I have watched these instructions several times and I am still puzzled at
what appears an error in the method. I am sure everyone knows what I am
talking about, but I will recap the instructions:

1) Find edge of the hole at a point A using an edge finder (wiggler was used
with a 0.250" ball). Set your Y axis to zero.
2) Move to point B which is diametrically opposite to A on the Y axis. Use
wiggler to find edge at point B.
3) Determine the distance from A to B by counting the turns of the handle
and add whatever figure corresponds to the partial turn on the dial.
4) Divide this figure by 2.
5) Move to AB/2 on the Y axis. Lock Y axis.
6) Repeat the procedure on X-axis. The point arrived at is the centre of the
hole.

My problem with this method is that it ignores backlash.
If backlash is B and the AB distance is the true diameter of the hole then
the actual dial reading is going to be B+AB. On reversing to reach the
half-way point on the Y-axis the backlash has to be taken up again (I assume
the same amount, B) but now we are moving the table by a distance of
[(B+AB)/2]-B. Thus the actual distance the table travels is going to be
(AB-B)/2, i.e the centre is going to be off by half the value of the
backlash.

I would be happy if anyone pointed the fault in my reasoning or outlined the
correct procedure.
I know there is indicators but that is cheating :-)

Thanks,

--
Michael Koblic,
Campbell River, BC

I think the instructions simply require YOU to take the backlash into
account. Before I got a DRO for my old mill, I had to keep a notepad
beside the mill all the time, recording the numbers for going left and
right the included calcs for the backlash. I thought that everyone did
that before the days of ball screws and electronic "linear scale" counters.

Pete Stanaitis
-----------------

Michael Koblic wrote:
I have watched these instructions several times and I am still puzzled at
what appears an error in the method. I am sure everyone knows what I am
talking about, but I will recap the instructions:

1) Find edge of the hole at a point A using an edge finder (wiggler was used
with a 0.250" ball). Set your Y axis to zero.
2) Move to point B which is diametrically opposite to A on the Y axis. Use
wiggler to find edge at point B.
3) Determine the distance from A to B by counting the turns of the handle
and add whatever figure corresponds to the partial turn on the dial.
4) Divide this figure by 2.
5) Move to AB/2 on the Y axis. Lock Y axis.
6) Repeat the procedure on X-axis. The point arrived at is the centre of the
hole.

My problem with this method is that it ignores backlash.
If backlash is B and the AB distance is the true diameter of the hole then
the actual dial reading is going to be B+AB. On reversing to reach the
half-way point on the Y-axis the backlash has to be taken up again (I assume
the same amount, B) but now we are moving the table by a distance of
[(B+AB)/2]-B. Thus the actual distance the table travels is going to be
(AB-B)/2, i.e the centre is going to be off by half the value of the
backlash.

I would be happy if anyone pointed the fault in my reasoning or outlined the
correct procedure.
I know there is indicators but that is cheating :-)

Thanks,

spaco wrote:
I think the instructions simply require YOU to take the backlash into
account. Before I got a DRO for my old mill, I had to keep a notepad
beside the mill all the time, recording the numbers for going left and
right the included calcs for the backlash. I thought that everyone
did that before the days of ball screws and electronic "linear
scale" counters.

This was a Part 2 of an instructional DVD series. In Part 3 they sort of
mention that the backlash "cancels out" using this method. Not according to
my calculations. Even if one goes past the mid-point and returns to it from
the same direction there is still going to be half a backlash error (in this
case half a backlash too far).

Theoretically one can ignore the backlash if one a) knows the diameter of
the hole accurately and b) performs the measurment *exactly* on that
diameter. The benefit of this method is supposed to be that one does not
have to measure exactly on the diameter as the mid-point of the measurment
will always lie on the X-axis diameter. I suppose one can compensate for
known backlash but can one really measure backlash accurately? Does it
change over time?

--
Michael Koblic,
Campbell River, BC

Theoretically one can ignore the backlash if one a) knows the diameter of
the hole accurately and b) performs the measurment *exactly* on that
diameter. The benefit of this method is supposed to be that one does not
have to measure exactly on the diameter as the mid-point of the measurment
will always lie on the X-axis diameter. I suppose one can compensate for
known backlash but can one really measure backlash accurately? Does it
change over time?

if you're working to this accuracy you're over due for DROs. You'll
find its like turning the light on in the room after working in the
dark for years.

But, it sounds like you have valid points on the backlash issue.

I've only used a wiggler for really accurate placement after first
getting within a few thou by other means. This would lessen backlash
concerns.

Karl

On Jan 20, 8:00*pm, "Michael Koblic" wrote:
I have watched these instructions several times and I am still puzzled at
what appears an error in the method. I am sure everyone knows what I am
talking about, but I will recap the instructions:

1) Find edge of the hole at a point A using an edge finder (wiggler was used
with a 0.250" ball). Set your Y axis to zero.
2) Move to point B which is diametrically opposite to A on the Y axis. Use
wiggler to find edge at point B.
3) Determine the distance from A to B by counting the turns of the handle
and add whatever figure corresponds to the partial turn on the dial.
4) Divide this figure by 2.
5) Move to AB/2 on the Y axis. Lock Y axis.
6) Repeat the procedure on X-axis. The point arrived at is the centre of the
hole.

