At least enough so to start construction.  I would figure out the details later.

I wanted to work out first the accumulation of the balls that would keep track of the hours.  The top section of the assemble will do that.  The lower section will be for the magazine of striker balls.  Once triggered the cycle would move the selector until it contacted the balls on the upper rack.  When stopped, a slide would separate the same number of balls on the bottom rack.  At full cycle, the balls on the bottom rack would be released into the striker mechanism.  (Not made yet)

Early on, there was a diverter than would route one ball on the hour to a different path.  I needed to receive that ball so it had some place to go.

The next aspect was the collect balls for a supply magazine for the striker.  I had excess balls not used for the quarter hour magazine coming down every minute.

The trick is to divert the needed balls from the exit path to the striker magazine.

Once the striker magazine was full, then allow unneeded balls to pass by and return to the second lift.

So this post is only about making the structure to hold the hour balls and the magazine for the strike.  The last photo shows balls in the upper rack for the hour count.  Now every hour a ball will end up in the upper rack.  Since there is only room for about 12 balls, right now the clock can only run for twelve hours unattended.

Next step will be to get the balls into the striker magazine.

Sometimes things get complicated.  Especially in relationships and ball clocks.

Having the clock run more often has allowed some gremlins to show up.  I started to notice that the reset for the upper flips was occurring at random.  This of course messed up the quarter, half and three quarter strike timing since everything for that quarter would start over.   I also saw that the balls that dropped through for the quarter trip would hang up on rare occasion.  Luckily, I happened to catch the extra cycle while working on another area.

It appeared that when the 12th ball would drop through the its flip, it would hit the frame by the clutch in such a way as to vibrate the clutch dog and releasing the clutch to cycle.

Now this was something that would happen maybe once every 8 hours or so.  Once was enough regardless.

Looking at the clutch dog, I noticed that it was resting against the latch that engaged the clutch.  That pressure would change slightly as the shaft rotated and at just the right time would allow the vibration to trip the clutch.  It also was causing erratic tripping of the clutch when the quarter, half and three quarter balls would drop through their respective flips.  It appeared that three things would need to be addressed.

1. The pressure against the latch had to be kept neutral.  I was able to accomplish this by notching the face of the clutch cam to create an index that held it in a home position until tripped.
2. The clutch dog needed to be held tripped until the clutch cycled.  This was because the home position of the clutch would allow the dog to return under the latch after the ball tripping it had passed through the flip.  So a latch was made with a notch to caught the dog.  Sounds funny huh?
3. The latch had to be reset during the cycle of the clutch so the clutch would only make one revolution.  This required another cam added to the end of the clutch with an arm that followed it to reset the latch.

That fixed it and the other issue of balls hanging up when dropping through the flips for the quarter hour trips.

That is one of the challenges of a project like this, to solve the unexpected and move on.  The other is coming up with new mechanisms.

Anyway, next will be planning for the hour strike and how to count the balls for that.

With the strikes for the quarter hours coming too quickly it was necessary to figure out a way to slow them down.  Fortunately (by chance) the exiting balls pass below the quarter lever strike mechanism on their way back to the second lift.  It was about 1 second after exiting the cup on the quarter lever that the ball would pass below the lever mechanism.  That seemed like a good delay time.

I put a bar across the columns to mount the trip lever assembly.  It also helped make the structure more ridged.

A pivoting piece was made with adjustments for throw and mounted on a block so that when a ball passed down the track, the upper part would push against a latch holding the quarter lever down, releasing it to accept another ball and start the strike cycle again.

This solved the issue and provided a nice pause between the strikes,

The latch the caught the quarter lever had to be weighted to suppress bounce and assure a solid latching to occur.

The movie is in slow motion and shows the action of the latch and release working.

So now I have balls going through the quarter magazine and ending up in the striker mechanism and then no where to go after that.

The excess balls that are not needed for the magazine replenishment also have no where to go.

It is time to make more track.

I liked the aesthetic of the spiral for the timing balls and decided to use the same effect on the opposite corner.

I learned a lot about making a spiral from the hiccup with the first one.

So this time it went much easier.

I had to deal with the balls from the excess drop of balls not needed to replenish the quarter magazine and from the quarter hammer lever after the strike and route them back down to the second lift.

A new section of track was made to loop the balls around and over the feed for the first lift.

I noticed pretty quickly that while the striker for the quarter hour worked well, it was almost too well.  The strikes were coming faster than I wanted and there was also some inconsistency.

It was obvious that I had more work to do.

The movie shows the hammer action and the balls exiting below to return to the second lift.