"Bill Schwab" wrote in message
...
Harold,

Once you've set a dial, don't reset it while you're running the part, or
series of identical parts. It's too easy to screw up by a few thou.
Naturally, when there's no alternative, do so, but be careful you don't
introduce backlash while spinning the dial to the new setting. Then
remember you've lost your datum point, so the dial will no longer be in
sync
with it.

When you say a series of parts, it sounds like you are assuming some
type of jig or setup block that will get the next piece in place,
otherwise the dial setting would be meaningless for the next part. How
does that work? Or am I reading too much into it?

Pretty simple, really. If you have more than one part you'd like to process,
it can be located adequately from a set of stops from the T slot for the
rear, or even a parallel clamped to the table and dialed true, and a single
stop on the left side. The T slot or parallel squares the part with the
machine axis, and the left side provides a reliable location of your part(s)
when installed in the setup so the part shares the same datum points with
the previous part(s). That stop should be a narrow spot, so the part is
always squared by the T slot, not the side stop. That assures the parts
always bank the same way when they're not perfectly square. It works
the same way from a vise with a permanent stop by any means available to you
for the left hand side. I have the back (fixed) jaw of my vise drilled
and tapped 1/4-20 for setting various stops, or even clamping a strange
setup of parallels if so needed. They're really handy, and get used
frequently. You can also set a stop from the table if you're using a
vise. What ever works and won't move when bumped. A piece of pipe or
tubing, faced parallel, works great when clamped to the mill table in the
appropriate place.


Thinking about it, if I am following you and if there is a robust way to
place pieces after the dials are set, it would be useful if the part
needs to be turned and reclamped.

Exactly--------or, as I've stated, drilling or otherwise processing more
than one part that shares similar features. For example, you can drill
and tap one plate, then drill the other for the appropriate screws. Even
when not drilling oversized holes for the screws, you can expect the two
plates to go together with no trouble---even when you've drilled a dozen
holes or more. This method is very precise as long as you don't screw up
with backlash, and you stop on the lines as required. Try that with a
layout and a center pop. Never happen, not even when you bust your butt
trying.



Yes, the scale serves as a lap counter in a sense. What you'll find is
that when you have to step off inches, you count turns, 5/inch. If your
machine has dials calibrated .200"/turn, it's real easy to work in
thousandths, which I assume you do. Smaller mills often have dials
that
are .125"/turn. They make it a little harder because you can't read
your
dimensions directly as you can with the .200 dials (for example, 7/16"
is 2
turns + 37 on the dial, whereas with a .125 dial it's 3 turns + 62 on
the
dial-----it requires a little more attention). Takes a little getting
used
to, and it's where you'll make some scrap initially, but once you've
mastered the dials, it's no big deal, and you can plot holes easily
within a
couple thou reliably, assuming your screw is in good condition.

I am tempted to say they are 0.1 per turn, but I might get home and wish
I had deleted this line. The screw seems to be in fine shape. So far,
I think there is more variability in my use of a dial caliper than there
is trouble with the screws.

Could be, but that's a pretty fine screw. Not likely if it's an Acme
thread, but who knows? What ever it is, you simply work accordingly.

Don't fall into the trap where you think a caliper of any kind is *truly* a
precision instrument. The potential for bad readings is so great that I
don't recommend them for anything closer than .005". While you'll find you
can make readings that are far more reliable than that, you'll also discover
(usually by scrapping a project you cherish) that they have a way of
letting you down when you need them the most. By their very nature,
they are subject to considerable error, particularly when you use them for
both inside and outside measurements. It's not uncommon for the jaws to be
apart in readings by more than double the tolerance you may be required to
hold. Establishing a proper slip or snug fit with them is taxing. Very
taxing, indeed. There are other methods of measuring that are far more
reliable. Small hole gauges, or telescoping gauges, for example, and, of
course, micrometers. The hole gauges require a little experience to become
familiar with them, but you can measure to tenths with care. Never happen
with a caliper. Not reliably.




What I do is use a wax pencil and make marks on the saddle/table as
reference points, noting the dial mark at that position. It's
especially
important to mark your "0" points, so you can go back to your reference,
or
starting points when in doubt. Assuming the screw on your machine is
good,
by so marking your machine, you can step off holes time and again,
stopping
at the same location without thinking. That's a nice feature if you're
drill
ing, tapping, counterboring, or otherwise going back to each location
for a
specific function that may not lend itself to being done while you're
there
the first time.

