I needed approximately 30 BF of walnut.  I checked all of the local stores, and found the availability of walnut in the quantities I needed was fairly limited. One store wanted $15/BF, however $8-10/BF seemed to be closer.  It wasn't terribly eager to spend $450 on wood.  The only available boards were < 10", therefore, I was going to have to glue up panels for the sides.

I scoured Craigslist, and even went to a garage sale that had 300-400 BF.  However, the boards were a bit too rough for my comfort level, almost all with live edges.  Again, getting wide enough and long enough boards was going to be a challenge.

I found a very assuming Craigslist post of someone trying to sell Walnut.  We exchanged a few emails, but I wasn't overly optimistic.  I borrowed the neighbors truck, and headed to his property.  It was nestled in the woods, with piles of wood and logs everywhere.  It was run by two gentlemen.  As soon as I got there, they had 6 boards, 5/4+ that were 9 3/4" wide and over 14' long.   It was exactly what I was looking for!  They had milled it that morning.  The wood itself has a beautiful grain pattern.  We did some quick math based on assumed yields (6 boards @ 12', 9" width, 1" thick).  Approximately 50 BF, he wanted $5/BF.  We agreed on $200.

We went to load them in to the truck, and quickly discovered that was going to be a disaster.  Fourteen feet boards just don't fit in a pickup. They offered to throw it on their trailer and drop it off at my house.  Very generous and convenient.

I also mentioned that I would love to get a 15"x15" piece to use for the clockface.  We walked around his yard, looking at various logs, and seeing if we could find something large enough.  We located a good looking burl that had good potential.  They tossed in on the mill and started slicing it up.  Unfortunately, there was a bit of rot on the inside, and a 15"x15" piece just wasn't in the cards.  They were confident they would find something that would work and let me know.

 

The wood made it home without any issue.  I now have it stickered and end sealed with polyeurethane.  The tough part is now going to be waiting for it to reach the proper moisture content.  It currently is in the 25-35% range.  I believe 12% is a reasonable target.

I did a quick recalc on the total volume.  14.5' x 9 3/4" x 8" (total thickness) = 94 BF!  That works out to $2.13/BF!  It is hard to complain about that.

Also, since the boards are 9 3/4" wide, I believe I can make the sides without gluing two together.  A 9" depth should work great.  Since they are 14.5' long, I can use a single board to make both sides, and another board to make the back.  Figure another board for the rest, and suddenly I have twice as much wood as I need.  My original figures (8 boards) were based on requiring glueups and 8' lengths.  I really scored on this one.  And I know I'll be back to visit these gentleman.

This post concentrates on the design of the wood case.  It is based on the dimensions determined in a previous post.

Autodesk Inventor was used for the modeling. See the Drawings for the details.

A couple of notes:

• All material from 4/4 black walnut.  I am still looking for a reasonable source of the material.  The best bet so far is from a local supplier with 4/4, approximately 7" wide and 7' long for $8.50/BF.  However, I am still looking around for wider or cheaper stock.
• It looks like all pieces will require a glue up of multiple boards.  I do not own a jointer, so planer and tablesaw jigs will be required.  I have never done panel glue ups, and that makes me the most nervous.
• All Dados will be 3/8" deep.  A router with a 1/2" upspiral bit will be used to create them.
• The mounting of the movement will require a specialized piece.  It will need to have holes for the weights, and not interfere with the pendulum or chimes.  There are taped holes on the bottom for mounting screws.  The layout of these holes is going to be a pain, and will be done closer to game time.
• I will design the doors after the case is built.

The Movement

The case will be built around the movement.  The critical dimensions are shown below.  The majority of them were measured with a tape measure.  Not exact, but sufficient for this exercise.

 

Face Width

• 2(Hand shaft to edge of lever): 2(6 7/16") = 12 7/8"
• Movement width: 7 7/8"
• Hand width: 2(4 3/4") = 9 1/2"

Use a 14" wide face (Case ID). For reference, Hermle 1161-853 round dials are 13" wide, with a 11" diameter timetrack.  This will yield a total width of 14" + 1" + 1" = 16".

Face Height

• Hand shaft to top of chime block: 10" - 2 13/16" = 7 3/16"

WIthout making the face wider, the face is not going to be able to be square.  We should also give ourselves some room in the top depending on the angle of the chime rods.  Therefore, go with 7 3/4" on top, and 7" on bottom.  Use a 14 3/4" tall face.

