View Full Version : Cuts going forward AND on the way back out?

04-25-2004, 06:32 AM
......My lathe does this. Usually only on heavier forward cuts. If I don't back the tool out but merely wind the carriage back the tool will take another small cut along the work.

I figured this might be an indication that the saddle was rocking across the bed (not up and down) causing the tool to advance into the work?

The lathe is an 11x36" Logan/Powermatic of 1981 vintage with hardened ways and in truly great shape. The 'texturing' or whatever it's called is still visible on the 'V's and flats, even up close to the headstock.

When I got the lathe in mid January of this year, after leveling I cut a test bar which came in showing only 2 tenths OOR over the 11" length I cut, and was only 7 tenths larger in diameter 11" away from the chuck.

I've tried lifting the carriage, and tried rocking it across the bed by holding opposing corners and I get no discernable movement.

Is what I'm seeing here the result of springback, or spring-away from the bit by the work while cutting toward the headstock, that's subsequently being cut when going away?

When making light finishing cuts it doesn't happen. Am I worrying about nothing? Is there another way to check the saddle for movement or rocking across the bed if that's a possible problem?


04-25-2004, 06:57 AM
The "texturing" you see on the lathe bed is scraping, once done by hand but I believe now done by machine is done to make it flat and parallel. What you are experiencing is "spring back" which is movement and/or deflection of the piece and the machine. No matter the size of the part or the machine they both bend or move when cutting. On the movement back after the cut the deflection is much smaller therefore a smaller cut is made. I sometimes take a free cut (a cut taken without moving the cutter into the piece) when I do more accurate work to ensure what I dial in is what I take off. Ensuring everything is tight and as rigid as possible will help to minimize this effect. This includes steady rests or follower rests if needed. The slower the feed (to a certain point) and the sharper the cutter will also help. Unfortunately sometimes it cannot be avoided and a free cut will solve this problem. And remember to make lots of chips!!


04-25-2004, 07:05 AM
It's deflection. No big deal, just keep it in mind when doing heavy cuts. I ease up about 50 thou under, then make a pass or two without changing the depth of cut. Then I check size and make a few fine passes to size.

Springback is when you make a cutting pass, and upon making another pass with out changing the tool depth, the tool continues cutting.

I wouldn't worry about the carriage. It sound like it's in your tool holder. If your're using a lantern style post, that's the likely source. This typically shows up when using a boring bar in a heavey/deep cut.


04-25-2004, 08:07 AM
As others have said, it's normal, giving one a greater appreciation of how rubbery and non-rigid metal really is.

Learning to deal with the phenomenon of spring-back is all part of the education in becoming a skilled machinist.

04-25-2004, 08:38 AM
Like the others said "normal".
Just remember to take the same depth of cut on the second last pass as what you expect the last pass to be. That way you have the same amount of deflection on each pass and can better predict how the diameter will end up. Example: if you have .020 left then cut only .010, then measure, if now only .009 left set for that. The deflection at nine won't be significantly different.

The same thing has to be dealt with on high priced cnc lathes. If the work piece has very tight tolerances I will leave .060 on then take .020 at a time. That way the last two passes should be nearly identical.
Super Dave

[This message has been edited by suprdvn (edited 04-25-2004).]

Forrest Addy
04-25-2004, 09:56 AM
Old time jig bore operators had a left index finger strong as trailer hitch from hooking it on the boring bar to deflect the tool away from the bore so he didn't scratch it on withdrawal.

Someone suggested everything is made of rubber, that is even the most rigid of structures will deflect under load to some degree. Skinny boring bars are a prime example.

Here's an experiment: Put a dial indicator on the end of your best boring bar extended about 6 diameters and push against the bar with a single finger. Surprising isn't it?

Deflection under load AND gravity (like taxes and speed traps) are everyday problems for the machinist.

I used to be a Nuke machinist which meant you did the same work as a regular machinist but dressed in a anit-C's, a wet suit, and air fed respirator, cotton gloves, and three or four pairs of rubber gloves in a containment on machine tools hwere all the controls are inside glovebags and other hindrances. We had a 3 Lucas horizontal we did valve work on. Anything nuclear is worked out in advance to the smallest detail complete with pictures. It's a great program even is it is a bit stifling with an inspector looking over your shoulder and verifying your every move.

The problen was to machine a weld remnent off a valve body canopy seal. You had to dial in to the valve internals and the instructions showed a Starret magnetic base and a Last Word indicator stuck out full length from the spindle face. I knew from long experience that this procedure would result in the work bore being 0.004" to 0.005" below the spindle centerline. The instructions allowed no latitude - this was an old instruction that had been in place for almost ten years. Hundreds of valves had been weld repaired by it. It was cast in concrete.

I had an out. I wrote a report against the instructions. These reports are standard procedure but worker level people seldom have need to write one. This brought the job to a halt until the report was resolved.

