A pox on the fly cutter!

[Ubarch]

A POX ON IT!


Seriously, I have no idea what I'm doing wrong here, but this thing is acting like the mill isn't squared, when it is in fact squared to the best of my abilities. Both "roll" and "pitch" of the Z-axis assembly (I'm using a Sherline model 2000 mill) square up just fine with respect to the table. However, the fly cutter always seems tilted. If I take 5 mils off of a piece of aluminum when the bit hits the metal in the initial 180 degrees, it's likely to take off nearly 20 mils when it hits the metal coming around the next 180 degrees. I'm fairly sure the work isn't moving around, and I've tried to square the milla s best I can. The bit has been used on aluminum and some cast iron.

The Sherline book makes a note about this behavior in fly cutters, but this is beyond what I would expect. Can anyone tell me the most likely sources of error? It seems I'm either not squaring the mill correctly, using a dull bit (I don't think it should be dull by this time... it is carbide), or there's some fly cutter technique that I'm missing.

Any ideas?

Thanks.

[HuFlungDung]

It sounds like flexure. You did not state what diameter you are working the tool on, but it does exert quite a bit of leverage, since it is most likely quite far from the spindle centerline and bearings. On a light duty machine, this is inevitable. Is this one of those machines which uses ordinary ball bearings in the spindle instead of angular contact spindle bearings? Do you have a means to lock the quill to help stabilize it?

[Mcgyver]

when the cutter is over the work, there's force involved and with a flycutter this force has some leverage. it is very normal, in fact its a good sign everything is working well to the get cross hatch pattern - ie cutting on both the forward and back 180. with everything right, the tool marks should be about the same.

rigidity of the machine & set up, the depth of cut, feed, speed, dia of flycutter, and the shape of the flycutter tool all play there part. you have to work these things out or tell us what you're doing and we'll help

what is a mil? if its a millimeter, i've found the problem how did you square the mill when - trammed with a good indicator? how square did you get it?

in general, reduce the cutting force and see what happens - slow it down, light cut, slow feed, hss bit with lots of rake, etc, at the same time tighten everything up; setup, gibs, quill lock etc. flycutting demands a lot from a mill, don't how well the sherline is up the task, but if you're flycutting a 6" slab of CI its a BIG difference than a 1" piece of AL

[Ubarch]

HuFlungDung:

I don't know the answer to either of the questions with confidence. If the quill is the rotating armature that holds the mill bit, then I do not have a means to lock it. I'll try to find out what kind of bearings are used; a quick google search didn't turn up anything immediately useful.

Mcgyver:

1 mil = 0.001".

I squared it with... squares. I have a set of precision squares. The table should be exactly orthogonal to the Z-axis column, so I tested that with the largest square I had; for both degrees of freedom applicable ("pitch" and "roll"), the column lined up well enough to prevent light from passing through the seam between itself and the square.

One thing I dont have are precision shims, which I would need to grip my work properly in this case. The work might be moving as I cut it.

I'll try reducing the cutting force, and report back if this works.


Thanks!

[Geof]

The simple answer is that the axis of your spindle is not perfectly perpendicular to the plane of motion of your table; if it was the flycutter would cut precisely the same depth around a full circle. It is unrealistic to expect you will ever get it perfectly perpendicular but you should be able to get it better than with .015" over the diameter of your flycutter, however probably not by using a square.

If you do not have a dial indicator that you can mount to the spindle so you can sweep out a circle maybe 4 inches in diameter and check the dial reading to the table is the same all the way round you can actually use the flycutter to true the spindle.

Bring the flycutter slowly very down to the work until it just touches somewhere around the circle; you should rotate it by hand to determine this. If it touches to your right when you are standing in front your spindle is out of true in a closkwise direction; touch to the left is counterclockwise, touch nearest you means the top is leaning toward you and touch furthest is top leaning away.

To adjust the spindle have the clamping bolts tight enough that you cannot move either pitch or roll but not fully tight. Now give a nice sharp tap at the bottom of the spindle with a plastic hammer or failing that a block of wood and retest the touch of the flycutter. If your nice sharp tap has moved things too far slightly tighten the clamping bolts and tap it back the other way. Eventually you will develop a feel for how tight the bolts should be and how sharp the tap should be to move the spindle a somewhat predictable amount. Incidentally you can define a nice sharp tap as that impact which applied to the side of your own head would give a good headache but not lay you out unconscious on the floor.

Using this technique it is possible to true a spindle well enough to get the cross hatched tool marks mentioned in an earlier post. If you can get it so the crosshatching is perfectly even that is extremely good and you should be willing to accept uneveness in the cross hatching.

