Trepanning and Parting Off

I think I've figured out a big piece of the mystery of what I generically called chatter in the "Clausing 5914 Chatter ..." threads.

Tightening the lathe dovetails up helped a great deal, and is a good idea in general, but didn't really solve the problem.

It came to me in the shower Wednesday morning - self-feeding would explain just about everything seen, and that if one reverses something, the sign of the mechanical feedback loop can be reversed. Now the feedback is positive, yielding the vicious loop that causes visible tilting, flake chips, and breakage. By contrast, negative feedback causes stability.

The key experiment was to try face grooving a 2.125" diameter 6061 aluminum slug held in the 4-jaw chuck, and using the BXA-6 toolholder with the huge overhang (~5" from cutting edge to toolpost center) but with the tool bit held upsidedown and the lathe running in reverse. Running backwards cases the self-feeding effect to change sign, becoming a self-unloading effect as the toolpost tilts away from the workpiece, reducing the bite of the tool bit.

Now I can peel very thin chips off the plate while face grooving, although it is still necessary to use the back gear, and there was mucho squealing. I did it dry, with too much overhang, and without properly shaping the tool bit. But it worked anyway. The difference is night-and-day.

So I now think it was the tendency to self-feed that caused all the problems, even if the self-feeding effect wasn't so powerful as to cause visible tilting of the toolpost. The squealing will go away with a stiffer method of holding the tool bit, and perhaps a better-shaped bit.

This also applies to parting off, and specifically explains why using an upsidedown blade coming towards the back of the workpiece, or coming from the front with the lathe running backwards, works. (BXA-7R)

I never really believed the theory that better chip removal was why upsidedown cutoff blades worked better, because I had problems even when there were no chip wads to be found and the groove sidewalls were clean. Actually, the galloping chatter tended to throw thick flake chips far and wide, so they were everywhere but in the groove.

To summarize, there are two elements that are necessary for face grooving, trepanning, and parting off on the Clausing 5914 lathe: Use the back gear (for tortional rigidity), and use an upsidedown tool bit (for stability).

As for use of a coaxial boring bar to hold a grooving/trepanning toolbit, it turns out to be a common method. I found it mentioned in US Patent 5,640,890 as prior art to be improved upon. The patentee is trepanning stainless steel hydraulic hose fittings in production, and needed a method that worked in a screw machine of some kind.

Joe Gwinn

Reply to
Joseph Gwinn
Loading thread data ...

[ ... ]

As am I. I've got the 5418 (similar bed profile, but manual change belts), and have very little chatter.

You know -- there is one other possibility which occurs to me. I presume that your lathe, like mine, has the hollow level adjusting screws in the feet of the bed through which pass the bolts to lock it down to the stand. If you have something like the near side bolt loose and the level adjusting screw a little clear of the stand, the bed would wind up under torque, which could give you similar behavior.

Did you get a proper sensitive level and adjust the bed to proper level at both headstock and tailstock end? (I did.) After that, did you make sure that the hold-down bolts were tight (and didn't disturb the leveling)? *This* may be where your flex is -- especially since you have a longer bed than I do (I believe), and thus more total flex possibility.

Get it chattering and feel each foot with the adjacent finger on the stand so you can detect relative motion between them. If you find any, you'll need to tighten the hold-down bolts, and probably re-level the bed.

Good Luck, DoN.

Reply to
DoN. Nichols

Do you mean the bolts between cabinet and the floor, or that hold the headstock to the bed? I assume you mean between headstock and bed.

I do have such a level (Starrett model 98-6), and did level the bed by adjustment of the leveling feet between cabinet and floor.

Hmm. This I did not think of, or check. But I will.

I don't see why bed length would matter, given that while trepanning all the action happens within a foot of the chuck face.

What I'm not quite sure of is which bolts you mean. On the 5914, the headstock is bolted to the bed with four solid 3/8-16 hex socket cap screws and two clamp bars, and the bed is bolted to the head pedestal foot, which is in turn bolted to the cabinet, all with solid non-adjustable bolts. The only hollow bolts I know of are in the leveling feet between cabinet and floor, which doesn't seem relevant to chatter (versus inability to turn a cylinder due to bed twist).

