Index Home About Blog
From: Robert Bastow <"teenut"@ hotmail.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: Choosing a new 16" gap bed lathe
Date: Wed, 02 Feb 2000 02:13:24 GMT

Reality Check time!!

I do hope you have lots of money.

A new lathe that will hold, and KEEP holding, those limits is going to cost a
LARGE bag of gold.

I am not sure that anyone has ever made a gap bed lathe of that size that can
hold what amounts to half of Schlesinger limits for a (Small) tool room lathe.

A well built CNC Turning center could hold this..mainly by dint of built in,
programable compensations for geometric deficiencies.

Now, What on earth do you want to make, of that size, that has to come off the
lathe (as opposed to say, a cylindrical grinder) with those sort of limits
"Guaranteed"

"J. Kimberlin" wrote:

> Need to buy a new 16" gap bed lathe at least 50" C-t-C capacity.
> It has to be guaranteed to 0.0002" accuracy.
>
> So far the Harrison/Colchester people seem to have a good
> product.  Anyone out there want to comment on the Harrison or
> suggest an alternate brand that meets/beats the requirement?
>
> TIA
>
> JerryK

From: "John (EBo) David" <ebo@eagle.west.asu.edu>
Newsgroups: rec.crafts.metalworking
Subject: Re: accurate hole technique
Date: Sat, 20 May 2000 21:54:32 -0700

Dave Baker wrote:
>
> >Looking for tips to make accurately sized and located holes on a BP.
> >Tolerance? As tight as I can make it.
> >
> >Locating is not to bad - I use the plug method to locate the spindle in
> >relation to the work. But no matter how nicely I locate, I cannot get a
> >'true' hole. I can get a nicely sized hole with a parabolic drill or a
> >reamer but the holes usually wander somewhat no matter how carefully they
> >are center drilled and then step drilled up to size, even using screw length
> >parabolic drills. I find that a 3/8" hole 1.2" deep can easily wander .0005"
> >or more, both axially AND from it's spotted location. And yes, I've set the
> >spindle vertical to my fixture within .0001" or so.
> >
>
> Not asking for much from a Bridgeport are you !! Half a thou over 1.2 inches -
> is this a piss take or a serious question ?
>
> You can get more run out on an R8 collet than that, plus the entire headstock
> will move relative to the table with hand pressure on the quill feed. And if
> you've set anything on a mill to within .0001" of anything else you're a better
> man already than anyone else on this group. I regularly set the head
> perpendicular to the table for flycutting and it's hard enough to get within
> half a thou - everything moves when you tighten up the bolts anyway. I bet it
> moves more than a tenth of a thou when you scratch your arse too close to the
> table or stare at it hard.

I've only *ever* worked on one milling machine that was capable of
producing this kind of precision repeatably.  That was a brand new
Deckel FP3NC (that is a 3 axis CNC mill that sported both horizontal and
vertical heads).  While the sales lit. boasted that it could hold .0001
over the ~1.5 cubic feet, my old mentor never believed it until one day
we got one of those odd little jobs that seemed to always find their way
to our shop...

We had a client walk in the door that wanted these two 10x14" plates
milled flat to .0001" with some groves, etc. milled in them.  Carl told
them that he could not guarantee that he could .0001" across the 14x10"
cast aluminum plates, but he was willing to try.  As he was the only one
in 500 miles that would even discuss it, they gave him the job.  As I
recall he started by heating the barrel of flood coolant and flooding
the part for awhile to get a predictable temp.  Then took several rough
cuts and removed the piece from the fixture each time to let it move if
it was going to.  When he was done, he checked it on the surface plate
and as far as he could tell it was good, but .0001" is pretty darn hard
to hold.  So, he sent the piece off and was hoping that it passed.

He gets a call the very next day from their QA person who starts off "I
though you said you couldn't hold a tenth on those plates?"  Carl
explained that he wasn't sure, but he did his best. the QA man then
explained that not only did he hold the spec., but the error between the
two plates measured out at .00005"!  I never did figure out how he did
it, but we trusted the machines capabilities after that.  But then
again, that was a brand new machine that was pampered since the day it
was uncrated, only had a 12x12x16" travel, was built like a tank, and
cost $80,000 new.  While I love Bridgports, I would never expect even a
new one to hold less than a thou.

