Center axis shift between collet and drill chuck

I'm getting a strange effect on my Bridgeport mill. I can zero an endmill on a part by cutting into the stock and setting that cut as my zero, but when I switch to my drill chuck and check the spindle zero with an edge finder, I find that it has shifted by .010". I'm using Lyndex collets and a Glacern R8 drill chuck; my endmills are YG and the edge finder a Mitutoyo. What can mechanically explain a spindle axis shift like this? I am using a DRO, so it cannot be operator error in reading the hand dials or lead screw error. I've racked my brain over every theoretical possibility but can't think of anything. My drill chuck runout is .002", which also cannot explain a .010" shift. The runout on my collets is .0002". Any ideas for me to explore would be appreciated.

If I understand correctly-
Did you measure the endmill to get your offset?, or just go with nominal value?


Sent from my iPad using Tapatalk Pro

After making the cut I moved to the centerline of the new edge by the cutter's radius. It's reported as, in this case, .500" diameter/.250 rad, and checks out at that with my caliper. The same effect occurs when I use the endmill while the machine is running to lightly scratch the edge of the part, which lets me know I'm at the cutter radius away from the part and the machine's center spindle axis, allowing me to move the cutter over by that amount and zero the DRO there. If I lock the table down and don't touch anything, and only swap out the collet for the drill chuck, the apparent center axis of the spindle will have shifted by .010" when I re-check with the edge finder. Not using an edge finder and simply drilling a hole, pinning it, and measuring the distance between the center of the pin and the part edge will also indicate that the offset has a .010" error. So, if I zeroed the endmill, swapped to the drill chuck, moved 1" over into the part and drilled a .250" hole and pinned the hole, I'd measure 1.135" from the edge of the pin to the wall of the stock that I originally zeroed the machine on. Minus the pin radius to get the center of the hole it'd be 1.010". I tried everything and am at the point of pulling what little hair I have left out

I didn't think of that! So even if the head is properly trammed in, there can still be a squareness error?

guest said:

I generally sweep the head with the quill extended, because it puts the pivot point as far from the indicator as I can get it. When the quill is retracted, the error will be reduced.

Click to expand...

It matters not where the quill is positioned. What matters is the distance from the head to the table (or whatever surface you are checking). For any given sweeping circle, the closer it is to the swing axis of the head the greater accuracy.

Thank you for the tips. I'll try to do that and see if I can eliminate the error. This will make my life a whole lot easier when working on this machine.

guest said:

The distance to the table is greater when you have to extend the quill to reach it.

If you sweep the table with the quill extended, that will be the greatest error. Greater distance to the table, greater distance from the pivot point. When the quill is retracted, the error will be equal or less because the pivot point is now closer.

That's what I want when I am squaring up the machine. I want to know the maximum error. I want to sweep as large a circle as I can, as far from the pivot point as I can get.

To put it another way- if I sweep in the head with the quill all the way up, and I am within .001", when the quill is fully extended the error will be more than .001".

In a perfect world it wouldn't matter- you are just measuring an angle. But BP's are far from perfect and having the quill extended magnifies the motion when you rotate the head, which makes it easier to get it squared up.

I don't tell anyone how to do their thing, I just say how I go about it.

Click to expand...

I don't agree. You are sweeping a plane with an axis at an angle to the plane. You will only get an exaggerated reading by increasing the diameter of the circle you are sweeping. Quill extension doesn't affect anything about tramming the head. Quill extension does affect tool position relative to a part datum if the tools have widely varying lengths, unless you get it all squared in, which is the point of tramming in the first place. But I never check a part datum when tramming, do you?

Is the table running true to the quill as it is raised and lowered? Are the table ways worn or is the table sagging? Just a couple things to check.

You also mentioned that you lock the table down. Don’t do that.
Lightly snug the table before edgefinding and leave it that way while you reposition. I’ve seen tables move drastically when the table is clamped.

Put it this way: if the indicator on the sweep arm shows .001 diff from one side to the other, and you lower the table and extend the quill, why is the indicator reading going to increase? It won't so far as I can tell. Extending the quill only changes the position on the table top where the sweep circle is positioned. It does not increase the sensitivity at all.

What you are arguing is that if you positioned the head at 45° to bore a hole, that the hole wouldn't be bored at 45° if you extended the quill to do the boring. That is circular reasoning that doesn't add up.

Drill chucks can be a bit of a problem in themselves. Check the runout of the edge finder once it's installed in the drill chuck to make sure that it is seated properly. I've found that drill bits, etc. can be "cocked" in a drill chuck, even a precision one, enough to make very bad things happen.

I doubt that there is a problem with the mill.

JMHO

-Ron

HuFlungDung said:

...What you are arguing is that if you positioned the head at 45° to bore a hole, that the hole wouldn't be bored at 45° if you extended the quill to do the boring. That is circular reasoning that doesn't add up.

Click to expand...

Well, no that is not what I was trying to say. Once the head is clamped the angle is what it is.

The context is tramming a BP.

I extend the quill when I tram in the head because it means the indicator swings a wider arc, hence more movement on the dial WHEN ADJUSTING THE ANGLE.

