Identify Damage to Lathe Spindle

Alright fellas, I need some help here.

Yesterday, I crashed my lathe. I was running an interrupted cut up to a shoulder, went one rev too far, BAM, tool digs in, ejects my part, and manages to unscrew the toolpost. The handle whipped around, right into the chuck jaws. Explodes the knob on the end and bends the handle. Toolpost otherwise looks OK. Miraculously, the cutting tool itself survived as well, as did my part (I was cutting acetal at the time, and the cutting tool only dug into the bit that I still needed to cut).

After that, I rotated the spindle by hand. Seemed to spin freely, without making any hideous noises. Seemed to run under power without issue as well. I chucked up a bar without noticing anything suspicious in the 4-jaw centering process. Once I made a cut, however, it became clear that something was going on. The spiral surface roughness pattern doesn't show up in the picture as starkly as it does in real life. The bar also has a repeating runout pattern, 0~.005~0.0005~.0055~0... I didn't check to see if it spirals around the part like the roughness pattern, but I did check a few points and it is consistent across the surface.

It was getting to be the end of the day, so I decided to call it quits once confirming that there was something serious going on. That said, based on the current information, anyone have some guesses as to the cause? Next steps? Remember, that runout pattern was over 5 thou on a cut bar, not the lathe spindle itself. Lathe is a 1968 19" Leblond Regal. Recommendations to scrap the tool will be unwelcome but considered, if worse comes to worst.

The only thing I can think of other than bearing damage or a bent spindle would be that when I went to ensure that the chuck was still tight, I inadvertently loosened it (L1 always seems backwards to me), then re-tightened immediately. I doubt that would show up as runout on a cut surface if I had entrapped something there though.

Meanwhile, with all the violence, the feed works in the apron are very likely quite unhappy - they were the party being rudely stopped or forced in the other direction.

Up stream the QC box may have suffered - and even on the way to the QC, the feed rod may be damaged

Regals are tender souls, not happy with such. Now if it had been big brother Heavy Duty, probably shrug off such happenings

johnoder said:

Meanwhile, with all the violence, the feed works in the apron are very likely quite unhappy - they were the party being rudely stopped or forced in the other direction.

Up stream the QC box may have suffered - and even on the way to the QC, the feed rod may be damaged

Regals are tender souls, not happy with such. Now if it had been big brother Heavy Duty, probably shrug off such happenings

Click to expand...

Hmm, good thought, but based on my memory of the event, not likely. I had my hand on the lever already, and if I recall correctly, I may have released it before contact was made on that last pass, but the 3/4 revolution where no cutting took place had advanced the shoulder into the danger zone anyway. If that is correct, the QC and feed rod would be spared.

If not, the force would have come as a single shock, building rapidly only until the toolpost unloosened, then a sharp thwack on the handle of the toolpost, which I know it was already released for.

I've got a used spindle and new bearings for a 15".

i think its likely that a spindle bearing has moved on slack in the adjusting nut,and that tightening the adjuster will push the bearing back into place,eliminating the slack..certainly try that before pulling the spindle out.....However G Bent s new bearings might be a good investment anyway ,if the lathe is in much use.

Not fmiliar with Regals, I would also check fit between the carriage and bed, saddle and compound etc. something bent and now rocking my cause irregular surface finish.

I replaced a costly Gamet roller wit a double row ball 50x90x30....which I already had.........however internal clearance in the C3 rated bearing was excessive,and resulted in a series of bands about 15mm wide in the finish........heavy cuts kept the bearing loaded.........cured by parting the inner race into two, thus making the bearing adjustable,and using the threaded adjuster to eliminate the clearance......the front double roller has its own adjuster.............Complete success,lathe like new......almost......Ive noted its far more likley a standard ball bearing will have very small runout than a standard tapered roller.

OK, first of all thanks for the feedback and some leads to follow.

I looked with a more critical eye at the carriage travel hardware, and it all still seems OK. That's good news. The other good news is that the spindle itself also seems OK, since the chuck body runs out around 0.001" and I had no issues trying to dial in my bar before I noticed the cutting issues. Axial also looks good, a little less than 0.001" with tailstock pushing firmly against it like a hard-fed drill, followed by my L1 spanner behind it to pull it back forward. I might delve into it a little more later, but it sounds fine and those numbers all seem normal (but what do I know) for a 50+ years old machine, so the spindle is no longer my prime suspect right now.