That afternoon all parties attended a meeting. The nuke coordinator's deputy said "your meeting" and handed me the podium. I was prepared. I referred to the instructions, stated my objections, and passed around my "show and tell" a 16" length of 3" square tubing with a 1/4" wall. On it I stuck the same type mag base and test indicator in the same manner as required by the instructions.

I demonstrated indicator deflection by setting a good zero with the indicator right side up and then turning the whole thing upside down to show what happend when the indicator drooped away. Eyebrows went up on foreheads all around the meeting as this demo was passed from hand to hand.

"Seems like a non-problem" said one engineer. "There's a theoretical problem none of us anticipated but we have a ten-year history of no difficulties." "I wouldn't say that," said the weld engineer. "We have to get approval to grind the new canopy to fit on every one I've dealt with. Takes about two weeks and about 20 hours of paper to do it." "So how do we fix it?" said the project engineer.

I was ready and showed off a short piece of 1" bar stock designed to go in the boring head that had a short extension turned down to take the clamp of a Last Word indicator. The rest is history; I did the job according to revised instructions and forgot about it.

A couple months later I got a meeting with the nuke superintendent and my boss and his boss who were grinning and nudging each other but were otherwise making sure I knew they had a big secret they were keeping from me. I got a speech, a handshake, and a certificite and a check for $3,000 - 7% of the savings associated with fixing the droopy indicator over three years system wide.

So that's the lesson: keep rigidity and gravity in mind. It will make your work easier and save you time and money.

[This message has been edited by Forrest Addy (edited 04-25-2004).]

04-25-2004, 10:03 AM
Nothing to worry about,I'll turn something in our 24x120 at work,then I'll turn something else in our 14x54 and forget I'm not using the bigger machine and get a stripe on the back feed.Then occasionaly I'll go over to a friends house that has a Little Logan 9" and REALLY forget it doesn't wiegh 5,000lbs make .050 facing cut and it pushes the carrage back about .050"

It depends on the rigidity of the machine,the rigidity of the part/setup and the ondition of the tool itself.

It helps if you use a roughing tool thats free cutting and swtich to a finishing tool with a fresh edge for the money pass.

04-26-2004, 09:01 AM
Just thought I'd drop in here and say how much this neophyte has learned from you fellows on this forum in the few days I've been lurking and reading. This thread explains a lot! http://bbs.homeshopmachinist.net//smile.gif

Thank you!

04-26-2004, 03:16 PM
There's a phenomenon I'll call spring-forward, and no, it's not related to saving daylight. Many times I have made cuts and then found that on the next pass, the cutter is clear of the workpiece. I have to dial it in more to take another cut. Depends largely on how the compound it angled. So much to learn.
SuperDave has an interesting point. You may have to take a finishing cut, measure what was actually removed, take another finishing cut, measure again. Many factors affect how much material comes off with a given depth of cut. One of the factors is the finish left by the rough cut. The first finish pass will remove a slightly different ammount of material than the next finish pass, even with the same depth of cut dialled in. Things change also with the diameter of the workpiece, and the type of material being worked. Compare machining a grade 8 bolt with a piece of leaded steel, and checkout some plastics as well. Turning a piece of nylon will be interesting.

04-26-2004, 08:16 PM
........I appreciate all the replies and suggestions. I figured that rigidity would be an issue. In addition to the regular crossfeed and compound, I have a dual toolpost cross slide I planned on using with the Palmgren. It's 3/4" shorter heightwise then the compound's 'T' slot, plus it has full length dovetails vs the other's half length dovetails.

Even so, the overhang of the milling attachment is at issue. So light it is and easy does it. Better then nothing in any event!

As far as a drawbar goes, could I use a length of 3/8" allthread from the endmill holder, and then through a piece of steel laid across the collet closer coller on the backside of the spindle? Put a nut on the rod and tighten it up to secure the holder?

Halfassed but workable, or forget that idea? How about machining a insert with a ledge to bear against the collet coller with a hole in the center for the allthread and nut? I'm just trying to keep from spending more money :-).


04-26-2004, 09:21 PM

FWIW I use a drawbar like you describe when I need to hold end mills in the lathe. I don't use end mill holders, (would if I had them), but 3MT collets direct in the lathe spindle. So far I've never had an end mill work loose either in the lathe spindle or the mill drill which uses the same system, provided the drawbar is tight. The drawbar I use in the lathe is just a piece of 3/8 or 1/2 in. allthread the correct length. On the outer end is a tapped stepped flange which locates the drawbar in the end of the spindle and bears on the outer end of the spindle. Outside that is a 3/8 locknut locktited onto the allthread which is also used for tightening the drawbar. Note that the thread in 3MT collets is not standardised - don't know about end mill holders. My imperial collets are tapped 3/8, but the metric ones are 1/2 x 12 TPI Whitworth, so two different drawbars are needed.