Take a yardstick, hold one end, and rotate it 5 degrees. The other end traces an arc. Now do the same with a 12" ruler- the moving end traces a smaller arc. If you fixed a test indicator to the end that is moving, the indicator on the yardstick will have more movement, even though the angle is the same. Longer hypotenuse.

guest said:

Take a yardstick, hold one end, and rotate it 5 degrees. The other end traces an arc. Now do the same with a 12" ruler- the moving end traces a smaller arc. If you fixed a test indicator to the end that is moving, the indicator on the yardstick will have more movement, even though the angle is the same. Longer hypotenuse.

Click to expand...

I thought you were right when you first posted to extend quill to get more accuracy. After letting it sink in I am sure it makes no difference if quill is extended or not. For the sake of discussion I will use your ruler and yardstick at 5º out of tram, let us also say that at 5º out the indicator is reading plus .050 and minus .050 on the low and high sides. (The other direction is in tram)
Now we spin the indicator around the ruler to tram head without the quill extended. The head is tilted at 5º and the indicator is reading +- .050.
So let us lower the table and extend the quill to yardstick length, the yardstick spins in the same axis with the same indicator set up. Tram has not moved so the indicator will read the same +- .050. If the table was dropped straight down without moving x or y the center of the indicated circle will be in a different spot, and that spot will be farther away from where the trammed center would be. But the same reading will be on the indicator if the quill is in either position since is showing the angle the head is tilted.

Rob F. said:

...But the same reading will be on the indicator if the quill is in either position since is showing the angle the head is tilted.

Click to expand...

Yes, the angle is the angle. If I am drilling a hole at 15 degrees, it doesn't matter how long the drill is- I can switch back and forth between a stubby and a jobber without losing location. As long as I don't move the table, I'm good no matter how much quill I use.

It's not about the angle changing with length (which is what people here are interpreting I said- I guess I am not articulate enough for this board).

Its ONLY about squaring up the head. When the quill is extended, the indicator is more sensitive to the turn of the bolt that tilts or nods the head. That's because the pivot arm is longer. A small tweak of the bolt gives a larger movement on the indicator. If the quill is retracted, the same tweak will result in less movement on the dial. The same thing happens when you tighten the bolts, which is why I like to have the quill extended. I can just "read" it better when I am getting it set, because the movement is magnified.

Once it's all tightened up, sweeping the table will yield the same indicator reading no matter where the quill is.

Think of a sine plate. You set a 30 degree angle on a 5" sine plate with 2.5" of gage block. If you want the same angle on a 10" sine plate you need 5" of gage block. Both sine plates are at 30 degrees, but the one with 10" centers needs more gage block to get there.

Two Z axis on a B-port and tramming your tabletop in will not get them true to each other if your machine is more than 2 days old.
The "square" tram on a extended quill will be different than a retracted quill due to the way the quill clamps and the necessary clearance so that it can move.
I make the assumption that you are working with better than tenths DTIs on your indicol.
A cylinder square can show you a lot as you tweak a compromise.

.010 is a lot, miles in my world, something is seriously out of wack.
Bob

guest said:

Yes, the angle is the angle. If I am drilling a hole at 15 degrees, it doesn't matter how long the drill is- I can switch back and forth between a stubby and a jobber without losing location. As long as I don't move the table, I'm good no matter how much quill I use.

It's not about the angle changing with length (which is what people here are interpreting I said- I guess I am not articulate enough for this board).

Its ONLY about squaring up the head. When the quill is extended, the indicator is more sensitive to the turn of the bolt that tilts or nods the head. That's because the pivot arm is longer. A small tweak of the bolt gives a larger movement on the indicator. If the quill is retracted, the same tweak will result in less movement on the dial. The same thing happens when you tighten the bolts, which is why I like to have the quill extended. I can just "read" it better when I am getting it set, because the movement is magnified.

Once it's all tightened up, sweeping the table will yield the same indicator reading no matter where the quill is.

Think of a sine plate. You set a 30 degree angle on a 5" sine plate with 2.5" of gage block. If you want the same angle on a 10" sine plate you need 5" of gage block. Both sine plates are at 30 degrees, but the one with 10" centers needs more gage block to get there.

Click to expand...

OK, I see that the farther away the indicator is to the pivoting head the more the indicator will move with the same head adjustment since it is swinging a larger arc. Has nothing to do with rotating the spindle.

What you are seeing is typical with knee machines with DRO's mounted on them. A lot depends on how much clearance there is in the ways and spindle bearings. Just because the DRO shows no movement that does not mean there is none. For example the X scales are mounted so that any movement in the X axis between the table and the base it fit's in BUT when tightening the Y axis the entire X base moves including the scale and the reader so the DRO indicates no movement. Same happens in Y. Though .010 is quite a bit .005 is about normal on a good used machine. Add movement of the Knee which will move the entire upper ways and the DRO will not be able to detect that movement! Then you have the quill, it's movement can be checked with an indicator. Then the collet holding the mill. if it is running slightly out the endmill will cut bigger than its size. To add to the problem drill chucks are not very accurate and they too can run out. To compound the problem you are using an edge finder that in not very accurate and a digital caliper adds to the error. I have a Cincinnati ToolMaster mill that is far more rigid than Bridgeport and I detect movement. On the other hand I have 2 Moore Jig Borers and get no movement error!
What your seeing is a little movement in the ways that the DRO can't read including the spindle. They all add up. You might try tightening the ways and using an indicator rather than an edge finder.