Mach2's suggestion to look at the carriage and slides proved more revealing. I grabbed the toolpost and leaned back, noticing ~0.005 on the compound and cross slide, but less than 0.001 on the carriage (yes, I did remove all the backlash first). Cross-slide to carriage was also small. Those measurements were all perpendicular to the spindle axis, parallel to the floor. I checked the carriage against the spindle in the other plane, parallel to the wall. Little to nothing on the operator side, 0.010"+ on the far side. Looks like we have our winner. The whole carriage is rocking back and forth around the front bed way.

Now, what specifically should I look at to address this? Should I start a new thread in the machine reconditioning section?

Alright, I'm stuck. I took a look at the carriage and saddle, and adjusting the gibs made a huge difference on the deflection readings at the carriage and toolpost. More work to do if I want it "perfect" but the problem is that it made NO difference in a machined part. Still 0~0.005~0.0005~0.0055~0... every other rev, on a finished part. Surface finish looks GREAT. The wonky pattern is completely gone, but the parts are still not coming out round.

So, I went back to the spindle. Put my dial on the chuck body while it was running (350 RPM or so) and noticed significantly more wiggle than .001, it was around .004 or so. Turned off the motor, and I watched the needle wobble decrease, settling down to around 0.001" TIR as it slowed down. So, it's definitely the spindle. It has me a little puzzled why it changes so much under load vs freely rotated by hand.

I opened up the headstock to take more readings around the spindle on various spots. Outside the headstock on the taper itself, we have sub-thou runout. On the other side of the casting, nearest the bearings, we have .005" runout. Moving towards the far end of the spindle, there's a mid-shaft bearing that reads .003" on the spindle nearby. Inside the headstock, farthest from the taper, I got sub-thou again.

I just can't seem to nail down why this is happening like this. Less than .001 on the taper, but .005" just a few inches away? On the other side of the same bearing, the suspect one? But runs out the full .005" when loaded? The linear change in runout makes sense along the three points I checked within the headstock, but the way it shifts so abruptly going to the outside of the headstock itself is puzzling.

If you're willing to pull the spindle from the machine, you can make up a pair of press-in plugs for the ends of the spindle, with 60 degree female tapers on the outsides. Try for best concentricity you can with these, it makes testing easier in the next step.

Then mount the spindle/plugs on a bench center, and using multiple test indicators "map out" the TIR readings on the critical diameters - bearings journals, chuck mount surface, etc.

Look for uneven readings that don't sync, errors that move together high and low just could be from the plugs not being perfect, but readings that go hi-low along the spindle could show a bend or other error that leads to cutting patterns. Even an "egg shape" could show up on a diameter from a hard, single hit.

If you don't want to remove the spindle, use the indicators on the chuck mounting surface and map its errors while in the headstock. You can press and pull on the spindle with the headstock cover removed and see how that influences the readings.

Runout every second revolution is certainly related to rollers in cage.....they rotate at 1/2 revs....your runout may be loose rollers climbing the races ,then back under load........have you tightened up the bearing adjusters?.......a bit at a time ,with the spindle rotating until you get a drag......give it a bit of a hit both ways too ,make sure bearings arent hung up .ready to move under load.

What happened when it crashed? Did the spindle stop? You say your hand was on the feed handle or the on / off spindle handle? When I looked at the turned part I thought spindle bearing issue, chuck jaw damage and it wiggles the part when your turning. Or bent or cracked gear tooth, but you should be able to feel vibration or hear a weird noise. Then I started to think tool bit. Have you used a mag glass to check your cutting edge or replaced the tool? I am assuming your using carbide and they can fool you with a small chip on the edge. Chuck up some soft cold roll and a new carbide and see what happens. Have toy pulled the chuck and tested the spindle nose? Drive down to a auto parts store and buy a oscilloscope and listen to the spindle bearings running at the rpm where it happened. I would say to listen when the spindle is ramping down when you shut it off, but you have a magnetic brake? Does your machine have the 2 position switch so you can let it coast to a stop? If so You can start with the cover on and putting the probe on the headstock casting If you remove the cover I would drain the head oil or it will sprat all over. You can run the spindle for a minute and listen then. Also you may have cracked the compound slide. It is easy to pull it apart and check. I am taking some wild ass guesses, but a crash can cause some hidden issues. Rich

Milland,

Unfortunately, I'm not well set-up to take your advice, as much as I would like to. Without my lathe, I won't be able to make those plugs, and I don't have enough indicators to take several readings at once. I could try to index the spindle and do it in place, but that would take a lot more thinking to map out. I still might try it though.

