PDA

View Full Version : Face milling cutter question



Spookydad
01-11-2012, 10:41 AM
This is another beginner question....

I have a CDCO 2.5" indexable face milling cutter.

I have noticed that when I make a cut, the cutter will cut deeper on the left side of the cut. So if I am cutting left to right, the trailing side cuts deeper than the leading side. Cutting right to left, the opposite happens. This is most noticeable after all the material has been removed to level out the face being cut. It doesn't matter how deep a cut I am making. I have been keeping it to 2 or 3 thou just to be safe.

I have measured the vise and it is out by 0.001" across the 5" of the vise jaws. This will affect the out of square of the piece but during the cut, it should be level.

So are face milling cutters designed to cut this way?

Is the base of the face mill not quite perpendicular to the central vertical axis?

Thanks in advance

Neil

Tony
01-11-2012, 10:49 AM
your vertical axis (spindle) is not square to your table.
Do a search for "tramming" or "dialing in" a mill.

gnm109
01-11-2012, 11:32 AM
What Tony says. The mill is out of tram.....

Spookydad
01-11-2012, 12:48 PM
Thank you. I will tram the mill.

Neil

hornluv
01-11-2012, 12:58 PM
As an aside, .002-.003" is WAY too shallow a cut for carbide (and for HSS to be honest). It generally needs a deeper cut than HSS as most inserts are not terribly sharp. I make the minimum depth of cut the same as the radius on the insert as a general rule. .020" is a good starting place if you don't know the radius.

PixMan
01-11-2012, 07:03 PM
A general rule-of-thumb for indexable insert milling cutters (or turning inserts for that matter) is to alway have a depth of cut equal to or greater than the corner radius of the insert.

BTW, I wouldn't ever buy any of those CDCO insert tools because they've got a reputation for poor quality and worse performance. I don't know how much HP your machine may have, but those cutters usually have a very straight-up helix angle for the inserts. That results in a higher spindle load than many home-shop machines may have available. You can reduce the load, get known-quality inserts and smoother cutting action with other name brand tooling.

http://img.photobucket.com/albums/v466/kenm10759/Dads%20shop/IMG_1369-r.jpg

Notice the shear angle on that one, and compare to the CDDO, Shars or other Chinese import product. ;)

beanbag
01-11-2012, 08:53 PM
As an aside, .002-.003" is WAY too shallow a cut for carbide (and for HSS to be honest). It generally needs a deeper cut than HSS as most inserts are not terribly sharp. I make the minimum depth of cut the same as the radius on the insert as a general rule. .020" is a good starting place if you don't know the radius.

This advice is from the old days when carbide wasn't sharp.
My Sandvik face mill can peel off .0002" slivers using aluminum inserts.

PixMan
01-12-2012, 12:11 AM
Which Sandvik milling cutter do you have? I use a Walter cutter for aluminum that also has razor-sharp inserts, and it's incredible. A gloat is that I picked up the 4" cutter for about $70 off of Ebay, was brand new.

http://img.photobucket.com/albums/v466/kenm10759/Dads%20shop/IMG_1224-r.jpg

Spookydad
01-12-2012, 12:53 AM
I spent a couple of hours working on tramming the mill. I started working on adjusting the tilt of the head. Left to right, the table is now only out 0.0003" across 16" of travel. The trouble was in the front to back as the column is leaning forward.

I put an axial indicator in the drill chuck. I was having trouble getting consistent readings due to the grooves on the table. I decided to put my granite plate on the table. This allowed me to do a full sweep on the longer axial indicator arm. This confirmed the front to back lean of the column.

I loosened the bolts on the bottom of the column. I don't have any proper shim stock so I decided to use two of the cheap razor blade scrapers. (the ones withe the one edge covered up for safety.) By luck, this happened to be the exact thickness that I needed. Once the bolts were tightened up again, the coaxial indicator was showing a fluctuation of only 1 division on the full sweep. According to the dial that is only 0.0005". It is an import coaxial indicator so the accuracy is a bit suspect.