I looked the the 5418 manual. Now I understand. The designs of 5914 and 5418 are very different in this area. But there is no harm in making sure that all those bolts are nice and tight, especially those holding headstock to bed.

I suspect that on the 5914 they just machined the mating surfaces of cast iron components close enough that adjustment wasn't needed, and that the two pedestal feet isolated the bed and headstock sufficiently from the sheet metal cabinet and chip pan that adjusters were not needed. Unlike the 5418 manual, in the 5914 manual, there is no procedure for adjusting how the cast iron stuff rests upon the sheet metal stuff.

Joe Gwinn

Reply to
Joseph Gwinn

[ ... ]
[ ... ]

I see that you figured out what I was talking about from examining the manual for the 5418. Since I don't have the manual for yours, I had not realized how different they were.

Hmm ... not nearly as sensitive. The 98-6 (and the rest of the

98 series) have a sensitivity of 0.005"/foot, while the No. 199 "Master Precision Level" has a sensitivity 0f 0.0005"/foot -- ten times the sensitivity. But, of course, the 98-6 gets you close enough if you then do the "turn two rings on a single bar and measure them" operation afterwards.

O.K. As long as you are that close, it would not really matter. Where it would matter is turning some distance from the headstock, or running something like a 1" drill bit in the tailstock -- especially with a long workpiece extension from the chuck.

O.K. Quite different. The 5418 is supposed to have the cabinet bolted firmly to the floor (I don't), and the bed leveled relative to the stand and chip tray.

Agreed. Check the tailstock end too -- because that is what would flex when the torque is transmitted from the spindle to the carriage. It would wind up the bed (a little, at least).

BTW For parting there used to be a gooseneck parting tool. It went in a lantern style toolpost, came out, turned up, formed an Omega shaped arch, back down to where the parting tool is actually clamped. The result is that excessive cutting forces tends to move the tool away from the workpiece, thus eliminating chatter. But your machine *should* be rigid enough to not need this, especially with an Aloris style quick-change toolpost, and the T-profile parting blades.

Perhaps the bed feet were machined flat on the bottom, then bolted to a thick flat plate (say 4" thick or so), machined flat on the top, the bed mounted on this, and the ways finish-ground this way, so it could get all the twist out of the bed by adjusting the cabinet feet.

Hmm ... another thing to check. While you have chatter, see whether there is any relative motion between the headstock and the bed. If it is not clamped down firmly enough, or if there are chips trapped between the headstock and the bed, that could introduce enough give to create problems.

O.K. This assumes that the sheet metal stuff is rigidly bolted to the floor, I guess.

Good Luck, DoN.

Reply to
DoN. Nichols

Ahh. Yes. It will take a while, but I may scan my manual, which does not appear to be copyrighted.

I have not yet done the bar turning test, but intend to, once the more immediate problems are at least understood if not resolved.

I probably should repeat the leveling exercise, as based on the other kinds of ignorance-induced problems I've already found, it's likely that the lathe was left unleveled and thus twisted for years, and may have taken a set that needs to relax out.

While using the 1" drill, the bed will most certainly wind up a bit. I guess that what would resist permanent twist would be the leveling, with the lathe resting on a concrete floor, as it now does.

The 5914 manual also speaks of bolting to the floor, but few people actually do this I suspect.

The tailstock now clamps pretty firmly to the bed. This is one of the first things I cleaned and adjusted, mainly because it was easy and could be done without the manual.

I vaguely knew (from reading old books) that there was such a tool, and that it involved a gooseneck, but I had the "picture" upsidedown in my mind, and couldn't see how it would work. I just googled it, and found a book from 1910 that explained the principle as applied to use in a planer. They did understand the self-feeding effect, saying that the gooseneck would eliminate gouging the work, but a sufficiently rigid machine didn't need gooseneck tools.

It occurs to me that a negative rake cutting tool could be set up to develop enough outward force to at least partially counteract the tendency to self-feed. Perhaps this is why one commonly stated cause of chattering is too-sharp tools. When one carefully blunts such a tool, one puts a little bit of very negative rake right at the cutting edge.

It's certainly a plausible approach, at least for the rough machining. But the machining will release stresses, causing warping when the bed is unbolted from the 4" slab. The bed ways are ground, so what may have been done is that well-aged castings were rough machined and then further aged and/or normalized, and then were finished on a surface grinder. The exact procedure may have been a trade secret, as it was exactly such recipes that were the secret sauce of lathe manufacture.

