Re: [CAD_CAM_EDM_DRO]Backlash & Wobbly Spindle on my Bridgeport

Alex Canale wrote:

> Let's see..... how can I start.
>
> I'm drilling 3 (.012" dia.)holes on a .0238"dia. bolt circle 120degrees
> apart. I'm using solid micrograin carbide because I'm drilling into
> stainless 303. I've got the pecking cycle (G83) figured out and the machine
> drills beatifully. The problem comes in when I try to actually get the holes
> on the right spot.
>
> Allow me to explain myself. I use a co-ax indicator to center the piece
> since it's cylindrical in shape. Once I get the center I actually jog to the
> right spot and press the enter key to let the machine due it's pecking
> cycle. When that hole is done I jog it to the next hole location and press
> enter again, and so on and so forth with the last hole. When the part gets
> taken out of the vise it's visually off. The holes are shifted and don't
> even look like they are on a bolt circle rather they look like they were
> drilled eratically.
>
> So I decided to approach the project in a different manner. Rather then
> jogging it from the last hole location I would back the table up a good inch
> and a half and jog back to the new position (at this point I was thinking
> that the backlash was causing the problem). Unfortunately same problem, the
> holes are misplaced.
>
> So then I decided to put some dial indicators on the table. I would move the
> "X" .500" and back and I would register "0" backlash.

To measure backlash, you have to approach the same coordinate from
the OPPOSITE direction. So, go to X=1.000, then to x=0, and set up the
indicator, and observe reading. Then, go to X=-1.000 and back to X=0
and see if the reading is different.

> Then I moved the "Y"
> .500" and back....same thing, no back lash. So then I moved the table .010"
> "X or Y" and before I could move it back I noticed on the dial indicator
> that the table only moved .005"!!!!!! It seams like the table is accurate in
> distances larger than 1/2" but anything smaller the tolerances are out the
> window.

I can't tell from your description, but you may not be measuring backlash
the right way.

> Since I had all the indicators out I also decided to measure the
> eccentricity of the spindle. I used a .500" edge finder and put the
> indicator on it. It measured .0025"!!!! That's .005" off if I were to mill a
> slot the same size of the tool!

I'm not sure an edge finder is very useful, here. You might try a short
length of .500 hardened and ground shafting of known circularity and
straightness. You also need to know the collet is in good condition.

After checking runout, you can use the dial indicator to check radial play.
Apply a force of around 50 Lbs radially to the spindle nose, while checking
for movement with the indicator. (You may want to use a magnetic mount
and put the indicator on the main head casting, to separate ram deflection
from the spindle.) A small change is not a big worry, the lube in the
bearings take up at least .0005" when running.

Does this machine have a tiltable head? If so, your head might be out of
tram. If it is not perfectly perpendicular to the XY plane of travel, then
switching different length tools into the spindle will locate to different
points on the work. What I do is use a small end mill and plane the top
of a scrap of aluminum, and then move the spindle to the cneter and
sweep a dial indicator in a circle over the piece. If you get consistant
differences between left and right, or front and back, then the spindle is
not perpendicular to the XY plane of motion. This can even be true
if the spindle is perpendicular to the table top! (This would mean the
table is not parallel, top to bottom.)

I have a CNC program that mills a circular path and then goes to the
center of the circle. This makes is much faster to do this check.
Not only can I adjust the tram, but I can see the wear in the bottom
center of the table reproduced in the top of the piece! It is a saddle
sort of shape, due to the central part of the table ways being worn.

> I bought this machine used from e-bay for $4000.00 at the beginning of the
> year it has the Ahha controller. It's a series I 2hp Bridgeport with a
> taper 30 spindle. It's propably 25 to 30 years old. Is this ok? Can it be
> fixed? Or am I trying to split hairs with an ax here? Maybe I just got the
> wrong machine for the job!

Do you have gauge blocks? You can clamp a machinist's square on the
table, aligned with the axes, and a dial test indicator in the spindle. You
can move around in rational steps, and then place blocks of the same
dimension in the inside corner of the square to measure the actual
movement. Then, you can apply forces to the ends of the table, and
see if looseness of the ways is contributing to the error. You can also
apply force axially to the table to see backlash.

If you can answer some of my questions, you may immediately understand
the problems. If not, write back, and we will see if we can help determine
where the problems are. But, yes, you probably DID make a mistake
ASSUMING that a 30 year old machine would be able to maintain
this sort of accuracy! A new machine would have to be in good shape
to do this, most of the import mill-drills would not do it out of the box.
And, it is asking a bit much for a WELL used machine. I'm sure it
can be fixed, but until we determine what parts are worn, it is pretty hard
to know how bad it is going to be.

Jon