An indicator on the table shows it to be nearly dead flat so I am happy with that. I put a block in the vise and took a test cut. Left to right, the cut is now smooth but front to back, the cut seems worse. I started adjusting the jaws of the vise and got them level but then realized I should have checked the bottom where the parallels sit. So I measured the casting of the vise itself and found it to be out slightly left to right but badly, front to back. At that point I decided to call it a night.

I don't have access to a surface grinder so I am going to have to figure out some sort of shim stock to attach to the bottom of the vise.

Neil

macona
01-12-2012, 02:06 AM
Whoa. Stop.

A granite surface plate is not parallel in thickness. All you need is a couple 1-2-3 blocks and you can tram it in that way. I recommend am indicol or knockoff to hold the indicator. It allows you to adjust the swing of the indicator to pass over the two blocks.

beanbag
01-12-2012, 03:33 AM
If I may give some unsolicited advice:

You should tram the mill relative to the table's XY movement, NOT the table's surface. (These two planes are not necessarily parallel).

To do this kind of tram, put the granite block back on the table, or better yet, granite block + sheet of glass. Don't worry if the granite block is not parallel or level.

Put in an indicator coaxial to the spindle, or somewhat close. Go to a point in the center of the block, and call this height zero, (X,Y) location (0,0). Now measure the heights at the nearby points (-4,0), (0,-4), (4,0), (0,4) by turning the handles and moving the indicator there. Crank the xy back to (0,0) and put in an indicator on an arm with 4" swing. Swing the indicator across these points and note the heights. If they match, you are in tram.

beanbag
01-12-2012, 03:45 AM
My sandvik is one of those R245 models. I got it for $25 off ebay, but it was used and crashed. I had to replace one of the anvils and with some futzing around, eventually got everything within .0003. I also cut down the shank from 1.25 to .75 so it can fit in an R8.

J Tiers
01-12-2012, 08:27 AM
If I may give some unsolicited advice:

You should tram the mill relative to the table's XY movement, NOT the table's surface. (These two planes are not necessarily parallel).



While I may agree with this, I sure would not discount the table surface squareness vs tram....

If the table is a "wedge shaped" thing relative to the ways, then anything you put on top and flatten with any cutter will also become 'wedge shaped" to match unless you pack up one side of the work with shims.

Otherwise it would be like putting shims under one side of the work on a mill which has perfect alignments and tram.

photomankc
01-12-2012, 10:23 AM
Frankly, with my facemill, I prefer the appearance when it's just a hair out of tram vs perfect cross hatch. Besides the fact that I almost never get that since almost any serious cutting will slightly shift my little machine's tram the look of the single repeated arc is just much nicer to my eye. I have maesured the surface left behind and having the tram out by 0.0005 to 0.001 over 6" doesn't leave any trough that I can measure. It just means I always cut in one direction instead of back and forth.

Definately don't do the normal 'swing left/right and get to 0' tram meathod on a surface plate. I found that out the hard way too. They are flat, they are not not big, heavy, stone parallels.

BTW: My Glacern 45 degree face mill will shave off such small chips that they float in the air like little Aluminum confetti strips with the AL inserts.

Carld
01-12-2012, 01:41 PM
I don't see how moving the table from left to right and front to rear is going to sweep the head in.

To indicate the head you have to rotate a dial indicator in the spindle and take readings at left, right, front and rear 90 degrees at four points.

Also, don't ever trust a granite plate or any surface plate to measure the same from top to bottom. The top surface should be the only surface that is flat and not necessarily parallel to the bottom surface.

EDIT: beanbag, I read the post again and I think I understand what your saying. Your canceling out any deviations in height of the granite block. I still don't think it's a good idea to use anything laying on the table to sweep the head in. However everyone has ways of doing things. I suppose that laying a perfectly parallel plate on a table and sweeping the head in would cancel out any irregularities in the table surface if your table surface is terrible.

