Re: Metric Thread Dial
I don't think I have the right words to explain.
Too long, don't read warning....
The thread dial indicates the phase of the cutter with respect to a thread being cut.
With the half nuts unlocked moving the carriage or rotating the
spindle changes the phase.
With the inch lead screw cutting inch threads the spindle the
ratio between the spindle and the lead screw is simple. More
importantly, you can lock the half nuts anywhere the appropriate
mark lines up and be in sync with the thread you are cutting.
About every 1/8th inch or so. I think. Need to check. In any event
-- places you can lock the half nuts are separated by a fraction
of a lead screw turn. HTRL has some notes on which marks you can
use for which threads per inch (multiples of 2, 4 or 8tpi make a
difference). The marks where you can engage the half nuts are the
ones where the spindle has rotated back around to where you can
pick up the thread.
Add the metric transposing gears, or one of the approximations, and the ratios get really messy. So messy that there is no simple, usable, relationship between the phase of the spindle and the thread dial indication. Spindle rotations don't divide evenly into lead screw rotations.
There are three solutions to cutting metric gears that I have heard of:
1) Don't disengage the half nuts while cutting threads. When done with a pass run the lathe backwards to return the carriage to make another pass. This is the most commonly suggested method. Somebody is sure to pipe up that they prefer doing this with inch threads even if they have a thread dial. This definitely works.
2) I have heard (but not tried) that starting with the carriage at the same position (perhaps using a carriage stop to the right of the carriage, or the tail stock as a carriage stop) AND always using the same mark on the dial (perhaps the 0 mark) will work. This makes sense to me mathematically - I won't bore you with the details)
3) I have also heard (but not tried) of a method where you turn off the lathe at the end of a pass, note the thread dial position, disengage the half nuts, move the carriage right to a position where the thread dial is the same as where it was when the half nuts were released and engage there. Then turn the lathe back on. I think there may be more to it than that -- it seems at least semi plausible to me.
The safe thing is to go with #1.
I'm amazed if anybody has managed to follow all of this. The math
of this fascinates me, but if I start blathering about
Stern-Brocot trees, prime factors, smooth numbers, least common
multiples and largest common factors I'll deserve to be banned
from this list.
On 1/3/21 10:56 AM, john kling via groups.io wrote: