Home Articles Forums Gallery Marketplace Videos

WoodCentral Forums

Knowledge You Can Build On — Est. 1998

Personalized Wooden Watches, Engraved Free

Precision question

Posts

Precision question

#1

Precision question

Ruth Niles

>You may have read where Ed Moore bought one of my mandrels that didn't run true. I sent him a replacement, it ran true, he sent the other back. When I put it in my Jet Mini, it didn't run true... I cleaned out the spindle before installing it....but when I put it in my General, it did run true.

I don't have precision instruments, I just put the tool rest tight to the threads and used my mag glasses then turned the lathe at super slow speed.

How is this possible? When a metal part is machined on a metal lathe, doesn't it have to be true? Aren't all headstock spindles trued?

Re: Precision question

#2

Re: Precision question

john lucas

>Ruth I'm still working with that question. When I cut threads on my Baxter theader I turn the wood and it's running true to within .003". I put the threader on and it's off as much as .015". When your only cutting a thread .035 inch that seemed like a lot. I took it back to Vic and he checked it and it his threader is accurate to a few ten thousands. He tried to explain how my powermatic is off and I guess I didn't quite grasp it. I'm not the machinist he is. I measure the spindle run out and it's within a thousandth. I've even tried checking run out with a piece of steel in the jaws of my chuck. When I cut threads on my Jet it's less than .070" off and on my NOva Comet it's inbetween the two. So I'm still confused. It runs perfect on the Jet mini he uses in the shop. We tested the threader on his metal lathe which I guess is simply a lot more accurate than our wood lathes.

So to answer your question, damned if I know.

Re: Precision question

#3

Steven Antonucci

Errors in their Measurement tools is my guess

Steve Antonucci

>Even the tightest micrometer will still have error. When person X checks work Y with their tool, it runs true. Person Z who made the lathe check the lathe with another measurement tool. If they swapped measurement tools, my best guess is that they would both find they were no longer running true.

I forget the exact term, but there is a way to calculate the "certainty" of precision (engineering mathematics). If I have a tool that measures out to .001, but is precise to 2 digits, I can see tolerances between .001 and .009 the same way, since I have no certainty around the last digit.

That's why I like wood. .001 makes no sense at all, since wood probably moves as much as .125! :-)

Steve

Re: Precision question

#4

Where are you measuring from?

Ed Moore in SW Virginia

>My assessment of the faulty mandrel was that the threaded portion was "bent" relative to the axis of the lathe and the Morse taper. I felt like a whack with a hammer could straighten it. I didn't try that, but that was the way it looked and the way it performed when I put a piece of wood on it.

I put the new mandrel in my Robust and it ran true visually and with the finger against the threads as it rotated with the lathe on. In every way it appeared to be perfect. I then put the faulty mandrel in and did the same thing. It was obvious visually and with the finger against the threads that it was not aligned properly. Previous usage had already identified it as problematic.

So, if the supposedly faulty mandrel ran true in the General, then I would worry about the General. Make sure no wood chip was in the spindle.

But, in any case, I thank Ruth for the new mandrel. We all know that she is a sweetie.

Ed

Re: Precision question

#5

Re: Precision question

Brian Hahn in SW Wisconsin

>There are two main sources of error: the mandrel and the lathe taper. The two can combine to cancel each other or they can add to each other.

Simple test: put the mandrel in the lathe and check the run-out. Remove the mandrel, rotate 90�, put it back in and check the run-out. Repeat at 180� and 270�.

If the run-out is the same regardless of the orientation of the mandrel, then the mandrel is at fault.

If the run-out changes, then the lathe is in error by 1/2 of the amount of change. The rest of the error is in the mandrel.

(It would be more accurate to measure the run-out at finer steps, but 90� will give you a rough handle on the situation.)

Re: Precision question

#6

To clarify "finger test" a little ...

Ed Moore in SW Virginia

>If you put your finger against a rotating spindle you can either sense wobble or you can't. My point is simple -- if you can sense wobble with your finger then it is there. You may not quantify it, but it will fall into one of two categories, acceptable or not acceptable.

