|
General Metalworking - All aspects of working with metal.
|
|
|
|
|
Posted by Joseph Gwinn on June 1, 2008, 11:51 am
Please log in for more thread options
> >
>
> [ ... ]
>
> >> > Hmm. What speeds? I've been thinking of getting a diamond wheel for my
> >> > HF #46727 carbide tool grinder (6", 3400 rpm).
> >>
> >> Do they *offer* a diamond wheel for that specific machine?
> >
> > Yes. The HF unit is a clone from Baldor, and I see suitable wheels in
> > catalogs and flyers all the time. Like Enco 391-5716, which I have seen
> > as low as about $80 on sale:
> ><http://www.use-enco.com/CGI/INSRIT?PMAKA=391-5716>.
>
> O.K. Then perhaps I could use one as well.
>
> > Although when I replaced one gray wheel with aluminum oxide, I had to
> > modify the wheel housing so the Norton wheel would fit, using a die
> > grinder that often outran my compressor.
>
> *That* would worry me -- because it would reduce the wall
> thickness needed to protect the operator when the wheel shatters.
The housing is cast iron, and quite thick, 7 mm at the outer edge,
thicker near the bade. It was the widening near the base that caused
the interference. I think I took a few mm off near the base, yielding a
cover that no longer widened so much near the base, but was never less
than 7mm thick.
Actually, are plated diamond wheels likely to burst at all? Aren't they
made of metal?
> >> >> And when grinding carbides, you are generating an inhalation
> >> >> hazard -- so I'll let others with supplied air do the grinding. :-)
> >> >
> >> > I think I'll most likely die of something else first. And, I do have
> >> > that respirator.
> >>
> >> O.K. Just be careful.
> >
> > I'm curious - why is carbide dust any worse than say HSS or stainless
> > steel dust?
>
> I *think* that it is the Tungsten matrix in which the carbide is
> embedded.
Actually, the matrix is cobalt, which glues tungsten carbide particles
into a solid mass, like cement gluing stones into a solid mass called
concrete.
<http://www.azom.com/details.asp?ArticleID=1203>
> [ ... ]
>
> >> > For a simple example, if one is trying to achieve 0.001" accuracy in
> >> > cutting a diameter, and the cutting force is 50#, the static stiffness
> >> > had better exceed 50/0.001= 50,000 pounds per inch.
> >>
> >> Yes -- but we were talking about motion which you *saw*, which
> >> is well above what we would test with such measures.
> >
> > But far less than the motion previously seen, so there has been progress.
>
> Agreed.
>
> >> If you want to see how flexible the machine is, clamp a "tenths"
> >> indicator in the collet and rest the tip on the underside of the tool
> >> tip. Then just press down with your thumb on the tool tip. You'll see
> >> motion on that dial indicator -- even though you would (or should) not
> >> have any chance at seeing it with the naked eyes.
> >
> > I'll try this. But I'm trying to hang a number on this, because the
> > lathe was designed and built to some required minimum static stiffness
> > value. I suppose I could ask Clausing, but it's unlikely that anyone
> > knows the answer anymore, even if they are willing to say.
>
> It may have been in the original specs in the catalogs -- or it
> may not have been.
I don't recall ever seeing such a number except in textbooks on the
design of machine tools.
> > All the methods discussed have the problem that spindle stiffness is not
> > included, and yet it matters in practice. However, for diagnosis, it's
> > best to be able to consider the parts in isolation.
>
> And the stiffness of the spindle is complicated by things like
> the length of the chuck, which can both add leverage to the spindle
> flexing itself, and flexibility in the chuck itself.
>
> >> > It's actually pretty easy to measure this. Clamp a ~36" steel rod or
> >> > bar in a tool bit holder. Stick a dial indicator to the ways, with tip
> >> > resting on the bar say 1" from the face of the holder (to represent the
> >> > amount of cutting tool that sticks out). Zero dial indicator. Hang a
> >> > known weight off the end of the rod, and note the indicator value. If
> >> > the weight is hung from the rod by a cord 30" from the center of the
> >> > toolpost, and the indicator rests on a point 3" from the same center,
> >> > then the force is (30/3) times the weight. Divide force by indicated
> >> > deflection to yield static stiffness.
> >>
> >> I suspect that you would get different figures measuring at
> >> different points along the bar. And remember -- up close to the holder
> >> you will be measuring give in all of your stack-up, while out near the
> >> end of the bar you will be measuring a mix which will be mostly the
> >> flexibility in the bar.
> >
> > With a 1" diameter bar, flex in the bar is less of a problem.
>
> It still exists.
Yes, but if one doubles the diameter the stiffness of the bar increase
by a factor of 8, so it's pretty easy to simply overpower the problem,
such that something else limits the accuracy.
> > The full strength version is to have a tommy bar that applies the force,
> > and a separate indicator bar that acts as the lever arm for the
> > indicator. This is the mechanical equivalent of a 4-terminal resistance
> > measurement.
>
> But a lot more bulky than a 4-terminal resistance measurement.
> I've got a nice rack-mount digital multimeter from HP which came from a
> hamfest some years ago.
Yep.
> > I came to the conclusion that a better way to summarize the measurement
> > is as deflection in degrees per foot-pound of torque. This can be
> > translated into deflection per pound of cutting force once the location
> > of the cutting edge is specified.
>
> O.K
>
> > With a 1" diameter 1018 bar mounted in a BXA-4 boring-bar holder
> > perpendicular to the 5914's bed way, a 5 Kg weight hanging on a cord at
> > 32.25" from the centerline (defined by the 5/8-20 mounting bolt) causes
> > a 0.015" deflection (with respect to the bed) at 7.25" from the same
> > centerline. This is 3262# per inch of deflection, at 7.25" from center.
> >
> > Banging through the math, this is 0.004 degrees per foot-pound of
> > torque, or 249.4 foot-pounds of torque to cause one degree of deflection.
> >
> > My suspicion is that this value is lower (floppier) than it should be,
> > but I have no comparison as yet.
>
> Remember that HSS is *not* normal steel and has a higher
> stiffness. Carbide has an even higher stiffness, which makes
> particularly good boring bars.
True enough, but the objective here is to measure the stiffness of the
machine tool holding the HSS or carbide bit, and not to measure the
stiffness of the boring bar. The use of a BXA-4D boring bar holder is
fortuitous - the holder accepts a 1" bar, and a 6' length of 1" CRS bar
fits perfectly and can be clamped as if it were a boring bar.
I do have some Criterion solid carbide boring bars (bought used for a
few dollars each), and they do a really nice job.
> > For people wishing to make such a measurement but lacking a calibrated
> > mass to hang off the bar, there is a simple solution given that we all
> > have vernier calipers. Iron and steel have a fairly constant density,
> > so one can simply measure a machined rectangular or cylindrical block of
> > steel and compute the weight. The density of mild steel is 7.85 grams
> > per cubic centimeter, or 0.283 pounds per cubic inch.
>
> Of course, I have a weight set which goes up to 5Kg weights.
Good, but they are not common. I had to get a set of test masses to
calibrate a scale that had gone seriously (~25%) out of calibration.
It's actually a lab utility scale (6 Kg by 0.5 gm), but my wife took it
over for cooking. So, this scale has to work.
> >> > This method isn't scientifically precise, but is pretty good. The main
> >> > problem is that it can be hard to decide exactly where the center point
> >> > is, as things flex and move, so the ratio (30/3) while good is never
> >> > quite right.
> >> >
> >> > The scientific approach involves a load cell, press (screw or hydraulic)
> >> > and a dial indicator, and measures force and deflection directly. Costs
> >> > more than the lathe.
> >>
> >> I think that you should measure each component of the flex
> >> starting from the carriage wings measured from the bed, then the
> >> cross-slide, measured from the wings, then the compound base, measured
> >> from the cross-slide, and so on up. This will tell you what *really*
> >> needs fixing. :-)
> >
> > I've been doing this by resting a finger on various gaps while I torque
> > the toolpost around. The fickle finger points at the cross-slide
> > slideway.
