How do you check out a small horizontal Mill ?

[David Powell]

My wife says it does not look " Square" I rather think that she is being deceived by the tapers in the unmachined castings. But this set me thinking, how do you check to see if everything is " Square".?
As yet the spindles are not fitted, but the slides are reassembled and adjusted just tight, but not too tight even at the ends of the travels.
As an initial check I mounted the round overarm ( 11/2 In dia) in its proper place, tightened the clamp bolts sensibly, moved the mill table up to almost full travel, put a dial gauge on the table with its mag base, tightened the locks on the x and z slides set the dial gauge pointer against the bottom of the overarm( by sight) and then wound the table to the column and back out to its most outward position.
I repeated this a couple of times and each time got about a thous movement each time, showing the table was dropping as it was wound outwards. It has 4 inches of travel. I tried leaning hard on the table and got about another half thous drop if that.
Where should I go from here? How can I tell if the X movement is really at 90 degrees to Y movement. What other fauts could bite me and how can I detect them?I am just a general machinist and here I am trying to pretend to be a machine tool fitter!!!!.
After all have to convince the wife I know what I am doing.
Regards David Powell

[Bented]

As with any other machine apply an indicator to the parts that move on ways, if the deviation is beyond what you require simply move on to the next target.

[J Tiers]

What do you have as "standards"? That makes a difference.

My actual first test is to check that the vertical ways are straight and the front surfaces of both lie in the same plane. I use a surface plate and some blue for that. You may not have that available, so you will have to assum that is true unless and until it is proved otherwise.

You need a couple accurate parallels of decent length that can be clamped to each other and to a surface (not the thin type). You need a decent sized accurate square. Whenever you are measuring, the ways that are not in use should be clamped so as not to move.

Using the bottom of the overarm is a reasonable idea. But it has issues. The overarm SHOULD be parallel to the spindle, and parallel to the table in and out movement. But it may not be perfect, and really does not have to be perfect to function. It may have bumps or whatever on it and a thou is not an unreasonable amount of error with it.

The first measurement is probably the spindle. It needs to be perpendicular to the vertical ways You can check this directly for right-left errors by having the spindle in place. then clamping an indicator to the spindle so that you can sweep it over the two vertical way faces. Reading should be the same.

Next, you need two parallels that are each known accurately flat on at least one surface. Since the ways do not extend past the spindle, you need to "extend" that surface. So by clamping one parallel to one of the ways, sticking up past it, and then clamping the other parallel to it so that the second is about the same distance above the spindle as the original check was below it, you can verify that the remaining clear way reads the same as the surface of the second (upper) parallel. if so, then the spindle is accurately perpendicular to the way the parallels were on.

If you then remove the parallels, extend the indicator farther, and check another spot on each way, you can check if they are parallel. The indicator should read the same on both there also.

For the table, vertically, you really only care that it does not rise or fall as it goes under the spindle. That means a slot or surface will be cut parallel to the table as to depth. Clamp your indicator to something solid on the upper part of the mill, with the tip on the table under the spindle. Wind the table back and forth. It should not change the reading. If so, then the table is likely moving correctly, but in any case, the table top is not rising or falling. That is X in one plane.

Horizontally, there are two things to check. Mount the indicator to the table, with a largish domed tip on it if possible. Bring the table in until the indicator is reading on one of the vertical ways. Now wind the table until the indicator is on the other vertical way. Reading should be the same. The large dome tip is just to avoid it "catching" on the edge of the way. If the ways are in the same plane, the reading of the indicator should be the same along the width of each as well. Check that, although if they are not you will be having a problem checking the same reading for both. This is the X movement in the other plane.

Now you are pretty sure the table is moving perpendicular to the spindle, so a slitting saw will not be forced to the side. the spindle is perpendicular to the vertical ways (trammed vs them) and the table is moving parallel to the front of the vertical ways. (If the way faces are not in line with one another, then most of this will not work. They should be in the same plane, but wear etc can mess that up).

next get out the square. Put it across the two vertical ways, sticking out over the table. make sure the leg across the ways is horizontal. Mount the indicator on the table, with the tip against the "blade" of the square, on the "square" surface. Crank the table in and out. The reading should not change if the table is moving square to the vertical ways. This checks Z in one plane.

Next, clamp the square to one way, with that leg vertical. Get the indicator to read on the proper surface of the square, which will now be a horizontal surface. Crank the table in and out. The reading should not change if the movement is perpendicular to the vertical ways. See if you repeat your 0.001 measurement. this is Z in the other plane.

The last direction is Y. Again clamp the indicator to the column. mount the square to the table perpendicular to the front of the vertical ways. Set the tip of the indicator against the square on the surface facing the vertical ways. Crank the vertical movement. you want the indication to be the same everywhere. It may not be, so check at several places, each time locking the vertical movement. All positions should read the same at least in the locked condition.

Repeat the measurement with the square mounted with the leg parallel to the vertical way front surfaces measuring on the same surface you have been using, for the indicator tip. Same deal, it should read the same everywhere, at least in the locked condition.

The last measurement is for your own convenience. Mount the indicator on the column, with the tip against the inside of one of the table slots. Wind the table back and forth, checking that the indication is the same all along. Check all the slots, and both sides of each. Now you know the best one to use for aligning a vise, or a workpiece.

Is this overkill? Well that depends. it's what I did when scraping the Benchmaster. I wanted it "right" and I think I got there. The same checks will tell you how your machine is as far as alignment. All the movements and so forth are checked.

you DO need a couple good standards. but you kind of need them anyhow, so hopefully you have them already.