Ed

Re: Precision question

#7

Re: Precision question

Barbara Gill, Remlik VA

>Ruth I don't have a clue. Why don't you take it to the machine shop.

Re: Precision question

#8

Re: Errors in their Measurement tools is my guess

Bob Macgregor

>I'm not sure whether you are referring to "significant digits" or what I call "false accuracy". Digital devices are especially problematic in this regard since they can display multi-decimal readings that exceed the significant digits constrained by the true precision of the measuring device.

When using a slide rule to do a calculation, it is frustrating how rough the answer seems to be; however a 12-decimal accuracy from an electronic calculator which is based on input numbers with single-digit accuracy (significance) leads modern-day students to routinely overstate the precision of their results. Of course, most of the time overstating calculated or measured precision has no practical consequence and, presumably, engineers and scientists are trained not to make this mistake!

BOB

Re: Precision question

#9

Some further points

Brian Hahn in SW Wisconsin

>These are obvious:

Both the taper and the socket must be clean and free from burrs.

The taper must be seated firmly into the socket. A quick, sharp flick the of the wrist is usually adequate. A light push is not adequate.

Also, not all sockets are created equal. On my Jet mini, the machined section of the socket extends in only 1/2 the way (1-1/4"). The remaining 1-1/4" is the the bore hole. This means that only the front half of the taper is supported - the back half is floating in free space. This can cause problems.

Re: Precision question

#10

I doubt it.

Brian Hahn in SW Wisconsin

>Machinist tools are quite accurate. 0.001" accuracy is on the sloppy end. Precision tools are accurate to 0.0001" or better.

Re: Precision question

#11

Re:We had a special tool for such....

Jamie D

>... alignment mismatches in the Army- 16 oz. impact alignment tool,handled, 1 ea. OD(Olive Drab!) If it doesn't work , the 22 oz. model was requisitioned! The alignment variation in 2 spindles is curious, and as Ed says, absolutely clean mandrel(without nicks and burrs) and morse taper is usually the cure.

Re: Precision question

#12

False accuracy (OT)

Alex Yeilding

>"what I call "false accuracy". Digital devices are especially problematic in this regard since they can display multi-decimal readings that exceed the significant digits constrained by the true precision of the measuring device."

Okay, you hit a pet peeve of mine! The one that really gets my goat is my bathroom scale that shows weight to 0.1 lbs, and shows exactly the same weight before and after I remove shoes and pants with pockets full of change and keys!! I haven't actually measured, but I think it is accurate to about 1KG.

"12-decimal accuracy from an electronic calculator which is based on input numbers with single-digit accuracy (significance) leads modern-day students to routinely overstate the precision of their results."

Oh, I don't think we can blame the electronic calculator for that! It might make it easier, but I still had a hard time learning that lesson with paper and pencil.

"Of course, most of the time overstating calculated or measured precision has no practical consequence"

I tried making that argument the first time I had a physics problem marked incorrect for showing the answer to four significant digits when one of my measurement devices was accurate only to two digits. I lost my appeal!

Now, don't you feel bad for raising my blood pressure to 148.7385/94.327?

Re: Precision question

#13

Re: Precision question

Alex Yeilding

>"If the run-out is the same regardless of the orientation of the mandrel, then the mandrel is at fault.

If the run-out changes, then the lathe is in error by 1/2 of the amount of change. The rest of the error is in the mandrel. "

I agree, given the assumption (probably valid) that the lathe is more accurate than the mandrel. But I think the same effect could be observed if the lathe was out by more than the mandrel.

Re: Precision question

#14

Steven Antonucci

Significant digits is the term I had forgotten-THX

Steve Antonucci

>

Re: Precision question

#15

Re: Some further points.....aha!

Ruth Niles

>Brian, that would explain why it did not run true in the Jet Mini but did in the General. But.......ok, that doesn't help, does it!