Spookydad
01-19-2012, 02:42 PM
I put my granite plate on the table and carefully shimmed it until I got it level. I tested the level by placing the indicator in a endmill holder, locking the quill, and moving the table back and forth under the indicator. I used my DRO's to make sure I was checking 2.5" out from the center in 4 directions. I got it as close as I could and marked the readings in chalk on the granite. I estimated the distance between the ticks on the indicator to get the 4th decimal place so that is not 100% reliable. I always started at 0,0 and zero'd vertical travel of the indicator at 0,0. The marks are shown in this picture:

http://i13.photobucket.com/albums/a273/SpookyDad/New%20lathe%20and%20mill/graniteplatetramtest01.jpg

The X axis shows a slight positive at both ends yet is zero in the middle. I tried getting a 0.0015" feeler gauge under a straight edge placed along the axis but it wouldn't go at all. Putting the straight edge on top of the feeler gauge caused it to rock back and forth. So I don't think it is the granite plate.

Next I put my axial indicator in the endmill holder and centered over 0,0 ( as measured by the DRO.) I set the arm to sweep the same 2.5" circle. I set the height of the head to make the needle go through a full rotation and back to "0". The needle moved 1 tick either side of "0" for the entire rotation at 500 rpm.

http://i13.photobucket.com/albums/a273/SpookyDad/New%20lathe%20and%20mill/graniteplatetramtest02.jpg

I think if I try to adjust the head at this point I will spend hours chasing that last little bit and probably never achieve it. Especially since the only truely precise piece in the entire set-up is the indicator that I got from my Dad. He bought it in the '50s during his tool and die apprenticeship. He tells me it cost him more than a week's pay.

Does anyone see any flaws in my methodology?

Should I just call it good?

Thanks

Neil

beanbag
01-19-2012, 04:11 PM
It sounds like your ways are worn, which is giving you the sag in the middle of travel. Also, make sure that when you move the indicator to each point via cranking the handles, you are moving in the same direction. This compensates for any slop in the indicator.

I think your tram is about as good as it's going to get.

photomankc
01-19-2012, 04:28 PM
Not sure if you mentioned what mill you are working on there but my smallish bench mill will show the same at the extremes of the long travel range in X since it's a narrow saddle and the weight is levering it up a bit. If I try to tighten that out it leads to stick-slip so I just avoid doing critical work out on the edges.

My only comment would be do you know with certainty that the table is dead flat?... because if it isn't and you shim the plate which is not perfectly parallel till the plate is flat then you won't be in tram to anything actually placed on the table.

My table is low in the back because the dovetail is a bit short on one side. so the table leans back a little. I shim my vice and tram to the vice bed but I know that I'll cut tiny steps if I bolt to the table without shimming first and the face mill will absolutely highlight the issue.

I don't chase 0.0005" around. It's amazing how little pressure it takes to move the head a little to one side in a cut. I can tell the difference in a heavy facing cut since the side pressure and lever action makes the wet noodles my mill is made of flex around and only cut on one side where the finish pass shows marks on both side of the cutter. A heavy facing pass will often move the tram a hair as will a bumpy cut that rumbles too much.

Spookydad
01-19-2012, 07:32 PM
photomankc: My mill is a Grizzly G0619 (SX3) and it is brand new. I am in the same situation with the stick-slip. I am reasonably confident the table is pretty flat.

beanbag: The movement is so slight, that I am not worrying about the ways being worn.

I just realized that I made the assumption that any differences in the relative height of the top of the surface plate were due to varying thickness of the granite plate itself. I had really just trammed the granite plate to the spindle rather than the other way around. So I put the coaxial indicator back in the spindle and put the long arm on it. Across roughly 12" of the x axis the deflection was only 2 ticks. I had to lift and pivot the arm instead of sweeping it with the spindle spinning. So it looks like the table surface was in fact matching the surface of the granite plate.

I haven't had to think this hard about geometry since college! :)

beanbag
01-19-2012, 09:24 PM
I just realized that I made the assumption that any differences in the relative height of the top of the surface plate were due to varying thickness of the granite plate itself. I had really just trammed the granite plate to the spindle rather than the other way around.

No, you trammed the granite block relative to the table's XY movement (in the first picture), which IMO is the right thing to do.

Then you checked the spindle tram relative to the granite block, which is also the right thing to do.

However, had you done things my way (ahem), then you could have still had a trammed spindle even with a lumpy, crooked granite plate.