>
> So attack that first. Operate on the assumption that when you
> fix all of the weak points, you will have a lathe which will do what
> Clausing designed it to do.
Actually, I attacked the spindle bearing preload first, to considerable
good effect.
A question: On your Clausing, if you pull the backgear pin, so the
spindle can turn freely, when you turn the spindle by hand, how much
drag do you encounter? And, is there an initial resistance to turning,
which resistance lessens once the spindle is turning? I'm wondering if
I over-tightened the spindle preload adjustment nut, but have no
comparison.
> >> [ ... ]
> >>
> >> >> > Or, one can provide a 0-10 volt
> >> >> > control signal to the VFD, using homebrew electronics to convert pot
> >> >> > signal to the 10 volt signal. And a low-impedance output is less
> >> >> > sensitive to noise. But it's bet to make sure it is short circuit
> >> >> > proof.
> >> >>
> >> >> Indeed -- though the effect should be stopping the spindle,
> >> >> instead of running it at an insane speed. :-)
> >> >
> >> > A short circuit to ground would cause the spindle to stop, and to +10
> >> > would cause the spindle to go at the max speed set into the VFD.
> >>
> >> A short circuit is assumed to be to ground. If the +10 is not
> >> properly insulated, look more for a short between it and ground.
> >
> > One can have a short between +10 and the slider on the pot, which would
> > both cause the machine to operate at full speed and possibly burn the
> > pot out.
>
> Yes -- it is is set to a low speed position, it is quite likely
> to fry the pot. The carbon film ones would be the most likely to go
> first, wirewounds next, and the carbon composition ones like the
> Allen-Bradleys would last the longest.
Yep. The speed pot in the X-axis drive on the Millrite blows out from
time to time, and the cause is always that the 10K carbon-film pots
cannot handle the current drawn from the wiper arm, causing opens where
the film is burned off the backing. The symptom is that the speed
varies from stopped to full speed in one big step. I talked to the
seller, and he complained that this happened a lot, causing a big
warrantee bill, and he had been changing vendors (all asian) to no
avail. I pointed him to the Japanese makers of heavier better pots, for
which he was happy, as he then had something to push back with. Don't
know how it came out.
I bet the original design (that was cloned) assumed a real carbon
composition pot. The pot and assembly is an oddball mechanically, a mix
of english and metric sizes, and retrofitting it is one thing that
awaits the lathe growing up to be a lathe.
Wirewound sounds workable. For instance, CTS 026TB32, like DigiKey
CT2159-ND, for $3.60 each, which is a 10K 5-watt wirewound in the
standard form factor of the traditional molded carbon pots.
> But -- this is more likely to hit the current limit of the VFD's
> 10V output.
And perhaps damage the VFD. I'm not sure if they are short-circuit
protected. You would think they would be, but no such claim is made.
An inline fuse may be a good idea. Not that I did any such thing, but
the series resistor can take a shorted zener.
> >> And IIRC the VFD (or at least the Mitsubishi ones) can be
> >> configured for +5V being full speed -- and -5V being reverse full speed.
> >
> > Turns out the Hitachi VFDs can do this too. Didn't think of it, but
> > could have worked.
>
> With a series resistor to drop the high end of the pot to 5V
> instead of 12V.
Well, 7 or 8 volts, not 12, but yes.
But I'd probably still use a zener diode, as they are simple and cheap.
> [ ... ]
>
> >> O.K. At least for the AF91. I still would want to check the
> >> manual for my Mitsubishi -- but I've been doing too many other things.
> >
> > There is no harm in using an ohmmeter on the input, so long as the
> > voltage control input is positive with respect to the common.
> >
> > The ohmmeter test signal may cause the machine to spin up, if the VFD is
> > programmed to pay attention to this input. The default is to ignore the
> > voltage control input.
>
> Of course -- but there may be conditions under which the input
> impedance is higher or lower, depending on internal conditions --
> including the possibility of the input impedance being
> voltage-sensitive. Given the typical digital multimeter, which switches
> current levels as it switches ranges, I would not depend on the measured
> input impedance being the worst-case input impedance.
Umm, it seems to me that any VFD that would be damaged by the test
signal from a modern ohmmeter is far too fragile for industrial or even
HSM use.
While the input impedance of the VFD may vary some, only the minimum
value is important, as it defines how large the pot resistance can be
without undue effect.
> [ ... ]
>
> >> >> > They say that they can generate about 60 inches or 5 feet (of water)
> >> >> > vacuum. This is (14.7)(5 feet/33.9 feet)= 2.17 psi.
> >> >>
> >> >> Vacuum -- but what about the pressure side? That may go higher,
> >> >> as the impeller would probably be more efficient at higher density air.
> >> >
> >> > They don't say, but it could be. But it won't be by a factor of 5.
> >>
> >> But I'm sure that a shop vac could produce enough pressure to
> >> push sufficient breathing air through a 1/4" line -- just as long as the
> >> demand regulator is made with a sufficiently large aperture.
> >
> > Probably. I have no idea what minimum input pressure typical demand
> > regulators are made to handle.
>
> I think that we would have to make our own demand regulator,
> with something like a 1/4" aperture covered with a soft rubber pad under
> diaphragm control.
Making a regulator to handle 10 psi should not be hard. The only thing
I would worry about is if a failure in the regulator can cause full
inlet pressure to appear at the facemask. The design should be such
that failure causes zero output.
> [ ... ]
>
> >> >> I suspect that the precision is not that critical -- you can
> >> >> always pack with shims until all three are equally tight before cranking
> >> >> down on the wrench.
> >> >
> >> > I'll have to think about it. There has to be a 17th century solution to
> >> > this.
> >> >
> >> > The 20th century solution is a surface grinder.
> >>
> >> Nope -- not unless you can be *sure* that the distance from the
> >> back of the jaws to the gripping face is constant for all three jaw
> >> positions. I would not like to bet on that.
> >
> > Hmm. How are gage blocks made?
>
> A surface grinder gets them to a certain dimension greater than
> the target dimension, and then lapping brings them to their final dimension.
>
> But gauge blocks are flat on *both* ends (excluding setting rods
> for micrometers, which have spherical surfaces at the ends), and most
> chuck jaws have a rounded end to match the curvature of the chuck body
> when the jaw backs are adjusted out to the surface of the chuck body.
> And this is just an approximation, since it is not a critical dimension.
> It is the position of the inner points relative to the scroll teeth,
> which are at a different position for each jaw.
OK. Why is lapping any more likely to yield parallel faces than a
surface grinder?
> >> Another trick which has been suggested at times is to drill
> >> three holes at the apicies of an equilateral triangle, and then bore out
> >> the center until the holes allow the jaws to fit in and protrude into
> >> the center area.
> >
> > This depends on the accuracy of the jaw's outside faces, but these will
> > typically be far less worn than the inside faces.
>
> Not the *outside* faces, but rather the angled faces -- the same
> ones which I was suggesting you use spacer blocks on.
>
> 1) There are three holes in an equilateral triangular pattern.
>
> 2) There is a hole bored in the middle which intersects all three
> outer holes.
>
> 3) The jaws are tightened until the angled faces press on the
> sides of the holes.
>
> Using the outside faces could be done with a single large hole,
> not four holes in a specific pattern. But -- There is no certainty that
> the outside faces loaded out will be in a fixed relationship to the
> inside ones loaded in. It depends on the individual backlash in the
> scroll teeth for each jaw.
I'm not visualizing this. The angled faces would not find flat surfaces
to rest upon in all those bored circular holes, so it would be hard to
know the effective spacings between jaws as tightened.
> Somewhat similar to the problems with trusting the backs of the
> jaws are positioned properly with respect to the inside points. If you
> are going to grind the backs of the steps true, you need to do it with
> jaws loaded out, just as you need to do the inner points with the jaws
> loaded in.
Yes.
> [ ... ]
>
> >> > I'll have to do this, but after seeing if the jaws are bell-mouthed.
> >>
> >> Put on new top jaws first.
> >
> > I tried this with a 0.5" drill blank, but no one socket yielded better
> > answers.
