Getting a lathe bed plumb level.

[ctwo]

0.020" / 48" = 417 PPM. (Your eyes plus plumb bob, needle, machinist scale, and 4 foot vertical)

oldvan, your analysis is absolutely Brilliant as much as it is Simpler...

Here is one reason why I inquired about setting (clamping) the plumb bob mast on the carriage and tailstock itself. Isn't that more important anyway rather than measuring localized imperfections in the bedways (which are assumed to be perfect)? Isn't the only thing that really matters how the carriage and tail stock sit on the ways? (of course I say that because my bedways are a bit beat and worn, but the concept would seem to apply equally regardless).

Another consideration is, rather than use my eye to measure the deviation, how about a DTI (of course with sufficient weight to overcome the spring)?

I now bet I can best that $800 level with items I have laying around the shop.

[Bill_Cook]

I now bet I can best that $800 level with items I have laying around the shop.

I think the next developmental step in a closed wallet super precision level would be to use a solid one piece pendulum with a ball on the bottom end, an old micrometer, and an ohmmeter.

The mike would be electrically connected to the base through the ohmmeter. For those of us who have let old leaky batteries take out the electrical tester, a low voltage, low amperage battery/bulb circuit would do - a little less elegantly.

The mike frame wouldn’t be needed, but might be handy for mounting. Don’t think a carbide tip would be good.

Since measurements would be comparative and under equal conditions, a tenths vernier could be trusted – possibly even split further.

With a 4 foot pendulum, the unit would definitely be in the realm of careful temperature control.

--

The manual for my old WW II Sheldon lathe says to use the tops of the "V's" to level the bed. On the other hand, the tail end of the bed has a spherical mount.

[tornitore45]

The mike would be electrically connected to the base through the ohmmeter. For those of us who have let old leaky batteries take out the electrical tester, a low voltage, low amperage battery/bulb circuit would do - a little less elegantly.

You are planning on a "touch" "notouch" detections, I suppose. Lamp ON; Lamp OFF within a1/1000 or so.

[ctwo]

My patience wore out waiting for the viscous damping of air to take affect... My lathe appears to be perfectly level. OK, I have not given up, yet...

[stevec]

I feel that we should find a better term than "level" when referring to mounting a lathe flat with the weight being born so as not to induce any twist to the machine which we know will affect the accuracy in time.
If the lathe were mounted on a dead flat and adequately supported, little or no harm would result if there was a slight slope in any direction.
I'm not advocating enough slope so as to result in automatic swarf/chip removal

[hammermill]

i often wonder if the phrase in the same plane is a better description of what is desired.

two thing with the plumbob mas will help dampen motion. a millright i worked with had long body plumbob filled with mercury and a glas tube to shield it from air movement. dont know if one could gather up that much mercury anymore.

[tornitore45]

"Level" is a conventional term understood by machinist to mean free of twist, it comes from the practice of obtaining such condition by means of a higly sensitive spirit level. Hence the the common espression. We all agree what it means, lathes are used on ships and you could mount one on the wall if you like it that way.

The thing I am still struggling with is how a severe twist translates in errors on the part.
I am not arguing with established facts and knowledge but I fail to see geometry of the effect.
If I analize a series of cross sections of the bed-part-tool mowing axially from HS to TS I see the tool at the same distance from the part center, in all cross sections. I assume the center of rotation of the entire assembly to be located in the middle of the ways at the flat slidind serface level.

Somebody got to have a clear explanation of the barreling effect of bed twist.

Thinking this further... a somewhat similar situation happens when the HS is not aligned, let's say is pointing up and toward the back. Turning cantilever woul give you a bad taper but turning between perfect (spherical or infinitesimally small) centers would still turn without taper because the part and the way are going in the same direction. (TS height is assumed correct).

[larry_g]

"

The thing I am still struggling with is how a severe twist translates in errors on the part.
I am not arguing with established facts and knowledge but I fail to see geometry of the effect.
).

If the bed twists then the carriage will tilt as it travels down the ways moving the cutting tool closer or further away from the material being cut. As for making things level, The level is the easiest tool available to measure the amount of twist in a bed. So if the ways of a tool are level then they are on the same plane by default.

lg
no neat sig line

[Patio]

"
The thing I am still struggling with is how a severe twist translates in errors on the part.

The way I understand it, and I have only been at it a couple of years now, so take what I say with a grain of salt.

If a part is held between the HS and TS, and the bed had a even twist, then in theory it would not effect a taper in the part.
But if the part is only held only in the HS, then the tool would become progressively farther or closer, depending on the direction of twist, to the part as it travels away from the HS.

[tornitore45]

Perhaps we should start a new thread since I have degenerated this one from the leveling method to the quantification of twist generated error.

Patio, I agree that a part held in the chuck alone will taper if the bed twist, however we hold only short parts in the chuck alone therefore the issue is very attenuated since the twist,presumably, develops progressively along the ways lenght.

Larry_g, I shared your thinking until I realize that, turning between centers, the TS also move in the same direction as the tool.
Therefore, moving the caariage from HS to TS the tool and the the part center moves the same amount in the same direction maintaining cylindricity. (is that a word?)

The old timers swear it is true so there must be a convincing explanation that can not be dismissed by my simple geometric model.

This is all very academic fo me because my late has bigger problems than that. For example clamping the TS is not repeatable, lining up the points has limitation, no matter how acurate is zeroed once with the clamp snug to allow radial motion, once you clamp it down it moves a couple 1000. If I really have a critical turning between centers I place my boring head in the TS and adjust for no taper before tacking to final dimension without ever unclamping the TS.

[Bill_Cook]

You are planning on a "touch" "notouch" detections, I suppose. Lamp ON; Lamp OFF within a1/1000 or so.

Yes, I do, and if I live long enough, I'd like to cobble one together.

I think an analog ohmmeter would be the best choice. It would give a better indication of an equal touch, and already has a low draw which wouldn't do much arcing dammage to the contact surfaces. Someone with more electronic ability could probably work out a capacitor circuit like in older automotive point ignition system to reduce arcing even more.

Some sort of brushes contacting the thimble or rod would give a more consistant reading than through the oiled threads. A longish, thin, flexible, stranded, uninsulated wire run through a straight insulating tube able to be clipped to the thimble when close to touching would probably do better.

Also, a 4' pendulum would probably far exceed the acaccuracy of the mike and electrical contact consistancy. Adding in slight temp changes, and achievable flatness of the base it could drive a sane perfectionist (if there is such a thing) over the edge.

If the unit is turned 180° on a close to level surface, the two readings could be averaged to make it actually a level.

_____

Years ago after a lot of discussion of precision levels on this board, I decided to build a quick level using a pendulum. It was crude and not meant to be useful.

Best viewed 'Original Size'. (Click the magnifying glass twice in the lower right corner.)

http://smg.photobucket.com/user/skytop1 ... ort=3&o=13

(Clicking "Previous" will show the spherical mount under the tail end of the old Sheldon bed mentioned in my previous post.)

The base of the level was 24" long. The frame was 3/16" rod and coat hanger oxy/acetalene welded.

The two levers multiplied the difference between having the .005" feeler guage under one end and not as seen in the insets. The camera was mounted on a tripod. I was satisfied with the results, and the fabricated parts hang on the wall of the shop sparking questions less than once in a blue moon.

_____

Another point made more than once by my father:

He kept his BP leveled. It was handy to be able use a level in some setups.

[ctwo]

I think there is, in some economical printers and scanners, a clear film that is marked similar to our DRO scales. The quadrature signal that can be interpreted may be such that an inexpensive encoder could be built to read position on a small scale.