I rarely care about the flatness of the table. When I face stuff, I either clamp it in the vise, or I clamp it down to the table with some kind of backing plate underneath (coz I often have to drill holes or mill a profile). I always check with an indicator the flatness of my clamped piece.

dp
01-19-2012, 09:54 PM
I put my granite plate on the table and carefully shimmed it until I got it level.

(snip

Next I put my axial indicator in the endmill holder and centered over 0,0 ( as measured by the DRO.)

Here's how I do it on my small mill:
http://www.youtube.com/watch?v=ZfioLDhBNBQ

In this short vid I have the rotor inverted, resting on the table with the hub surface. When I check tram I actually turn the rotor over and check one braking face against the other and ignore the hub face. The faces of brake rotors are incredibly parallel (at least this one is). This was a new rotor purchased for this use and has never been in service on a car. I think it is from a VW bug, and was the smallest rotor I could find. It turns out that the hub face is also remarkably parallel to the opposite braking surface and so it works well for this in either orientation. I generally tell this story for skeptics who will argue against the accuracy of the setup even when they confess they'd never tried it :)

Happy tramming!

BTW, the values you are reading on your coaxial indicator are meaningless - the thing to shoot for is zero deflection. To make the numbers meaningful you should convert the reading to degrees as that is what it is measuring (the degrees of error between the mill head and the table). In the mode you are using it it acts like a protractor. The length of the probe makes no difference.

http://www.youtube.com/watch?v=8ikS5aJeHG4

When using the coaxial indicator you need to test the instrument first by swinging the arm with it away from the test surface. It should show zero error in this mode. If it does show error then you need to find a clever way to back that error out or have it calibrated.

Spookydad
01-19-2012, 11:26 PM
However, had you done things my way (ahem), then you could have still had a trammed spindle even with a lumpy, crooked granite plate.


I thought I had done it your way. I don't have an arm that I can trust to be at right angles to the spindle. So that is why I went with the coaxial indicator.

I was thinking that everything had to zero out but I just realized what you were saying is the readings only had to be the same number in each location.

Neil

beanbag
01-19-2012, 11:43 PM
I thought I had done it your way. I don't have an arm that I can trust to be at right angles to the spindle. So that is why I went with the coaxial indicator.

I was thinking that everything had to zero out but I just realized what you were saying is the readings only had to be the same number in each location.

Neil

you came close to doing it "my way". The difference is:
1) You didn't need to shim the granite block
2) You have to sweep over the same 4 spots you measured in the first place.

Anyway, I think your mill is in tram so it's all good.

Spookydad
01-20-2012, 12:17 AM
BTW, the values you are reading on your coaxial indicator are meaningless - the thing to shoot for is zero deflection.


I understand the number itself is meaningless. That is why I said "ticks". I was trying to reflect that it was moving away from the common midpoint.

Neil

dp
01-20-2012, 01:57 AM
I understand the number itself is meaningless. That is why I said "ticks". I was trying to reflect that it was moving away from the common midpoint.

Neil


They can be calibrated in terms of angle. If you have a target that you trust (in my case, the rotor) and set it at a known angle by shimming it from below, you can quickly derive the degrees/marking on your dial. You can't do this easily by placing a shim under the prob on your target, for example as there is a cosine error that comes in to play. You have to put the shim under, not over the target. The coax indicator is an interesting but perplexing tool.

Black_Moons
01-20-2012, 02:25 AM
... Im gonna have to vote with the others in that a granite surface plate is not guarenteed to be parallel, and on a small mill may flex things.

Parallel is absolutely required, or you'll be traming it in wrong.

Either sweep the table, Or sweep something you can accurately measure to be of a *constant thickness* and flat. Don't just guess that something is flat or parallel.

This is the absolute most important thing..

However.. a good trick is to actualy tram your mill to your *vise* bed, As that will compensate for the angle your vise holds the work at, if your vise is not parallel (Most likey its not).. Untill you move the vise anyway :)

dian
01-20-2012, 08:14 AM
beanbeag, in you "method" you assume the z-axis is vertical to the others, right?

beanbag
01-20-2012, 09:00 AM
The exact direction of the z (spindle) axis is what you are adjusting when setting tram. It's mostly vertical.

dian
01-20-2012, 09:46 AM
spindle axis = z-axis = wishfull thinking.

