Self Winding Clock Co "Western Union" #2

© Brooke Clarke 2007

Fig 1	Fig 2
Fig 3 SWCC #4	Upper right = single sync coil middle = dual coil spring wind vibrating motor Bottom = Dual coil local sync battery relay
	Fig 4 SWCC #4 w/o 15.5" Convex Glass
Fig 5 (5 Aug 2007) new glass installed	Fig 6 Synchronizing Button

The first "Western Union" clock I got only had minute and hour hands and I wanted one with a big second hand. See that page for most the the info I have on these clocks.

Model	Dial dia"	Dial dia	Shape	Pendulum	Beat	Case	Back	Dial ctr to Mtg scr"
25	11	12				Metal	Metal	6
25	15	16				Metal	Metal	8 5/16
27	10	11				Wood	Metal	5  5/8
27	12	13				Wood	Metal	5  5/8
28	10	11				Wood	Metal	5  5/8
28	12	13				Wood	Metal	5  7/8
35	10	11				Wood	Metal	5  5/8
35	12	13				Wood	Metal	5  7/8
37	11	12	Round			Metal	Metal	6
37	15	16	Round			Metal	Metal	8 5/16
41				Mercury	60
42	11	12				Metal	Metal	6
43	11	12				Metal	Metal	6

Overall diameter about 19" by under 5" thick.  15" opening in case to view dial.
The dial is 16" in diameter.
Glass missing from case, maybe should be 15.5" OD.
Movement # 402449.  Subtracting the number on the other S.W.C.C. movement yields ( 402,449-39,580) 362,869.  It's very unlikely that by chance I have the first or last of the Western Union clocks, so an estimate would be that they made over 400,000 thousand clocks between 1934 and 1970.  Probably making most of them in the earlier years.

Henry says that the sweep second had was a request from the broadcast (radio & TV) business that needed to start their shows at the exact top of the hour.

The pendulum has a 1/2 half second beat or 1 second period so the escape wheel (60 T) advances makes one revolution per minute.
The pinion gear (8T) on the escape wheel advances one tooth each 7.5 seconds.
The first idler shaft gear wheel ( 60 T) takes 60 * 7.5 sec or 450 seconds ( 7.5 minutes) to make a turn.
The pinion gear ( 8 T) on the first idler shaft takes 56.25 seconds for each tooth.
The gear ( 64 T) at the front of the minute shaft is attached to it's shaft and takes 56.25 s * 64 = 3,600 ( 60 minutes) seconds to make a turn.
The gear (28 T) on the front of the minute shaft takes 60 min / 28 minutes to move 1 tooth.
The second idler gear (56 T) takes 56 * 60 min / 28 = 120 minutes to make a turn.
The second  idler pinion (12 T) takes 10 minutes for a each tooth.
The hour gear (72 T) takes 10 min * 72 = 720 minutes (12 hours) to make one turn.
The second idler gear drives the minutes hand gear (28 T) (not the minutes shaft) since both the minute shaft gear and the minute hand gear connect to second  gear and both have the same number of teeth ( 28 T) the minute hand moves at the same rate and in the same direction as the miunte shaft.

I've tried a number of ways to measure the anchor, and none of the prior methods worked for me.  But this method of photographing the anchor held on a corner of the frame and with the flats of the pallets parallel to the frame seems to give very good results.

The silver rectangle on the right is used as a spacer to get the crutch rod at the correct height and it's being pulled to the frame by rubber bands hooked on the end just to the right of the photo.

The distance between the pallet faces can be computed as SQRT(1.16*1.16 +0.274*0.274)=1.192"
The distance between the center of the anchor shaft and the escape wheel shaft center can be measured a number of ways and they all are within a few mils of 1.1635"


The interesting thing is the 14.1 degree angle of the crutch relative to the frame.  It was measured on the photo and so may not be accurate.  The arc tangent of .274/1.16 is 13.3 degrees.  So the crutch is very close to, if not exactly, square to a line between the pallet faces.
This is s/n 402419 which is one of the newer clocks fitted with a single coil synchronizer.

The escape wheel has 60 teeth.  The diameter across the points of the escape wheel is 1.865" and at the root of the teeth is 1.591".  90 degrees divided by 6 degrees per tooth is 15 degrees for a design that has no drop.  This anchor spans 13 teeth leaving 12 degrees for drop.

This anchor has a different outline from the older anchors, like on s/n 79006, but the pallet faces are in the same place.

I've heard that this escapement is different from the standard escapement where the wheel has 30 teeth.  For a 60 tooth wheel the design is different.  So the designs published in clock books for 30 tooth wheels are not the same.

The pallet faces have grooves worn into them that are very deep to the point of generating some recoil.  The other S.W.C.C. clocls I have do not have this problem.

This is a dead beat, or Graham escapement, it's just not working that way now.