That is an excellent argument for leaving the dials set. I have been
resetting them to make the immediate readings easier to understand, but
with overshooting to remove backlash and a "flight plan" (or your wax
pencil marks) with distances from the datum and dial readings to expect,
it should be fine.

It really shines when you have more than one piece. Otherwise, with care,
it makes little difference. Needless to say, if I have no reason to
revisit a location, I don't mark the table aside from often marking my datum
points. You never know when you might have to go back to 0 for some reason.
Saves you starting over with the edge finder.

I encourage you to do what ever you end up doing, to do it consistently, so
it becomes routine. Only then will you make parts without making scrap. I
made a ton of bad parts getting on to drilling this way-----parts that were
worth a ton of money----flight stuff for a missile. Virtually all of it
would have been avoided had I practiced what I preach about using a scale
for each move. I was young and had other things on my mind, as you might
imagine. I ended up divorcing her years later. Best move I ever made.
(Next to learning to trust my 6" scale, anyway). :-)



Each edge finder seems to have its own personality, at least those I've
used. Yours appears to be no different. I think you're on the right
track by trying it at various speeds, even if you see it with each one
tried. One of them will be more "clean" than the others. Won't take
you
long to figure it out. When you find the magic number, you'll notice
it's
more reliable than the others, tending to be more sensitive.

The good news is that I am probably running too slow for Aluminum
anyway, so if I get the belts the other way around, a temporary speed
increase would be be a single belt move. With the dials sitting in one
place, I should find myself needing to use the edge finder only once per
clamping or head movement, which hopefully the ER set will make rather
rare.

Along with good planning. That's really key to running a machine like
yours----but it's true across the board, even on machines that are more
forgiving.

If you're machining aluminum with cutters under 1", it's highly unlikely you
have enough spindle speed to be a problem. I often run a 1" end mill @
4,200 in aluminum. If the cut is deep and fast, the problem I encounter
is lack of power(2 HP Bridgeport)------not over speed.



Sounds like things are coming together for you. Can't tell you how
nice it
is to share my experience with someone that is willing to learn instead
of
challenging each and every comment. You've been a pleasure.

Likewise, it is a pleasure to get sound advice from someone with
experience. Thanks for sharing it. I am certain I will get around to
"challenging" you soon enough - it is a poor student who fails to do
that from time to time.

I welcome intelligent questions, and don't mind trying to discover why my
advice may have been bad----it's the guy that thinks he knows more than
someone that has dedicated a life time to his trade that troubles me. I
don't claim to know everything, but my years on the machines were a
resounding success. I did sub-contract work for Litton Guidance and Control
for 16 years, mostly tooling, with a little product along the way. Tough
work, all defense oriented, much with tolerances in tenths. 5
rejections in 16 years. Yeah, I'm proud of that. I'd have been even
more proud if I could have had none, but I'm only human! g


Come to think of it, my "flight plan" idea might be my first rebellion
Some of it will depend on what I do for CAD. In the past few years,
the things I have done that required any precision were circuit boards,
and I fell back on a very simple homegrown framework that allowed me to
ultimately drive a printer to get a 1:1 layout of the traces. I did
that only when pre-printed/drilled boards ($3 or so at Radio Shack)
would not work, so it was rare. That same setup could perhaps kick out
the tedious part of a sequence of waypoints, absolute coordinates, dial
readings on my mill, etc.

If anyone cares to recommend an inexpensive but flexible CAD program (2D
should be fine), I am willing to consider it. I have some $30 thing I
bought a while ago, but it's a real pain to use. The snaps do not work
well, and editing objects to change length is a problem because it seems
to "think" only in scaling at that point: I have yet to find a way to
simply set the lengh of a line after it's been drawn. Getting it right
from the start is challenging because of the flaky snaps.

Thanks!!

Bill

I don't know if Intellicad is still offering a free download, but it's a
pretty nice program, especially for free. www.cadopia.com

I'm not wise enough to know if it would serve your purpose, but it's worth a
look.

Harold