Paul Alciatore
01-20-2012, 01:24 PM
I don't see how moving the table from left to right and front to rear is going to sweep the head in.

To indicate the head you have to rotate a dial indicator in the spindle and take readings at left, right, front and rear 90 degrees at four points.

Also, don't ever trust a granite plate or any surface plate to measure the same from top to bottom. The top surface should be the only surface that is flat and not necessarily parallel to the bottom surface.

EDIT: beanbag, I read the post again and I think I understand what your saying. Your canceling out any deviations in height of the granite block. I still don't think it's a good idea to use anything laying on the table to sweep the head in. However everyone has ways of doing things. I suppose that laying a perfectly parallel plate on a table and sweeping the head in would cancel out any irregularities in the table surface if your table surface is terrible.

Putting a flat, parallel object on the table accomplishes two things while adding a additional source for error. The first thing it does is it eliminates the bumps to the indicator as it is swept across the table's slots. This is not a serious concern if you are careful and sweep slowly. The second thing it does is average out any irregularities in the table's top surface. This is serious and I have measured +/-0.002" on import mill tables. However, it adds any errors of it's own to the process. If you are off in thousandths and have something like a large bearing race that is ground accurate to sub tenths, this is not a problem. But using a granite surface plate is going to be a LOT worse than that.

As I understood his method, you ARE swinging an indicator on a 4" arm to the four directions and taking comparative readings there. I don't particularly like this method because it may work, but it adds unnecessary complications that can only add errors to the process.

The only way to account for a lack of parallelism between the ways and the table top is to move the table with the indicator as it is being swept. So, while sweeping the indicator on an arm, you simultaneously move the table to bring the SAME spot of that table to the position of the indicator. That way any irregularities in the table's flatness and parallelism to the ways is removed from the equation. His initial sweep to four other points on the table is not really needed. Nor is the added surface plate or other parallel object. And since you are using the same spot on the table for all measurements, any lack of flatness in the table is completely out of the process.

Yes, you do need to consider both the column/head tram, performed as above AND the flatness/parallelism of the table top. Import mills are probably made with fresh castings and can age and warp after being machined. The hot hold of a ship crossing the ocean probably makes an excellent oven to relax all those stresses out of the already machined castings. And of course, the table top may not have been machined parallel to the ways to begin with. If you want to do accurate work, an import mill is a starting point, not a ready to use tool.

I have noticed several problems with my import mill and plan to do the following when it is reinstalled in my new shop.

1. Insure the spindle axis is parallel to the column ways. It is important to do this first or you are just wasting your time. This is done by mounting a piece of shafting in a collet and moving the head up and down while observing readings on the side of that shafting. It is rotated 180 degrees and a second set of readings are averaged with the first. Then the head is adjusted to make the axis parallel to the up/down movement. The column MUST be locked down in a uniform manner for each point where a reading is taken both here and in all further steps.

2. Tram the column as I detailed above.

3. Check the table top with an indicator in the spindle. NO sweep arm, just use X-Y table movements.

I fear the third step will show that the table is NEITHER flat nor parallel to the ways. The only thing I can think of to correct this third condition is to use a face cutter to trim it over the entire surface. I may need to do some scraping after using the face cutter. Or perhaps tape some fine sandpaper to the top of a small granite surface plate and use it to finish the top. I need a larger surface plate anyways.

Paul Alciatore
01-20-2012, 01:30 PM
I thought I had done it your way. I don't have an arm that I can trust to be at right angles to the spindle. So that is why I went with the coaxial indicator.

I was thinking that everything had to zero out but I just realized what you were saying is the readings only had to be the same number in each location.

Neil

The arm does not need to be at a right angle to the spindle, it simply needs to be rigid.

Paul Alciatore
01-20-2012, 01:35 PM
The exact direction of the z (spindle) axis is what you are adjusting when setting tram. It's mostly vertical.

NO, there are TWO Z axises: the column itself and the spindle. These two must both be aligned to the plane formed by the movement of the X-Y ways. See my post above.

