PDA

View Full Version : the amazing apparently twisted lathe bed - help!



1200rpm
10-20-2012, 08:09 AM
almost every day for a month or so i`ve gone to my Sheldon, stuck a piece of 3/4" 12l14 steel or aluminum about 4" out and took a skim pass...
and always get .002 taper getting smaller toward the headstock.

so i monkey around with the levelers, use a two collar bar with identical size collars, until an indicator in the tool post reads the same on the collars.
then take another test pass and voila!!! i get no taper! :)

BUT.... then the very next day the exact same .002 taper will be back! always the same!!

i feel like i`m stuck in a time loop and repeating the same day over and over.

at this point i can only conclude the bed has taken a "set" and overnight keeps going back to where it was.

any thought or help here is appreciated, i really like the machine but can`t keep doing this every day.
is re-grinding the only answer?? twist the snot out of it in the opposite direction??
get a big hammer?? set it on fire?? try to find someone that needs parts with .002 taper?

start making lathe mandrels??

i don`t know... anyone sucessfully dealt with this problem?

rode2rouen
10-20-2012, 08:22 AM
What size Sheldon?
Is the lathe mounted to a bench, or is it on a Sheldon cabinet?
What is the condition of the floor?


Rex

1200rpm
10-20-2012, 08:31 AM
sorry... 10" Sheldon on the Sheldon cabinet, levelers on both the cabinet and the bed feet(3 points on each foot).
floor is concrete in good condition.

Rosco-P
10-20-2012, 09:31 AM
Sheet steel cabinet or cast iron cabinet? Under-drive at the headstock end? Might consider bolting the cabinet to the floor (roughly level) and shimming the bed feet. How sensitive a level are you using for setup?

RussZHC
10-20-2012, 09:49 AM
Sell me the lathe and let someone else deal with it? :rolleyes:


twist the snot out of it in the opposite direction, risking potential flaming...often advice is to not deliberately put a twist in but IMO, is not the primary concern of turning to get a true cylinder when you are done? And given that more than one of us have equipment that has been through who knows what simply due to age...let alone how it came out of the factory in the first place. If it needs it what other choices do you have?

Again, just an opinion, I think you have to start from the floor and work up, sort of at each level (floor to feet, top of cabinet/stand to feet/base of lathe) and at each stage reduce the amount of degree of adjustment (e.g. if the floor is 2" out, get rid of say 1 3/4" of that at that connection, not further up the chain).
I don't know Sheldon cabinets or later models than mine re: adjustment points of the lathe feet or bed. Is there a way of holding down as well as pushing up as it were? What I mean is mine is just bolted to a steel plate but I have a nut both above and below the attachment point both of which need to be changed when I need to adjust, both of which need to be re-tightened when adjustment is complete. Basically, for my situation, there can still be movement if a nut only holds "down" or holds "up" since there are more than enough forces to move along those studs.

If factory cabinet, could one of the attachment points be weak? e.g. rusted through so not providing support on that corner

J Tiers
10-20-2012, 10:03 AM
almost every day for a month or so i`ve gone to my Sheldon, stuck a piece of 3/4" 12l14 steel or aluminum about 4" out and took a skim pass...
and always get .002 taper getting smaller toward the headstock.

so i monkey around with the levelers, use a two collar bar with identical size collars, until an indicator in the tool post reads the same on the collars.
then take another test pass and voila!!! i get no taper! :)

BUT.... then the very next day the exact same .002 taper will be back! always the same!!



Of course there is a tendency for that type taper to occur even if the machine is perfectly aligned. The pressure of the cut can let the free end bend away and result in a taper where the free end is bigger. This does not HAVE to be what is happening, but that tendency is always there, especially when the work is stuck out a tad more than 5 diameters (0.75" diameter, out 4").

If the taper keeps coming back, then something is happening to shift the machine somehow.

The adjustment may be not tight, so that it can settle back with vibration. Lock nuts on leveling feet, or the like, are good to have to prevent this.

The adjustment may be pulling a base 'foot" off the floor, so that the entire base eventually settles back, re-applying a twist to the bed. You would need to check that the machine is still solidly set on all 4 feet. Small lathes don't usually have the weight necessary to settle down by themselves.

