Last modified: 2023-12-20 20:27:20
< 2023-12-19 2023-12-21 >I've parted off 3 discs, will try and cut teeth on them using the superglue arbor.
I think having a second collet chuck would be helpful, then I could keep one on the lathe and one on the 4th axis. Would speed things up a lot.
I had a look in the feeds and speeds calculator to work out what speeds and feeds I should be using. I couldn't quite get it set up properly, because it doesn't have a setting for silver steel. But I think it's saying a HSS tool of broadly this shape wants to be spinning at only 1600 rpm, so I'm using 1600 instead of 2200 rpm, not convinced it will make a lot of difference though.
One advantage of cutting teeth on a disc instead of a bar is it easier to see the shape that is being formed.
I missed the centre-line pretty badly again, must try harder.
It completed the cut without damaging the cutter, excellent.
Lol, the collet nut wasn't even tight. Remarkable that this worked as well as it did.
These teeth are starting to look good, except for the fact that the cutter was not centered:
Also that the drilled hole is too big. OK, next one: more careful with the centre drill, and align the cutter on centre.
Re-centering the cutter, I found it to be off by 0.64mm, which is about 2 entire teeth! Though note Z backlash is huge, so could easily be +/-0.3mm from that, maybe even more.
I have another possible explanation for the angled teeth: it looks like each cut bends over the adjacent tooth a little bit. The gap between the teeth is largest on the most recent cut, and then that gap closes up slightly during the next cut. Is the brass too soft? Is the cutter too blunt? Is the cutter fitted at an angle, so that the side relief is actually a "side wedge"?
Also it looks like the positional accuracy of the rotary axis is not good enough, as some teeth seem smaller than the others.
For the third disc I'm going to hammer it flat to see if that can work-harden it at all.
Actually the only tooth that looks notably smaller is the one formed by the combination of the very first and very last cut. Does that lend credence to the theory that the teeth are being bent over by the cutting?
Second one is better than the first:
Straighter teeth, smaller bore.
I should make something that stands at the centre height of the 4th axis, so that I can set it on the table and compare it to the tip of the cutter. Actually I can use the little model marking gauge for this! Good idea. That worked really well.
I opened up a couple of the pocket watches from the box of scrap watch stuff, and discovered that the teeth on those gears are actually bigger than the ones I'm making now.
Three gears made, cutter still looks good.
I took the pocket watch apart as far as I could conveniently go (didn't take the balance spring off, didn't undo anything that was riveted).
I cleaned everything in alcohol and put it back together with oil on the pivots.
It still doesn't run, not sure what the problem is. I can't tell whether the balance has conical pivots, or if the pivots are actually broken off.
This website calls them "cone pivots": https://www.weclarkwatchrepairs.co.uk/blog/types-of-pivot/
The pivots turn in cups which are made into the end of a screw. Over a number of years this type of pivot wears out – and as it does so, the area of contact at the end of the pivot increases, which increases friction and causes even more wear. The timepiece will eventually stop in certain positions as a result.
And there is indeed a screw I can adjust to set the end shake. So maybe the problem is that the cups are too worn?
Maybe I should use cone pivots in my watch.
I reckon I didn't put enough oil in the mainspring barrel. But I really don't want to open it up again because it is so hard to put back together. Also I'd have to let the power down first.
I opened it up a bit without releasing the spring and put some more oil in, maybe it helped a bit.
I've left it running on the bench, and it is running but it has amplitude only about 90 degrees so there is obviously something wrong. I noticed the pallet fork pivot is slightly loose, not sure if that's the main problem or not.
Tomorrow I should take some measurements from the pocket watch movement:
Because I am thinking some more about using a more conventional barrel in my watch. You could still have the "bezel" have internal teeth for winding the barrel, and still have the movement held in the case with a ratchet so that you set the time by turning anti-clockwise. We don't really need any stop-work, I don't care about that.
So then you could imagine lifting the entire mainspring barrel out of the pocket watch and putting it in my watch. You could let the mainspring barrel do the same number of complete turns, but let that be 12 hours of runtime instead of 40 or whatever. Then there's less gearing-down required for the train to the escapement, and just 1 gear to drive the hand at the centre. In principle the gear for the hand could be trapped between the top of the movement and the bottom of the dial, like a cannon pinion (?), so it could pass over the top of other pivots so it doesn't even really take up any space inside the movement. It just adds height. Or the top plate could be stepped, with a hole, maybe that is better. That way it can mesh with a gear on the barrel arbor (which can have 2 gears if it helps) that is "in between" the barrel arbor's pivots, even though there are other pivots underneath the hand gear.
So: