Why your Nonus could be stopping after 2 or 3 cycles.
  • The crucial phase with the Nonus is when the drop weight hits the pendulum pallet.

    In order for it to work correctly the first point of contact when the drop weight rollers hit the pallet has to be on the top of the pallet.

    image

    The pendulum continues to advance slighty with the rollers pressing down on the pallet then gravity kicks in and pulls the pendulum into its return swing, this is when the rollers slip off pallet top and run down the pallet impulse face giving the pendulum the necessary momentum for the next 60 seconds.



    SO WHAT CAN GO WRONG ?


    Instead of the first contact point being pallet top it can be pallet impulse face.

    image

    What happens here is that the pendulum is still continuing its swing but is braked by the rollers hitting the impulse face, the pendulum continues its swing but loses a lot of its momentum by lifting the drop weight slightly before it starts on its return swing.

    image

    Then the drop weight impulses the pendulum as normal, but as the drop weights impulse force is only enough to keep the pendulum swinging it can not compensate for the lost momentum, and the Nonus stops.

    You can see this happening in the following video.

    NONUS VIDEO

    It is in slow motion and if you look closely, and I mean really closely as it is only a couples of frames, you will see the drive weight hit the impulse face and then be pushed upwards slightly by the pendulums momentum before completing its impulse cycle.

    SO WHY DOES IT GO WRONG ?

    One cause could be a lack of pendulum momentum, if the pendulum is moving too slow at swing extreme it will not have the time to move under the rollers after the triggering of the drop weight, lack of momentum can of course come from the count wheel absorbing too much force over the 60 second period.

    Another cause is that the drop weight is being triggered too soon, the reason for this is that the trigger arm is not correctly aligned to the count wheel, easily done as the alignment is a bit of a hair trigger.
    If it is only a small amount the trigger arm faces as well as the pawl face could be sanded back so that they connect slightly later, but be cautious as too much sanding reduces the length of both parts, which could result in them not contacting at all.
    Another method would be to increase the length of the pallet, by simply gluing on a thin piece ply/wood onto the impulse face.

    image

    The best but slightly more adventurous method would be, with a fine saw, cut through the spacer (10B) between trigger arm and count wheel, and re glue with the trigger arm turned slightly in an anticlockwise direction, causing the drop weight to be triggered slightly later.



  • I have that timing perfect (as far as I can tell). I cut and re-glued the trigger mechanism twice. Still not enough energy to get the clock to keep moving. I tried adding weight to the stop mechanism (as someone mentioned) and even tried using more drive weight. Still it fails. Someone must have gotten this to work, right? What did you do?
  • Hi Brian, I had four things to work through.

    1) the pendulum and count wheel were one issue. I re-polished the sliding surfaces with the highest sandpaper (1200) and 0000 steel wool. I then used a dab of paste wax to further polish the wood (don't know if this was needed). I also carefully reviewed the perpendicularity of the pendulum, and the alignment of the impulse pallet to the drop weight. Without any other mechanism on the frame, my pendulum continues to move the count wheel for over 4 minutes. Early on I had tried adding weight to the drop-weight body. Since doing 1,2, and 3, I've removed that weight without impact.

    2) I re-polished the arbor holes (0000 steel wool) of the gears, and made sure the small gear surfaces were smooth (the 6- and 8-tooth gears on the drop weight and count gear arbors).

    3) I needed a couple of extra polished washers for spacing. I did everything by hand, and perhaps over-sanded some parts, particularly the gears' arbor holes. Some gears do not rotate perfectly parallel and brush each other—the most difficult for me to find was the count gear rubbing on the pawl arm that prevents backward movement of the count wheel (subtle, about 1 cm out of the whole rotation).

    4) The amount of weight. I'm using stones in an antique Japanese kettle for an 8-day nonus. For simplicity, I started with a 5lb ballast ball on a 4-day setup. After fixing the #3 problems, it worked fine. Switching to 8-day setup, I started at 4.4kg (a bit under double the 5-lb that worked for the 4-day setup). There seemed to be a problem with power transfer from the drive arbor to the drop-weight arbor. More weight didn't help. I temporarily made it a 4-day again while I experimented with 1&2 above. Worked fine (actually, better after 1 & 2)! I played with the weight amount and found it would run really well (good quick response of the stop arms) at only 1.61kg. Back to 8-day set-up again, it won't run at even 4kg, but has been running a week at 3.52kg. Too much weight stressing the system and adding friction?

    Hope something in there is useful!
    Good luck,
    Antony
  • PS:
    For the record, just saw the discussion on the amount of sanding today—and can see rougher surfaces in some of Dave's videos. Minimal sanding, but the clocks run fine, so he must be right (and he knows his clocks). I don't regret the sanding I did, but sure slowed me down! (180, 300, 600, 1200, steel wool progression. Did I get that from earlier conversations in the blog, or from my Dad who did antique restorations? :-)

    When you *do* sand, be careful of the effects on the edges, and from my perspective (#3 in my earlier post) especially the gear's arbor holes. If you have a drill press, a friend recommended just using the perfect-sized drill bit, and running it for a few minutes in the hole to do the polishing.
    Antony
  • I always preach if your clock is not working and everything is assembled correctly (I get a lot of pictures with back to front anchors) then all that is left is friction.
    Apart from the Duodecimus all of my clocks have involute teeth.
    image
    As you can see the contact point of the teeth moves in relation to their rotation so they never really slide against each other but rather maintain a constant pressure at a constantly moving point.
    Which is why sanding the teeth to such a high level is certainly not counter productive but also not necessary, I do indeed just flick over the teeth with a 180 grade so they are optically clean, no more than that.
    Also when I have been looking for friction problems then I have never started polishing the teeth.

    Now the arbors and bearings are a whole different kettle of fish and as far as I am concerned you can never over polish or apply too much attention, if you think you have polished enough then do some more :-)

    Anthony, thanks a lot for your break down, real hands on from a clock builder, always welcome.

    Love the idea of using an antique Japanese kettle for a drive weight, you have so got to post a photo !
    Involute_wheel.gif
    256 x 256 - 326K