[David Powell]

Thankyou for your very detailed reply. I will try all the tests you suggest tomorrow Meanwhile I simply clamped a 123 block to the bed indicated it true to the Y movement then put the indicator to read the X movement. I get an indicator travel of 2 thous for 1" movement of the table. Is this a valid test? Regards David Powell.

[J Tiers]

If I am visualizing what you did correctly, it should be. But since you got a measurement that makes little sense, either there is a problem or the dial is screwey.

A picture would help. (And I agree pics would help my description as well.)

What are your definitions of X, Y, and Z? I am using axial to spindle as Z (in and put for a horizontal mill), table longwise travel as X, and Y as up and down. So I'm not sure what you did.

[oxford]

What are your definitions of X, Y, and Z? I am using axial to spindle as Z (in and put for a horizontal mill), table longwise travel as X, and Y as up and down. So I'm not sure what you did.

I would have called it the same as a vertical mill which I think it technically wrong which leads to what do you label axis on a normal lathe?

I always thought long travel was X and cross slide was Y.

[David Powell]

HI , Sorry for any confusion, facing the mill X is the crosswise travel of the table, y is the travel towards or away from me and Z is the vertical travel of the table. I borrowed the vertical head from my compact 5 outfit. The mill deserves much better, it is a bit worn, but certainly fully usuable, Regards David Powell. Time for bed!!!

[J Tiers]

I would have called it the same as a vertical mill which I think it technically wrong which leads to what do you label axis on a normal lathe?

I always thought long travel was X and cross slide was Y.

Conventionally Z is along the spindle axis.

http://eng.kuraki.co.jp/mother/boring/fraise/index.html

[J Tiers]

HI , Sorry for any confusion, facing the mill X is the crosswise travel of the table, y is the travel towards or away from me and Z is the vertical travel of the table. I borrowed the vertical head from my compact 5 outfit. The mill deserves much better, it is a bit worn, but certainly fully usuable, Regards David Powell. Time for bed!!!

Are you moving in Y and measuring the movement in X? I guess I am not seeing this correctly.

Anyway, if there is 2 thou error in X when moving an inch in Y, then either the block is not as you think (not made square, or actually not aligned square), or the Y movement is not correct vs the spindle or at least vs the plane of the vertical ways.

You have to have a known accurate square reference. The 1 2 3 block "SHOULD" be one, but some of them are not so good.

Then with a known good square, you have to know what it is referenced to. I reference everything to the vertical way front surfaces, and the first things I make good are those surfaces.

That's why I suggested using them. But if they are not in one plane, then all bets are off.

You can, if you have a good flat surface, preferably a good surface plate or granite flat, check two blocks or squares for being good. Ideally you hav ethree, but two are possible.

Set them on the surface, and butt the blades together. If they meet accurately everywhere, they may be good. If they do not, then all you know is that one at least must be wrong.

So then if they do butt well, then you turn one around, and see if both will butt to a flat surface accurately. If they do both, then they are likely good. If there is any failure, then at least one (you do not know which) must be wrong, or else both are wrong.


Pic if setup would help. Even a sketch. Sketch might be better, because you can write on it to explain what was done.

[Mcgyver]

Thankyou for your very detailed reply. I will try all the tests you suggest tomorrow Meanwhile I simply clamped a 123 block to the bed indicated it true to the Y movement then put the indicator to read the X movement. I get an indicator travel of 2 thous for 1" movement of the table. Is this a valid test? Regards David Powell.

The test is valid for checking the squareness of the two axes, assuming the block is square. Unfortunately that its rather out of square. If you try and twist the table, what happens to the indicator reading? Just wondering is there lots of slop, a gib loose etc. The other thing to check is the X to the Z (spindle). Mount an indicator in the spindle and using the block you already indicated to the X, revolve the Z so that you catch both ends of the block. If they are the same reading, the X is square to the table. If so, that might be good enough as a lot of work on the horizontal (especially if you have a vertical) is just big cuts along the X. If not, time to start tearing it apart and surveying it

btw, Z is always the the spindle axis. I don't think I'm overly pedantic, but that's an important one so we all stay on the same page. But you mentioned got vertical head on it? In that case your original description of the axes is correct.....I was going off the title "horizontal mill"

[Richard P Wilson]

X movement being square to Y movement isn't as critical on a horizontal mill as it is on a vertical, because you can only cut in one direction on a horizontal.

[Bented]

I would have called it the same as a vertical mill which I think it technically wrong which leads to what do you label axis on a normal lathe?

I always thought long travel was X and cross slide was Y.

On every lathe that I have used in North America the "Z" axis is a negative number from part 0.000 towards the spindle which makes sense because left of 0.000 makes the part shorter.

The "X" axis is a positive number moving toward the operator from "X"
0.000 and a negative number on the far side of "X" 0.000

If a 1" diameter by 1" length is desired the finish position would be X 1.000, Z -1.000, if turning from the back side of "X" 0.000 the position would be X -1.000 Z -1.000
This is a CNC lathe in DRO, in this case the tool position is .341" diameter and .358" away from the end of the part in absolute coordinates (ABS)

[Bob Engelhardt]

"Horizontal Milling Machine Testing Procedures" : http://web.archive.org/web/20150308110620/http://shopswarf.orconhosting.net.nz/hormill.html

[J Tiers]

X movement being square to Y movement isn't as critical on a horizontal mill as it is on a vertical, because you can only cut in one direction on a horizontal.

That is true to an extent, UNLESS you use it like a vertical, or like an HBM, with an end mill in the spindle, and the surface being worked on facing the spindle (and oriented vertically). But if you ever want to use a vertical head, you will want it right.

[Richard P Wilson]

Agreed