Question On Aligning The Lathe Spindle And Bed Twist

Quality, could you be more specific on point 3 in post no. 28 please? What does 'mess' around and 'tweak' the level involve?
 
Once you know your headstock is in alignment by the test bar method and your bed is level for no twist, then you can move on to doing a two collar test using tailstock. If there is any change in diameter, you adjust the side travel in the tailstock to compensate for it, not the bed. This is where people get in trouble doing this test. And I wish a certain person out there that I know would quit teaching this technique because it gives a false perception on correcting taper in a part. When you "twist" a bed to remove taper, this turns into a permanent twist that cannot be removed easily if in time a lathe bed was to be re-ground. The people that do the grinding have to shim the bed on the grinder and grind out the twist, which causes them to remove more material than they would if the bed never had a twist in the first place.
 
Anyone that has a Starrett 99 level can do a decent job of leveling their lathe bed as long as they carefully place the level on the bed or the carriage. Take your measurement carefully, you should get it within .0000-.0010" in the bubble readings. yeah, I know, each division is .005" in 12". Cut that in half at 6" that would be .0025". A 199 level is too sensitive for us H-M. You can walk across the floor and make it read differently! I prefer putting the level on the bed myself. Once the bed is leveled, put the level on the saddle and move the carriage up and down the bed and verify the same readings. You should get the same readings. IF you don't, go back and recheck the bed. Please understand, if you have a lathe bed that has wear, and most do unless new, you will see a difference in readings from that taken on the bed vs those taken on the saddle. Now what do you do?
 
lathe beds are cast
after that they are put out and left to season for a year or so
or they are taken through heat up cool down cycles to artificially season them
after that they are put on a planer/mill and are roughly brouht to size and shape
another round of seasoning follows
after that they are fine planed/milled
finally they are hand scraped to perfect if needed

that is the way it ued to be....

nowadays in China
a bed is cast
the week after it is put on a mill and milled to form and size
finished

the moment the clamps at the mill are released the form and size of the bed change
in the time thereafter this proces continues

by the time the lathe is received by the customer form and sizeof the bed have changed, and will continue to do so for considerable time

so in order to get it up to specs the bed will have to be checked

the ways need to be straight and parallel
straightness is checked by means of a straightedge to the side of the ways and a level to the top of the ways
parallelism by means of a micrometer measuring the distance between the outsides of the ways

measuring is done while the bed is put freely on three points
two at the headstock and one at the tailstock, the latter being a pivotpoint at the heart of the bed

if all measurements check out the bed is ok
if not, one may try to twist the bed back into the alignment it had when beeing milled/ground

the level is the only measuring instrument that doesnt need a datum to be precise, the level will measure with a precision no indicator will even come near to

checking the alignment of the headstock, be it by means of a testbar or by means of the two collar method , can only be done after it is established that the bed, on which the saddle, and therefore the indicator runs, is straight and parallel
 
Ken covered my mess around with and tweak.

Also, the lathe in questions is new, so no issue there.

And I completely agree with the fact of the 199 level at .0005 per foot. It is very accurate, but so sensitive. If you think about what .0005 over a foot is.

Lots of good information coming up here.

I do also strongly agree, Never twist a bed to get the taper out especially on a new machine. The machine in question is only about 1300 pounds, so to really twist the bed, it would have to be anchored down anyway, it is going to resist moving a lot without doing that, one foot would just end up in the air.
 
It is a no brainer to agree that the objective is to remove twist in the ways and not create it.

That a new lathe once installed in our shop can be assumed to have straight ways with no twist is debatable. So how do we test for twist and how do we correct it?

On one hand we have said that twist can be checked with a precision level but that using one properly is probably beyond our ability. Add the fact that we don't all own a precision level and we start looking for a plan B.

If a bar is chucked up (not between centers) and turned as mentioned in the original post and ends up tapered, what does that show? To my mind it shows that headstock misalignment AND/OR bed twist is present. At this point we should neither straighten the ways nor adjust the headstock because it is not clear exactly where the problem lies.