The links (called cocks) have a pin that can be removed simply by pressing on the cock while the pin is on a hard surface.  With the pin out of the cock the anchor can be lowered deeper into the escapement wheel to see how it works.  This will have to wait since now, 27 Jun 2007, the clock is completely apart and has been cleaned.



The wear to the back is on the dead face and probably is not too important, but the wear on the narrow front face has changed the shape and now causes the escape wheel to turn backwards a little.

Can this be repaired?  ans: Yes but not easily.
A better thing to do is move the escape wheel so that it meets the pallet on the side that is still good.  It's almost as if the factory setup does not center the escape wheel on the pallet but instead puts it on one side just so that it later can be moved to double the time until a repair is needed.

Pallet Adjustment

The Graham style anchor is held not by the frame but rather by a link on both the front and rear frames.  The links are attached by both a screw and a pin to the frame.  The hole in the link that passes the screw is a tight fit.  This means the link is not an adjustment, but rather a way to remove the anchor - crutch part, like to release the main spring.  If the link screw hole was enlarged then the link could swing about the pin allowing some up or down adjustment of the anchor.  This might allow balancing the drop on both sides.

The front and rear frame are almost identical stamped parts.  I think the link location was a seperate step that probably was done using a fixture on each frame plate to set the anchor shaft hole directly above the escapement wheel shaft the correct distance.  Then the holes for the screw and pin were drilled.

This means that the anchor crutch and escapement wheel are interchangeable between clocks.

The bushing that is driven by the main spring on the front has black goo around it.  That's consistent with the pallets being worn  (see above).

This clock got a lot of use.


The stub shaft probably should not be removed since the holes for the two screws are oversize to allow adjusting the stub shaft position.  That's why there are flat washers under the screw heads.  This shaft holds an intermediate gear that mates to similar gears above and below it.

I've seen 4 different ways a second hand is handled on the Self Winding Clock Co. clocks:

• This clock has a sweep seconds hand (a hand marking seconds on a timepiece mounted concentrically with the other hands and read from the same dial as the minute hand).  But in this case the all the hands are mounted on the seconds shaft.
  
• A regular second hand (a hand marking seconds on a timepiece not mounted concentrically with the other hands and read from it's own scale) maybe an inch long located between the dial center and 12:00 with a hole almost two inches in diameter.  The second hand serves the normal purpose of a second hand.  The hole serves the purpose of allowing someone to see the contact points of a master clock to confirm that they are separated from each other, i.e. are not fused together.  If they are fused together all the slave clocks are stopped.
• A regular second hand maybe an inch long located between the dial center and 12:00.  This second hand is on the 'cape shaft and is a simple thing to add.  The  Minute and Hour hands are on the hour shaft.
  
• No second hand is also very common.  Although the 'cape shaft is just behind the dial.  Virtually all of today's digital clocks that operate from a quartz crystal do NOT have a seconds display.  That's because they are not accurate enough.
  

Contacts on the clock close the circuit that connects the local battery to the winding motor.  The motor is the classical two coil type where each coil is aprox 1" dia x 2" long.  About 5 Ohms resistance.  The 4 screws that hold the motor to the frame were not holding it in the proper place.  I changed the position of the coils so that when the armature is at the top of it's travel it would just hit the top point of the coil core.  This involves both the left to right and the rotational position of the coil relative to the clock frame.  This cured the problem of the one coil switch drive pin hitting the insulating washer on the end of the coil.  The other drive pin is missing.  After the coil alignment to motor is running nicely.

Winding can be manually done while the leaf switch is pressed.  The main spring has an auto stop brake and you will hear the motor stop winding.
When one of the 12 teeth of the winding cam lifts the leaf switch and it contacts the other leaf power is sent to the winding motor.  But it's not clear how there leafs are prevented from shorting to the grounded cam?  Also what disconnects the winding motor from the battery?

Adjustments

• 4 screws that attch the dual coil assembly to the frame.  Loosen and friction tighten then adjust coils so that bottom of armature just touches the top of coil cores at top of swing when upper spring has been removed.  I could not do that on this movement because the insulating cylinder that opens the electrical circuit hits the coil insulating washer.  So for not set coil bace a little.  Probably should grind off some of the washer so it clears the contact sleeve.
• Top spring can be rotated by loosening screws on front and back frame that hold it's support shaft.  Top of armature should just touch spring when it's 1/16" below top of core.
• Bottom spring should just tough armature when top of armature is even with bottom of core.  Again the support shaft can be rotated to adjust.
• The contacts should be open 1/32" when the armature just touches the top spring and should sit in the center of the contact metal when at rest.
• The front are rear springs should "make" and "break" at very close to the same place. (I'm missing the small screw and plastic sleeve so can not tune both.
  

The idea is that when not powered the armature rests on the lower spring and the contacts are closed.  When the circuit is closed the armature swings up and opens the circuit just as it touches the top spring but inertia will carry it a little further.  Then the spring throws it back down and about when it's free of the top spring the electromagnet is energized and starts pulling it back up.