John,

I think that will be my next step here, considering that it runs out more when spinning freely under power and when under load than when it is simply rotated by hand. My expectation is that things will either get much better right away, or that it will start going .005 all the time. I hope for the former. Only reason I didn't get around to that yesterday is my damn welding helmet stopped darkening, so I couldn't finish my home made spanner for the adjuster nut.

Rich,

I can answer some of your questions already, but some of them I will need to look into.

Spindle did not stop. Part was ejected, toolpost loosened and spun the handle into the spinning chuck. Toolpost handle got thwacked real good. It was knocked out of the way and the lathe kept on going, not contacting anything.

Gears are all OK, at least in the headstock.

Tool bit used on the test bar was not involved in the crash, and the same one used yesterday when I got the pattern to go away. Re-tightening the rear carriage gibs got rid of the pattern, but did not get rid of the out-of-round condition.

Pulled the chuck and checked the spindle nose, it runs out around 0.0005 on its own, rotated by hand. The chuck body does around 0.001 under the same conditions, but when rotating under power at 350 RPM (a little above crash speed, this is the speed I used for the test bar cut) I noticed the needle wiggling through .004-.005 on the chuck body. I can let it coast to a stop, and as I did so, the needle on the indicator slowly calmed down to 0.001 like it does when rotated by hand.

For noise, I think there is a very light sound, like something dragging along a smooth-ish surface. It's faint, but maybe if I pick up a cheap stethoscope I will be able to hear it better.

Thanks everyone for the input, I will chase down some more of these leads today. If you want to know a little about me, I have approximately zero formal training in machining. Anything I didn't learn the hard way, I learned here. Almost all the machining I do is to support engineering, inspection, and testing of rubber parts, and in my arrogance I decided that I would go straight to self-sufficient mechanical support in parallel with the setup of the test lab. It's a lot to bite off, but I'm starting to feel confident that with your help I will get this lathe cutting accurately enough to satisfy my needs (typical tolerance is ±0.002). Good news is that we are still trying to get our customers interested in the new services, so I don't have any active projects being held up by this issue so I don't need to kill myself working to get back up to speed.

I really appreciate you guys taking the time to help me out.

A tip....run the tail stock and use it on those short jobs. As the others have said try snugging up the spindle bearings. We did a thread some where on here a few months back on how to do that. If you can find long dowel pin and slowly tighten it in the jaws and try wiggling it. You could also spray some Dykem layout blue and tighten it in the jaws and lossen it and check where the blue has worn away.

Look at page 22 & 23 http://vintagemachinery.org/pubs/2102/3414.pdf I like how your being a detective! Good luck :-)

We have made some progress here, but there's more to do.

I finished my homemade spanner, got the bearings tightened down some, and it made a big difference on the circularity. We're now down to 0.001~0.0015 on the out of round condition. However, the surface pattern is back, and more distinct than ever. Seems like fewer marks than before, but each more stark and visible. The other distinct thing is that the marks are mostly at the same angle, but move across feed revolution lines. The out-of round appears that it is not a single consistent oblong shape, but can be seen clearly from the way the light reflections wobble.

Noise on the spindle near the nose is kind of like a seesaw - a back and forth sound with each revolution. I don't think that is very promising.

At this point, I think I'll take a closer look at the chuck jaws and compound slide, and if I can't get more improvement there, I'm going to start wondering if that's the best it'll be until a more thorough rebuild can be done. I could always jump ship and start making music boxes now, though

Try turning some brass,or get a piece of free cutting steel.....your latest pics look like a result of false edge effect in soft draggy steel..

Last update from me: IT WORKS!!!

I tightened down the bearings even more today, and then hoping against hope I took the advice of Thermite from an old thread I dredged up. Apparently, in some cases if the bearings are repeatedly stressed and destressed, the can work the rollers back into agreement with each other. So I took some moderately heavy interrupted cuts in an old cast iron chuck backing plate that I wanted to convert to a face plate. I don't care if it's a stupid idea or not, because the repeated stress apparently shook the bearings back into compliance with each other.

The lathe cuts round again! WOOOOOOOOO!

As I mentioned,the bearings must be turning to adjust them,or you get a false setting........you can see this in truck wheels where a newbie adjusts the bearings tight,and next day there is a 1/4" endfloat and the brakes are full of oil.....Dont adjust loose lathe bearings all in one go ,sneak up on the correct adjustment after running the machine for a while.....this allows the new setting to wear in .....adjusting tight in one go is a sure fire way to damage the bearings permanently.