Machinist level vs. Master precision level

I need some advice. I'm bringing home the Logan I bought from shop at work to my basement. I want to set it up correctly. Looking at various catalogs, (MSC, Travers, Enco) Starrett has a 15" master precision level, 10 sec accuracy for $578. They have an improved level, 80 sec acc., for $150 - 200. Aside from the obvious, anyone care to give some insight on which to buy. Can a 10" x 24" lathe be leveled to an acceptable degree of accuracy with a level with graduations of .005"? I'd like to avoid shelling out $578 if I can avoid it. I'm not in business, just a recreational still learning machinist. Thanks

Starrett #199 Master Precision levels are constantly available on Ebay. For some reason, there is a glut of these levels and there has been every time I ever looked. I just did search of completed items, and 8 of these sold in the last two weeks, ranging in price from $90 to $280. Most of them look to be in very good condition, some virtually new.

Fowler makes an equivalent product that is less common, but always sells for close to half what a Starrett sells (Starret is to machinists like Gucci is to ..)

Note that the much cheaper ($50-$100 on Ebay) Starrett model 98, .005/ft level is more accurate than the spec might suggest, because it's .005 per bubble graduation, and the graduations are about 3/32 apart, so you can easily interpolate between them.

I'm a total amateur, but I suspect that accuracy better than what can be obtained with a $75 level is only relavent in ultra high-precision environments (including temperature controlled), and totally meaingless to a hobbyist. I think there's a lot more rewarding things you can do with $600 bucks than buy a level you'll use once or twice and that has an accuracy far exceeeding your requirements. My suggestion would be to get a model 98 on Ebay for under $100.

If you don't have $478 worth of "I"m not buying an import"....

MSC has $99 precision levels at the moment....and I think they are east european origin as opposed to chinese.

The chinese ones may be OK, but the one I got from another source (didn't know it was chinese until it arrived) was a banana. Totally unacceptable non-flatness of base, and I had no inclination to scrape it into flat....that was THEIR job.

BTW, I've been using a $15 6" Starrett 0.005 per foot level for several years.....interpolated you can get to at least 3X the sensitivity. Sensitivity is measured "per division".

If you are gonna pay for a decent level, I'd go with the Polish or Russian 0.0005" level. But you will be chasing that sucker around....the bubble is very active. Just walking around the machine can move it.

I picked up a import precision level on ebay from Pratt & Whirney Surplus last year for around $50.00. It supposed to be accurite to .0005 in 10". I bleive it, It is very difficult to get a machine dead on. The slightest change make a big difference in the reading.

Do NOT go out and buy any kind of precision
level to set this machine up - at least initially.

Us the two-collar method as described in the
South Bend booklet, which only requires a piece
of stock and a micrometer.

You will get better results this way than if
you spend hundreds on the level.

Jim

When we level machine tools or turbine parts, we start with a Starrett 98 level- the basic precision level. To start right off with the 199 level would be insanity as it is simply too finely graduate an instrument. We "rough level" with the number 98 levels and let things sit or "settle in" for a few days. Castings that have been rigged/lifted with a crane or jacks and then trucked- even soemthing so massive as a heavy duty US Built Engine lathe's bed- will need to "settle in". We recheck/relevel with the number 98 level. When things are dead nuts accurate with the number 98 level, we put on the Starrett 199 master precision levels. What appeared DNA level with the 98 will often show some error with the 199.

For a Logan lathe in a home shop, a levelling with a Starrett 98 level ought to be good enough. A Logan is not a Hardinge or Weiler toolroom lathe. Unless you are doing some kind of really precise gauge work and have a very tight new toolroom lathe, there should be no need to go to a Starrett 199 or equivalent master precision level. IMHO, you will be chasing your tail to try to level a used Logan lathe with a master precision level. It is also a question of what the lathe is sitting on. If it were on cast iron legs on a reinforced concrete floor, then you might stand a chance of levelling it DNA with a 199 Starrett level. If the lathe is on some wood bench or the floor isn't a solid, stable surface, or all the "feet" on the base or legs of the lathe aren't bearing solidly, then all bets of getting a good leveling are off.