The "feet" may have rubber under them, which means they never are solidly on the floor. Rubber is OK, but the thickness on most leveling feet is way too much. My Logan has 2" square feet with very thin rubber glued to the bottom. Might be 30 thou thick of fairly hard rubber. which doesn't compress much at all. Just enough "rubbery-ness" to keep it from sliding around.

Floor might be shifting.... probably not if concrete, but temperature changes and moisture under a slab floor can affect it. My garage floor moves with the seasons.

Some other cause????

JCHannum
10-20-2012, 10:11 AM
Of course there is a tendency for that type taper to occur even if the machine is perfectly aligned. The pressure of the cut can let the free end bend away and result in a taper where the free end is bigger. This does not HAVE to be what is happening, but that tendency is always there, especially when the work is stuck out a tad more than 5 diameters (0.75" diameter, out 4").

My thoughts also. While it is possible the bed has taken a set, I find it doubtful that the same set would return after biasing it with the leveling adjustments.

QSIMDO
10-20-2012, 10:22 AM
I had the same problem with a wet noodle Asian lathe until I used the tail stock on anything over 1.5" out of the chuck (stock diameter factored in).

Mcgyver
10-20-2012, 10:29 AM
not daily prehaps, but floors do move...concrete moves. I guess you have to start looking at what's changed, room temp, an electric heater behind it, adjacent to a window covering it in sunlight.

I've always consider lathe twist to mean slight changes in shape induced by its own weight vs twist to a wood worker meaning the the boards at rest shape. One could say the lathe could have a real twist to the bed, like a board might, but how's this cause it to spring back?......and if that were the case, why don't all the properly done beds just go "boing" back to the correct shape without any twist?

I think that's why all the Q's on base and floor....that too is where i'd be looking first.

I also don't get the slow reaction; a deflected piece of metal goes back to shape right away; not overnight. It suggests something the operates more slowly, ie change temp.

do you have a precision level, like a starrett 199? might help to start to eliminate things. without doing that first, its hard to know if the taper is tailstock or twist induced.....if it is tailstock and forcing the bed out of shape to compensate, maybe it is springing back some, especially if its just sitting there vs bolted.

Another idea occurs on how to isolate things; set things up on three points. try lathe on three points on the stand, then stand on three points. Might eliminate if twisting action is coming from chance in the floor or stand. This wouldn't be a safe set up for permanent use but could be done for experimentation with light cuts.

justanengineer
10-20-2012, 10:47 AM
As others have suggested, the only way to find the issue is to start leveling and eliminate possible issues. I would start with the cabinet itself getting it as close as possible, then move to the bed. Once you have that, you can begin to consider wear and adjustment issues. Keep in mind that not only does the tailstock move forward and back, but also can move up and down. It can also be cocked/angled slightly in 4 directions, in which case you will find no taper with the tailstock extended at a single extended point, but taper at all others.

Considering your machine is likely 40+ years old, you need to be rather systematic about this, not only looking for wear, but also an inaccurate "fix" or "regrind" at some point in the past.

chipmaker4130
10-20-2012, 10:51 AM
I have a 13" Sheldon and had the same problem. Turns out it wasn't the lathe at all. I did use a center in the tailstock, and after putzing around for an hour seemingly unable to get the tailstock tru I discovered that my carbide insert in the cutting tool had loosened! Couldn't see or feel it, but I turned the retainer screw half a turn and the problem was gone.

loose nut
10-20-2012, 11:07 AM
I had the same problem with a wet noodle Asian lathe until I used the tail stock on anything over 1.5" out of the chuck (stock diameter factored in).

+1 on that.

4 times the diameter is about all you can turn with out a center in it without getting some amount of taper.

Paul Alciatore
10-20-2012, 02:10 PM
One thing I have learned, the hard way of course, is that concrete floors are NOT flat. And the people who do concrete work do NOT even have any idea of what a thousandth of an inch is, much less work to that kind of tolerance. A concrete floor can vary by 32s or even 16s of an inch over a fairly short distance. The FIRST thing you must do for any lathe is lock it down to one and only one position on that concrete floor. If it moves, even a fraction of an inch, it WILL change it's level.

The next thing I would check is the size of the leveling feet. Concrete will abrade and if the feet are too small, they can wear down into the concrete from normal machine vibration. It is also a good idea to run a brick or small flat block of concrete over the area where the feet will sit to be sure there are no high points there (in other words, sand it flat). These high points can wear down from vibration and that leg will settle over time.

