Lathe Leveling

[Muskt]

Greetings Group

I am a relatively new member; but not new to machining. I do not claim to be an expert, either.

I have migrated from the Yahoo 9x20 group where I participated actively for over 10 years. My DSU (Designated Spousal Unit) & I are both recently retired, and plan to relocate from Anchorage, Alaska to the United States. Most likely to Delaware--still undecided on that point. I have sold almost all of my home shop equipment, and plan to re-establish a small shop after we find a new home. I am currently leaning toward the PM 12x36 and the PM 932. I also dabble in welding with Mig, Tig, & OA.

Anyway, on to the topic that I have been thinking about---Lathe Leveling. Over the years, there were many (sometimes heated) discussions concerning leveling the lathes on the 9x20 group. Some of them got pretty exciting, with some members loudly voicing the opinion that the entire machine must be totally "level" in all directions. Others (me among them) voiced the opinion that the only real requirement was for the ways to be straight and parallel. This means no twisting or sagging, etc. A machinist level is mandatory for this check and adjustment--at least I used to think so. I found the following YouTube video this morning, and really enjoyed it. To me, it makes perfect sense, and the machines can be checked after fabbing a relatively inexpensive jig without the expense of a real machinist level. Here is the link:

http://www.youtube.com/watch?v=2qIdsnl5vpg

I hope you enjoy it & find it of some value.

I also have a website primarily devoted to the 9x20, but there are a few articles & links that could apply to any lathe:

http://www.akpilot.net

Well, that is about it for now. I am beginning to go through withdrawal over not having any tools to play with.

Jerry in (for now) Anchorage

 

[Shadowdog500]

Jerry,

Thanks for for the video link. I plan to level my lathe this week, and have watched everything I could find online on the process, including the video you referenced. Right now I'm waiting for an order of half sized hockey pucks to come in to act as feet under my leveling screws. I have a borrowed machinists level, and plan to follow the guidance given by Oxtools.

if you do move to Deleware you may want to shop around for quality used machinery. There is plenty to be had in the north east for reasonable prices. If you are in northern Deleware you may want to check out cook machine in NJ, they buy and sell used machine equipment and have a bunch of equipment in their warehouse. There is also Fazzio that has good prices on metals, and just started getting into the used machine sales as well(but I would check cooks first for equipment)

Chris.

 

[Ray C]

Hi Jerry...

Glad you're posting and we'll get a chance to talk in public...

As you know by now, we keep discussion on this list civil so, no worries there about offering different views about this common topic.

Anyhow, in my view, the lathe does not need to be level and matter of fact, mine is intentionally "unleveled" just enough to insure fluids in the chip tray migrate to a place that's easy to reach. The lathe has to be aligned properly and there are age-old techniques called "The Rollie Method" and "Two Collar Method" used to accomplish it. These work for me and you might be interested in looking into them.


Ray

 

[Muskt]

Hello Ray.

I am very familiar with the RDM. I wrote a fairley extensive treatise on its application for the 9x20 several years ago.

I also wrote a document that ended up detailing the effect of my high-powered halogen shop lites on the alignment (bed twist due to uneven heating) of the little 9x20. If memory is correct, I experienced about 20% change in the twist of the bed after letting the temp stabilize over night. My twist was pretty small to begin with, but a 20% change was significant. If anyone is interested, I'll attempt to dig up the document & post it.

We are supposed to get the "For Sale" sign in the yard this afternoon. Houses up here are selling like crazy. Anchorage has approx 300,000 people, with less than 150 homes on the market.

I am getting very anxious to get some toys and begin this retirement thing in earnest.

Jerry in Anchorage

 

[Ray C]

Hello Ray.

I am very familiar with the RDM. I wrote a fairley extensive treatise on its application for the 9x20 several years ago.

I also wrote a document that ended up detailing the effect of my high-powered halogen shop lites on the alignment (bed twist due to uneven heating) of the little 9x20. If memory is correct, I experienced about 20% change in the twist of the bed after letting the temp stabilize over night. My twist was pretty small to begin with, but a 20% change was significant. If anyone is interested, I'll attempt to dig up the document & post it.

We are supposed to get the "For Sale" sign in the yard this afternoon. Houses up here are selling like crazy. Anchorage has approx 300,000 people, with less than 150 homes on the market.

I am getting very anxious to get some toys and begin this retirement thing in earnest.

Jerry in Anchorage

Jerry,

I'm about to call it quits in the area I now live. -Pretty sure it's a question of "when" and not "if". Similarly, with Ft. Meade, the NSA, The Naval Academy etc all nearby, homes here are only on the market for 1-3 weeks. I need more space and to be further away from next door neighbors. Problem is, I haven't a clue where I want to go.