I levelled my old Southbend Heavy 10" lathe with my 98 Starrett elvels. My old Southbend has some wear on the bedways, so I kind of averaged things. That old lathe cuts within 0.001" in 6 or 8", so I am not arguing.

I have an old friend who is a toolmaker. In his shop he has a Southbend Model A 9" lathe and a Brdigeport. He levelled both machines using a starrett 98 level and does jig and fixture work for the grinding of prisms for medical optics. OTH, when we levelled the rebuilt LeBlond 25" x 96 heavy duty engine lathe at the powerplant, we DID use the Starrett 199 levels- it was a heav y duty lathe, freshly rebuilt to better than toolroom standards, and it was a long bed lathe. It took a few iterations with the 98 levels before the lathe "settled in", and two iterations with the 199 levels. That big LeBlond cuts within 0.0003" in 7 feet. Can't argue with that either.

My own opinion: Use the Starrett 98 precision level and make sure it is adjusted. Interpolate the readings as noted in other posts for this thread. If you can get the lathe DNA by number 98 level, it will be fine for anything you will do with it.

Joe Michaels

Thanks for the responses. BTW, the 820 is mounted on stand as opposed to cabinet. My basement floor is fully "seasoned" reinforced concrete.

DNA?

OTH?

Thanks.

Joe's right - I forgot to add - the most
important thing is to make sure the machine
is mounted to a solid, stable surface. All
the tweaking and trimming will be for nought
if the floor moves a few inches.

DNA = "real close" or "right on."

I would set your machine up with a carpenter's
level and then do the two collar test to
trim the bed in. All you need is a micrometer.

Jim

I recently bought the Enco precision level. Made in Poland. Nice quality. The only thing that annoyed me was that the description said it was hand scraped, but it was ground instead. This doesn't really matter, I just get annoyed when pictures and/or descriptions don't match the product.

Over all, still a very good value.

Jim,

Can you describe the two-collar method ? I have not heard of this one.

TMD

Probably the deal of hold a bar or tube parallel and concentric with spindle. Put it in the spindle taper preferably, but since you don't have a test bar, most would put it in the chuck.

Near nose and at far end are narrow areas larger in diameter.

Take a cut across both (very light). Measure diameters, they should be the same if everything is alined.

Since actually I don't think this measures level, but instead measures the resultant of leveling AND any headstock misalinement, I don't know if its good as the ONLY measurement. Depends on whether you think there maight be any headstock alinement issues.

its basically the Rollies Dad's method, done a different way. That has been argued into oblivion.

If there WERE any headstock issues, and you tried to correct by twisting bed "into level", to get teh collars right, I think you'd have a pretzel.

I used my 6" starrett 98 and THEN did the two collars test. It came out close enough that I was pretty sure it was just tweaking level, which I then did.

A few years ago, after disassembling a lathe to move it a few hundred miles, I read some posts on "Rollie's Dad's Method" (RDM) and tried it after re-assembly.

Pretzel is the right description - that poor lathe would turn those dumb-bell collars between centers OK (but what happened in between the collars was anybody's guess).

The problems were eventually sorted out, most were of my own making, but I lost interest in RDM and went back to leveling, indicating the tailstock, the dumb-bell test and so on ...

There may not be a single "right way" to align a lathe bed, but I think that you should start out with leveling before refinement. (I still use RDM as a quick check for spindle alignment, BTW.)

Two-collar method as described by South Bend's
booklet:

1) place a piece of stock in the chuck.
Do not use the tailstock at all. As long
as the chuck holds the work securely it
does not have to be terribly accurate. No
precision test bar is required, nothing has
to be ground to fit the spindle taper.

2) turn down the center part of the stock
to a reduced diameter between the far end
and the end near the chuck, leaving two
collars on the stock. Rough this out.