Hmm. Offhand, I don't see any safe way to do this, as my hand would need to be very close to the spinning chuck, or to various gears and the like.

I think I'll just torque all the clamp bolts, to see if any are loose. I doubt that the headstock was ever unbolted, at least not by the people who caused all the ignorance-induced problems. Taking the headstock off would have frightened them, at least one fondly hopes that it would. In any event, chips can only get in there is the headstock is unbolted.

The headstock is located on the bed ways by careful fitting of headstock bottom to ways, plus a pair of steel dowel pins to prevent sliding.

That would make sense. In the 5914, these adjustments have moved to the cabinet-floor feet.

Joe Gwinn

Reply to
Joseph Gwinn
´*x?¥Z]?Û?}^þ?©ºUÖnBrW±¤è#?³²{£?-%*?û0??ÂKÑu|Îé??åRv*zP?XLOO?>ÝÃ7µúÃÕÕÓÅÕãÅÕæêï?¦P_ïl]«ÏÊîÅÏÎløå/?«2í»emºÏ6Ý µk]g?Ïþdÿ,_oj¥ÛÎf¥?e¾lëÍÃò?GËú*_æµÍ Wú¿¼?®õ2_è:_äKåêù?.í?Qê³½ë^|å¾[ªïºðD$¢î? J:ü,¯¼ùï{$ïô?n{µÜÿ|?ù?`vö¿j¹\ªÿ?I?g×^éJÝ,?ºypkÔÆuª+?züèáSuîmeK??ÉU?µ-Í\­úNUºîuy¯Òò!+t½1XWvþb® ?*4Ä?v¯JÛu°Þé:Ó.¤=Hgøõ7Ç%?¹^mk·S?Ôe½rµQ?ßT?+ UöjW??ÊeYßzÕ9U?c?70?ñ}e UwjïúV?2ç8×?ÊÖøXh/f?{JèP?[S*ÿÜûÎ?ÓÒ}?W¶?Õkc:åÖò?.è?Öõ?"*Üh6Y9X???.Û*??»]͹Ê#sº}ÍC÷xW?®??á,|­6ð¿·yØ[ YNo@Gs??s)=¸¾¤Ôx2Qd>rçú2?½ò2W}£ú:?;×~èaï`??KÕÆâ$±"???ƾ6?¯?¯±??ØNÍT¯ cò?B)>º? P ¶Å ???T^í?§\?êÝc&üÐÛ?D9V9éÛþoÐâ µéx¬®G? U)ÜØÓ? §%?G? ¦µP?{ÌÐÛÉ?°êZ?)ú?{ø(µ º/C>?µ??=Ð/ýL³1?x"Pg]/ ? VAOÔ3Áº¸ãH?øýX?ÊlÀ. Ýð?6+æ¡R§¹æxìCõt??Ú}!?³ÝRrSf¨ ,pY?@)÷Èc«¦Ç.V[ÙZÃc?´m?á$¥Ö÷?ÏÒ|
Reply to
DoN. Nichols

...

...

Just for reference, here's a working link to the same page, with non-essential parts (ie, the last two-thirds) cut away:

formatting link
The "Download PDF - 36.3M" link on that page fetches a pdf file with 1100-odd scanned pages from what appear to be pamphlets #41 through 110 of Machinery's Reference Series, ca. 1908-1913.

Also, regarding cut-and-paste of broken-up URL's like that, I usually highlight the whole mess, paste it into Firefox's URL box, then delete the return characters to fix it, as opposed to cutting and pasting three separate parts.

-jiw

Reply to
James Waldby

That's right.

The cut&paste exercise is not necessary if the poster provided the start and end brackets needed by the newsreader (or email client) to know exactly what is the URL string.

I do provide those brackets, so when you see something like "", just click on it, even if it wrapped to multiple lines. It's the < and > that do the job.

The exception is when the URL is in a quoted message, and so extra quote mark characters have been inserted.

Reply to
Joseph Gwinn

That's actually a valid link.

Reply to
Jim Wilkins

OK. I'll probably post it to the metalworking archive.