More on topic though... It's important for a lathe to be properly set up but equally as important that the operator knows how to operate it. The alignment issue comes-up a lot and I think people get obsessed with it at times. I know for sure that my lathe settles a little bit and the concrete floor underneath it most certainly changes with weather and seasonal variances. I've spot-checked the lathe from time to time using the two-collar trick and in fact, it shows to vary just a tiny bit. Also, I've used two-collar rigs were the collars were further apart and got slightly different readings... Still though, when I focus and do my part, I can (and regularly do) knock-out boat propeller shafts (2-3 ft long) that are dead-on in terms of average diameter with variances usually held +/- 0.0003". Holding half thou is cake-walk.

On a manual lathe -no matter how good or expensive, I've never seen a shaft produced that didn't vary (up, down and everywhere in-between) by 2-4 ten-thous -and that's under the best circumstances. Often times, the variances are even greater. I suspect the variances are due to differences in material density and other factors outside of normal control. If higher precision is needed, manual lathe turning is not the answer. -TPG, maybe -but more likely, dedicated grinding equipment.

I know this may sound flippant but many of my customers call with all kinds of concerns about the apparent alignment of their new equipment -and my usual response is "how precise is a shaft that's been spun between centers?". In many cases, they don't know how to do that yet or, they attempt it on a 1/2" diameter shaft (which of course bends like a rubber band in the middle) held in a fixed-jaw chuck at one end and a poorly centered tailstock at the other end.

... BTW, I know that you know all this -and hope you don't mind if I try to inform others. Alignment is needed but, good technique and procedure is needed more-so.

Ray

 

[jererp]

Love the video. Makes perfect sense to me. I had a production lathe manufacturer tell me over 30 yrs ago that it didn't matter if the machine was level, as long as the ways were straight and parallel, we would make a good part.

 

[Marco Bernardini]

I also wrote a document that ended up detailing the effect of my high-powered halogen shop lites on the alignment (bed twist due to uneven heating) of the little 9x20. If memory is correct, I experienced about 20% change in the twist of the bed after letting the temp stabilize over night. My twist was pretty small to begin with, but a 20% change was significant. If anyone is interested, I'll attempt to dig up the document & post it.

Hello Jerry, if it happens to you to find that document, it could be a good reading.
I think not a lot of people take in account the heat generated by lights, but metal could be pretty sensitive to it, as you reported.

 

[Muskt]

I located my document pertaining to the method that I used to check for bed twist on my 9x20 about 5 years ago. I think that I did the calculations correctly. Anyway, this is not a definitive procedure--moreso, it is how I tried to determine if my little 9x20 was straight or twisted. What I found interesting, was how much the twist changed after turning the lights off for the night and then checking again in the morning. My garage stayed very nearly at the same temperature all the time, so the biggest impact was from the 2 lights which were approx 24 inches above and to the rear of the machine.
Once again, this is not a definitive "how to" it is just ramblings from a hobby machinist.

Jerry in Anchorage




{Since there are several new members to the group, I am reposting a blurb on "leveling" that I initially posted in April of 09. I do not claim that it is a definitive "how to", only a description of how I went about it.

Before I go any farther, I need to clarify one important point. The term "leveling" is generally meant to mean to check the bed of the lathe for parallelism (meaning that it is straight & not twisted). Being level really has little meaning for a lathe, other than being a starting point for the check.

The following is my earlier post:

"All the discussions about "leveling" has gotten me to thinking about my machine.

I have a Starrett 98-6 level. It is advertised to give accuracy of 0.005 inches over 1 foot per gradation.

I went out to the garage/shop & removed the plinth from my cross slide.

I cleaned it (the cross slide) well and placed the level in the center (perpendicular to the ways) and moved the carraige till the level was just under the chuck jaws.

Next, I shimmed the far end till the level read "zero".

I marked the near way and then moved the carraige 18 inches toward the tail stock end (I had removed the tail stock.).

I read the level, & it showed that the far side way had risen two & one half (2 1/2 ) gradations. At 0.005 inches per grad, I determined that the bed was twisted approximately 0.0125 inches over the 18 inches of travel.

Not good, I think. Twelve & a half thous is a lot. Time to do some thinking.

So far, I just have a number that looks large to me.

How that number translates into real live turning is the important thing to me, not just the number.

Here are my thoughts & reasoning behind my conclusion (which I haven't revealed yet).

This is a bit convoluted, so bear with me.