3) set the cross slide to take a one or
two thou skim cut across the collars, with
a dead-sharp tool. Do NOT move the cross
slide while doing this.

4) mike the two diameters thus formed. If the
bed is adjusted properly they will be the
same diameter. Shim the legs or adjust the
bed, and retest until they read the same to
the desired degree of accuracy.

I like to use about one inch or so stock when
doing this, make it about as long as you want
the machine to turn accurately for.

If the bed is unworn then when you are done
it will be adjusted to turn and bore true
over its entire length. If the bed is worn,
you have adjusted the twist in the bed to
give two spots along its length (near the
headstock, where most work is done) where
it turns and bores true. Even if the bed
is *badly* worn.

No bed adjustmend will make a bed with wear
turn and bore true over its entire length,
this gives the best approximation for accurate
work where most turning is done.

What happens between the two collars? Turn
three, and mike all of them, you will find out.

This is the test that south bend suggests be
done *after* the machine has been set up with
levels, etc. as it is the only test that shows
how the lathe actually performs during real
use. Because of its simplicity it works well
for small lathes found in home shops. For
setting up large, long bed lathes which are
new from the factory, the level may be used
to advantage. But my 10L south bend only has
*one* adjustment for the bed.

Performing the two-collar test and adjustment
allows it to be made as accurate as can possibly
be.

Jim

I don't want to add to the confusion about leveling, but I wondered if anyone has tried to make their own level? Seems to me a glass tube and some mineral oil can be used to make your own 'sensor'. Imagine sealing one end of a tube, filling it almost completely with oil, then sealing the other end. Then you take a piece of stable bar, lay a shim at the center of it, then lay the glass tube on the shim, and pinch the ends of the tube down to the bar at the ends. There will be a very slight curve to the glass tube. Anything more than this and the tube would break, of course. But that small upward curve would let the bubble drift toward center, as in a real level. You would calibrate it by turning it end for end, planting the feet in the same place on a surface each time you turn it. A foot at one end is adjustable. When the bubble rests at the same spot whichever way you turn it, it's calibrated to read actual level.
For leveling a machine, you don't need actual level. All you need is to get the bubble to the same spot for each of the lathe's ways. Put your own marks on the glass, and just bring the bubble to these marks when adjusting the bed.
Maybe I'm way off here, but I think that would work. I have no way of knowing how much curve in the glass tube would be needed, or how much it could tolerate, but the less curve, the more sensitive it is. And a glass tube will bend a small amount without breaking.
Just throwing this idea in, could make a good metal project if there aren't any insurmountable obstacles. I'm probably off on the mineral oil, maybe a colored alcohol would be better, or just plain alcohol, like methyl hydrate. Maybe a few grams of your favorite vodka even.
The other thing that brings this idea up is the fact that when leveling, you would be trying to get the readings to be as close as possible for each of the lathe's ways. You don't need to know how far exactly in half-thous per mile or whatever the reading is off. That info doesn't help you, when your chore is to adjust until the bubble hits the same mark for each of the ways. That's what the real goal is. It isn't getting to know how far in thou per foot the bed is twisted.
Feel free to shoot me down on this if that's what I deserve for presenting this idea. I'd like to hear of anyone's experiences with this if they've tried to do this themselves.
By the way, I've been saying glass tube. No reason a plastic tube couldn't be used, and sealing the ends might be easier. Lexan or acrylic tubing is available. You could have a fill port on top at one end, with an o-ring on a screw to seal it. Might be easier to get a proper sized bubble this way, and with plastic the breakage factor would be less.

Alrighty then, I'm gonna get a level to setup my SB 10L after I take delivery. I'm not gonna go nuts and drop a bunch of money, but I do need to know how big I should go. Is 8" enough, 10"?

C9-
DNA: Dead Nuts Accurate
OTH: (OTOH) On The Other Hand

-Ben

I think glycerin is used, and the glass tube has a small bend in it. at least on some of them.

