Friday, December 23, 2016

Clock case

The clock I made has been attached to a set of shelves in my workshop.  That's really the best spot to temporarily test a clock, where it can be held up with clamps.

But now that it is done, it really ought to move to a more permanent home both for its own sake and so that I can build another clock.  That requires the clock to either be mounted to a wall or to be free standing.  Since it's fairly heavy (10 lb weight plus a few pounds for the movement itself), wall hanging doesn't seem like a good idea.

So it needs a free standing, tall case.  Especially with the two-fall pulleys now in place, the clock should run for about a day if it's mounted appropriately.

  • The minute wheel makes 12 rotations per day (recall that the minute hand makes a rotation every two hours!)
  • The winding barrel is 1.8125 inches in diameter
  • A two-fall pulley is used
Given that, the weight drops 12 * 1.8125*3.14159 / 2 = 35 inches per day or about 1.4 inches per hour. Located where it is, there is about 46.5 inches available for drop, so the clock should run for about 32 hours.

Building a case for a clock is essentially an exercise in carpentry, and can be utilitarian or artistic or somewhere in between.  My style tends to be more utilitarian, mostly featuring straight lines and circular arcs.

I obtained wood for the case from Home Depot. I had hoped for birch to match the movement, which is birch plywood.  None of the birch plywood they had was in particularly good shape (nor was it high quality).  As a second option, I'd go with maple, since the color and grain is similar.  However, the selection of sizes wasn't sufficient.  So instead, I went with oak.  Overall, oak is not a bad choice: it's easy to work, sufficiently strong, and looks nice.  However, once finished, oak will be darker and will have a more prominent grain than birch.

I cut out the parts for the case with a jigsaw, held mostly freehand.  Though for the long straight cuts I used fences.  After rough cutting, I used a router to transfer the oval cutout from one of the sides to the other.

The most critical part is where the two sides (the longest parts) interface with the movement.  This needs to be a fairly tight friction fit, so that the movement stays where it should during running, but the can be removed easily as well.  I cut the slots small with the jigsaw and then filed to fit.  I first filed the sides separately until a scrap of birch plywood slid into place with some friction.  Once satisfied, I tried each side individually with the movement itself, filing that to fit.  Then I assembled the rest of the unfinished case.  Once assembled, only then did I try both sides together with the movement.  This ensured that I didn't overdo the filing and that everything was aligned properly.  Although filing to fit in this way is a time consuming process, it worked well enough.

To install the clock, it is lifted up and over the back of the case, which is open, and then slipped into the slots.  To remove the clock, the process is reversed.  The fit is tight enough so that the movement does not shift when the weight is applied or removed, or when the clock is wound.

The front of the case is a door.  This is mostly for show, since none of the oval holes cut in the front or sides is covered with windows.  The holes serve a dual purpose: to show the pendulum and to allow the pendulum to swing slightly out of the sides.

In order to operate the door, I made a small oak handle and a brass catch on the lathe.  The catch is threaded on the outside to fit the handle and on the inside to accept a screw for a wooden finger.  I'm dubious about the wooden finger.  Although it seems strong enough, the finger might wear out or break.  If this happens, I'll replace it with a metal one.

To hold the door shut, the finger on the catch engages a slot on the left side of the case.

After trying the fit of all parts, I assembled the case, drilling and countersinking all screw holes carefully.  The video above shows the assembled (but unfinished) case with the door removed.  I used brass screws with Phillips heads throughout, which may or may not be a good idea.  Oak can be a bit rough on brass screws.  Sadly, Home Depot was out of flat head brass screws, which would have matched the screws on the movement better.

Once confident that everything was in order, I disassembled the case, reattached the clock to the shelves temporarily, and spent the better part of a day sanding the case parts.  After all parts were smooth, I polyurethaned (three coats) all the parts. I left the insides of the slots where the clock fits into the side panels unfinished, so that the fit to the movement isn't disrupted.

And, here's the case completed!  Although the case looks quite nice (the picture doesn't do it justice), it now is clear that the movement needs some finishing as well.  Numerous wooden clock makers warn about finishing the wheels or anything that moves, it should be safe to re-sand and polyurethane the plates of the movement.  But that will need to be a task for another day...

Sunday, November 20, 2016

Clock drive updates: Chain sprocket failure and ratcheting friction drive

The friction drive I devised earlier for my clock has generally proved to be somewhat unreliable.  The drive barrel ended up getting slippery, so I glued sandpaper around it to grip the drive cord.  This worked well enough to drive, but winding was then a problem.  Although you could release the friction enough to slip the cord upwards to wind, this slowly eroded the cord... and eventually the cord would break! 