  EBo --


From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Thu, 15 Feb 2001 18:28:11 GMT

Most machine tools have ground bedways today. To answer the old complaint
that ground beds don't hold oil as well as scraped ones, many manufacturers
apply a light scraping to the surface, done without gaging but only as a
superficial treatment, called "flaking" or "frosting". On mass-produced
machines, this is often done with a Dapra power scraper.

However, grinding offers no advantage in accuracy. In fact, when you get
down below 0.0001"/ft, production grinding won't do it. You have to scrape,
and that's what's been done for years on the best grades of toolroom
machines. Someone mentioned Moore jig grinders or jig borers. They were
hand-scraped to 0.00004"/ft, or 0.00002"/ft on their measuring-machine-grade
machines. These machines were used by many countries' bureaus of standards
for certifying gages and instruments. When you work to these levels of
accuracy you run into another problem, which is induced stress. Grinding a
bedway actually stresses the surface, and the iron won't be perfectly
stable. In ordinary machining it's irrelevant, but, at the outer reaches of
accuracy, it can be an issue.

With scraping you can check your work as you go with straightedges and
surface plates, which are self-checking instruments. In other words, they
can be made in such a way that they don't have to be calibrated against
another instruments. So you can make them as accurate as you want, or as
accurate as you have the patience for, without referring them to some
*other* instrument, which would involve a chain of calibration back to a
national standard. That's expensive and tedious.

A five-foot way grinder will cost upwards of $200,000. A hand scraper costs
around $30. That can be an issue, too. <g>

Ed Huntress



From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Thu, 15 Feb 2001 20:52:45 GMT

>>They get around the stress issues in the hardened prisms they use as ways
by bedding them in accurate scraped ways, by grinding the prisms in
temperature controlled oil, and by hand-lapping with a master lap after
final assembly. The said the lap was good for maybe one or two machines
before it wears out.<<

That's the way Moore made the most accurate machine tools ever built -- at
least until the '70s. I don't know of anyone else who did it that way,
before or since.

Hang on to that book. There's nothing like it anywhere. It went out of print
once, and it may not always be available. It's like Howe's books on
gunsmithing, in one sense, only without as much instructional material.

Ed Huntress




From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Fri, 16 Feb 2001 01:00:22 GMT

>>Where can this book be obtained-if at all???<<

The book is "Foundations Of Mechanical Accuracy" by Wayne Moore. Originally
it was vanity-published, I believe, for Moore Special Tool Company of
Bridgeport, CT. The company name is now Moore Tool Company. The book *may*
have been republished by a commercial publisher, or else Moore himself
republished it.

If you don't find a listing for it on Amazon or your bookseller of choice,
I'd contact Moore Tool Co. and ask them.

Ed Huntress




From: Bucky Goldstein <bg@netcare.lucent.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: 16 Feb 2001 04:44:37 GMT

Ed Huntress <mike234@bellatlantic.net> wrote:

> The book is "Foundations Of Mechanical Accuracy" by Wayne
> Moore. Originally it was vanity-published, I believe, for Moore
> Special Tool Company of Bridgeport, CT. The company name is now
> Moore Tool Company. The book *may* have been republished by a
> commercial publisher, or else Moore himself republished it.
>
> If you don't find a listing for it on Amazon or your bookseller
> of choice, I'd contact Moore Tool Co. and ask them.

See

	http://www.mooretool.com

There's a listing for the book and it doesn't say "out of print"
next to it as it does next to some of the other titles.

--
Bucky Goldstein

	Be happy.  Be root.