I found that setting up the motor based on the dimensions is a good starting point, but it's running weaker than it could be.  So I tried various adjustments while the motor was running and find that the top spring adjustment is best "tweaked" while the motor is running and the movement is in a upright position.  The latter is very important since gravity pulls the armature down.  When the top spring is way too high or removed the motor will run but it's weak.  As the top spring is lowered the motor speeds up and sounds stronger.  Too low and it stops.  Once the top spring is set the motor will run with the movement upside down.

This is sort of like dyno tuning a car compared to doing it by setting things statically. Dyno tuning results in more power.

The lower spring does not seem to do much.  Maybe it needs to be longer or shorter, not sure.  If you could get an increase in power like the top spring provided then that would really be something.  Door bells and buzzers have the armature supported on a leaf spring that acts in both directions and they can be installed in any orientation relative to gravity.

Electrical

When a couple of "D" cells power the coil pair and then the current is stopped, the kickback is about 40 Volts.  The shunt resistor is marked blue gray black silver which translates into 68 Ohms 10% tolerance.  40v/68 ohms is 588 ma.  An ohm meter across the combined coil and resistor reads 5.6 Ohms.  With a voltage across the coil of 2.5 volts the current is about 446 ma.  The acutual cycle is for no current to be drawn from the battery between windings which may be 5 minutes or 60 minutes minus say 15 seconds of wind time.  During winding on the up stroke while the contacts are closed the current tops out around 600 ma. then as soon as the contacts open the current does not change because the coil generates a 40 volt spike

It takes 824 us for the 33 volt pulse to recover to 37%.

Winding Mechanism

At the left is the armature.  When the electromagnet is activated it's shaft rotates CCW lifting the arm that piviots about the ratchet wheel.  The pawl on the pivot arm turns the ratchet wheel clockwise.  When the electromagnet is deactivated the armature shaft rotates CW, lowering the pivot arm.  Now the pawl in the upper right of the photo holds the ratchet wheel and the pawl on the pivot arm moves with respect to the ratchet wheel. 

The ratchet wheel shaft drives a pinion gear, a part of which is just visible in the photo, which in turn turns the main spring housing, the rear part being the smooth wheel in the photo.  But the large gear behind the main spring housing is not turned by the ratchet wheel.  The large gear is on the hours shaft.

Mainspring Shaft

By making the ratchet teeth very fine on the winding ratchet gear the amount of dead space is minimized making for a more efficient winding mechanism.
These videos were made prior to disassembly.
SWCC2AlmostWinding.avi - but stops - as clock was received
SWCC2Winding.avi - after adjusting the motor
SWCC2WndMech.avi - frame like shown in winding mechanism photo above, armature moved manually

Main Spring

Strength

Removing Tension

The anchor is held in place by adjustable bars on the front and back frames. By removing the bar on the rear frame the anchor and attached pendulum drive crutch can be moved free of the 'cape wheel and by light finger pressure you can control the speed of the wheel to allow the spring to unwind.

A cam located on the hour shaft activates a leaf switch that turns on the winding motor about every 6 minutes.  The manual winding switch (near wires) is connected in parallel with the leaf switch activated by the cam.  Note that by winding more frequently the tension of the spring on the 'cape wheel will be more uniform and because of that the clock should keep better time.

This is very different from the single lobe cam used in the prior Self Winding Clocks that should only wind once per hour.

At the left of the photo you can see the 'cape wheel (seconds shaft) and the rod used to drive the pendulum is visible below the back plate.

The silver shaft in the center is the minutes and the spring is on the hours shaft to the right.

Laminated Core

My guess is that the laminated core was the key technology needed for a single coil to replace the dual coil setup that was used since about 1840.  Since the "Western Union" clocks started around 1934 and ended  about 197? the laminated core coil may be a fairly new development.
  For comparison look at the Sync coils on an earlier Self Winding Clock.


Marked in white paint box:

C.E. 6088      F100A

(C.E.  0088?)

 This is the first time I've seen a single coil in this application.  All the others have been dual coils.  The single coil has a larger diameter (1  11/16" dia x 2 1/2" long) and the most important feature is that it uses a laminated stack of metal to form the return magnetic path.  All the dual coil setups are using single chunks of soft iron for the magnetic path.  The problem with a single chunk of metal is that when the magnetic field changes Eddy currents cause losses which get fixed when laminations are used.

Note when A.C. mains power became available Eddy currents quickly became understood.  But in the days when DC was the only kind of electricity that was being used no one saw the need for laminations.  But you would notice that the efficiency of an electromagnet would suffer when there was any kind of a magnetic path outside of a coil.  Note that in all these electromagnet applications the E-M does not just sit there doing nothing.  The only time it has value is when it's pulsed, i.e.  turned on and off.  During the changes is where the need for the laminated core comes into play.