I guess I'm wondering how metal turned stuff made on a metal lathe cannot run true. Wood I understand, but metal?

Barbara, if I take it to a machine shop, won't that just confuse me more because now it's another machine's opinion?

Re: Precision question

#16

Re: Precision question

paul waddle

>Hi John,you bring up an interesting question. Do you keep the box or lid in the same chuck and just unscrew it form the lathe spindle and screw it on the Baxter and you're off by .015?

Are at least two places where the error can be induced; 1. Shaft on the Baxter is not lined with lathe shaft. (this where I believe the error is) 2. The threads on the Baxter may not be perfectally cut true.The first item maybe at how its bolted to the lathe bed,the 2nd part is not that easy to chk. You can dial indicate at different sections of the Baxter thread,that might tell you where the error is.

Here's what I would do: make a centering shaft(6-10" long) and put a center in it. Put the dead center in the head stock and put the Baxter threader on the lathe. Put the chuck on the Baxter and ease the shaft into the chuck,running on the dead center at one end and in the Baxter at the other. Tighten chuck only so much as the rod can turn.

Indicate at both ends;should run true. Now tighten chuck,if every thing is not aligned,you will get an error.Tighten and loose the the bolt to the bed... Bet there's the error pt.pw

Re: Precision question

#17

Re: Some further points.....aha!

Dick Coers

>There are many conditions on the metal lathe that can affect accuracy. Scroll chucks are less accurate than collets, dull cutting tools can cause excess cutting pressures and metal deflection (also off center cutting if the tool isn't sharpened correctly), burrs and metal chips in the holding device, excessive speed or feeds of the cutting tools, etc......In my opinion, there are only two differences between turning wood and metal; one is the wood movement (from atmospheric humidity, stresses released that were imposed from the wood drying process, or improperly dried wood), and tooling following the grain of the wood. After that, you can get inaccuracies when turning either.

Re: Precision question

#18

That definitely was a simplification.

Brian Hahn in SW Wisconsin

>If I were trouble-shooting the problem, I'd measure the run-out as a function of both the spindle and the mandrel orientation and I'd use much smaller angular increments, more like 30� to start with.

The headstock run-out can be checked independently with a simple drive center at several orientations. Indeed, I was assuming it was a small factor.

Re: Precision question

#19

Ahh, the gentle persuasion device!

Brian Hahn in SW Wisconsin

>According to some of the mechanical engineers I worked with, it was one of the three essential tools for a mechanical engineer. (The mechanical technicians on the other hand had a tool chest and roll-around filled with their essential tools.)

I've occasionally had to resort to the 15-pound persuader.

Re: Precision question

#20

Re: Some further points.....aha!

Brian Hahn in SW Wisconsin

>Dick's list of potential machining errors is good... but add to that operator error.

A machine shop can measure the mandrel and tell you where the problem lies. They have all the tools to do it quickly and accurately.

Re: Precision question

#21

Keith Tompkins

Re: Precision question

keith tompkins

>In most cases, a bit of run-out won't matter. For example, take a chuck with an obvious wobble... mount a blank in the chuck,and turn a bowl. Will the bowl wobble or be out of round? No, although you'll have a tough time putting it back in the chuck the same way. Put a blank in the mandrel, and turn it. Will the blank wobble or be out of round? No, the runout will be compensated when you turn the piece. The only problem you will notice with the finished piece is that the base will be slightly out of square...you will probably never see it by eye. As long as the spindle isn't bent, a bit of runout won't matter all that much. If you want to see this for yourself, take a faceplate, mount a blank, then drive a wedge under one edge, tighten the screws, and start the lathe. It will wobble like crazy. Face off the blank...then, with the lathe running, touch a pencil to turned face; it will touch evenly all the way around. Measure with a dial indicator...it will be right on. For most work, runout won't be much of a big deal. Problems will develop if the piece needs to be swapped end-for-end in the chuck or mandrel, the amount of runout may be multiplied by two. In that case , accurate work cannot be produced.