From: Pete Somebody <pete@netcare.lucent.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: 22 Feb 2001 13:50:35 GMT

Ed Huntress <mike234@bellatlantic.net> wrote:

>> There's a listing for the book and it doesn't say "out of
>> print" next to it as it does next to some of the other titles.
>
> That's good to know. When it became difficult or impossible to
> obtain, around 1980 or so, a lot of people squawked. I was
> looking for it today at my publisher's office. We don't have
> it. 8-(

The following appears on

	http://www.moretool.com/about.htm

**********

 PUBLICATIONS

 Contact sales@mooretool.com to purchase any of the following books:

        "Precision Hole Location" (Not available)

             J. Robert Moore - 1946

        "Holes, Contours and Surfaces"

             Richard F. Moore & Frederick C. Victory - 1955

        "Foundations of Mechanical Accuracy"

             Wayne R. Moore - 1970

			 Classic text on precision engineering with over 550
			 photographs and engineering drawings.  "Foundations
			 of Mechanical Accuracy" has been translated into
			 seven different languages with over 15,000 copies
			 sold.

**********

So there ya have it.  They're not exactly forthcoming when you
contact them, though.

--
Pete Somebody

	Be happy.  Be root.


From: Pete Somebody <pete@netcare.lucent.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: 22 Feb 2001 14:06:34 GMT

Pete Somebody <pete@netcare.lucent.com> wrote:

> So there ya have it.  They're not exactly forthcoming when you
> contact them, though.

Wait, wait!  I spoke too soon.  I left Moore a message yesterday
and just got a call back.  So okay, here's the real deal:

Precision Hole Location is unavailable, period.

Holes, Contours and Surfaces is available, it's in stock, it
costs $49.

Foundations of Mechanical Accuracy is available, in stock, $97.

Shipping within the US is $10 per package.  If you order both
books at once, you only pay one shipping fee.

No credit cards.  You have to snailmail a check for book(s) plus
shipping to

     Moore Tool
     800 Union Avenue
     Bridgeport, CT 06607

Be sure and direct your letter to the attention of Ruby Martin.

There.  When I told Ruby the kind of money people have
occasionally paid for the Foundations book, I think she almost
fainted.  "Well, tell them to send me a letter and I'll send 'em
the book!" sez she.  "Yes, ma'am, I'll do that, ma'am," sez I.

--
Pete Somebody

	Damn.  There goes another hundred fifty bucks.


From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Sat, 17 Feb 2001 03:05:32 GMT

>>I just bought the book, it worth the money (approx $109).<<

Ouch! Have they no mercy? It was $45 last time I looked. But that was 1977.
<g>

Keep it in good shape. It will be valuable some day. The photos, BTW, were
shot by a former LIFE magazine photographer, who went out with the wash when
LIFE folded for the first time, around 1971. It's some of the best black &
white photography I've ever seen in manufacturing, along with that of Joe
Ruskin, who did award-winning photography for "American Machinist". Those
days are over.

Ed Huntress



From: Rich Williams <rwilliam@engr.wisc.edu>
Newsgroups: rec.crafts.metalworking
Subject: Moore Books  was Scraping - Why, when you can grind?
Date: Mon, 19 Feb 2001 10:03:32 -0600

Precision Hole Location   Pub. by Moore Special Tool Co.  copyright 1946
Holes Contours and Surfaces ,  Moore Special Tool Co. copyright 1955
Foundations of Mechanical Accuracy, Moore Special Tool Co. Copyright
1970

All appear to be self-published by MSTCo.  Together, they are quite a
trilogy.  I've been to the Alibris site, and if the prices they quote
for books holds true, I'm gonna retire tomorrow.  I think that the
asking prices in most cases are just outa sight.


Rich Williams

Ed Huntress wrote:
>
> >>I've always been fond of the photos in Moore's first book  "Precision Hole
> Location"...<<
>
> AHA! That's the title I've been trying to think of. Do you have the book,
> Rich? If so, would you tell me the publisher and date of publication?
> Thanks.
>
> Ed Huntress


From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Thu, 15 Feb 2001 21:00:36 GMT

>>when a high spot is encountered with a grinder two things occur, the
spindle deflects and the rate of wear on the wheel increases. Of course,
after every pass the high spot is less high, but it will be there. Shapers
and planers actually create a straighter surface than a grinder.<<

This may be true with production grinding, but in precision grinding you
always spark out, so that any high spots get ground flat with the rest of
the surface. Of course, machining bedways is a production job. But someone
may get the impression that a shaped or planed surface is flatter than a
toolroom-ground surface. That's not the case.