In Charles R. Underhill' book "Solenoids, Electromagnets and Electromagnetic Windings" first edition 1910 second edition 1914.
Laminated cores are mentioned in relation to AC plunger solenoids, and nowhere else.  There are many mentions of "Iron clad" solenoids and electromagnets, but never laminated ones.

Laminated Core Electrical Steel (Wiki)

In October of 1890 Steinmetz was asked to calculate the loss in iron motor cores.  By 1892 he was publishing papers on the loss.  He said the hardness, saturation and hysteresis are the three properties of the core material. (Ref Science Vol. XX No. 509, 1892)
Allegheny Technologies - Electrical Silicon Steel - Grain Oriented Silicon Steels - when the DC is turned off these have lower remanent magnetization than soft iron cores.  They also have higher saturation magnetization thus can be smaller or  use less current.

The laminations are about 0.030" thick.  At the left of the photo you can see that there's a small end gap between two different laminations on the same layer.  There is a similar gap at the left face of the coil.  So to get this "U" shape core the laminations are "L" and "I" shapes.
The central core is 0.533" x 0.420" for an area of 0.224 sq in. (144 sq mm).

Sync Coil Data

The wire is enameled about 0.029" dia, or maybe AWG 22 wire.  6 Ohms would take (Cooner Wire) 6/16.2 = 370 feet.
The coil might have an ID of 0.68" and OD of 1.6" with a length of 2.3"  this gives a rectangle for the turns that's 0.46" high x 2.3" long or 1.058 sq inches.  If each wire was a square 0.03" on a side then 1,175 turns would fit.  The mean radius is 0.522, mean circumference is 3.27" which times the number of turns is 3,850" or 320 feet, not too far from the computed 370 feet.

Wedging the armature in the open position and using the FLC-100 magnatometer to measure the filed at the end of the core while the coil current is varied from 0 to 250 ma produces the following plot.  The coil is clearly saturated at 250 ma.

Marked in white paint box:

C.E. 6088      F100A

(C.E.  0088?)

1873659 Process of Treating Silicon Steel, August 23, 1932, 148/110 ; 148/111; 29/17.2 - aka Magnetic steel
1714038 Process of Treating Silicon Steel, May 21, 1929, 148/110
Tool steels are made with the minimum possible amount of silicon, very different from magnetic core silicon steel.

So have not found a patent to cover this.  But I was looking in the 1890 to 1940 time frame, but now think it's in the 1934 to 1970 time frame.

6  Ohm sync coil at 3 volts draws about 450 ma for a pair of "D" cell batteries.  Does actuate but hard to say if enough force to move hands far.

11 June 2007 - Sync coil testing - The coil measures 150 Ohms  and 135 mH using the HP 4332 LCR meter.  This is an AC measurement, not DC. These numbers are probably bad since the coil is 6 Ohms DC.

Tried remeasuring the DC resistance with a Fluke 87 DMM and get wildly varying readings.  The AC volts output from the sync coil jumps around hundreds of mv if the armature is wiggled.  The proper drive is going to be a voltage much higher than 3 volts from an even still higher voltage loop with current limiting. 

Based on the DC power supply the coil resistance is 5.9 Ohms.  Armature just pulls in at 0.71 V @ 0.105 A.  Reasonable sounding action at 3 volts, movement standing up, either polarity.

11 July 2007 -
2.5 volts across the coil gives current of 417 ma.  With a 327 ohm resistor across the coil the kickback voltage is about 100 v.
.417 A * 327 Ohms = 136 Volts just after the instant of turn off.  After 125 us it has recovered to 100 volts.  The time to recover to 50.5 volts (37% of total change) is the time constant, and is 710 us.  If this is just an L/R time constant then the self inductance is about 232 uH.  I doubt this is the case.

If we take 3 time constants as the time to settle, i.e. 2.13 ms then:
V = L * dI / dT, so L = V * dT / dI = 136 v * 2.13E-3 / .417 = 0.694 H or 694 mH

Using a 1k resistor across the coil the instantaneous kickback voltage will be 417 volts.  37% is 154 V.
272 us is the time constant.  If L/r then L is 88.9 mH.  Using Self Inductance:
L = 417 v * 3 * 272 us / .417A =  816 mH

A method of testing is to put the HP 54501 scope in triggered single shot mode.  Connect the two batteries to the synchronizer coil.  Then clear the scope display (which arms the trigger) and quickly disconnect the clip lead from the coil terminal.  This generates the negative spike cleanly.  With a 10X probe the max volts/div is 50 * 8 divisions is a max on screen range of 400 volts.  The one shot bandwidth is not that great on this scope.
It's not clear what the value of self inductance actually is.

Placing a diode across the coil slowes down the recovery time to a little more than 157 ms. This is a real bummer.