Be sure that your levelers are locked down. Sounds basic, but it is important. I like a leveler that has the lock down nut ABOVE the threaded hole that it sits in. If the nut is below, it will pull the leveler out/down when you lock it and that destroys the work you have just done to level things. If the locking nut is above, it pulls it up and since the machine weight is already pushing it in that direction, it will not move very much.

Oh, another point: if your concrete floor has any kind of tile or vinyl covering on it, the levelers' feet can settle into that tile or covering. It would be best to make cut outs for the feet to sit directly on the concrete. If that is not possible, I would suggest using steel plates that are at least 1/4" thick and at least 8" in diameter for the levelers to rest on. I used a strategy similar to this for a lathe in a trailer.

Only after you get a ROCK SOLID footing can you even start to level a lathe.

darryl
10-20-2012, 02:48 PM
Yeah, I'm thinking that if you are continually making the adjustments, the lathe bed is remaining fairly true but you are flexing the stand. The bed does not want to be 'twisted', and is overcoming that by forcing the pressures elsewhere- the stand, the floor, the feet- anywhere that a movement can happen, including in the floor itself. There is no guarantee that a concrete floor won't flex.

rohart
10-20-2012, 03:01 PM
Firstly, while I agree that foru or five times diameter sticking out is sort of acceptable, there's a vast difference in deflection between a 20 thou cut with carbide and a 2 thou shaving cut with a sharp HSS tool.

Secondly, if you're convincing yourself that your bed has a twist, give the bed the freedom to show you. I mean that you should support the bed at three points - one front and one back at the headstock end, and only one point at the tailstock end. Where shoul I put the support at the tailstock end, you ask. Well, perform three tests, with the support at three positions, at the front, in the middle and at the back.

If all three tests are identical, then your bed weighs so little that it doesn't mind if it's supported at the front or the back. I suspect any lathe would should some difference. The question is, which position begins to, or completely, removes your taper.

Black_Moons
10-20-2012, 03:16 PM
A few good points I would like to rehash and add my own..
Rubber under the feet can creep, and it 'twisting' the bed back will put lots more pressure back on the rubber that just creeped, causing it to creep again. Use *very* thin amounts of rubber over a wide area with lots of metal support behind it.
I use 1/8" thick neoprene gasket material under a 2" diameter flat washer I welded onto 1/2" all thread rod. (Welded moreso to a nut on the all-thread rod above it, with welds in the center ground down, just to keep the nut from ever spining).

Vibration can loosen feet. This can be machine vibration, nearby trains, etc. Needs nuts that clamp the rod down to the machine, Not jamnuts that lift the adjusting nut up further. Large threads have lots of clearance. Also good idea to have a wrench on the foot itself, to prevent it from rotating as you tighten the top locking nut.

3/4" stock isent something you can test a 4" cut on. Get something at least 2" in diamiter to test a 4" cut on. Expensive yes, but you don't need to take much off for each test, You could do hundreds of tests before you end up with a chunk of 1.5" diamiter stock to put back into your stock shelf.

Your tool dulling can also cause taper:
the tool that cuts straight today, will require extra pressure to cut when it dulls tomarrow, causing more taper from deflection of the workpeice.

darryl
10-20-2012, 03:18 PM
As Richard said, but also pay attention to how tightening the headstock hold-down bolts might affect things. You could be inducing some warpage just by that, depending on how the stand is, even if you are only supporting the tail end on one leg.

There's a method I like when arranging a single tailstock end leg- it involves using a pivoting mount that can be left loose while two bolts mount it to the bed, and two bolts mount it to the stand. Essentially the mounting is secured to both the stand and the bed, but tail end of the bed takes whatever angle it wants to- then you tighten the pivot bolt. I realize that this lets the bed float and doesn't give a method of applying any 'anti-twist' directly, but you'll see how you can still do this anyway. The idea of the pivot bolt is to allow a stand or floor misalignment to exist, while the bed can be made 'four bolt mount secure'

oldtiffie
10-20-2012, 07:54 PM
almost every day for a month or so i`ve gone to my Sheldon, stuck a piece of 3/4" 12l14 steel or aluminum about 4" out and took a skim pass...
and always get .002 taper getting smaller toward the headstock.

so i monkey around with the levelers, use a two collar bar with identical size collars, until an indicator in the tool post reads the same on the collars.
then take another test pass and voila!!! i get no taper! :)

BUT.... then the very next day the exact same .002 taper will be back! always the same!!

i feel like i`m stuck in a time loop and repeating the same day over and over.

at this point i can only conclude the bed has taken a "set" and overnight keeps going back to where it was.

any thought or help here is appreciated, i really like the machine but can`t keep doing this every day.
is re-grinding the only answer?? twist the snot out of it in the opposite direction??
get a big hammer?? set it on fire?? try to find someone that needs parts with .002 taper?

start making lathe mandrels??

i don`t know... anyone sucessfully dealt with this problem?