My former company made what is probably the most accurate surface grinder
ever built, the Wasino Wing Ace. We could hold 4 millionths of an inch per
foot flatness (0.000,004", 0.0001 mm) with that puppy over almost any
surface. You can't scrape to 4 millionths.

It was a nice toolroom grinder. However, at $350,000 each, we only sold four
of them in all of North America, and two of them went to Kodak.

Ed Huntress



From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Sat, 17 Feb 2001 02:55:42 GMT

>>But indeed nowhere do they simply slap any of their parts onto surface
grinders and then call them 'done.'<<

Speaking of which, does the Book tell you how they accurized leadscrews? I
don't know where I got that info, as there were two previous books by
Wayne's father (Richard Moore) and I had a ton of literature from them when
I covered toolmaking for AM.

Anyway, they cast a lead lap around them, had a lead profile made and
enlarged as a full-size graph, and then pulled and pushed the lap as the
leadscrew turned, following the graph, in order to lap it to near
perfection. You have to do this in both directions, as entirely separate
operations.

Thus, the cost of a Moore Jig Grinder or Jig Borer. Thus, the 36 months
waiting time. Thus, the world-wide reputation.

Ed Huntress





From: jim rozen <jrr0@watson.ibm.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Sat, 17 Feb 2001 00:21:55 -0500

Ed Huntress wrote:

> Speaking of which, does the Book tell you how they accurized leadscrews? I
> don't know where I got that info, as there were two previous books by
> Wayne's father (Richard Moore) and I had a ton of literature from them when
> I covered toolmaking for AM.
>
> Anyway, they cast a lead lap around them, had a lead profile made and
> enlarged as a full-size graph, and then pulled and pushed the lap as the
> leadscrew turned, following the graph, in order to lap it to near
> perfection. You have to do this in both directions, as entirely separate
> operations.

There's an entire chapter about linear measurement and lead screws.  The
process is pretty involved, and starts with roughing the thread, and then
hardening.  I think they nitride them while hanging vertically, in a chamber of
ammonia gas.  Then they are precision ground in filtered, temperature
stabilized oil.

Next they are lapped, as you say, with some sort of accurate measuring
device to check the real lead.  Might be optical?  Not sure, have to look
that up.  But as you say the deviation from a perfect lead is plotted, and
then lapping is done to null the error.

Finally an individual nut is tapped to progressively larger and larger sizes
till it just about fits the leadscrew it will be mated to - and then that is
lapped to fit, till a final check is passed:  a lathe dog which is clamped
to the lead screw will cause the screw to just barely spin till the dog
is pointing down.

Jim



From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: Scraping - Why, when you can grind?
Date: Sat, 17 Feb 2001 16:46:38 GMT

>>NC Machines have ground ways...<<

Yes, but they usually have pressure lubrication, too. There's an unending
argument whether the oil-retention quality of a scraped or frosted surface
helps toolroom machines, with no pressure lubrication. But production
machines and modern CNC machines aren't scraped. Scraping reduces bearing
area too much and it's useless when you have pressure-fed lubrication.

Ed Huntress



From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Tue, 20 Feb 2001 03:25:36 GMT

>>However, it would probably be better to use square ways, mounted with one
corner up, and mounted in V-shaped footings to the cast base -- probably
with several additional supports along the length, unlike the Unimat's round
ways.  If done right, it could allow an extra pair of bearings on the upper
half of the two lower faces, to keep the carriage from lifting under cutting
forces.<<

This is basically what a Schneeberger way is. And high-quality surface
grinders are made with an inverted V machine way and a V that projects
downward from the table. They run on roller bearings because of the need for
fast, easy rapid traverse.The smaller Brown & Sharpe and Wasino machines
(Wasino built B & S's small surface grinders) were/are built that way. (B &
S is the "were"). So are many others, although I forget now which ones have
which type of bearings.