Case Depth

The depth of the case, neglecting the door.  The slimmer the profile the better.  The door will add about 1"

• Tip of winding shaft to back of hammer: 1 1/2" + 1 7/16" + 4 1/4" = 7 3/16"
• Back of hammer to back of chimes: 5/8"
• Minimum depth: 7 13/16"

Use 9" deep case.  

Case Height

• Face: 14 3/4"
• Bottom for square: 14"

Use 78" tall case.  This was determined based on the space.

 

See the complete design here.

The Setup

I finally got tired of having a movement, and not having it in action.  However, a reasonable case is way to far away, so I put together a test rig.  The rig consisted of some 2x4s, and a cedar top plate that allowed the cables for the weights to pass through.  I picked up some lead shot (for making shot gun shells) for weight.  The left two weights are 7.7 lb mason jars.  The right weight is a 9.9 growler.  The horizontal bar was required to keep the weights from interfering with the pendulum.

It actually took me more time than I care to admit to attach the pendulum (a 1x2) to the leader.  Getting a rigid connection proved to be a challenge.  The best solution ended up slotting the pendulum, and sliding it on the leader.

How it Works

It runs!  It ticks! It tocks!  When I put hands on it, they even move.  I spun the minute hand a little bit and the hammers started firing.  A quick beat count with a stop watch yields it is 14% fast.  Adding a clamp to the bottom of the pendulum got it to within 4%.  Not bad for no calibration.

The pendulum is the most important parameter in keeping accurate time.  Therefore, it is critical to get the pendulum design correct.

The Period

The only parameter governing the period of an ideal pendulum (on earth), is the length.

The 1161-853/94 is geared for 3960 beats per hour [Hermle Service Manual].  In order to make the math work out correctly, I presume that each half-cycle of the pendulum is a beat.

$f = \frac{3960 beats}{hour} \cdot \frac{1 hour}{3600 seconds} = 1.1 \frac{beats}{second}$

$f = 1.1 \frac{beats}{second} \cdot \frac{1 cycle}{2 beats} = 0.55 Hz$

$T = \frac{1}{f} = \frac{1}{0.55Hz}$

$\boxed{T = 1.\overline{81} seconds}$

Therefore, we are going to have to target a pendulum design that maintains a $1.8181 second$ period.

Let's see how that compares with an ideal 94cm pendulum.

$T \approx 2 \pi \sqrt{\frac{L}{g}} = 2 \pi \sqrt{\frac{94 cm}{981 cm/s^2}}$

$T = 1.95 seconds$

Well, it looks like an ideal pendulum isn't going to do the trick at 94 cm.  Therefore, what we really want is a pendulum with an effective length, $\bar{L}$, of

$T = 2 \pi \sqrt{\frac{\bar{L}}{g}}$

$\bar{L} = \left( \frac{T}{2 \pi} \right)^2 g$

$\bar{L} = \left( \frac{1.81818181s}{2 \pi} \right)^2 (980.665 cm/s^2)$

$\boxed{\bar{L} = 82.1175 cm = 32.3 in}$

Therefore, what we really want, is a pendulum with an effective length of 32.3 in!

Physical Design

The pendulum will consist of two components.  See the Introduction for the inspiration for this design.

Bob

A bit arbitrarily, the bob was selected to be a 4" diameter, 1" thick, 6061 aluminum disc.  Assuming a density of $0.0975 lb/in^3$, the bob will weight 1.23 lb.

Rod

The rod will consist of a 3/8" x 3/8" x 1/16" U-Shaped 6061 aluminum channel with a walnut inlay.  Using the same assumed density of aluminum, and a density of walnut of $0.0243 lb/in^3$, the rod will weigh 0.012lb.

Length

I have to admit, I put together a fairly comprehensive spreadsheet to calculate the effective length and period.  I attempted to account for the moment of inertia of the rod and the bob itself.  However, at the end of the day, the eccentricity of the bob outweighs the other factors.  Therefore, the distance to the center of the bob should be 32.3 in.

Connection to Movement

The movement has a pendulum leader that is 5.5" long (should be included in the total pendulum length.  There are hooks available that are designed to interface with the pendulum that I will likely use. However, I still need to develop a way to tweak the length.