The longer the level, the more accurate it will be as it tends to "average out" any slight deviations in the surfaces it is placed upon. For levelling a Southbend 10L, either an 8" or 10" long level should work. What you will need is some sort of parallel to span the width of the bedways aslong with a matched pair of smaller parallels and shim stock. To corss-level the bed, you will have to span the bedways with the parallel, and put the level on top of it. You will have to get the parallel to bridge over the tops of the vee-ways and be supported off the flat ways. That is where the smaller parallels come into use. Once you have built your "bridge" over the vee ways, take a feeler gauge and check to be sure the long parallel is bearing solidly upon the tops of the two short parallels which are laid longwise on the flat-ways. If there is any difference in "elevation" between the flat ways on either side of the bed, then the long parallel making the span willnot bear solidly on the short parallels laid on the flat ways. You should not be able to get a 0.001" feeler under the long parallel. If you can get a feeler under it, shim the low side as required until a 0.001" feeler will not enter under the long parallel on either side. That done, you can then put your precision level on top.

Other advice: Use tissue paper wipes & contact cleaner to keep everything nice and clean when doing levelling work. Check for burrs or dings by running the palm/heel of your hand on surfaces where the parallels or level is to be landed. Stone these surfaces off lightly with a very fine Arkansas Hard Stone. A little pocket stone is all that is needed. A light rub with the stone and a wipe with tissue paper and conact cleaner will insure a good surface for landing the level or parallels upon. Once your bridge is made, mark the position of the parallels upon the bedways, and the level upon the top parallel with a "Sharpy" marker. If you break down the bridge and set it up at theother end of the bed, you can then return it to its first position to recheck after adjusting the levelling of the bed.

Make sure the level is in adjustment. I simply stone off the table of any handy machine tool, clean it with contact cleaner & take a "Sharpy" marker and make an outline of where I put the levels down. I turn the levels end-for-end when using to be sure they "repeat" - meaning the level is in adjustment. If they do not repeat when turned end-for-end, I adjust the vial. On the Starrett 98 levels, there is a set of small adjustment nuts which use a pin wrench. If a level is bought used or has been carried around on and off jobs, it should first be checked for adjustment before doing any real levelling work. clean the level, let it "normalize" to the temperature of the place you will be using it, then check it for accuracy. If the bubble does not repeat when turned end for end, then adjust the vial. Some very small adjustments to buck in the vial are all that should be needed. The surface the level is set upon need not be DNL (dead nuts level). As long as the surface is dead flat (or reasonably close to it such as a milling machine's table or a surface plate) and solidly supported, it can be used for checking and bucking in a precision level.

I would post a glossary of old time millwright, machinist, boilermaker and ironworker terms and abbreviations but it is not printable on this forum. 40 years around this sort of work in various parts of the USA and overseas has given me a good vocabulary of working terms not taught in engineering schools. With today's atmosphere of political and social correctness, to print such a glossary could be fatal. Still, the oldtime shop or field terms are a whole lot more descriptive than the bland, plain English and get the point accross without referencing some ANSI or DIN or ISO standard that most of us do not have access to. So, terms abbreviated into DNA or DNL, NG, NFG, and a few more will creep into the positings.
Joe Michaels

I bought a Russion precision level from MSC a couple years ago. It's 8" and accurate to .0005 (degrees?) Seems to be well-made. The thinnest tissue paper I could find sends the bubble off-scale. And it repeats reliably when reversed. Cost $55.
I used it to level my 9" Logan the other night. First I leveled the bed - starting at the headstock. I used thin alignment shims from the auto parts store. Then I removed the compound , set the level on the carriage and cranked it along the ways. A few minor tweaks to the mounting bolts got it where I wanted it.
Ran out of time, but the last check I intend to do goes like this: I have several shafts from used struts for a car. Very straight chromed shafts about 7/8" diameter. I intend to cut, face and center drill each end. Then mount between centers, checking each end for runout, then check with a DI on the toolpost.

Joe Michaels,

In regards to accuracy.
RCH
BCH
Have used these for years