>
> Interesting. Very little wear in that chuck, then.
Nor does the chuck look like it has seen much use.
I'm thinking that I will not attempt to regrind the chuck jaws, as they
test OK, and the increasing the preload on the spindle bearing had a
dramatic effect on the chatter, so the chuck was probably falsely
accused.
> >> Anyway -- those could be done on a
> >> surface grinder with a fixture which keys on the notches in the top
> >> jaws.
> >
> > I'm sure that's what the jaw makers do.
>
> Only for the two-piece jaws. No bets on how the one-piece jaws
> are handled.
Probably ground in a fixture that has been lovingly tweaked over the
years.
> [ ... ]
>
> >> >> I saw the article. Glad that it worked out. But you have
> >> >> obviously discovered one reason why I have multiple lathes. :-) The
> >> >> compact-5 would have been big enough to re-do that spacer with no
> >> >> problems.
> >> >
> >> > I imagine so. The spacer is 1.25" OD by 0.430" (now 0.399") long.
> >>
> >> Dead easy for the Compact-5. You could even have a Sherline
> >> which would be satisfactory for that.
> >
> > You know, in an earlier day, I did look at Sherline, but was put off by
> > the lack of steel in its construction.
>
> O.K. Look at the competition. Trying to remember the US name.
> The UK name is Peatol. It has a ground steel dovetail bolted to the top
> of an extruded aluminum bed, with cavities in the bed filled with
> concrete to dampen vibrations.
>
> I've got one of those (which may be why I can't remember its
> name? :-). I actually got it at a hamfest quite a few years ago, before
> it was sold under the current name here. Mine has *no* name marked.
>
> Anyway -- unlike the Sherline, which uses a hand-cranked
> leadscrew to move the carriage, this uses a rack and pinion, like the
> hand feed on a serious lathe.
I never heard of Peatol before. All that's advertised here is Taig and
Sherline, who are both relentless. I bet it's like soap - 1/3 of the
price is the cost of the advertising to pound it into your skull, to
blot out all other possibilities.
They do look good. <http://www.peatolmac.talktalk.net/>
But it appears that Peatol is made by Taig, and the difference is
unclear: <http://www.lathes.co.uk/taig/>.
> [ ... ]
>
> >> > Umm, the firestop need not prevent attaching a trolly track. It's OK to
> >> > have hanger rods coming from the beam through the drywall to the trolly
> >> > below, so long as the gaps are filled with an approved firestop caulking
> >> > material. This is done with wires and plumbing all the time.
> >>
> >> But a trolley track below this would require me to duck every
> >> time I passed under it. There is really not that much headroom here.
> >
> > Hardhat?
>
> Just makes it more likely that I will hit something. My
> hardhat adds at least another inch to the top of my head. :-)
But reduces the impact-induced IQ decrement.
> [ ... ]
>
> [ ... quite a bit snipped -- so I can perhaps read something
> else in the newsgroup before I go to bed. I've been doing too
> much else the past few days -- and lost about five hours to a
> thunderstorm-induced power outage -- plus another hour getting
> all the systems back awake and talking to each other. ... ]
This has happened once to me up here near Boston, over the last 20
years, and no permanent damage, so I count myself lucky. I do have a
pair of big UL-rated surge arrestors attached to the main panel. The
arrestors are made in Texas, where lightening is far more a problem than
here. <http://www.deltala.com/>
What I have are the LA-302 units, purchased from the local electric
supply house. <http://www.deltala.com/prod02.htm>
I did this in liu of getting a lot of surge protection power strips,
because these units are tiny (in terms of surge capacity), together cost
more than the two LA-302 units, and cover very little of the equipment
in the house. The only disadvantage (as seen by the general public) is
that you need to be able to install something in the main power panel,
which explains why the power strips are big sellers.
Joe Gwinn
|
|
Posted by DoN. Nichols on June 1, 2008, 9:48 pm
Please log in for more thread options
>
[ ... ]
>> > Yes. The HF unit is a clone from Baldor, and I see suitable wheels in
>> > catalogs and flyers all the time. Like Enco 391-5716, which I have seen
>> > as low as about $80 on sale:
>> ><http://www.use-enco.com/CGI/INSRIT?PMAKA=391-5716>.
>>
>> O.K. Then perhaps I could use one as well.
O.K. Harbor Freight's web site is back up -- and this is
designed to mount via four socket head cap screws to a flange, not to be
bolted with nuts and flanges directly on the spindle.
I see no RPM listed.
And I *do* see the following:
For Baldor & Other Popular *CARBIDE* Bench Grinders
^^^^^^^
My emphasis on the "carbide" part. I know that the carbide grinders
have driving faces instead of just clamp flanges. I *don't* know what
the RPM is.
>> > Although when I replaced one gray wheel with aluminum oxide, I had to
>> > modify the wheel housing so the Norton wheel would fit, using a die
>> > grinder that often outran my compressor.
>>
>> *That* would worry me -- because it would reduce the wall
>> thickness needed to protect the operator when the wheel shatters.
>
> The housing is cast iron, and quite thick, 7 mm at the outer edge,
> thicker near the bade. It was the widening near the base that caused
> the interference. I think I took a few mm off near the base, yielding a
> cover that no longer widened so much near the base, but was never less
> than 7mm thick.
>
> Actually, are plated diamond wheels likely to burst at all? Aren't they
> made of metal?
But you thinned the housing to fit a stone, not a diamond wheel
from what I read above.
>
>> >> >> And when grinding carbides, you are generating an inhalation
>> >> >> hazard -- so I'll let others with supplied air do the grinding. :-)
>> >> >
>> >> > I think I'll most likely die of something else first. And, I do have
>> >> > that respirator.
>> >>
>> >> O.K. Just be careful.
>> >
>> > I'm curious - why is carbide dust any worse than say HSS or stainless
>> > steel dust?
>>
>> I *think* that it is the Tungsten matrix in which the carbide is
>> embedded.
>
> Actually, the matrix is cobalt, which glues tungsten carbide particles
> into a solid mass, like cement gluing stones into a solid mass called
> concrete.
O.K. But there are MSDS sheets which come with each carbide
tool from MSC which warn about grinding them.
[ ... ]
>> > I'll try this. But I'm trying to hang a number on this, because the
>> > lathe was designed and built to some required minimum static stiffness
>> > value. I suppose I could ask Clausing, but it's unlikely that anyone
>> > knows the answer anymore, even if they are willing to say.
>>
>> It may have been in the original specs in the catalogs -- or it
>> may not have been.
>
> I don't recall ever seeing such a number except in textbooks on the
> design of machine tools.
O.K. Then those should give you the ability to calculate the
deflection -- if you take the time to measure everything carefully. :-)
[ ... ]
>> > With a 1" diameter bar, flex in the bar is less of a problem.
>>
>> It still exists.
>
> Yes, but if one doubles the diameter the stiffness of the bar increase
> by a factor of 8, so it's pretty easy to simply overpower the problem,
> such that something else limits the accuracy.
Does it? I don't have the formulas to hand, but I would expect
only a factor of four.
[ ... ]
>> > With a 1" diameter 1018 bar mounted in a BXA-4 boring-bar holder
>> > perpendicular to the 5914's bed way, a 5 Kg weight hanging on a cord at
>> > 32.25" from the centerline (defined by the 5/8-20 mounting bolt) causes
>> > a 0.015" deflection (with respect to the bed) at 7.25" from the same
>> > centerline. This is 3262# per inch of deflection, at 7.25" from center.
>> >
>> > Banging through the math, this is 0.004 degrees per foot-pound of
>> > torque, or 249.4 foot-pounds of torque to cause one degree of deflection.
>> >
>> > My suspicion is that this value is lower (floppier) than it should be,
>> > but I have no comparison as yet.
>>
>> Remember that HSS is *not* normal steel and has a higher
>> stiffness. Carbide has an even higher stiffness, which makes
>> particularly good boring bars.