>>I'm still not sure that it would be strong enough for heavy cuts, however.
The flat bearing surface does have a lot to recommend it.<<

The best lathes and mills are built with sliding ways, except in very small
sizes, and except for rapid-traverse, high-spindle-speed machines. Flat ways
last longer, chatter less, and will take heavier cuts. When you use good
pressure lubrication with flat or V sliding ways, they'll go on nearly
forever. I've seen square-way Wasino lathes that had run without special
attention for 15 years, 3 shifts/day, that still held 0.0002" accuracy. When
I asked the owner of one of these how he found Wasino's service, his reply
was, "I don't know. I've never had any."

Ed Huntress




From: Ned Simmons <neds@javanet.com>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Wed, 21 Feb 2001 23:50:24 -0500

In article <steve-2102011433070001@newtyr80.tyler.net>,
steve@puzzlecraft.com says...
> In article <12u79t8acnmo5788h6e9tco2353mdjohqr@4ax.com>, Gary Coffman
> <ke4zv@bellsouth.net> wrote:
>
> > A very flexible manufacturing robot with an unorthodox configuration is
> > hard to swallow. In effect Stewart machines like the Hexapod replace
> > brute cast iron with intelligently used electricity. That's a hard idea for
> > some to grasp, but the theory is sound, and the means to put it in practice
> > inexpensively are coming into our grasp.

Unfortunately actuator technology hasn't kept up with the pace of
increases in the speed of number crunching. While the speed of servo
controllers has improved dramatically (using the common metric of servo
update rate), the actual response of servo systems really hasn't
increased that much in the past 20 years. Though the ability of high end
controllers to update the servo loop has dropped well below 100
microseconds (>10kHz), the bandwidth of industrial electro-mechanical
and electro-hydraulic servos is typically in the low 10's of Hz range.
Not fast enough to implement adaptive control for a machine tool, which
is what the hexapod will need to be competitive.

By adaptive control, I mean the ability to compensate for shortcomings
in the hexapod structure, especially the damping of resonances. I
believe NASA has done work in this area for the purpose of stabilizing
ultra-lightweight structures in space, and I suppose you could include
the active stabilization schemes in some skyscrapers as another example
of this sort of system, but the response of these systems is probably a
few orders of magnitude below what's required to maintain accuracy and
surface finish in a high speed machine tool.

New actuator technology is required to make this all practical. Piezo
maybe?

>
>
> This approach is being taken. The residual errors in the Stewart Platform
> are being analyzed and mathematically modelled. The derived equations are
> then simple to code in standard programming languages. See the following
> link for some details:
>
> http://www.enee.umd.edu/medlab/papers/robautforsem97/robautforsem97.html

I have no doubt that the math required to calculate toolpaths is
tractable.

>
> The invention of the desktop computer by Apple was successful in replacing
> heavy metal, the Linotype machine, with intelligently used electricity.
> More advanced systems have replaced the use of film and even plates in the
> lithographic process with whirling bundles of electrons and photons.

Well not exactly. I worked on Heidelberg's direct imaging process about
10 years ago, there was plenty of hardware involved and a skilled
operator was still required.

>
> The potential advantages of the Stewart Platform are enormous. Extremely
> simple construction, extremely small and light weight shipping, very large
> sizes and unparalleled precision, speed and versatility.

The concept is simple and elegant, the implementation is anything but
simple. Six active links and their attendant joints whose motions must
be limited to the desired degrees of freedom while allowing no free play
in the remaining axes. And the integrity of these constraints must be
maintained in a nasty environment.