The taper (if uniform) is 0.0005" per inch.

If "half a thou per inch of taper" is acceptable for most of your work - use it as is.

If it were me on my lathe and I wanted it "closer" I'd turn the "small" (head-stock) end to say "half a thou to a thou over" and use a good fine single-cut file (and some "fine "wet and dry" cutting "paper" (at Auto shops) to remove the taper (and perhaps improve the finish as well).

If you've got the time to keep on "trying to fix it" and you want or "need" to do it - then keep on trying.

1200rpm
10-20-2012, 09:32 PM
some good suggestions.

one in particular caught my eye - "lock the levelers"

the Sheldon has a pretty nice system for leveling - it`s on a heavy cabinet(300ish with an almost 3/8" chip pan.
there are 4 points of adjustment at each end of the cabinet. (4 corners at each end)
then each bed foot has 3 points of adjustment.

my "procedure" is to use a .0005/10" level to get the top of the chip pan level by laying the level in the middle of the chiip pan( across a couple parallels) and adjust the cabinet levelers. i use the levels cross vial to get it level longitudinally.

then level the bed by using the level across the bed at each end over the flats( the level sitting on parallels)
then tweak it using the identical collar bar with an indicator in the tool post.
my two collar bar is a piece of 1" 6061 aluminum held in an ER40 collet,sticking out maybe 4".
i`ll turn two identical collars(say .990") about 1/2" wide at each end - each time i mount the bar i`ll re-true the collars, they get a little smaller each time.
then with an indicator in the tool post check the collars and tweak until the reading is the same.
immediately afterwards it will cut true and will continue to cut true for that day.
then somehow something changes overnight and the .002 taper is back.

this has always worked for me with my other lathes - admittedly my other lathes were much more flexible than this one. :)
today i did it again, this time i overcompensated a little - it is cutting a .001/ in about 3" taper, getting BIGGER toward the headstock- the thinking being maybe it will pull back to true overnight.
and this time i locked the cabinet levelers!

it wouldn`t bother me so much if i didn`t know it was possible for it to cut true...

Jaakko Fagerlund
10-21-2012, 02:12 AM
Just a question, but what is a such work that you need that sort of precision over a long length?

J Tiers
10-21-2012, 02:24 AM
Just a question, but what is a such work that you need that sort of precision over a long length?

How about "any" sort of shaft work... or threading a long acme screw, or threading a long anything.... Even just turning a long part, if you want it parallel and not tapered. Any appearance part" may need not to be tapered several thou per foot, and something made of thinnish tubing could surely be an issue over 600mm or so.

Sure you can take out the gross errors by tailstock adjustments, but that isn't necessarily parallel to the error..... although it probably will be close. And it involves "special setting" by trial and error to get the machine to do what should be a normal task.

I have no problem with a person trying to get the machine adjusted so that any sort of work can be done on it "offhand", without having to "special adjust" for that work.

oldtiffie
10-21-2012, 02:34 AM
If I were in that situation - and assuming my lathe bed was straight (no bends in the horizontal plane) I'd opt for "between centres" with a travelling "steady" and re-locating the fixed steady as well to suit.

"Touch-up" with a good singel-cut file and "wet and dry" "cutting paper" just needs a "lick" and there I am - done.

If finish was a "problem" as it often is ("chatter" etc.) on long thin (and not so thin) jobs it might justify mounting a good/fair tool-post grinder.

Without "steadies" there is bound to be some "up-lift" due the the reaction in the shaft to the cutting tool up-thrust - the job will/may tend to "climb up" (onto) the tool.

Too large (any?) a lead angle will also force the thinner jobs "inward" as well.

Well-placed and well-set-up steadies largely eliminate these problems.