>
> True enough, but the objective here is to measure the stiffness of the
> machine tool holding the HSS or carbide bit, and not to measure the
> stiffness of the boring bar. The use of a BXA-4D boring bar holder is
> fortuitous - the holder accepts a 1" bar, and a 6' length of 1" CRS bar
> fits perfectly and can be clamped as if it were a boring bar.
Hmm ... that would not work on *my* lathe. The bed is not long
enough to support a 6' bar by the middle without it having to pass
through the spindle. (Hmm ... 1" bar, 1-3/8" bore -- I guess that you
could do that a bit at least.
But -- I am interested in seeing the difference when you rn that
test at right angled to the bed and parallel to the bed. I'll bet that
it is a lot worse parallel to the bed, which is the kind of load you
would have with the offset of the tool holder on the BXA toolpost.
> I do have some Criterion solid carbide boring bars (bought used for a
> few dollars each), and they do a really nice job.
Solid carbide with non-interchangeable cutting surfaces? Mine
is a Kennametal, with the bar 1/2" diameter (with a pair of opposed
flats for clamping it in the tool holder) with the very end steel brazed
on, and triangular inserts. I know that they are Kennametal, because I
had to dig through the overstuffed drawer beside the lathe to find them.
I finally wore out the third point on the insert which I have been using
for some years. :-)
[ ... ]
>> Of course, I have a weight set which goes up to 5Kg weights.
>
> Good, but they are not common. I had to get a set of test masses to
> calibrate a scale that had gone seriously (~25%) out of calibration.
> It's actually a lab utility scale (6 Kg by 0.5 gm), but my wife took it
> over for cooking. So, this scale has to work.
:-)
I've got a two-pan balance with beam weights (picked up at a
hamfest) plus the set of weights (picked up at an estate sale. I wish
that I had been able to meet the previous owner. he had a fascinating
collection of stuff. :-) And I have a Metzler single-pan balance. The
other "pan" is really a three-legged spider onto which various diameters
of metal rings are lowered or lifted clear by a set of knobs for the
two MSDs, and a set of tiny rings hung from a tiny beam to handle the
LSD (not counting the knob which applies bias to the whole thing.) That
also came from a hamfest. The pan is enclosed in three glass panels,
two of which slide open for access. The tilting of the beam is picked
up by a mirror, through another couple of mirrors, to project a moving
scale on a ground-glass window.
[ ... ]
>> > I've been doing this by resting a finger on various gaps while I torque
>> > the toolpost around. The fickle finger points at the cross-slide
>> > slideway.
>>
>> So attack that first. Operate on the assumption that when you
>> fix all of the weak points, you will have a lathe which will do what
>> Clausing designed it to do.
>
> Actually, I attacked the spindle bearing preload first, to considerable
> good effect.
O.K.
> A question: On your Clausing, if you pull the backgear pin, so the
> spindle can turn freely, when you turn the spindle by hand, how much
> drag do you encounter? And, is there an initial resistance to turning,
> which resistance lessens once the spindle is turning? I'm wondering if
> I over-tightened the spindle preload adjustment nut, but have no
> comparison.
Mine may not be truly equivalent, since mine has three belts
applying stress to the side of the shaft. But I get what feels like a
3:1 or perhaps 4:1 ratio of starting force vs moving force. It pegged
the force gage which I was able to find -- at about 1.5 KG applied at a
point 3.125" from the center (on a chuck jaw at the OD of a 6-1/4"
chuck.)
The heavier force gauge has gone into hiding for the moment. I
last used it to check the match of a newly purchased clock weight to the
original one (time was still there, gong was missing, but I remembered
them as being the same dimensions.)
[ ... ]
>> >> > A short circuit to ground would cause the spindle to stop, and to +10
>> >> > would cause the spindle to go at the max speed set into the VFD.
>> >>
>> >> A short circuit is assumed to be to ground. If the +10 is not
>> >> properly insulated, look more for a short between it and ground.
>> >
>> > One can have a short between +10 and the slider on the pot, which would
>> > both cause the machine to operate at full speed and possibly burn the
>> > pot out.
>>
>> Yes -- it is is set to a low speed position, it is quite likely
>> to fry the pot. The carbon film ones would be the most likely to go
>> first, wirewounds next, and the carbon composition ones like the
>> Allen-Bradleys would last the longest.
>
> Yep. The speed pot in the X-axis drive on the Millrite blows out from
> time to time, and the cause is always that the 10K carbon-film pots
> cannot handle the current drawn from the wiper arm, causing opens where
> the film is burned off the backing.
Hmm ... It looks as though State electronics has taken over the
production of the Allen Bradey carbon composition pots. Look for "Type
K". Check with:
<http://www.potentiometers.com/>
or
<http://www.state-elec.com>
and look for "Series K".
I was considering going with them until I discovered that active
circuitry will offer me more flexibility.
> The symptom is that the speed
> varies from stopped to full speed in one big step.
It would be. I've seen enough of this sort of failure, though
not on a speed control pot. :-)
> I talked to the
> seller, and he complained that this happened a lot, causing a big
> warrantee bill, and he had been changing vendors (all asian) to no
> avail. I pointed him to the Japanese makers of heavier better pots, for
> which he was happy, as he then had something to push back with. Don't
> know how it came out.
O.K.
> I bet the original design (that was cloned) assumed a real carbon
> composition pot. The pot and assembly is an oddball mechanically, a mix
> of english and metric sizes, and retrofitting it is one thing that
> awaits the lathe growing up to be a lathe.
>
> Wirewound sounds workable. For instance, CTS 026TB32, like DigiKey
> CT2159-ND, for $3.60 each, which is a 10K 5-watt wirewound in the
> standard form factor of the traditional molded carbon pots.
That sounds as though it would work -- other than eliminating
the finest of the control -- but you don't need that fine a control. ;-)
>
>> But -- this is more likely to hit the current limit of the VFD's
>> 10V output.
>
> And perhaps damage the VFD. I'm not sure if they are short-circuit
> protected. You would think they would be, but no such claim is made.
> An inline fuse may be a good idea. Not that I did any such thing, but
> the series resistor can take a shorted zener.
O.K.
>> >> And IIRC the VFD (or at least the Mitsubishi ones) can be
>> >> configured for +5V being full speed -- and -5V being reverse full speed.
>> >
>> > Turns out the Hitachi VFDs can do this too. Didn't think of it, but
>> > could have worked.
>>
>> With a series resistor to drop the high end of the pot to 5V
>> instead of 12V.
>
> Well, 7 or 8 volts, not 12, but yes.
Did I say 12? Sorry -- I was thinking about a 9812 regulator
chip from earlier.
> But I'd probably still use a zener diode, as they are simple and cheap.
O.K.
>
>> [ ... ]
>>
>> >> O.K. At least for the AF91. I still would want to check the
>> >> manual for my Mitsubishi -- but I've been doing too many other things.
>> >
>> > There is no harm in using an ohmmeter on the input, so long as the
>> > voltage control input is positive with respect to the common.
>> >
>> > The ohmmeter test signal may cause the machine to spin up, if the VFD is
>> > programmed to pay attention to this input. The default is to ignore the
>> > voltage control input.
>>
>> Of course -- but there may be conditions under which the input
>> impedance is higher or lower, depending on internal conditions --
>> including the possibility of the input impedance being
>> voltage-sensitive. Given the typical digital multimeter, which switches
>> current levels as it switches ranges, I would not depend on the measured
>> input impedance being the worst-case input impedance.
>
> Umm, it seems to me that any VFD that would be damaged by the test
> signal from a modern ohmmeter is far too fragile for industrial or even
> HSM use.
I wasn't suggesting that it would *damage* the VFD. Just that
it would be difficult to be sure that you got a reading which truly
indicated the minimum resistance.
> While the input impedance of the VFD may vary some, only the minimum
> value is important, as it defines how large the pot resistance can be
> without undue effect.
Of course.
[ ... ]
>> >> But I'm sure that a shop vac could produce enough pressure to
>> >> push sufficient breathing air through a 1/4" line -- just as long as the
>> >> demand regulator is made with a sufficiently large aperture.
>> >
>> > Probably. I have no idea what minimum input pressure typical demand
>> > regulators are made to handle.