Since the timer ratchet mechanism seemed to work so well, I figured that might be a good idea. But holding the friction barrel with a ratchet wasn't the first thing that jumped into mind for whatever reason, although in the end, it's what I've chosen to go with.  First, I embarked on an experiment to see whether I could use a chain drive with a ratchet mechanism.  I found some "sash chain" at the hardware store that was apparently rated for 35 lb, which is more than enough for the clock's weight of 10 lb. 

Oddly, no one seems to carry sprockets for sash chain, because it's really for window sashes, which don't actually need to be driven.  (There are sash chain pulleys, though.)  So I attempted to make my own sprocket, shown above, by cutting a barrel, drilling holes for nails and then pressing nails into the holes for teeth.  The nails were originally finishing nails that I cut to size using a cutoff wheel and a jig to ensure they were all the same size. 

Although the above picture is just an indication (don't spin with the nails installed!), I cut the barrel to a nice round circle on the lathe.  I find I'm using the lathe more and more to ensure accuracy of circles.  I find it helpful to grip a large bolt in the three-jaw chuck to use as an arbor for the work. 

Here's a picture of the mechanism in place, with a click spring and the chain installed.

Here's another picture with the clock assembled.  Notice the large gap to the left... if the winding barrel slips too far to the left, the click disengages and the weight falls! 

So, I installed a brass pipe...

... that keeps the barrel pushed to the right against the click.

Unfortunately, the mechanism was nothing but trouble.  The placement of teeth is way more critical than I imagined, and the chain kept jumping off the sprocket.  Initially, the chain would lift and jam on the gear above it, which was annoying, but not hazardous.  But as I got the teeth a little closer, the chain would jump off entirely, sending the weight falling.

So, back to the original friction drive...  Here are the parts: a ratchet, barrel, end cap, and brass pipe to keep the barrel in the right spot.

Again, I used the lathe to make perfect circles, and indeed to drill and bore all holes.

Here is the mechanism installed.  Time will tell if it proves reliable!

Sunday, October 30, 2016

Balance staff (next step)

Continuing with my balance staff repairs...  I have a watch that is in pretty bad shape: no hands, no case, etc. and it has a broken balance staff.  So, I'm trying to fix it by replacing the balance staff.  The rest of the movement appears to work -- though it's very dirty.

The first step is to dismantle the balance to get at the staff...

Here is the movement without the staff.  It's a six-jeweled wristwatch.  Seems to be well-enough made, at least on the surface.  Edges are generally not filed smooth -- ok....

But surprisingly, on the back of the balance cock it appears that the drill for the cap jewel skidded across the surface!

Anyway, here's the balance assembly on its way out.

And another shot of the other side...

Here's a view of the spring collet.  It's easily removed with a screwdriver...

...and the roller table.  I also was able to pry it off with a screwdriver.

And indeed, one pivot is broken off (the bottom one in the picture above)...

After about 4 tries, I managed to turn half of a new balance staff.  It's very small:

Here is the view through the microscope.

You can see the upper pivot (appears at the bottom of the frame, and is too long) and that I've yet to turn the lower pivot. In all the watchmaking books I've found, turning balance staffs the the lathe requires flipping the staff around.  This causes lots of trouble, it seems especially since I don't have a small enough collet.  Previously, I tried a pin vice, but this is very off center when gripped in a collet.  For whatever reason, the three jaw chuck does a better job.

I was able to shorten the staff a bit (the pivots are still to large, but need polishing) using the turns/lathe I made.  Here is a view of the staff, the pulley for driving it, and a dummy for checking length.  The dummy is just a hair too short...

Ansonia "La Duchesse" initial examination

This clock is from my sister's landlady.  Here's a link to a similar model.  Inside the clock, behind the back plate is handwritten in pencil "H. Paul, June 26/85."  Presumably this means that this clock was completed on June 26, 1885.

The outside of the clock isn't in good condition.  The case is rusted, with much of the original black enamel having fallen off.

 One of the hinges has broken off, and it seems that the owner used tape to hold the glass in place as a result.

The dial surface is porcelain, which cleaned easily with dish soap and a toothbrush.

Also, it seems that the back of the bezel rusted onto the dial's gilded edge.  Rust is remarkably difficult to remove from gilding, at least without damaging the gilding...  I've removed what I could by carefully brushing with a toothbrush.