Heard from Henry that the line voltage was 120 and the current was 250 ma.  No more than 25 series connected clocks on one circuit.  The panel has an adjustable resistor to set the current.
My example:
20 clocks where each has 6 Ohms DC resistance so 120 Ohms clock resistance.  The wire resistance might be 150 Ohms for a 1 mile loop, more for longer loops.  So the current would be = 120 Volts / (120 + 150) Ohms or .44 amps.  So to get down to 1/4 amp the sending station resistor would be adjusted until the current was 1/4 amp which would take about 210 Ohms.

Now look at the impact on the charging time constant.   With 120 Volts and 1/4 amp the loop resistance will be 400 Ohms.  The mix of clocks, wire and adjustable resistor may change but that's the total.  So now the time constant of any clock is made faster by 400/6 or 66 times.  This is a very big improvement in the speed of operation.  It applies to both charging the inductor and discharging.

Strange Armature

The armature is constructed from materials I don't recognize.
The side away from the coil looks like aluminum (lead?), but a magnet sticks to it.  On the other side it looks like asbestos.  On the coil side a thin iron strip is attached with a screw (hole tapped into "aluminum") that has a "V" pointing to the bottom of the coils central  laminated core.  Trapped under the iron strap is a thin sheet of brass that extends up higher than the core, so when the coil is charged and the armature closes the brass prevents the asbestos from touching the central core.  But when the coil is installed in a clock that can't happen because of the connection to the clock mechanism.

Synchronizing Relay

This is a new item that's probably part of the single coil setup.  In the older Self Winding Clocks the synchronizing coil was driven from line power.  But in this case there appears to be a classic design dual coil relay that the line activates and then uses the local battery to drive the single coil. I haven't traced it out, but that's what it looks like.

Both the older and newer clocks use the local battery to light the red "lightening bolt" lamp each time the clock gets synchronized.  When the lamp is dim it's a clue the battery needs to be replaced.  The other clue is the winding takes longer and sounds different (just a guess).

12.7 Ohm relay coils is the load on the line feeding the clock.

Synchronizing Button

In Fig 6 above is shown a Synchronizing Button.  It's simply a doorbell button with one wire going to the battery terminal for +3 volts and the other wire going to the left Fahnestock Clip.  A short wire goes from the -3 volt battery terminal to the right Fahnestock Clip.  Push at 1 second till the hour brings the minute and second hands to 12:00 and release at the hour to allow them to move again.

I tried this with a long wire and it did not work.  That's consistent with what I've been learning about the No. 6 Dry Cell being capable of delivering currents in the area of 20 Amps.  When I was testing using an Agilent (HP) E3617A bench DC power supply that's rated for only 1 Amp the relay action was not what one would hope for.

The battery adapter I'm using the the 37SS clock shown in Fig 6 above is a special double "D" cell adapter where they are connected in series and no springs are being used in order to  minimize the internal resistance. 

This might explain the relay in the synchronizing circuit of my other 37SS.  That would be the way to synchronize a clock from another clock where the loop voltage was much lower than the 120 VDC used on the Western Union time circuits.

Sync Wiring

S.W.C.C.#4 clock (the one that came with burned out motor contacts) The prior owner misconnected the sync coil wires.  See the two photos below.

Bad wire connection from Coil to triangle blocks	Good Fahnestock Clip wiring. Sync button wiring.

The three screws that go through the sync binding post triangle plates are electrically connected to the movement frame and are NOT connected to either of the sync coil terminals.  Both plates are electrically insulated from the clock movement.  It's best to only make electrical connection using the top screw both to avoid this type of mistake and to minimize the loop resistance.  I don't remember touching the coil to plate wiring since that's not needed to remove the movement from the frame.  Maybe the prior owner did that for some reason.  It puts a short across the coil which has no effect if the coil is not connected to anything.  It has the effect of presenting an open to the sync circuit preventing it from working.

The battery adapter on this clock was made using No. 8 hardware for the terminals and with two series connected "D" cells inside for a 3 Volt battery so the right side battery cup is not used.  Also the crossover cable to the right battery is not used.  This lowers the circuit resistance.  Positive battery to lower movement terminal.

The sync button is wired with a gray wire from battery positive to left Fahnestock Clip.  Red doorbell wire to right Fahnestock Clip.  Black doorbell wire to battery negative.
It's best to install the minute hand in the correct position on the square shaft.  That way the synchronizing will work at the top of the hour.  If you have the minute hand installed in one of the three wrong positions, that's OK, just try synchronizing at 15, 30 and 45 minutes past the hour and make a note of which one of those works and just use that time for sync.

29 Sep 2007 - When the sync button is pressed the minute hand will move to some position between 00 seconds past the minute to 59 seconds past the minute.  This is because the seconds arbor is NOT reset by sync on clocks that don't have a second hand.  (Anyway that's what I'm thinking after trying to sync the clock at 11:00 am but not getting the minute hand straight up.)