>>
>> I think that we would have to make our own demand regulator,
>> with something like a 1/4" aperture covered with a soft rubber pad under
>> diaphragm control.
>
> Making a regulator to handle 10 psi should not be hard. The only thing
> I would worry about is if a failure in the regulator can cause full
> inlet pressure to appear at the facemask. The design should be such
> that failure causes zero output.
If you get full inlet pressure to the facemask, it would simply
float clear of the face with lots of air escaping all around -- which
would *still* keep the undesirable components out. If you had straps
tight enough to build up a pressure dangerous to the lungs, you would
probably also be limiting blood circulation to the scalp. :-)
[ ... ]
>> >> > The 20th century solution is a surface grinder.
>> >>
>> >> Nope -- not unless you can be *sure* that the distance from the
>> >> back of the jaws to the gripping face is constant for all three jaw
>> >> positions. I would not like to bet on that.
>> >
>> > Hmm. How are gage blocks made?
>>
>> A surface grinder gets them to a certain dimension greater than
>> the target dimension, and then lapping brings them to their final dimension.
>>
>> But gauge blocks are flat on *both* ends (excluding setting rods
>> for micrometers, which have spherical surfaces at the ends), and most
>> chuck jaws have a rounded end to match the curvature of the chuck body
>> when the jaw backs are adjusted out to the surface of the chuck body.
>> And this is just an approximation, since it is not a critical dimension.
>> It is the position of the inner points relative to the scroll teeth,
>> which are at a different position for each jaw.
>
> OK. Why is lapping any more likely to yield parallel faces than a
> surface grinder?
It isn't -- but it can produce both a better finish (You can't
wring blocks which are just surface ground together), and it can remove
very small amounts of material at a time -- allowing approaching the
final dimension with less chance of overshooting. (And remember that
the grinding increases the temperature a lot more than lapping does.
Does anyone finish lenses by grinding? All that I know are
finished by lapping. The surface finish is an important part, as well
as the final dimensions.
>> >> Another trick which has been suggested at times is to drill
>> >> three holes at the apicies of an equilateral triangle, and then bore out
>> >> the center until the holes allow the jaws to fit in and protrude into
>> >> the center area.
>> >
>> > This depends on the accuracy of the jaw's outside faces, but these will
>> > typically be far less worn than the inside faces.
>>
>> Not the *outside* faces, but rather the angled faces -- the same
>> ones which I was suggesting you use spacer blocks on.
>>
>> 1) There are three holes in an equilateral triangular pattern.
>>
>> 2) There is a hole bored in the middle which intersects all three
>> outer holes.
>>
>> 3) The jaws are tightened until the angled faces press on the
>> sides of the holes.
>>
>> Using the outside faces could be done with a single large hole,
>> not four holes in a specific pattern. But -- There is no certainty that
>> the outside faces loaded out will be in a fixed relationship to the
>> inside ones loaded in. It depends on the individual backlash in the
>> scroll teeth for each jaw.
>
> I'm not visualizing this. The angled faces would not find flat surfaces
> to rest upon in all those bored circular holes, so it would be hard to
> know the effective spacings between jaws as tightened.
Apparently, it is close enough so the deformation of the points
left by the joining of the holes allows the jaws to all reach the same
preload force (which may or may not be with equal spacing between the
faces, BTW).
Do a web search on "truing chuck jaws" and you will find several
mentions of this system. Maybe one of them can answer your questions
better than I can.
[ ... ]
>> Interesting. Very little wear in that chuck, then.
>
> Nor does the chuck look like it has seen much use.
>
> I'm thinking that I will not attempt to regrind the chuck jaws, as they
> test OK, and the increasing the preload on the spindle bearing had a
> dramatic effect on the chatter, so the chuck was probably falsely
> accused.
I suspect that you are right.
>> >> Anyway -- those could be done on a
>> >> surface grinder with a fixture which keys on the notches in the top
>> >> jaws.
>> >
>> > I'm sure that's what the jaw makers do.
>>
>> Only for the two-piece jaws. No bets on how the one-piece jaws
>> are handled.
>
> Probably ground in a fixture that has been lovingly tweaked over the
> years.
Not sure of that. The proper positioning is also dependent on
the scroll plate, so they may be finished in a chuck body with the
scroll plate with which they are to be used.
[ ... ]
>> > You know, in an earlier day, I did look at Sherline, but was put off by
>> > the lack of steel in its construction.
>>
>> O.K. Look at the competition. Trying to remember the US name.
>> The UK name is Peatol. It has a ground steel dovetail bolted to the top
>> of an extruded aluminum bed, with cavities in the bed filled with
>> concrete to dampen vibrations.
>>
>> I've got one of those (which may be why I can't remember its
>> name? :-). I actually got it at a hamfest quite a few years ago, before
>> it was sold under the current name here. Mine has *no* name marked.
>>
>> Anyway -- unlike the Sherline, which uses a hand-cranked
>> leadscrew to move the carriage, this uses a rack and pinion, like the
>> hand feed on a serious lathe.
>
> I never heard of Peatol before. All that's advertised here is Taig and
> Sherline, who are both relentless.
And the Taig was the name which I was struggling to remember.
> I bet it's like soap - 1/3 of the
> price is the cost of the advertising to pound it into your skull, to
> blot out all other possibilities.
:-)
> They do look good. <http://www.peatolmac.talktalk.net/>
I like the Taig/Peatol. Among other nice features, it is
covered with T-slots (headstock, tailstock, etc), making it easier to
fixture up for strange tasks.
> But it appears that Peatol is made by Taig, and the difference is
> unclear: <http://www.lathes.co.uk/taig/>.
The spelling of the name, and who gets what share of the profit,
I suspect. :-)
O.K. Mine is not like the one at the bottom of the web page, so
it is a later one -- lots of T-slots instead of a cast-in "Micro Lathe"
logo, and a much longer bed.
The 4-jaw chuck is an excellent one, and I modified mine to fit
the Compact-5 -- and then ordered a replacement for the Taig. The 3-jaw
chuck has two-piece jaws, but the tommy-bar drive of the scroll plate
does not allow you to tighten it as much as the pinion gear on a normal
3-jaw chuck. Also, there is no keying to assure that the jaws replace
precisely. The screws which secure them offer what positioning you can
get.
[ ... ]
>> > Hardhat?
>>
>> Just makes it more likely that I will hit something. My
>> hardhat adds at least another inch to the top of my head. :-)
>
> But reduces the impact-induced IQ decrement.
I tend to learn quickly for things like that. I am tall enough
so I tend to spot things likely to hit my head, while shorter people are
not accustomed to looking for overhead hazards. Not that I am *really*
tall, but tall enough. :-)
>
>> [ ... ]
>>
>> [ ... quite a bit snipped -- so I can perhaps read something
>> else in the newsgroup before I go to bed. I've been doing too
>> much else the past few days -- and lost about five hours to a
>> thunderstorm-induced power outage -- plus another hour getting
>> all the systems back awake and talking to each other. ... ]
>
> This has happened once to me up here near Boston, over the last 20
> years, and no permanent damage, so I count myself lucky.
We get an outage long enough to require shutdown of the systems
before the UPS runs out of charge at least once every other year. Often
it is during a thunderstorm which blows tall trees into the HV three
phase going along the street near our house which typically causes the
power to shut down, then retry after about five seconds. Sometimes this
happens frequently enough to damage things. We lost an HP Color
Laserjet 5 to that a couple of years ago. The power was going down and
coming back too often, and something in there died. I was never able to
find an external fuse, and once I started tearing it apart to make room
for the HP 4600dn which replaced it (the LJ5 from a hamfest, the 4600dn
from a local computer recycler) I found five fuses in there -- all good.
I never did identify what failed. But it was amazing how much I had to
remove to find those fuses. :-)
I used to lose a Frame Relay CSU/DSU about every other year,
when lightning struck either the power line or the phone lines not too
far from the house. Since they replaced the wiring, I have not had a
repeat of that problem -- so far. But since I went to a T1 line, there
is a board of electronics at the demark (just outside the house) with is
the phone company's responsibility, and (so far) that has not gone out
either.