A "box tool" (as from a turret or capstan lathe) would be even better as you can turn long bent shafts "on size" (but still "bent") - to be straightened by hand later.

loply
10-21-2012, 04:23 AM
When you come back the next morning and find a taper is being cut, have you let the spindle warm up fully first?

oldtiffie
10-21-2012, 04:34 AM
Good point.

Machines and jobs and measuring tools should be "warmed up" - to shop ambient temperatures before (and during) use.

If nothing else it will tend to reduce/normalise/equalise "size creep" due to different tool etc. temperatures and unequal temperature co-efficients of expansion.

Similarly, jobs should be allowed to cool (heat too?) to ambient temperatures before measuring - and sometimes during "setting-up" too.

There isn't much that's colder then a surface plate first thing in the morning that has been covered over-night and is still nearly at night ambient and takes ages to warm through to shop ambient temperatures. Any error (if any at all) will be very slight but it may be there never the less. Surface plates are calibrated at 20 deg C (~68 deg F) but their co-efficient of expension (granite more so than cast iron) is very low.

1200rpm
10-21-2012, 05:06 AM
ok, maybe some hope!

when i left it yesteday i was getting .001 taper in about 3" with it getting bigger toward the headstock.

just now(4:40 AM) i started it up, left it run a few minutes while my coffee brewed(took the suggestion of letting the spindle warm up)
took a test cut and got .0008 taper in 3" getting bigger toward the headstock - a slight improvement!
but mainly i did not get the same .002 taper(getting smaller) this time! maybe the cycle is broken!

it`s been driving me nuts because i know this machine is capable - just something on my end has been lacking.
if it drifts back to the .002 taper i`ll try bolting the cabinet down - i have never had to do that before but i`ve never had a machine with a chip pan made with 3/8" plate either!

oldtiffie
10-21-2012, 05:53 AM
It might take more than "a couple of minutes" for the temperature to be at ambient and stable - try 30 minutes warm up and see what happens.

Your micrometer needs to be checked for zero at ambient temperature/s as well.

1200rpm
10-21-2012, 06:57 PM
when i got home from work today i did another test cut.... it had creeped again.
not quite .002 this time, but still.:confused:
so i decided to take the machine loose from the cabinet and think i may have found the culprit!

a few months ago when i put it together i replaced the hardened, broken up gasket that went between the foot and the chip pan with some thick automotive gasket material.(insert dope slap)

i switched it to brass shim washers under each foot - time will tell but it seemed to be easier to dial in.
here`s a pic of the Sheldon`s foot for anyone not familiar-
and a pic of what not to use.
http://i1078.photobucket.com/albums/w494/hobbymachinistpics/1350857861.jpg
http://i1078.photobucket.com/albums/w494/hobbymachinistpics/1350857934.jpg

jkilroy
10-21-2012, 09:07 PM
I doubt that gasket material has the power to move things, but it certainly was not offering you a solid surface. Thin shims of brass are good, but once you get to stacking them together they can deform under moderate pressure. Stainless makes a great shim material.

Jaakko Fagerlund
10-21-2012, 11:57 PM
How about "any" sort of shaft work... or threading a long acme screw, or threading a long anything.... Even just turning a long part, if you want it parallel and not tapered. Any appearance part" may need not to be tapered several thou per foot, and something made of thinnish tubing could surely be an issue over 600mm or so.
The reason why I asked is that if the OP is not doing long work, why try chasing that error.

1200rpm
10-22-2012, 05:17 AM
it`s a personality flaw. :)
OCD.

i`m a happy boy now!
5 AM - no shift! it stayed the same!
test cut on 1" diameter 6061 - maybe sticking out a little far but i have a really sharp bit.
mic readings every inch(black marks)
using a standard 1" mic w a vernier i got the following -.9943" , .9946" , .9945" , .9944" , .9942"

http://i1078.photobucket.com/albums/w494/hobbymachinistpics/1350864556.jpg

J Tiers
10-22-2012, 08:14 AM
Good for you.....

You may want to check every so often if this is an issue for you... as noted, floors shift, machines shift, etc, etc.



The reason why I asked is that if the OP is not doing long work, why try chasing that error.

So far, maybe not.... tomorrow? who knows?




I have no problem with a person trying to get the machine adjusted so that any sort of work can be done on it "offhand", without having to "special adjust" for that work.