Well, if I do bolt it down, I'll probably do all four feet. But I'm reluctant to bolt it down, because I may move it. Nor is a basement slab all that thick and rigid, compared to the floor in an industrial plant.

I'm still looking for a faceplate. One can buy a new faceplate plus L00 back from Bison for about $300, but the need is not immediate, and so I've been watching eBay.

In looking at the 5418 manual, I see that it was also sold without cabinet stand, for mounting on an owner-provided bench. Thus, the adjusters had to be between cast iron stuff and the base (bench or cabinet).

Ahh. Now it makes sense.

In the original posting, you should be able to just click on the URL, even though it appears wrapped, because I provided start and end delimiters (< ... >). In quoted postings, this won't work, and editing is required.

I have the gibs adjusted so there is much drag, but the cross-slide backlash is about 0.020", and the compound slide backlash is about

0.006".

The compound slide backlash was ~0.055" when I got the lathe, and so I replaced the screw and its bronze nut. What a difference it made. I have no idea why it was so worn compared to the cross-slide; perhaps someone replaced the cross-slide screw and nut.

The cross-slide wasn't too bad (after being cleaned and adjusted), so I left it alone. At least for now.

Ahh. I see what you mean. One can get at the back of the lathe safely enough. The problem is that I cannot then reach the handwheels to advance the toolbit into the work. I'll have to cogitate on this.

It's possible, but I'm pretty sure that I am the first HSM owner of this lathe, so this is probably its first basement gig. And my basement has level entry, so no disassembly was needed.

What I called dowels are in fact #6 taper pins, and they fix the headstock to the bed way, preventing motion along the bed way.

There are in addition four pointed bolts that attach the bed to the cast iron pedestal foot. I thing these studs are pointed for ease of manual assembly in the factory.

Yes. Not that the defense is perfect. I have pulled swarf out of just about everywhere.

Joe Gwinn

Reply to
Joseph Gwinn

So it is. Is nothing safe?

Joe Gwinn

Reply to
Joseph Gwinn
[ ... ]

O.K. That (the full PDF) is dowloading as I type.

O.K. Possibly other intersting things in there, then.

O.K. But I normally don't have Opera up and running, so I invoke it by a script "op" which does a cd to a place where I tend to save things, then invokes opera via the URL which I cut-and-pasted into the comand line -- and the command line sees the newlines as end of a command line and the start of another, which of course fails to reach the place desired and then tries to interpret the other two as separate commands, which also normally fail. :-) So I have to cut and paste it one line at a time, and surround the whole thing with double quotes to keep any stray '?', '%' or similar characters with special meanings to the shell from having those meanings.

Obviously, if I used a web brower for a newsreader I wouldn't have this -- but I would have other possible problems.

Thanks, DoN.

Reply to
DoN. Nichols

[ ... ]

1122 pages worth of them :-)

That only works in a newsreader which is designed and configured to directly invoke a web brower -- or which is part of a web browser. It does not work in mine, and that is the way I want it -- so a malicious URL can't be accessed without my active participation. And yes, I *have* seen malicious URLS in usenet articles.

Or -- when you are using a system which is not configured to do it, as I am. I *think* that I could configure this newsreader (slrn) to do it, but I will not.

Enjoy, DoN.

Reply to
DoN. Nichols

You mean to the dropbox ? That should do -- or perhaps to the archive of manuals which Iggy is building up.

[ ... bed twist ... ]

O.K. Mine is not bolted down -- but I don't see the kind of chatter that you do. Perhaps it is because most of the industrial service life of the machine was as a turret lathe, where most of the motion was in the turret ram.

[ ... ]

O.K. So was I -- until I found one at a flea market/picnic held by the local metalworking club (which, BTW, is also where the dropbox is. :-)

With a somewhat different headstock which allowed the belts to run back to a rear-mounted motor instead of down into the pedestal.

Yes -- but before that, it was common to adjust between the feet and the cabinet (or table) with stacks of shims.

[ ... ]

Good.

[ ... ]

Nope! *I* can't -- because I don't use a GUI-based newsreader. I can click on the url and get at most a single line of it to paste to a browser or another window with a command line.

People with GUI-based newsreaders (especially those built into web browsers) can do so. Frankly, I am glad that I can't, because it makes it more difficult to accidentally click on a malicious link.