I am temped by the FIOS fiber optics -- but until they can offer
me things like a Class-C static IP block, I won't play. I'm not that
interested in the cable service which they are pushing, and I like the
copper-based phone system still -- it keeps working during a five-hour
power outage. I'm not sure how long the charge on the FIOS battery
backup will last.
> I do have a
> pair of big UL-rated surge arrestors attached to the main panel. The
> arrestors are made in Texas, where lightening is far more a problem than
> here. <http://www.deltala.com/>
I grew up in South Texas -- and in those days, electronics were
tube-based, so a bit more forgiving of surges. :-) But the lighting rods
occasionally took a hard hit, and you could hear them thrumming through
the whole house. :-)
> What I have are the LA-302 units, purchased from the local electric
> supply house. <http://www.deltala.com/prod02.htm>
That looks good. Do you need two to protect the typical 240 V
US service, or does one unit protect both sides. It sounds as though
you only need one -- but you mention having two.
And where is it mounted? It looks as though you poke the wires
in through a knockout and route the wires to the incoming power point --
perhaps after the main breaker so you can shut off while you connect.
:-) Is it intended to be visible so you can easily detect if it has been
hit hard enough to take it out of service?
All of the servers, and the router are powered by a 2KVA BEST
FERRUPS which will keep all of the servers and the router running for 45
minutes (now verified by the logs from the last outage), using four 12V
large automotive sized gel-cells in series. It also monitors for weird
power line frequencies, and stabilizes the power from either the line
or the batteries. BEST made really high quality UPSs, but they died or
were bought out some years ago.
> I did this in liu of getting a lot of surge protection power strips,
> because these units are tiny (in terms of surge capacity), together cost
> more than the two LA-302 units, and cover very little of the equipment
> in the house. The only disadvantage (as seen by the general public) is
> that you need to be able to install something in the main power panel,
> which explains why the power strips are big sellers.
They sound a like a good idea for the things which I can't hang
on the UPS. The two client computers (my wife's and mine), and the two
laser printers (you really don't want a fuser hung on the UPS. :-)
Anyway -- the power went down just as we were leaving for a
get-together with someone from out of the area -- so we had to trust to
the UPS and its software to shut the computers down cleanly. Apparently
they worked as desired.
Enjoy,
DoN.
--
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---
|
|
Posted by Joseph Gwinn on June 1, 2008, 11:27 pm
Please log in for more thread options
> >
>
> [ ... ]
>
> >> > Yes. The HF unit is a clone from Baldor, and I see suitable wheels in
> >> > catalogs and flyers all the time. Like Enco 391-5716, which I have seen
> >> > as low as about $80 on sale:
> >> ><http://www.use-enco.com/CGI/INSRIT?PMAKA=391-5716>.
> >>
> >> O.K. Then perhaps I could use one as well.
>
> O.K. Harbor Freight's web site is back up -- and this is
> designed to mount via four socket head cap screws to a flange, not to be
> bolted with nuts and flanges directly on the spindle.
True. It's a knock-off of the Baldor carbide tool grinder.
> I see no RPM listed.
>
> And I *do* see the following:
>
> For Baldor & Other Popular *CARBIDE* Bench Grinders
> ^^^^^^^
>
> My emphasis on the "carbide" part. I know that the carbide grinders
> have driving faces instead of just clamp flanges. I *don't* know what
> the RPM is.
The nameplate on mine says 3400 rpm, 1/2 hp.
I don't understand the part about driving faces and clamp flanges.
With HF, I find it best to download and read the users manual.
What grinder do you have, that you would be mounting a diamond wheel on?
> >> > Although when I replaced one gray wheel with aluminum oxide, I had to
> >> > modify the wheel housing so the Norton wheel would fit, using a die
> >> > grinder that often outran my compressor.
> >>
> >> *That* would worry me -- because it would reduce the wall
> >> thickness needed to protect the operator when the wheel shatters.
> >
> > The housing is cast iron, and quite thick, 7 mm at the outer edge,
> > thicker near the bade. It was the widening near the base that caused
> > the interference. I think I took a few mm off near the base, yielding a
> > cover that no longer widened so much near the base, but was never less
> > than 7mm thick.
> >
> > Actually, are plated diamond wheels likely to burst at all? Aren't they
> > made of metal?
>
> But you thinned the housing to fit a stone, not a diamond wheel
> from what I read above.
True, but how is that relevant to the question? And both wheels are the
same nominal diameter, 6". With the HF carbide tool grinder, all the
action happens on a flat face, like a cup wheel, and the cylindrical
outer edge is not used.
> >> >> >> And when grinding carbides, you are generating an inhalation
> >> >> >> hazard -- so I'll let others with supplied air do the grinding. :-)
> >> >> >
> >> >> > I think I'll most likely die of something else first. And, I do have
> >> >> > that respirator.
> >> >>
> >> >> O.K. Just be careful.
> >> >
> >> > I'm curious - why is carbide dust any worse than say HSS or stainless
> >> > steel dust?
> >>
> >> I *think* that it is the Tungsten matrix in which the carbide is
> >> embedded.
> >
> > Actually, the matrix is cobalt, which glues tungsten carbide particles
> > into a solid mass, like cement gluing stones into a solid mass called
> > concrete.
>
> O.K. But there are MSDS sheets which come with each carbide
> tool from MSC which warn about grinding them.
I get from MSC a MSDS every time I buy a steel bar, telling me of the
dangers of iron.
> [ ... ]
>
> >> > I'll try this. But I'm trying to hang a number on this, because the
> >> > lathe was designed and built to some required minimum static stiffness
> >> > value. I suppose I could ask Clausing, but it's unlikely that anyone
> >> > knows the answer anymore, even if they are willing to say.
> >>
> >> It may have been in the original specs in the catalogs -- or it
> >> may not have been.
> >
> > I don't recall ever seeing such a number except in textbooks on the
> > design of machine tools.
>
> O.K. Then those should give you the ability to calculate the
> deflection -- if you take the time to measure everything carefully. :-)
I'm also hoping to motivate people to test and publish the results.
> [ ... ]
>
> >> > With a 1" diameter bar, flex in the bar is less of a problem.
> >>
> >> It still exists.
> >
> > Yes, but if one doubles the diameter the stiffness of the bar increase
> > by a factor of 8, so it's pretty easy to simply overpower the problem,
> > such that something else limits the accuracy.
>
> Does it? I don't have the formulas to hand, but I would expect
> only a factor of four.
Time to look into Roark, Formulas for Stress and Strain, 1954.
Table III, page 104:
The max deflection of a simple cantilever beam is Max
y=-(1/3)W*l^3/(E*I), where W is the applied force, l is the length, E is
the stiffness of the material (modulus of elasticity), and I is the
section moment of inertia. Assuming that only the size and shape of the
bar cross-section varies, only the moment of inertia I matters here.
Table I, pages 74-75, on the properties of cross-sections:
Circle: Moment of inertia I=(Pi/4)*R^4, where R is the radius.
Rectangle: Moment of inertia I=(1/12)b*d^3, where b is the width and d
is the depth, with the bending being in the depth direction.
So we are both wrong. Doubling the diameter of a round bar will
increase the stiffness (decrease the max deflection) by a factor of 2^4=
16, while doubling the depth of a rectangular beam (the width remaining
constant) will increase the stiffness by a factor of 2^3= 8.
> [ ... ]
>
> >> > With a 1" diameter 1018 bar mounted in a BXA-4 boring-bar holder
> >> > perpendicular to the 5914's bed way, a 5 Kg weight hanging on a cord at
> >> > 32.25" from the centerline (defined by the 5/8-20 mounting bolt) causes
> >> > a 0.015" deflection (with respect to the bed) at 7.25" from the same
> >> > centerline. This is 3262# per inch of deflection, at 7.25" from center.
> >> >
> >> > Banging through the math, this is 0.004 degrees per foot-pound of
> >> > torque, or 249.4 foot-pounds of torque to cause one degree of
> >> > deflection.