A precision test bar in the spindle taper (likely MT5) is the best way to go about isolating headstock alignment. I can make a MT5 test bar but I can't make a precision level. How to adjust the headstock accordingly is particular to that lathe.

Once the headstock is aligned then the 2 collar test (not between centers) is the best way to see if there is twist in the ways. It is not unreasonable to consider it a good second check even if the precision level method was used properly beforehand. It is correct to adjust the feet of the tail stock cabinet or shim between the lathe and cabinets to REMOVE any twist in the ways.
 
Last edited:
I will add that I checked my PM1340GT headstock alignment the other day after the discussion here, each chuck had a different run out at 6" out. I use a 3/4" Tight-Tolerance Rod Precision Ground 12" put in the chuck. Level the machine as Ken describes. I use a 4J Bison chuck to get as close to 0 TIR reading at the chuck as reasonable possible, the dial indicator I use is indicated at 0.0001" per division and barely moves. At 6" out, I take the average of the swing, which is +/- about 0.0002" on either side of zero. If you do not take into account the swing of the shaft due to to different chucks or the shaft not being perfectly straight your head alignment will be off. What I had read was to take the average of the high and low dial indicator readings at different points on the rod, the 0 point should swing evenly around 0 for alignment to be correct. My 5C chuck is worse, and runs around +/-0.0005" at 6" out. The alignment is pretty much spot on after tweaking the head alignment. A 5MT test bar would be great, they sell a 5MT on eBay for around $115, they say they are accurate to 0.0002" or better. But a lot to spend, but can also be used for your tailstock alignment. I am still a newbie at this, but it has worked well for me.
 
Time for an update.

I redid the test bar setup using a 1.5" dia steel bar I had in my shop. It was long enough I could have 2 areas 7" apart that I could machine and measure. For reasons I'll cover later I leveled the lathe end to end and front to back with the level near the headstock. I then made sure the leveling feet had fairly equal pressure and made a .002" cut. I measured .005" difference in dia at the two ends with the tailstock end bigger. I setup a dial indicator at the tailstock end of the bar and then loosened the 4 bolts that attach the headstock. There are 3 socket head capscrews above the motor, I screwed in the one further from the chuck until I got .002" shown on the indicator. After snugging the mount bolts (not tight) I did another test cut and got .0032" taper over the 7". I moved it another .003" on the indicator and that test cut was .0004' taper so I stopped there and fully tightened the bolts. The final test cut ended up at .0006" taper so I stopped there with the headstock alignment. It ended up being fairly easy to do using that setup.

Throughout the setup process I have been getting an oscillation in the lathe at certain speeds. Adjusting the leveling feet when trying to level the lathe would make it worse while getting them fairly even would make it better. After doing many test's trying to isolate where it was coming from I think I have it narrowed down to either the belt and/or pulleys. I have a Fenner link belt on order that I hope will get rid of the resonance I have been getting. I will also check the pulleys with an indicator as that might be the issue. The last test I did today was to swap the belt to the high speed position and was unable to duplicate the resonance I have been seeing. Jim Dawson was a big help today, he pointed out several possible causes and hit on what I hope was the root cause. Thanks again Jim.

On this lathe it is very difficult to make a big change to the test bar results by adjusting the leveling feet. The best result I got left me with one of the tailstock feet off the floor 1/4" which wasn't a solution. That was why I ended up making sure the leveling feet at the tailstock end had fairly even pressure before adjusting the headstock. I'll update this thread again when I solve the oscillation issue, hopefully Sat.
 
The stock belt takes a set on the pulleys, you can feel it catch in one spot, this can cause vibration on the machine. Also you may want to try different belt tension, I run mine a little on the looser side. I had better results with Gates Tri-Power notched belt (BX24 and BX25), link belts also seem to work well by many others. The pulleys seem to run very even on my machine, but might check the belt alignment.
 
You must log in or register to reply here.
Back
Top