My problem with this method is that it ignores backlash.
If backlash is B and the AB distance is the true diameter of the hole then
the actual dial reading is going to be B+AB. On reversing to reach the
half-way point on the Y-axis the backlash has to be taken up again (I assume
the same amount, B) but now we are moving the table by a distance of
[(B+AB)/2]-B. Thus the actual distance the table travels is going to be
(AB-B)/2, i.e the centre is going to be off by half the value of the
backlash.

I would be happy if anyone pointed the fault in my reasoning or outlined the
correct procedure.
I know there is indicators but that is cheating :-)

Thanks,

--
Michael Koblic,
Campbell River, BC

When you approach the sides from within the backlash affects both
edges symmetrically, as though the hole was simply larger. Consider
the limiting case where the edge finder fits snugly in the hole. If
you 'find' opposing edges the dial readings will be + and - half the
backlash from the true center.

This assumes the backlash is constant. It may not be for a large hole
on a worn milling machine. This old brain needs another cup of coffee
after visualising all that.

I use the pointed wiggler and a magnifier to get close, then an
indicator if the job justifies it, which is rare. I can usually get
close enough quickly by centering a 6" ruler* over the hole and moving
the wiggler point to the 3" line. The lines are 0.003" wide and I can
just about center on them.

Jim Wilkins
* Some call them scales, but to me a scale measures weight.

This is why real milling machines (and not a mill drill) have dials on BOTH
ends of the X axis table. You set one to zero going in one direction and the
other to zero going the other direction. And yes this assumes that the
screw is worn evenly and the backlash is constant.

And as been said the answer to the original question is it DOES cancel out.

And as also been said, get a DRO and forget about worn screws and backlass
forever "und jus vatch das blinken glowie lamphenfiber".

--.- Dave


Before I got a DRO for my old mill, I had to keep a notepad beside the
mill all the time, recording the numbers for going left and right the
included calcs for the backlash. I thought that everyone did that before
the days of ball screws and electronic "linear scale" counters.

Pete Stanaitis
-----------------

Michael Koblic wrote:
I have watched these instructions several times and I am still puzzled at
what appears an error in the method. I am sure everyone knows what I am
talking about, but I will recap the instructions:

1) Find edge of the hole at a point A using an edge finder (wiggler was
used with a 0.250" ball). Set your Y axis to zero.
2) Move to point B which is diametrically opposite to A on the Y axis.
Use wiggler to find edge at point B.
3) Determine the distance from A to B by counting the turns of the handle
and add whatever figure corresponds to the partial turn on the dial.
4) Divide this figure by 2.
5) Move to AB/2 on the Y axis. Lock Y axis.
6) Repeat the procedure on X-axis. The point arrived at is the centre of
the hole.

My problem with this method is that it ignores backlash.
If backlash is B and the AB distance is the true diameter of the hole
then the actual dial reading is going to be B+AB. On reversing to reach
the half-way point on the Y-axis the backlash has to be taken up again (I
assume the same amount, B) but now we are moving the table by a distance
of [(B+AB)/2]-B. Thus the actual distance the table travels is going to
be (AB-B)/2, i.e the centre is going to be off by half the value of the
backlash.

I would be happy if anyone pointed the fault in my reasoning or outlined
the correct procedure.
I know there is indicators but that is cheating :-)

Thanks,

On 2009-01-21, Dave August wrote:
And as also been said, get a DRO and forget about worn screws and backlass
forever "und jus vatch das blinken glowie lamphenfiber".

DRO certainly works very well in this regard, indeed.

i

Dave August wrote:
This is why real milling machines (and not a mill drill) have dials
on BOTH ends of the X axis table. You set one to zero going in one
direction and the other to zero going the other direction. And yes
this assumes that the screw is worn evenly and the backlash is
constant.
And as been said the answer to the original question is it DOES
cancel out.
And as also been said, get a DRO and forget about worn screws and
backlass forever "und jus vatch das blinken glowie lamphenfiber".


Ah, but this DVD series relates *specifically* to the mini-mill.

I spent a long afternoon today trying to prove/disprove my assertion: Being
an engineering mutt I did everything at least 10 times. In the end, using a
dial indicator, I was able to show (conclusively I think) that I was
correct. The centre was consistently out by about 0.010".

Interestingly I spent a lot of time yesterday on the 'net trying to find out
if anyone else has commented on the same issue. Not a thing!

Some interesting findings as a by-product of the trial:

1) My edge finder (the cylindrical one) seems more accurate than the ball
wiggler (the 0.250" one) - by about 0.005".
2) Cheap digital calipers from Canadian Tire are surprisingly accurate - I
got the same inside dimension using them, a dial indicator and a telescopic
gauge with micrometer (also cheap).

It seems that lacking a DRO, more dial indicators, preferrably ones with
travel longer than 1" would be a good thing.

--
Michael Koblic,
Campbell River, BC