> >> >
> >> > My suspicion is that this value is lower (floppier) than it should be,
> >> > but I have no comparison as yet.
> >>
> >> Remember that HSS is *not* normal steel and has a higher
> >> stiffness. Carbide has an even higher stiffness, which makes
> >> particularly good boring bars.
> >
> > True enough, but the objective here is to measure the stiffness of the
> > machine tool holding the HSS or carbide bit, and not to measure the
> > stiffness of the boring bar. The use of a BXA-4D boring bar holder is
> > fortuitous - the holder accepts a 1" bar, and a 6' length of 1" CRS bar
> > fits perfectly and can be clamped as if it were a boring bar.
>
> Hmm ... that would not work on *my* lathe. The bed is not long
> enough to support a 6' bar by the middle without it having to pass
> through the spindle. (Hmm ... 1" bar, 1-3/8" bore -- I guess that you
> could do that a bit at least.
I see. I have confused things by talking of boring bars. In the
present test, the boring bar holder is clamped to the left side
dovetail, and the bar is perpendicular to the bed ways.
I have also done the test with a shorter bar parallel to the bed ways.
For this, I removed the tailstock so the bar could project out off the
end of the bed, with the weight dangling just beyond the end.
> But -- I am interested in seeing the difference when you run that
> test at right angled to the bed and parallel to the bed. I'll bet that
> it is a lot worse parallel to the bed, which is the kind of load you
> would have with the offset of the tool holder on the BXA toolpost.
What is reported above is the perpendicular-to-bedway value. I will
repeat the parallel-to-bedway test.
> > I do have some Criterion solid carbide boring bars (bought used for a
> > few dollars each), and they do a really nice job.
>
> Solid carbide with non-interchangeable cutting surfaces? Mine
> is a Kennametal, with the bar 1/2" diameter (with a pair of opposed
> flats for clamping it in the tool holder) with the very end steel brazed
> on, and triangular inserts. I know that they are Kennametal, because I
> had to dig through the overstuffed drawer beside the lathe to find them.
> I finally wore out the third point on the insert which I have been using
> for some years. :-)
They are solid carbide, cutting tip and shank, with a cylindrical steel
shank sleeve brazed to the carbide.
> [ ... ]
>
> >> Of course, I have a weight set which goes up to 5Kg weights.
> >
> > Good, but they are not common. I had to get a set of test masses to
> > calibrate a scale that had gone seriously (~25%) out of calibration.
> > It's actually a lab utility scale (6 Kg by 0.5 gm), but my wife took it
> > over for cooking. So, this scale has to work.
>
> :-)
>
> I've got a two-pan balance with beam weights (picked up at a
> hamfest) plus the set of weights (picked up at an estate sale. I wish
> that I had been able to meet the previous owner. He had a fascinating
> collection of stuff. :-) And I have a Metzler single-pan balance. The
> other "pan" is really a three-legged spider onto which various diameters
> of metal rings are lowered or lifted clear by a set of knobs for the
> two MSDs, and a set of tiny rings hung from a tiny beam to handle the
> LSD (not counting the knob which applies bias to the whole thing.) That
> also came from a hamfest. The pan is enclosed in three glass panels,
> two of which slide open for access. The tilting of the beam is picked
> up by a mirror, through another couple of mirrors, to project a moving
> scale on a ground-glass window.
Those are real scientific scales. Perhaps a bit fussy for cooking,
though.
> [ ... ]
>
> >> > I've been doing this by resting a finger on various gaps while I torque
> >> > the toolpost around. The fickle finger points at the cross-slide
> >> > slideway.
> >>
> >> So attack that first. Operate on the assumption that when you
> >> fix all of the weak points, you will have a lathe which will do what
> >> Clausing designed it to do.
> >
> > Actually, I attacked the spindle bearing preload first, to considerable
> > good effect.
>
> O.K.
>
> > A question: On your Clausing, if you pull the backgear pin, so the
> > spindle can turn freely, when you turn the spindle by hand, how much
> > drag do you encounter? And, is there an initial resistance to turning,
> > which resistance lessens once the spindle is turning? I'm wondering if
> > I over-tightened the spindle preload adjustment nut, but have no
> > comparison.
>
> Mine may not be truly equivalent, since mine has three belts
> applying stress to the side of the shaft. But I get what feels like a
> 3:1 or perhaps 4:1 ratio of starting force vs moving force. It pegged
> the force gage which I was able to find -- at about 1.5 KG applied at a
> point 3.125" from the center (on a chuck jaw at the OD of a 6-1/4"
> chuck.)
Ah! That sounds exactly how it now feels like when turning the spindle
by hand. It sounds like I did not over-tighten it.
> The heavier force gauge has gone into hiding for the moment. I
> last used it to check the match of a newly purchased clock weight to the
> original one (time was still there, gong was missing, but I remembered
> them as being the same dimensions.)
Use a fish scale and a cord?
> [ ... ]
>
> >> >> > A short circuit to ground would cause the spindle to stop, and to +10
> >> >> > would cause the spindle to go at the max speed set into the VFD.
> >> >>
> >> >> A short circuit is assumed to be to ground. If the +10 is not
> >> >> properly insulated, look more for a short between it and ground.
> >> >
> >> > One can have a short between +10 and the slider on the pot, which would
> >> > both cause the machine to operate at full speed and possibly burn the
> >> > pot out.
> >>
> >> Yes -- it is is set to a low speed position, it is quite likely
> >> to fry the pot. The carbon film ones would be the most likely to go
> >> first, wirewounds next, and the carbon composition ones like the
> >> Allen-Bradleys would last the longest.
> >
> > Yep. The speed pot in the X-axis drive on the Millrite blows out from
> > time to time, and the cause is always that the 10K carbon-film pots
> > cannot handle the current drawn from the wiper arm, causing opens where
> > the film is burned off the backing.
>
> Hmm ... It looks as though State electronics has taken over the
> production of the Allen Bradey carbon composition pots. Look for "Type
> K". Check with:
>
> <http://www.potentiometers.com/>
>
> or
>
> <http://www.state-elec.com>
>
> and look for "Series K".
>
> I was considering going with them until I discovered that active
> circuitry will offer me more flexibility.
I'm just looking for a simple fix for the X-axis power feed, and the
circuit diagram isn't available. (A diagram did come with the unit, but
the provided diagram makes no sense electronically, and has no
connection to what's on the circuit board.)
> > The symptom is that the speed
> > varies from stopped to full speed in one big step.
>
> It would be. I've seen enough of this sort of failure, though
> not on a speed control pot. :-)
>
> > I talked to the
> > seller, and he complained that this happened a lot, causing a big
> > warrantee bill, and he had been changing vendors (all asian) to no
> > avail. I pointed him to the Japanese makers of heavier better pots, for
> > which he was happy, as he then had something to push back with. Don't
> > know how it came out.
>
> O.K.
>
> > I bet the original design (that was cloned) assumed a real carbon
> > composition pot. The pot and assembly is an oddball mechanically, a mix
> > of english and metric sizes, and retrofitting it is one thing that
> > awaits the lathe growing up to be a lathe.
> >
> > Wirewound sounds workable. For instance, CTS 026TB32, like DigiKey
> > CT2159-ND, for $3.60 each, which is a 10K 5-watt wirewound in the
> > standard form factor of the traditional molded carbon pots.
>
> That sounds as though it would work -- other than eliminating
> the finest of the control -- but you don't need that fine a control. ;-)
A 10K wirewound will use *very* fine wire.
> >> [ ... ]
>
> >> >> But I'm sure that a shop vac could produce enough pressure to
> >> >> push sufficient breathing air through a 1/4" line -- just as long as
> >> >> the
> >> >> demand regulator is made with a sufficiently large aperture.
> >> >
> >> > Probably. I have no idea what minimum input pressure typical demand
> >> > regulators are made to handle.
> >>
> >> I think that we would have to make our own demand regulator,
> >> with something like a 1/4" aperture covered with a soft rubber pad under
> >> diaphragm control.
