Taris

Sounds like you have the dog clutch running at the wrong speed.

To maintain synchronisation in both forward and reverse drive the dog clutch has to run at the same speed as the spindle.

Adjust the clutch drive gear sizes to get it running at spindle speed.

Easy way to check is to mark both spindle and clutch drive gear at some convenient point, eg at 12 o’clock on both, and verify that both marks remain in synchronisation as the spindle is turned by hand.

Clive

Have a look at the manual for the machine, particularly the exploded diagram of the headstock, to see the relationship between the spindle and the gear below it, that is providing the input to the leadscrew (and clutch).

If it is not 1:1, there’s your problem lady.

The historical precedent for leadscrew dog clutches is lightly discussed in Cleeve’s book.

On 4 August 2025 at 07:54 JasonB Said:

 

Is it possible the dog clutch is engaging in more than one position which would be a bit like disengaging the half nuts when cutting metric on an imperial lead screw?

Yep thats what happens if the dog clutch and spindle aren’t running at the same speed to remain in rotational phase synchronisation.

The point about the single tooth is that it defines the relative rotational positions between spindle and lead screw when engaged so it always picks up the thread in phase. If the speeds are different it effectively acts as a multi tooth clutch with several possible engagement positions so phase is lost. The screw cutting gears keep the spindle and lead screw in synchronisation at whatever difference in speed is necessary to cut the thread and the dog clutch ensures the relative phase is correct on engagement.

Mathematically it’s no different to how the numbers on a screw cutting dial helps you pick up the phase so successive cuts follow each other. But a single tooth clutch is right once per revolution. So you don’t have to wait

DC31K

The Colchester manual / parts books I have aren’t very helpful at identifying the drive gears between spindle and the shaft carrying the drive gear for the change wheel train. But it’s almost certainly not a 1:1 ratio. I imagine it’s easier to engineer things on the screw-cutting drive side if it isn’t 1:1. On my P&W model B, which has a dog clutch set up as standard, arranging the 1:1 gearing appears to make things significantly more complex.

Fortunately phase shift is easy to measure and, because a gear train is involved its all whole numbers so, in principle, compensation is easy.

Clive

To put it clearly (I hope), the dog clutch must be in a constant phase or orientation with respect to the spindle. The only place for the dog clutch is where it can remain thus synchronised. Allow that to change, and the pickup from cut to cut will be lost. It means in your initial picture that the steel and brass gears at the top necessarily need to be the same number of teeth, and further must never be allowed to loose mesh with the spindle. The Colchester lead screw reverser must never be disengaged or changed either.

It irritates me that the Colchester lead screw reverser could so easily have been no more complex overall, but worked like the Hardinge design! I have toyed with a conversion to achieve that, but I am too idle to have got on and done it. However, it might be a worthwhile project for someone.

Martin Cleeve’s book is called Screwcutting in the Lathe. It is one of the Workshop Practice Series by Tee Publishing and is widely available. Well worth reading.

But you can’t do better than take up Graham Meek on his offer above of help. Straight from the horse’s mouth, so to speak.

The thing to grasp is that a single dog clutch at the spindle (or on a 1:1 synchronised gear chain) can be disengaged, then irrespective of how many revolutions the spindle makes, the synchronisation will be exactly the same when the clutch is eventually re-engaged.

However, the leadscrew must also be kept in synchronism, though this is usually achieved by moving it back to a reference start position (eg a stop) before each pass.

Any other configurations of gearing before the dog clutch spoils this. The reduction after the dog clutch merely sets the pitch of the thread, and can be any ratio and does not affect synchronisation.