>
> Another benefit that will seem horrible to many machinists is the
> elimination of skilled machine operators with large paychecks. The
> software will do it all someday. Anybody who doesn't believe this will
> happen simply needs to look at the printing industry where all skilled
> trade positions have been eliminated by software.
>
> When a new idea like this comes along, wonderful opportunities for the
> inventor open up. How many millionaires were created simply from writing
> software for the desktop computer at a time when the prevailing attitude
> was there was absolutely no market for a personal computer?
>
> The Stewart Platform is a brilliant application of Kinematic Principles.
> What other high precision specialized devices can be cheaply made
> utilizing these ideas? I suspect just about any metalworking device can be
> greatly improved with kinematics and software.
>
> Will the next home built machine revolutionize the industry? Has the
> revolution already begun? Who's going to make the first 3D desktop Stewart
> Platform 6 axis mill and sell it for under $1000?

If you're looking for a revolution, I'd put my money on metal
*deposition* rather than metal removal. Rapid stereolithography that can
produce end products with no consumable tooling and no chips (waste) is
a much better analogue to the advances in publishing you cite.

Ned Simmons


From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Thu, 22 Feb 2001 05:28:52 GMT

>>New actuator technology is required to make this all practical. Piezo
maybe?<<

Or maybe magnetostrictive, adapted from submarine sonar actuators, which is
being used on a new elliptical piston-turning device that we pioneered at
Wasino over the last couple of years. It's good for 20 kHz, with some
serious force -- much more force than piezo. (But don't start programming
yet. The hysteresis is a bugger to program around.)

Obviously you have the practical experience to know where the shortcomings
lie in Steve's "kinematic" approach. I'm biting my tongue, because I don't
want to damp anyone's enthusiasm and inventiveness for new ideas. There are
probably some good ideas for machine tools that could come out of some of
the work he describes. But not, probably, the Hexapod. Too many engineering
problems, which are being solved only slowly, while bridge-type mills are
now positioning to absolute X,Y,Z coordinates +/- 50 millionths of an inch.
It's a Zeno's paradox for the Hexapod.

The essential problem is that the thing *is* a kinematic device -- the
classical definition of kinematics being the study of dynamics "without
regard to force or mass". Force and mass are the central issues. The Hexapod
is a Fulleresque structure that gets you to a position and then pretends
that modulus of elasticity doesn't matter.

BTW, as you noted, there are trends and countertrends at work on
closed-loop, adaptive control. The response time is one issue. Another is
the "six-sigma," or Cpk 2.0, requirement of a growing number of large
manufacturers now. If your machine can't stay well within the process
tolerance envelope without the benefit of closed-loop control, you're lost
anyway. We actually took closed-loop control off of some of our systems to
get better performance and better accuracy. But that's a long story.

Ed Huntress



From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Thu, 22 Feb 2001 15:42:54 GMT

>>Ballscrew mechanisms are much too slow for compensation, but perhaps you
could employ ballscrews for gross motions and use linear motor or piezo
methods for shorter range compensation. That would remove most of the
scaling limits at the cost of a more complex control system.<<

The news from the front is that you don't *want* compensation. You want
smooth, accurate, continuous motion, which you get by building accuracy into
the system from the beginning.

You're dealing with large masses here, whether you move the spindle or the
work. Positioning compensation done before the fact of moving something, as
in leadscrew compensation, is fine. That all goes on inside the computer
before the machine components are actually moved. Compensation done after
the fact, as in closed-loop adaptive control, means that you're accelerating
a huge mass at extraordinary G-forces, wearing out the mechanisms and
running into the elastic properties of your system -- a decidedly unpleasant
phenomenon to deal with.

There actually is a countertrend away from closed-loop adaptive control for
machine tools, and towards getting it right in the first place.

Ed Huntress





From: "Ed Huntress" <mike234@bellatlantic.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Fri, 23 Feb 2001 05:13:13 GMT

>>That's what I had in mind re the piezo actuators, use them to actively
damp vibrations/resonances in the structure, not for gross movements. But
think about the driving frequencies you'll encounter with a multi-flute tool
turning at 20KRPM.<<

OK, they're used for that, and electromagnetic counter-vibration units have
been used in a few applications since the '60s. When Etrema approached
Wasino a few years ago about using magnetostrictive devices on machine
tools, counter-vibration was the application they had in mind.