I forgot to mention earlier all this adjustment should be done with the table at the center of its travel in both axes and to get your cross hatching do not move more than one flycutter diameter each way from this position. It is very likely that at the extremes of table travel you will see the crosshatch pattern change due to the table sagging with the weight all at one end of the travel.

[One of Many]

A few questions/suggestions for mill inspections that I use:

How big is your indicator swing radius? A makeup mirror a bit larger than the table will give a better result than a smaller radius. A sticky indicator can also screw up the result.

If you tram it in then move the head, quill or knee(for those that have these features) during part setup. There may be cases where other sliding members shift or no longer run square/parallel to the expected plane(s), that can alter the tram.

Check and/or tram the y axis ways to the spindle before or after it is trammed to the table. Set up some 1-2-3 blocks or some other standard that is above the table resting on the y axis ways and indicate off them front rear, right left of the saddle. Notice any difference between the 2 surfaces? If so, the problem can be in the saddle, table surface or column and likely a combination of the three. Seek out the clues!


Another approach. I would set up on a piece with the fly cutter and get it to cut with a nice cross hatch in both the x and y travels. Make adjustments to bring it in until it does cut equally about the fly cutter swing. Then indicate the table to see how far out the table surface is to the ways. Cheesy but it does correct for anomalies that are not easy to rectify.


Test bar method. Find a known flat ground steel bar(Preferably less than .0005, but at least record any bow). Mount the bar mid-point to a shank that can be gripped in the spindle. It is easiest if the bar is mounted on the side of a flat or slot cut into the end of the shank so it can swivel and be locked in place. The bar should be no longer than what can swing between the column and spindle. Use an indicator mounted on the table(mag base) to check that the bar can rotate 180 deg by hand and be set equal at each end between the table and bottom face of the bar, then lock it up tight. It will not matter at this point if the head is trammed during this setup. The point is to know the test bar is running perfectly perpendicular to the spindle center line and the bar is flat. Prior to the next step it should be trammed as close as possible. Stick the mag base indicator on the test bar at the extremes with the table centered.

Put the indicator back on the table. Now you can move the table with the test indicator running against the bottom face of the test bar to detect drift and list in the way surfaces relative to each axis. This inspects the running surfaces and not the arbitrary table top that may not be aligned to anything. Is what we are looking for is loose gibbs, worn ways indicated by listing due to weight shift, dips and arching along the travel and at the travel ends. A straight line can be compensated, sudden increasing drift cannot. Determining the relevant cause can take a bit of experience to discover. If you think about it long enough, you can develop tests to prove or verify the logic of square and parallel to each axis in relation to a spinning centerline. If anything else, if you map the reliable envelope of tolerable drift. At least you'll know what should be avoided.

Lastly........I have heard of some of the small mill tables can get significant bow when tightening down on small part using oversized clamping. It checks fine with it off again, but the tweak returns when trying a second attempt with the same setup. The lesson is to make sure the hold down clamps are close to the work. I'm sure there is a ratio of clamping length to table strength, but I'd think that to be obvious.

edit: And now I see that you are using a square to tram. EEH Gads@!

DC

[DareBee]

EEH Gads - exactly, definately not good enough to square up a spindle for milling.
I am really curious to know why you call a "Thou" (0.001") a "mil"

[Ubarch]

I am really curious to know why you call a "Thou" (0.001") a "mil"

A mil is a real unit of length.

Here's a length conversion utility that will let you convert one inch into mils:

http://www.onlineconversion.com/length_all.htm


Anyway, there have been several promising techniques posted, which i have the proper equipment to try. I'll report the results.

Thanks!

[turmite]

Noun, 1. mil - a unit of length equal to one thousandth of an inch; used to specify thickness (eg, of sheets or wire). linear unit - a unit of measurement ...

This is from the free online dictionary. You typically see mil dimensions used in proplylene sheeting, at least that is the only place I have seen it used. I once rebuilt pianos for a living and the wire I bought for that was in thousandths.

Mike

[Geof]

Originally Posted by DareBee I am really curious to know why you call a "Thou" (0.001") a "mil"

My guess is that he has a strong background in Physics which is quite compatible with being an electrical engineer.

Hint to Ubarch; you gotta learn the lingo or you will get dumped on.

[WhiteTiger]

I suspect this pic of the machine in question will pretty much explain the problem with the flycutter.


Tiger

[ahren]

Urbarch, I work with high accuracy pick and place robotics for a living, and we use "mil" all the time to describe 0.001", as in "This machine is placing at +/- 2 mils at 3 sigma right now". The guys at the machine shop next door use this terminology as well. I'm with you .