[ ... ]

O.K.

Well ... my cross-slide had 0.070" backlash. But I know why. The power feed was used for parting off workpieces after the turret does most of the work of shaping them. So the carriage just sat there, the compound was untouched, and the cross-slide cycled under power for every part made.

O.K. But when you are experiencing chatter, you might also try placing a finger where it can sense motion along the cross slide.

Isn't there access above the quick-change gearbox? The gearbox is bolted firmly to the bed, so you could sense motion between the top of the gearbox and the front of the headstock.

[ ... ]

[ ... ]

O.K. That helps greatly.

O.K. That is as it should be.

Hmm ... I've not checked how my bed feet are attached.

[ ... ]

It does tend to get everywhere.

Enjoy, DoN.

Reply to
DoN. Nichols

Yes -- the Clausing carriages have wide wings.

[ ... ]

Actually -- the wear is normally on the harder of two mating surfaces. This is because harder objects (grit, etc) embed in the softer surface and cut the harder. This is how a lap works.

Enjoy, DoN.

Reply to
DoN. Nichols

...

...

Add the following near the beginning of the script: URL="`xclip -o | tr -d '\n'`"

This will set URL to the contents of the clipboard with newlines deleted. Adapt as necessary for shells other than sh or bash.

-jiw

Reply to
James Waldby

I'm not sure that cast iron is soft enough for much embedment to happen, unless one is using a toolpost grinder. And while the gibs are visibly and heavily worn, the dovetails are somewhere between pristine and lightly worn to the eye.

Anyway, the gibs both had something like 0.003" to 0.004" of wear on one side and another 0.001" on the other side. This doesn't sound like much, but it approaches to exceeds the adjustment range of the gib. For example, given a taper of 5/32" per foot (0.013021" per inch), 0.003" of wear allows the gib to move 0.003/0.013021= 0.230", a major fraction of the adjustment range. This plus the bunged corners forced me to flip the gib over. However, now the gib rests at an angle on the dovetail, as the surface is stepped (the wear not being uniform, instead leaving a step running the full length of the gib). So, I threw in the towel.

Joe Gwinn

Reply to
Joseph Gwinn

I meant the dropbox, although Iggy would be welcome to a copy as well.

OK.

It helps to be lucky...

Yep. I wanted a lathe with integral cabinet stand, on the theory that it would all fit together better.

If you are not on Windows, why would this be a problem?

I figured out how better to adjust things, and got this down to ~0.003".

Yep. I tried the same improved adjustment method on the cross slide, but the backlash remained at ~0.020", so this is probably inherent to the screw and nut.

Did you ever figure out how much of the 0.070" was the screw versus the nut? I think I recall you saying that the Acme threads were worn down to sharp points, which would imply that screw wear was the larger contributor.

Not without some disassembly.

And it turns out that there is only one taper pin, not two, so the two models are about the same here. I recalled that there were two, but when I cleaned the dirt off, only one.

In the shoes is the worst. And chattering tools generate particularly annoying swarf - millions of sharp little steel needles. Coolant helps by wetting the whole mess down, keeping it from blowing around in the shop.

Joe Gwinn

Reply to
Joseph Gwinn

Also right. I wonder if anyone offers this on CDs.

What's wrong with newsreaders that correctly handle URLs? If one is not on Windows, there is little danger from URLs.

Joe Gwinn

Reply to
Joseph Gwinn

Katana:dnichols 0:53:00 > which xclip xclip: Command not found.

How about for systems which don't have xclip? :-)

O.K. I've found the source, compiled and installed it without any drama. Now to see what I can do with the script modification.

Hmm ... "Reads from standard in, or from one or more files,"

But the script is invoked as:

op URL-goes-here

and when URL occupies three lines, this would still be:

op FIRST-LINE-OF-URL (opera barfs because of incomplete URL) SECOND-LINE-OF-URL (shell barfs, because it is expecting a new valid command) THIRD-LINE-OF-URL (shell barfs, because it is expecting a new valid commnd)

So -- unless you can see a way to drop the cut-and-paste as standard input for the script, I don't see how this can work.

Still -- xclip looks like an intersting tool to play with.

Thanks, DoN.

Reply to
DoN. Nichols

PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.