> >
> > Making a regulator to handle 10 psi should not be hard. The only thing
> > I would worry about is if a failure in the regulator can cause full
> > inlet pressure to appear at the facemask. The design should be such
> > that failure causes zero output.
>
> If you get full inlet pressure to the facemask, it would simply
> float clear of the face with lots of air escaping all around -- which
> would *still* keep the undesirable components out. If you had straps
> tight enough to build up a pressure dangerous to the lungs, you would
> probably also be limiting blood circulation to the scalp. :-)
OK.
> [ ... ]
>
> >> >> > The 20th century solution is a surface grinder.
> >> >>
> >> >> Nope -- not unless you can be *sure* that the distance from the
> >> >> back of the jaws to the gripping face is constant for all three jaw
> >> >> positions. I would not like to bet on that.
> >> >
> >> > Hmm. How are gage blocks made?
> >>
> >> A surface grinder gets them to a certain dimension greater than
> >> the target dimension, and then lapping brings them to their final
> >> dimension.
> >>
> >> But gauge blocks are flat on *both* ends (excluding setting rods
> >> for micrometers, which have spherical surfaces at the ends), and most
> >> chuck jaws have a rounded end to match the curvature of the chuck body
> >> when the jaw backs are adjusted out to the surface of the chuck body.
> >> And this is just an approximation, since it is not a critical dimension.
> >> It is the position of the inner points relative to the scroll teeth,
> >> which are at a different position for each jaw.
> >
> > OK. Why is lapping any more likely to yield parallel faces than a
> > surface grinder?
>
> It isn't -- but it can produce both a better finish (You can't
> wring blocks which are just surface ground together), and it can remove
> very small amounts of material at a time -- allowing approaching the
> final dimension with less chance of overshooting. (And remember that
> the grinding increases the temperature a lot more than lapping does.
OK. It all comes down to some kind of precise machine.
> Does anyone finish lenses by grinding? All that I know are
> finished by lapping. The surface finish is an important part, as well
> as the final dimensions.
In the IR, lenses can be machined with a diamond point. But in the
visible, lapping is required. Except for molded lenses. But this only
makes the lapping once removed - one laps the mold.
> >> >> Another trick which has been suggested at times is to drill
> >> >> three holes at the apicies of an equilateral triangle, and then bore
> >> >> out
> >> >> the center until the holes allow the jaws to fit in and protrude into
> >> >> the center area.
> >> >
> >> > This depends on the accuracy of the jaw's outside faces, but these will
> >> > typically be far less worn than the inside faces.
> >>
> >> Not the *outside* faces, but rather the angled faces -- the same
> >> ones which I was suggesting you use spacer blocks on.
> >>
> >> 1) There are three holes in an equilateral triangular pattern.
> >>
> >> 2) There is a hole bored in the middle which intersects all three
> >> outer holes.
> >>
> >> 3) The jaws are tightened until the angled faces press on the
> >> sides of the holes.
> >>
> >> Using the outside faces could be done with a single large hole,
> >> not four holes in a specific pattern. But -- There is no certainty that
> >> the outside faces loaded out will be in a fixed relationship to the
> >> inside ones loaded in. It depends on the individual backlash in the
> >> scroll teeth for each jaw.
> >
> > I'm not visualizing this. The angled faces would not find flat surfaces
> > to rest upon in all those bored circular holes, so it would be hard to
> > know the effective spacings between jaws as tightened.
>
> Apparently, it is close enough so the deformation of the points
> left by the joining of the holes allows the jaws to all reach the same
> preload force (which may or may not be with equal spacing between the
> faces, BTW).
>
> Do a web search on "truing chuck jaws" and you will find several
> mentions of this system. Maybe one of them can answer your questions
> better than I can.
Is this it: <http://www.loganact.com/tips/chuck-jaw.htm>?
> [ ... ]
>
> >> >> Anyway -- those could be done on a
> >> >> surface grinder with a fixture which keys on the notches in the top
> >> >> jaws.
> >> >
> >> > I'm sure that's what the jaw makers do.
> >>
> >> Only for the two-piece jaws. No bets on how the one-piece jaws
> >> are handled.
> >
> > Probably ground in a fixture that has been lovingly tweaked over the
> > years.
>
> Not sure of that. The proper positioning is also dependent on
> the scroll plate, so they may be finished in a chuck body with the
> scroll plate with which they are to be used.
Using a fancy fixture to hold the jaws in plane but under pressure as
for actual use.
> [ ... ]
>
> > They do look good. <http://www.peatolmac.talktalk.net/>
>
> I like the Taig/Peatol. Among other nice features, it is
> covered with T-slots (headstock, tailstock, etc), making it easier to
> fixture up for strange tasks.
>
> > But it appears that Peatol is made by Taig, and the difference is
> > unclear: <http://www.lathes.co.uk/taig/>.
>
> The spelling of the name, and who gets what share of the profit,
> I suspect. :-)
>
> O.K. Mine is not like the one at the bottom of the web page, so
> it is a later one -- lots of T-slots instead of a cast-in "Micro Lathe"
> logo, and a much longer bed.
>
> The 4-jaw chuck is an excellent one, and I modified mine to fit
> the Compact-5 -- and then ordered a replacement for the Taig. The 3-jaw
> chuck has two-piece jaws, but the tommy-bar drive of the scroll plate
> does not allow you to tighten it as much as the pinion gear on a normal
> 3-jaw chuck. Also, there is no keying to assure that the jaws replace
> precisely. The screws which secure them offer what positioning you can
> get.
OK.
> [ ... ]
>
> >> > Hardhat?
> >>
> >> Just makes it more likely that I will hit something. My
> >> hardhat adds at least another inch to the top of my head. :-)
> >
> > But reduces the impact-induced IQ decrement.
>
> I tend to learn quickly for things like that. I am tall enough
> so I tend to spot things likely to hit my head, while shorter people are
> not accustomed to looking for overhead hazards. Not that I am *really*
> tall, but tall enough. :-)
Stay away from warships. A 5' zip Chinese female colleague of mine told
me how she rung her bell a few times while installing systems in US Navy
ships. This was a great surprise to her - she is not at all used to
running out of headroom.
> >> [ ... ]
> >>
> >> [ ... quite a bit snipped -- so I can perhaps read something
> >> else in the newsgroup before I go to bed. I've been doing too
> >> much else the past few days -- and lost about five hours to a
> >> thunderstorm-induced power outage -- plus another hour getting
> >> all the systems back awake and talking to each other. ... ]
> >
> > This has happened once to me up here near Boston, over the last 20
> > years, and no permanent damage, so I count myself lucky.
>
> We get an outage long enough to require shutdown of the systems
> before the UPS runs out of charge at least once every other year. Often
> it is during a thunderstorm which blows tall trees into the HV three
> phase going along the street near our house which typically causes the
> power to shut down, then retry after about five seconds. Sometimes this
> happens frequently enough to damage things. We lost an HP Color
> Laserjet 5 to that a couple of years ago. The power was going down and
> coming back too often, and something in there died. I was never able to
> find an external fuse, and once I started tearing it apart to make room
> for the HP 4600dn which replaced it (the LJ5 from a hamfest, the 4600dn
> from a local computer recycler) I found five fuses in there -- all good.
> I never did identify what failed. But it was amazing how much I had to
> remove to find those fuses. :-)
>
> I used to lose a Frame Relay CSU/DSU about every other year,
> when lightning struck either the power line or the phone lines not too
> far from the house. Since they replaced the wiring, I have not had a
> repeat of that problem -- so far. But since I went to a T1 line, there
> is a board of electronics at the demark (just outside the house) with is
> the phone company's responsibility, and (so far) that has not gone out
> either.
So far, I have not had such dramas.
> I am temped by the FIOS fiber optics -- but until they can offer
> me things like a Class-C static IP block, I won't play. I'm not that
> interested in the cable service which they are pushing, and I like the
> copper-based phone system still -- it keeps working during a five-hour
&g | | |
Yahoo! 
Google 
Windows Live 
del.icio.us 
digg 
Netscape 