Coil Drive

In order to get snappy and sure operation of the synchronizing solenoid the applied voltage needs to be about 200 Volts and current limited with a resistor to provide the desired current.  A good way to generate the 200 Volts is by using a blocking oscillator.  The one from a used throwaway flash camera could be used (it may be closer to 350 Volts, but that's OK since the series resistor can be made larger.  These oscillators can be made to start when the input voltage is a fraction of a volt and will easily work from a 3 Volt supply.

Time Source

The ideal time source would be a radio broadcast that includes both the time code and Daylight Saving Time information, like WWVB in the U.S.  The most practical non broadcast method would be to use one of the Dallas Semi (now part of Maxim) clock oscillator chips.

Daylight Saving Time

The only contorl that's built into these clocks is the synchronizing pulse.  You can only use the sync pulse with the minute hand is within a few minutes of the hour.  So, to turn the clock back one hour the sync pulse would need to be sent 20 times seperated by 3 minutes  thus holding the hour and minute hands fixed.  To advance the clock one hour the sync pulse would need to be sent for 23 hours once each 3 minutes stoping the clock for just short of a day.

Note all of the above requires no change to the clock and so can be used with a string of clocks and they all would be kept exactly on time.

Radio Time Signals

I've heard that at some radio stations there was a relay associated with the clock that made the "beep" to go with at the tone the time will be xxx O'Clock.  I'm guessing this was a relay wired in series with the synchronizer coil and so would be closed at the top of the hour at 59 minutes and 59 seconds, held for a second then opened at exactly the top of the hour.  It would switch a 1 kc audio tone into the radio audio circuit.  If you know more details please let me know.

As current goes into an electro-magnet it charges the magnetic field.  If the current is stopped the magnetic field collapses generating a current in the opposite polarity but starting out at the same value.  If there is no snubber circuit across the coil the voltage can be very high.  This causes arcing on the mechanical points controlling the current input, wearing out the points.

By placing a resistor in parallel with the coil about ten times the coil resistance the voltage across the coil at the start of the back EMF will be 10 times the normal coil voltage and so eliminates the point arcing.  The early shunt resistors were made in the form of a coil wound on a non magnetic core where the wire gauge and length determined the resistance.  Newer versions use a carbon composition axial lead resistor.

If a diode is installed across the coil it will stop the arcing but also it maximized the recovery time of the coil.  So if a diode is placed across the synchronizing coil there  will be a delay is allowing the second hand to start.  If put across the winding motor coil it will weaken the motor.  That can be seen and heard on the video of the Electromagnetic Toy Engine.

A better snubber circuit would be to use the resistor but add a diode in series.  Now during the normal charging part of the cycle the diode is back biased and the resistor is disconnected thus saving battery power, but when the contacts open the diode turns on putting the resistor in the circuit limiting the back EMF and protecting the contact points but doing it without slowing down the coil recovery.

I think the color is battleship gray.
An approiate color for a USNO clock.
Missing glass as received.
Top hole for alignment pin and notch for top wire access.
Bottom threaded hole for missing attachment screw.

Glass Attach

Glass Attaches with 5 clamps lined with what may be sheet cork?
The opening in the metal case is 14 & 13/16" so the glass must have a larger diameter.
The distance from the opening edge to the inside metal bracket is hard to measure exaclty but using the minimum rradial distance of  0.512 gives a max diameter for the glass of (14.8125 + 2* 0.512 =) 15.8635".

The cork (?) material is about 0.050" thick so the glass should be less than 15.8635 - 2* 0.050 =) 15.7635" diameter.

The black goop was holding sound deadning material to help cut down the noise for use in radio and TV studios.

Naphtha (Wiki)  (aka: White Gas, Coleman Camp Gas, Lighter Fluid, Benzin, Petroleum Spirits, Ronsonol, Ligroin) or a commercial clock cleaning solution can be used.  Many of the commercial clock cleaners contain ammonia which can be harmful to brass if not thoroughly rinsed.

The electrical and fiber parts should not be cleaned in any of these solutions since they will be harmed.  This means disassembling the movement.  That means how do you get it back together.  This is where a digital camera and maybe making some drawings is a good thing.

The minute hand shaft shows some slight wear.  In the worn bushing photo you can see black goo (oil plus worn brass bits) on the end of the shaft which in a normal clock would be for the minute hand.  This clock is unusual in that there are no hands on this shaft and instead they all are centered on the second hand shaft.  The post to the left of the minute shaft holds gearing.  In the background you can see the escapement wheel.  The large gear driving the pinion gear on the minute shaft is part of the main spring shaft assembly.

At this point I've removed all the items that have fiber or plastic parts getting ready for cleaning.
Also making drawings to be sure it can be put back together.  The tricky part is that there are adjustments that need to be made during the assembly process which I'm still trying to understand prior to dismantling the frame.

The frame has been opened.  And more photos and drawings being made.  The small washer that fell off the back frame plate came from the armature shaft.  The two screws that were removed to take off the back frame are sitting on it over the holes they came out of.