The two lions on either side of the movement appear to have been painted with a gold paint of some sort (apparently not gilding).  They were nearly black with gunk/rust.  After spending most of the afternoon in the ultrasonic cleaner, I think most of the gunk is off, but it has revealed that the gold paint is mostly gone.

The movement is missing its pendulum.  I think because the markings on the movement indicate "3 3/4", this means that the pendulum ought to be about 3.75 inches or so.  Nothing else seems amiss on the movement, though we'll see!

Here are a few other views of the movement...

Ansonia "Prism" Crystal regulator

I recently cleaned an Ansonia Crystal Regulator clock "Prism".  The clock had been sitting in my sister's landlady's basement, apparently unused for many years.  Although the markings on this clock are a bit ambiguous, it shows up in this old catalog which is dated 1905.

I made a video overview of the movement to help remind me of critical part locations.  This came in handy because I forgot to mark the striking train wheel positions, so the striking was initially off when I reassembled the movement.

The case is made of polished brass and thin glass windows.  I polished each of the brass parts after disassembly, mostly using Brasso.  However a few parts were really corroded, and needed steel wool first.  After everything was nice and shiny, I used spray lacquer to prevent oxidization.  None of the movement parts were lacquered!

I cleaned the dial (which appeared to be some kind of early plastic) with dishsoap and a toothbrush.  This seemed to do a good job on the gilded metal parts as well, but there were some places on the dial where the gilding had peeled off.  I left those alone.

I cleaned each of the four glass windows with warm soapy water, which worked very well.  Except for one thing -- the soapy water makes them slippery!  I accidentally dropped one window, which promptly shattered on my concrete floor. I spent the next few hours making a replacement window from a thin lexan sheet (32in x 44in x .093in Polycarbonate Sheet).  After cutting the lexan to size, I beveled the edges on a beltsander to match the bevel of the glass windows.  Then I polished the bevel with increasingly fine sandpaper (up to 600 grit), then steel wool, then with polishing rouge, and finally with toothpaste.  Although replacement window is pretty sharp, it isn't exactly the same as the others, so I put it in the back after reassembly.

The movement showed very little wear of any kind, and each pivot hole was still quite round.  I did take the time to polish each pivot by hand, by gripping the arbor in a pin vice and spinning the pivot in a small piece of fine sandpaper.  Since this could conceivably score the pivots, I checked each one under the microscope to verify that they were indeed mirror-polished.

Assembling and oiling the movement wasn't particularly challenging, since I had previously made a pivot locator as suggested by Mark Headrick.  Once the movement was together, it did require a bit of thought to get it back in the case.  There is very little room to maneuver once the glass windows are installed, and I didn't want to scratch or fingerprint anything.  After getting the movement inside, I realized that the intended assembly order is:
  1. Assemble the case (no side windows),
  2. Install the dial (no movement),
  3. Attach the movement to dial,
  4. Install the side windows,
  5. Install the inner top,
  6. Install the chime,
  7. Install the outer top.
The clock runs rather smoothly, and has a satisfying "gong" sound that chimes once at each half hour, and chimes the hour at the top of each hour.  After a little bit of adjustment, it seems to keep time to about one minute each week.

Saturday, February 27, 2016

A drive pulley for the turns

Since I had sooooo much trouble driving the turns I made, I had to think of other options.  I realized that it might work to take the idea of a super-glue arbor for the lathe, and anchor the pulley to part with super glue.

Anyhow, I turned a very small pulley from 1/4" brass rod.

Here it is attached to the part with a tiny drop of super glue, viewed under the microscope and mounted in the turns.

It's still quite difficult to operate, but seems much less troublesome.

I added an electric motor drive that seems to make it easier to use.  I guess it's now a lathe...

Making a gouge

Edwin wanted to carve some details into his pinewood derby car.  So we made a gouge for him so that he could carve by hand.

We started by turning a wooden handle on the lathe

Then we moved on to metal; we took a 1/4" steel rod and turned it down to the width he needed for his detail work.  Then, we drilled out the center

once the center was drilled, we cut a bevel for the edge.  The bevel angle was taken to match a wood chisel.

After cutting the bevel angle, I polished it with 600 grid sandpaper.

Once polished nicely, I cut the tip into a semicircle with a file.  This was kind of rough, because I knew I would get it flat with a sharpening stone.  Then we parted it off

I reversed the part in the lathe chuck and beveled so that it would self-center into the handle.
Then I ground the cutting edge flat on the sharpening stone...
... and Edwin pressed it into the handle using the lathe tailstock quill.