It's not the recoil that drives the pendulum. In fact it's the pendulum that drives the recoil! Some designs of escapement have no recoil at all. Which of J.W.'s designs is your clock?

If the spring is the wrong width have you also checked it's thickness? That is more likely to give lack of driving power if wrong.

What have you used for lubrication? Are the pallet faces lubricated? A common problem is too much oil on the pallet faces.

Russell

Deleted double post…

Edited By John Haine on 09/08/2018 12:07:24

[ Update ]

For convenient reference, here is a frame-grab from that excellent Clickspring video:

.

MichaelG.

Lack of recoil will not stop the clock from running, many escapements are designed without recoil at all and some with a minimum amount. In the case of the anchor escapement such as yours it would however cause poor time keeping. Given that your clock is only running for the period of the pendulums natural decay I would suggest, as has already been touched on, that the problem is lack of impulse. If the movement is free running and there's adequate side play on the arbors you may have to re-check the pallet geometry. If, as already commented on, the pallet face is radial to the pallet arbor then no impulse will be given. Filing metal from the pallet may facilitate some degree of impulse but it will be very small, that's why clocks with worn out anchor pallets are only capable of small and unreliable pendulum amplitude. It will also create excessive drop. Although it's necessary for reducing friction and wear in a working movement, if a new movement will not work without oil on the pallets there is an underlying problem. Arbitrarily attacking the pallets with a file is likely to finish you up in the madhouse so I think you should carefully study the pallets you've made and consider making a replacement set.

How smooth are the faces of the escape wheel teeth? I assume everything has been polished. Look at the points of the escape wheel teeth, they shouldn´t be too sharp. However carefully you cut the escape wheel you might have to spin the wheel in the lathe and apply a fine file to the tips of the teeth.

Before you start to cut metal again, has this design got an eccentric pivot for the arbor holding the anchor? If so, have you tried adjusting the amount of engagement?

If you have made the little depthing tool that John Wilding advocates, check the anchor. In MichaelG´s frame grab the distance between the tips of the exit and entry pallets is five and a half teeth.

Having once assembled an escapement mirror inverted, don´t ask me how, corrected it and fiddled around bending anchors and filing away until there was no alternative but to make a new anchor,I can give 1 vote for Michael m´s comment that "Arbitrarily attacking the pallets with a file is likely to finish you up in the madhouse "….

PS. Bazyle -Very nice explanation.

Whilst appreciating that all advice is well intentioned I do feel that some comments are ether gratuitous or ill-considered.

The OP has already discovered that oil won't make a faulty clock run. It wouldn't, no matter how administered.

The OP has told us that drop is 1/3 of the tooth thickness. Assuming a tooth thickness of .015" which would be reasonable then the drop would be some .005". All else being OK the clock would run. So shortening the pallet arbor and escape wheel centres may reduce an already tiny drop but won't alter the impulse faces. That's not the function of that adjustment. (Assuming it exists on that clock). Agreed it is done by poor repairers in an attempt to compensate for worn pallets but is a poor bodge and affects the overall escapement action.

The OP has told us that the pallets and wheel teeth have been brought to a mirror finish. Polishing the wheel teeth may not be advisable on a first build given the risk of affecting the tooth profile for minimum benefit on an anchor escapement.

Filing the escape wheel teeth in the lathe is I think risky advice for someone new to clock making. The teeth are delicate. Does he have a suitable file? I appreciate it's done but normally to correct an inaccurate or damaged escape wheel, not to try and compensate for an innacurate pallet. Given his small drop it's unlikely there's a proud tooth or it would manifest itself by catching on the pallet and he'd be aware of a short tooth skipping through. If he succeeds in accurately reducing the diameter of the escape wheel then he's upset the whole escapement geometry and a new made perfect pallet may not work. He's now introduced a second variable. What's he to do then? Bend the pallet arbor?

Anchor escapements are incredibly tolerant of wear and abuse and the fact that this clock fails to make any attempt to run suggests that there is a radical fault in the manufacture.

I concur with John Haine who has summed it up thus "I fear the solution most likely to succeed is to make a new anchor"

You could initially of course make one more easily from aluminium or brass as a try out before going for carbon steel.

Dick, those points have been covered. There's a direct observation that there's no recoil which would be consistent with at least one of not both impulse faces being on the wrong centre.

Good luck!

Russell

+1 here, I've just read the complete thread where I started out with it all whistling over my head and ending up with knowing a little about how a clock works, I recently bought a 1940's ish pendulum wall clock which quietly ticks away in the hall and I admit to watching it trying to see what makes it tick.

Martin P

Oh dear. Recoil happens as the anchor drives the escape wheel backwards just after the impulse. If there is a seconds hand attached to the escape wheel arbor you see it move backwards a little just after it jumps forward. The impulse is delivered at that jump forward. The train is, somewhere, absorbing energy from the pendulum – how much goes where depends on many factors such as play in the pivots and various other things. Some might even slightly raise the driving weight. With luck most of the recoil energy will be delivered back to the pendulum, but if things are worn and or inaccurate maybe not. The level of escapement error is also determined by the recoil (amongst other things) so wear etc that changes the recoil changes timekeeping.

The Graham deadbeat escapement has no recoil as its name implies, so no energy is extracted from the pendulum by recoil.

Harrison's Grasshopper is very good at returning the energy especially if it has a remontoire, and recoil is an important aspect of his clocks' operation. Many gravity escapements inherently have recoil, in that the pendulum picks up a weight before it reaches maximum swing, "recoiling" it, then the weight is taken off the pendulum at a lower level by a stop and is then returned to the higher level by some sort of mechanism.