This is the "F" movement, just like the first S.W.C.C. clock.  But this frame has additional tapped holes to support that stub axle shown in the photo above that are not on the more common "F" frames.  There may be other differences.

The mainspring barrel assembly after slowly releasing the tension, it took 4.0 turns, which will be the amount required for reassembly.  The wheel leg that has the pin is pointing to 12:00 on both wheels in the photo and that's the leg that has the outer end of the spring attached.  It's hard to see but on the left wheel at 2:00 is a non threaded pin to the outside of the spring, i.e. the spring is inside the 3 wheel attachment posts and the pin.    There is a loose part in the center of the left wheel that connects the spring to the shaft.  The shaft has a pin that drives a slot in the "C" part which in turn has a ball ended pin that grabe the inner end of the spring.

Inspection

On of the things that needs to be looked at is the slop on the hour wheel where the black goo was.  I tried using digital verneer calipers, but it seemed I needed a hand to hold the clock down, another hand to hold the capipers and another hand to work the zero button.  So changed to using a dial gauge that has it's own stand.  Don't bother trying to rotate the dial to zero or using the little pointers (you can see one of them in the upper right of the photo) since the zero will change.  Just note the reading before and after reaching through the hole in the frame and lifting the shaft.  In this case it moves about 0.007 mils.

I'd call it a little worn, but not enough to replace anything.





More Cleaning
The first cleaning with Naphtha removed the caked on black combination of old dried up oil and metal particles, but left a white film underneath where they were.  It took a couple of cycles in the ultrasonic cleaner to get them clean.  The image at the left was made using a flat bed scanner.  These work very well for flat objects, so the shaft ends most forward are in very good focus and I can see the scratches on the bearing surfaces.

Abrasive methods of getting rid of the scratches are not a good idea since that makes the shaft diameter smaller.  The preferred method is to burnish the pivots.

Need to shop for a manual burnishing tool and pin vise that can hold a 1.8 mm pin on Monday.

This is a process of deforming the metal on the pivot bearing surface to remove scratches or worse.  Unlike abrasive methods of polishing, burnishing does not remove any metal, it just moves it.


The brunishing tool (Time Savers 17526) is just a rectangular metal bar about 2.75 x 6.5 x 180 mm.  All the corners are sharp 90 degree angles.  There's no way you can hold this in your hands and working with it without cutting  yourself.  So I have mounted it in a file holder handle. 

The Pin Vices are not the proper size. The Large Pin VIse ( TS 16537) and the Pin Vise (TS 13424) both have about the same size (hold 1.98 to 2.35 mm) collet.  But will  holdneither a  1.72 mm pinion bearing surface nor a 3.05 mm arbor shaft.

The opposite end has no round hole and is designed to hold very small diameter items.

Oils and greases can provide less friction between moving metal surfaces.   A bearing should be designed so that the loading per square inch is way below the limits of the softest metal used.  But as the bearing area increases so does the normal operating friction.  So some balance needs to be made.

If the oil dries up (lighter oils evaporate faster than thin oils and they are faster than grease) then the metal to metal friction can do damage.  The other thing that happens to a clock that's in an open atmosphere (a watch is in a sealed case) is that airborne particles find their way to the oil surface and become abrasives that grind up the bearing surfaces.  I'm guessing that a little of both effects had started working on this bearing when the clock was taken out of service.

After some web surfing I think the main spring will get Mobil 1 Synthetic Gear Lube LS 75w-90 along with the main spring arbor pivots.  This is an all synthetic oil (good for not growing stuff) that has a proven ability to reduce bearing friction.  So a thin coat on the spring should make for smooth and free movement.  This can be checked by manually winding and slowly releasing the tension and watching how the spring moves.



For the pivots Nye Clock Oil 140B.
Nye seems to be an up to date company not only conversant with the approiate specifications for oils but also developing completly new types of oils.
TAI Lubricants - U.S. distributor with kits:  Damping Grease Kit, Squeaks & Rattles Kit, Hobby Kit, Optical Coupling Kit + individual products

Mainspring Barrel

Here the spring has been looped onto the center mushroom pin and is sitting loose on the barrel half.  I'm going to practice the winding on a dry spring, then lube the spring and do it again.

Construction of Western Union Clock Battery Packs by N7CFO.
If you're interested in getting a kit to build this adapter let me know.

By putting two "D" cell batteries in one adapter the blue jumper wire in the clock becomes redundant.  This also lowers the circuit resistance.  The parts used inside the adapter were also chosen for low resistance.  The clock now winds much stronger than when two seperate "D" cell battery holders were being used.