I know that my bed is twisted, so how can I determine the effect on turning (leave the HS alignment completely out of the equation for now).
I must know the amount that the cutter moves into or away from the work, not how much the bed is twisted. I can calculate that amount by determining the height of the cutter from the cross slide (the point that the measurement was taken from). I placed a dead center in the chuck and measured the height from the cross slide to the center of the dead center and determined it to be 3.687 inches. Now I construct a right triangle (mathmatically--not a real triangle) with a base leg of 12 inches and a vertical leg of 3.687 inches. If I set the triangle on the cross slide with the 90 degree angle exactly below the center of the spindle axis and the base leg on the cross slide pointing away from me and then raise the far end of the triangle 0.0125 inches, the point that is under the spindle center will move toward me.

HOW MUCH?

I set up a proportion equation and arrived at 0.00384 inches. WOW, that is quite a bit. OK, so I didn't account for radial movement of the theoretical points. I will just accept that the curve of the very small arc will be a straight line--so if anyone wants to argue about that --- DON"T.

Continuing on.

Remember, though, that that number (0.00384) isn't arrived at at the spindle, it is 18 inches toward the TS.

Here is my conclusion.

If I divide the 0.00384 deviation by the 18 inches , I arrive at 0.00021 per inch of deviation due to bed twist. That is not nearly so terrible as I had initially thought.

The story continues......

When I made the measurements, last evening, I had been at the lathe for at least 30-45 minutes with my "work lights" on. My lights are 2 150 watt halogens that shine on the lathe from above and to the rear. I know they put out a lot of heat. I live in Alaska, & my garage/shop normally (in the winter)(yes, it is winter here) remains at around 50 degrees F.

I went out this morning and quickly tried it again (I had left the level & shims in place when I quit for the evening).
The level still showed "zero" at the chuck end, but only 2 grads high on the TS end. Wow, that is a 20% difference just due to the possible effects of the lights warming the lathe unevenly.

My lathe is not bolted on the TS end, it just sits on a riser with a bolt acting as a pin to prevent sliding.

I really don't know where I was going with this entire exercise, I just felt like doing something.

So, to you the members, the practical machinists, the theoretical machinists, and the home shop tinkerers, have at it. What did I do that is wrong/stupid/incorrect? I can handle the criticism without getting upset.

The next project will be to check/adjust the HS alignment, again (I haven't checked if for a couple of years)."

Jerry in Anchorage}

 

[Marco Bernardini]

Thanks, Jerry!
I guess if it would be possible, then, to use hot lights to get a "negative twist", i.e. to straighten the lathe.

 

[Ray C]

... Stuff snipped...


I really don't know where I was going with this entire exercise, I just felt like doing something.

So, to you the members, the practical machinists, the theoretical machinists, and the home shop tinkerers, have at it. What did I do that is wrong/stupid/incorrect? I can handle the criticism without getting upset.

The next project will be to check/adjust the HS alignment, again (I haven't checked if for a couple of years)."

Jerry in Anchorage}

Jerry,

Totally good and valid observation.


I view manual machines, which are seemingly static, as highly dynamic but at a level humans don't perceive very well. All kinds of things change in a lathe (or mill) as you're using it. The headstock warms up and just like your halogen bulbs, it certainly must have an effect on head alignment. I've had shafts warm-up and expand while being held between centers and it ever so slightly bent the shaft and I ended-up with a lop-sided barrel-shaped shaft. Toward the center, it actually had an inverse belly on one side and an outward belly of a couple thou on the other side, all located toward the center of the shaft. The front and end 3" were just fine. -That one had me scratching my head! In essence, the bearings warming-up in your live center are more likely to cause a problem than a small amount of expansion somewhere in the body of the lathe.

The whole issue of a machine expanding and contracting is a long going battle. The designers/engineers know all about it. The CNC mill I'm looking at has refrigerated coolant water running through the bed ways, column ways and through all the lead screws. It's a $15,000 option. Back in the day, the Hardinge HLVs and Monarchs had their issues with hot/cold cycles. I don't have personal experience with this but, I recall my father and uncles discussing it.

To me, it's an absolute given that things expand/contract -and do so sometimes without uniformity.
The real question though, is how much difference does it make when cutting? (Rhetorical question)... I contend/suggest that once a lathe is basically setup properly, the deleterious effects of machine dynamics is largely offset by following a plan and good procedure when performing the work. If you're trying to do work that is so critical and requires ultra high precision, a typical manual lathe is the wrong tool.

In any event, the topic of science that covers this is a mix of reliability engineering and failure analysis. The standard method of analysis is to first consider the ideal case then, add imperfections to the system and predict/measure their effect. Once this is known (assuming you can model all the possible scenarios) reliability can be estimated and likelihood and outcome of failure can be predicted. In the case of a manual lathe, I don't think it's an exercise worth doing because I suspect (and my experiences support) other macroscopic issues far greater influence potential outcome.


Ray