Debugging my lantern pinion

[dewbane]

It's amazing how far out in the weeds I got on this. I'm still messing around with getting the mill leveled. Adjust. Settle. Get bored. Get distracted. Go back an hour later. Adjust. Settle.

When I eventually get it level in some number of hours, then I will whip out the goodies to check the tram, and I'm still betting $5.00 it will be fine.

 

[johnnielsen]

Yeah, it's probably like picometers or whatever is smaller than picometers. I have snapped carbide stuff that was like 1/2" in diameter, and this is about 0.040". However, I drilled four holes uneventfully. Every time the bit snapped, it always did so upon entering the lower flange. Maybe I just got over-confident after four holes, and turned the fine quill feed knob faster. Also, I used high gear and the highest speed possible, which was something like 1800 RPM. That's probably half what it should be.

Yeah. I poinked some numbers into a speed and feed calculator, and this drill should be running at 19,000 RPM. Okay, that's not happening today.

I still haven't checked the tram. Got side tracked making a fancy holder for my ridiculously fancy center bashing rod. I painted it to match the lathe, which prompted me to realize my paint can is almost empty, and go looking for another one. Hammerite Deep Green used to be $80 a gallon, and now it's over $300 a gallon. OUCH!

This statement could be telling (Every time the bit snapped, it always did so upon entering the lower flange.)
Perhaps you have a miniscule amount of vertical end play in your spindle so the spindle drops a couple few thou when you drill though the first piece and start to drill the second. The drill would buckle a bit before starting to cut causing breakage.

 

[dewbane]

This statement could be telling (Every time the bit snapped, it always did so upon entering the lower flange.)

You could be right. The farther the bit extends, the worse the play gets. It's traveling pretty far to reach the lower flange. It's 0.200" top to bottom, after going through 0.060" + 0.080" + air gap. It's close to 3/8". I should check the quill for slop, and see if there is any adjustment I can make to tighten things up. I do have a suspicion that something could be off.

At least things should be better on the lathe side of this operation. I got the ISCAR MT4 ER32 and had to face my first problem. It was set up for through coolant, and while it would be useful to be able to extend a small diameter part all the way through, it would be hard to knock out without a beater backer. I ordered an off-the-shelf beater backer from McMaster-Carr, and when it came in, I got an M6 instead of an M16, because I'm stupid.

So I makeded one.



I cheated and just smoothed the end on a belt grinder. Then I whacked it in, and set up the tenths indicator. This is WAY better than the Chinese one I had before. It's still a lot of TIR compared to what I could achieve in theory, but I went from 0.0160" to something like 0.0004". Sometimes, spending money on good stuff actually rewards you.

I did enhance that image a little in the GIMP, but I actually recorded several Super Slo-Mo videos with my phone, and the worst I saw was 0.0008" TIR, and it was only that bad because the needle reached that far once in 18 seconds.

For a Chinese lathe and Chinese collets, I'm going to call this a gigantic improvement, and a huge win. I could probably get it even better if I shelled out for Rego-Fix or Techniks collets, but I might not do any better at all. It's a piece of "tight tolerance" O1 drill rod from McMaster-Porn, and it's just a chunk lopped off the end in a bandsaw. It's obviously not perfect, because when I go near the cut end, the needle goes all to hell. My lathe may not be the most accurate, but it's not THAT bad.

Anyway, it's the lathe I own. I can't afford to shell out for a Taig or a Sherline or whatever right now. I have to do this with what I have, or I can't afford to do this at all. They used to make clocks by hand filing the teeth by eyeball, and I think it will all be totally fine in the end. I think I can, I think I can, I think I can.

No, I still don't have the damn mill leveled yet. I'm close. I'm taking a break to let the levels settle. Yeah, the levels. It has nothing to do with the scotch I'm drinking.

Public Service Announcement: DO NOT OPERATE MACHINERY WHILE UNDER THE AFFLUENCE OF INKAHOL. YOU COULD END UP CALLING AN AMBULANCE TO TAKE YOU TO THE ER TO HAVE PARTS SEWN BACK ON, BECAUSE YOU WERE TOO DRUNK TO DRIVE THERE YOURSELF, YOU ALCOHOLIC LOSER!

But tweaking some wrenches by hand, and checking some bubbles. That's fine. Enjoy your vacation!

 

[dewbane]

Here is the video. Maybe I am telling myself what I want to see. How do you read this?

 

[jwmelvin]

Here is the video. Maybe I am telling myself what I want to see. How do you read this?

Yes about half a thousandth. There seems to be some motion that is not once-per-rev, which is indicative of bearing runout (because the balls take more than one revolution to come around).

To eliminate the collet and stock influence, you could indicate the taper inside the ER adapter. And compare that to the spindle taper.

 

[dewbane]

To eliminate the collet and stock influence, you could indicate the taper inside the ER adapter. And compare that to the spindle taper.

Hmmm. I didn't check the spindle taper with the tenths indicator, but it was pretty obviously barely moving a thousandths needle. Seemed fine Surprisingly fine, considering its Chinese roots.

Collet and stock errors are part of making real parts in the real world. I was getting 0.016" with a similar setup, and I got that down to at worst 0.0008". Like you said, there is some feature on the part that is worse than everything else. It could be an eyebrow or nose hair, or maybe the gaseous deposition of chemicals from a strong unicorn fart. Realistically, even if it's a lousy 0.0008" TIR, that still beats the pants off of 0.016" TIR.

My dog just farted. That probably explains the discrepancy.

 

[jwmelvin]

Realistically, even if it's a lousy 0.0008" TIR, that still beats the pants off of 0.016" TIR.

I couldn’t agree more - great improvement. It can be interesting to me to better understand where deviations come from, especially when the indicator is set up.

 

[dewbane]

By the way, I owe someone $5.00

I finally got set up to check the tram on the mill. I assumed it had to be fine, because I was very thorough when I dialed it in, and every check I've done since then just confirmed a job well done.

Until now. Nope. This is not even close to being in tram.

 

[homebrewed]

With all the problems you've run into, I think you could have made a conventional pinion gear by now. Are lantern pinions (somehow) supposed to be easier to make? Not being familair with horology I'm probably asking a dumb question but curious minds want to know .

 

[dewbane]

With all the problems you've run into, I think you could have made a conventional pinion gear by now. Are lantern pinions (somehow) supposed to be easier to make? Not being familair with horology I'm probably asking a dumb question but curious minds want to know .

I could have just bought one six hundred times over, or just bought an entire complete, working clock movement, but that's why I started with this part. I knew it would be the most challenging thing for me to make, and if I could make one, I could make all the rest of the parts.

I think the plan does use them, because they're supposed to be easier to make. Clock pinions have a weird tooth profile, and the small ones are apparently the hardest to cut. I don't know about any of that, really, I'm just trying to follow the print in the book.

It's funny how the one thing I should have checked first has turned out to bite me square in the butt. Oh, I'm SURE the tram is good. I'm so sure, I'm not even going to check that!

I was overdue to move the milling vise anyway. It left a mark on the table, although there was no serious, heart-stopping corrosion. I pulled everything off, and I stoned the table. It really needed it. Lots of little random spots were proud of the surface a little. Now I need to make a new centering setup for the rotary table, that fits into an ER32 collet, and I need to indicate the vise back on. Oh, and fix the tram before I do any of that. It never ends.

This has been more fun than my last vacation. I spent 10 days working 12 hour days to rip out and rebuild one wall of my shower.