Sound Deadending material to go into front cover.  The battleship gray cover has the material, but it's been removed from this clock.
1 each  need the thumbscrews to hold the cover on.
2 each screws to attach dial to movement

The box arrived crushed by a couple of inches.  A box this size probably gets put on the bottom of a stack that's 8 feet high so needs to be able to carry that weight, either based on the strength of the box itself or based on packing it so it can not be crushed, or both.  I suspect there's plenty of bubble wrap around the clock but not enough to really fill the box.  The rule of thumb for radios is to have at least 4" of bubble wrap on all sides that's taped tightly on each layer where the layers are at right angles to each other.  Then the box should have stuffing at the corners and sides so that the item can not move around.  Then the box can be cut down so the top just covers the item with more packing in the corners.  When the box is all tapped up and shaken there must not be any movement of the contents.  If there's any movement open the box and pack it tighter.




This is poor bubble wrapping.  There's only one layer and it's loose, not tight to the clock.  Thickness of bubble wrap varies from 0 to maybe 1 inch, should be 4".  The rest of the box was filled with plastic peanuts.  That would be OK if the item had 4" of bubble wrap and if the box was "tight" so that nothing can move inside.  If there's any movement the peanuts will crush with each shake, so it's a matter of time until they are completely gone.  With heavy items the peanuts can be completely gone after a few hundred miles of shipping.

Note that shipping insurance is for protection from loss or mishandling by the carrier, it DOES NOT cover damage caused by this type of poor packaging.  Many people don't understand that it's the responsibility of the sender to properly package the item, not the shipping company.

10 July 2007 - received today via Fedex ground.  This is not the fault of the carrier, it's unquestionably the fault of the shipper.

This clock (s/n 59132) looked like it might actually run on eBay and had the very hard to get glass installed. 

In the photo you can see a fold to the right side of the box down maybe 6".  This is because the box was not full.  Another sign of bad packaging.

Notice the packing above included a couple layers of bubble wrap.  But this one only had some small pieces of bubble, but not wrapped or taped, just stuffed in.  The main packaging material was Styrofoam.

I specifically asked about the pendulum being bolted down to the clock, since the eBay ad showed the pendulum sitting inside the clock, which was laying on it's back, and no bolts were visible.  I was reassured that the pendulum was bolted down.  As you can see it was not.  Sort of like putting a hammer in the same box with something delicate and NOT securing them so the hammer can beat up whatever else is in the box.

I've notified the seller, we'll see what happens.  This is a real shame since the glass was in one piece when it was boxed.  As far as I know you can not get this large glass anymore.

15 July 2007 no word from the seller in response to my emial and neither of my two phone calls has been returned.  On the first phone call I left a message.  On the second call the phone never answered.

18 July 2007 - the seller has called, but we're having trouble exchanging email.

Somehow she felt that Fedex would pay for the damage caused because she did not properly anchor the pendulum.  I left negative feedback for mickiecat1 but it cost me a retaliatory negative feedback.  The only one I've ever received.

The only worse packaging I've seen was when two lead acid batteries were put into one box with Styrofoam.  In that case there was no Styrofoam left, it had all been crushed.

Note I did not unwrap the clock, this is how it came out of the box, in fact it's still partially in the box.  To the right and down out of the frame there are sheets of styrofoam, not bubble wrap.

Please, if you don't know what you're doing, do not ship items that you will end up destroying.


The bob was tied down with a piece of wire when put into this box.   Only a half turn was used to join the wires and with a bob weighing 24 ounces, about the weight of a framing hammer for a big carpenter, you can be sure it will work it's way loose.

I put the "knot" back together.  There are many scuff marks around the knot where the wire fought it's way out of the bob.
The clock frame and bob have two clearance holes for a 1/4-20 bolt and using a wing nut makes it easy to install.



Fig 3 at top of page shows this clock (s/n 59132) running.  The pendulum rod was bent, but straightened out.  It has the single coil synchronizer.  No synchronizer relay.  A once per hour winding cam.  This clock winds less than a minute before Self Winding Clock Co. #1 winds at 38 minutes past the hour.  So it looks like the once per hour wind occurs at this time.

New Way to Know the Rate

17 July 2007 - When you have two of the Self Winding Clocks running in the same room one acts as an alarm clock each time it winds.  By noting the start and stop time of winding for two clocks you can check their rate.  SWCC #1 is within a few seconds per day, but #4 seems to be running about 10 seconds per hour slow.
10 Sep 2007 16"45 - 3 min 11 sec slow in 56 days or 3.4 sec/day slow.  3/100 of a turn on the rating nut is very difficult.

Self Winding Clock Co "Western Union" #1
Standard Electric Time Co Slave Clock 
Self Winding Clock Repair Class, class was taught by Mary Bess Grisham
W. Kapp Collection - Hairspring replaces pendulum - prototype
NWACC - Electric Clocks -
Construction of Western Union Clock Battery Packs - The Burlingame Cell -
Antique Clock Pendulums -
Time Savers - parts
Meadows and Passmore - parts for clocks & barometers
Two "D" cell battery adapter - PVC pipe.