But that's a minor issue today. Modern machine tools damp most vibration
quite well. I saw some astonishing examples of this last week, watching a
Cat 50 tool shank, Rc 60, being turned at high speed, right across the 1"
drive slot. Yike! There's an interrupted cut for you. But the big Hardinge
just hummed, and the finished tolerance was +/- 0.0002".

>>Which I take to mean that as conditions change, long term wear and near
instantaneous changes in the cut, the system must "learn" how to respond.<<

That's a challenging goal, although mapping the vibrations in a long
production run probably is doable and may be worth the effort. Maybe it's
even being done, I don't know. One thing I can tell you for certain: trying
to make on-the-fly corrections with a modern high-speed mill will kill the
tools. They can't tolerate much variation in chip depth, nor can they
tolerate even a fraction of a second of pause on their cutting path. You'll
burn the coatings off if you try.

You may be interested that all of this discussion has missed the real issue
in precision machining today, which is thermal control. There are mills that
can position to +/- 1 micron (40 millionths of an inch), and lathes that
position to half that tolerance. But vary the environmental temperature by 5
degrees F, and their accuracy goes out the window. No ordinary shop can
control ambient temperature that closely. To work at their designed
accuracies, these machines need a temperature-controlled room. Some are
operating in near metrology-lab environments.

Dick Moore started the trend, first with a clean room for building his
machines, back in the '30s, which had a temeprature gradient of less than 2
degrees F from floor to ceiling. He used Invar, a
low-temperature-coefficient iron alloy for the heads in his machines. Then,
in the Moore Jig Grinder, he put an electric resistance heater to stabilize
temperature at something above the natural running temperature of the head.

Now you'll see other high-precision machines that use heaters (Bostomatic)
and spindle-oil coolers (Roku-Roku, Makino, etc.). Wasino's precision lathes
have both heaters and refrigerant-based coolers stabilizing spindle oil and
X-axis ballscrew oil. If it's too cool, it gets heated. If too hot, it's
refrigerated.

This is where the action is. Building a machine to sub-tenths accuracy is no
magic trick today. Keeping it stable in a real shop environment is the magic
trick.

Ed Huntress



From: "Ed Huntress" <huntres2@optonline.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Fri, 09 Mar 2001 16:55:58 GMT

>>My main point was that there is a counter-example in use that disproves
the claims that a kinematic mill similar to the Stewart Platform is
incapable of exceeding the tolerances of a heavy metal machine.<<

The Bridgeport was designed in the 1930s, and has changed only little since.
If you're going to make a comparison, you have to compare with a Mikron,  a
Roeders, or a Roku-Roku mill. What you'll need to beat is +/- 50 millionths
of an inch absolute at 100 m/min. traverse rate, all day long, every day,
for several years.

Good luck. <g>

Ed Huntress




From: "Ed Huntress" <huntres2@optonline.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Fri, 09 Mar 2001 16:58:48 GMT

>>...at 100 m/min. traverse rate...<<

Egad. That should have been "100 inches per minute traverse rate..."

So, you'll have a little easier time of it. <g>

Ed Huntress




From: "Ed Huntress" <huntres2@optonline.net>
Newsgroups: rec.crafts.metalworking
Subject: Re: designing a machine for hobby use.
Date: Sat, 10 Mar 2001 00:50:52 GMT

>>Is that as fast as that type of machine moves? I know there are smaller
machines that have *feed* rates as high as 1200 inches per minute. I done't
know what the following accuracy is but they are much faster than 100 ipm.<<

No machine with a feed rate of 1200 ipm will contour in three axes while
holding +/- 50 millionths, no matter how much you slow it down. It requires
a fast pitch to do that -- or linear motors, which are an utter disaster for
this kind of work (they heat the machine up too much, right where it hurts).
These machines will traverse at 400 - 600 ipm, but the rest of the
requirement is to do it while machining a mold cavity for a statue of Venus,
at 30,000 rpm with a 0.004" dia. end mill. <g> On the average, these
machines will do that at around 100 ipm actual traverse rate.

Ed Huntress

Index Home About Blog