On a round column mill you can only align the spindle axis.

gnm109
01-20-2012, 01:59 PM
I used to have the same issue as the OP here, Spookydad, with regard to using a face-milling cutter. If the part was wider than my 2" cutter, I would get a raised line where the cuts met in the middle. I saw elsewhere on the internet that some people use a disc rotor for tramming the mill. Having a spare 1998 Dodge Dakota disc lying around which was left over from a brake jobs some time ago, I decided last year to press it into service to use on top of my mill table (10X50 Taiwanese Webb, 3,500#)

With this in mind, I also made a little holder out of 6061 scrap and used a 1/2" Grade 8 bolt for a shaft. A few set screws and some cosmetic work with a ball mill and I had a usable tool.

The indicator is a small .001 with a rear shaft. That way I can view the dial all the way round without using mirrors or moving to the back of the table, which I can't reach anyway. Setting both X and Y Axes is now very simple. Placing the spindle in neutral so I can easily move it, I can spot front to back and left to right as I adjust the tilt or the nod of the head. I do one axis at a time, of course. I can get it to the point where the dial sits on zero from front to back in under a minute on each axis.

I also readjust the dial holder so that I can sweep the rear of the vise and get that nice and square after setting the two axes.

Of course, I make sure to clean the table thoroughly before setting the disc down on it. For the record, I recall when I was refurbishing this mill a couple of years ago that I swept the entire table with an indicator and found that it was remarkably true.

I've also checked the disc with a Lufkin micrometer and it's almost perfectly flat, that is, I can't detect any irregularities.....the discs are nicely ground too so that the indicator won't dig in. A shot of WD-40 helps as well.

beanbag
01-20-2012, 05:45 PM
NO, there are TWO Z axises: the column itself and the spindle. These two must both be aligned to the plane formed by the movement of the X-Y ways. See my post above.

On a round column mill you can only align the spindle axis.

Yes, I know that, but I was assuming the OP had a knee mil, or one with a quill that extends. Then Z really is Z (excluding the knee part). I have done a similar procedure as the one you mentioned involving a rod sticking out of the spindle on a CNC machine where the spindle axis is not necessarily the same as the z movement.



As I understood his method, you ARE swinging an indicator on a 4" arm to the four directions and taking comparative readings there. I don't particularly like this method because it may work, but it adds unnecessary complications that can only add errors to the process.


What complications and errors?



move the table with the indicator as it is being swept. So, while sweeping the indicator on an arm, you simultaneously move the table to bring the SAME spot of that table to the position of the indicator. That way any irregularities in the table's flatness and parallelism to the ways is removed from the equation. His initial sweep to four other points on the table is not really needed. Nor is the added surface plate or other parallel object. And since you are using the same spot on the table for all measurements, any lack of flatness in the table is completely out of the process.

I kind of like this technique, but the part I don't like is how you have to somehow lift or otherwise disable the indicator as you are swinging it around and moving the table. I don't trust indicator measurements after such a movement. (This is probably more a reflection of my poor indicator workholding). Also, it is slower to re-check back and forth.

beanbag
01-20-2012, 10:57 PM
I'm going to pull an Al Gore and rescind my endorsement of Paul's technique. The main reason being that it is too slow to go back and forth if you are actually adjusting the tram. The head pivots at a point about a foot above the work, so once you tweak the tram, the indicator values are totally off, and you have to go back and forth to check.

Doing it "my way", when you adjust the tram, you only have to swing the indicator back and forth 180 degrees to check it. Doing it Paul's way, you have to lift the indicator, crank, crank, crank, crank, drop the indicator.

That's just my opinion, nothing personal.

Spookydad
01-21-2012, 12:47 AM
I was making a fixture plate today and I decided to use a fly cutter to skim the surface flat. This is for drilling holes and slots in 18 gauge sheet steel so it wasn't critical that it be dead flat. I offset the passes by half the width of the cutter (roughly 2" offset) and there was a distinct line left at the edge of the sweep. It was only a couple of thou difference so I am ok with that. I need to get some parts made!

Thank you for all the help and choices of tramming methodologies. :D

This has been very educational but, if you want to debate the merits of the differing methods, it might be better to start a new thread that would reflect that topic in the title.

Neil

Mcgyver
01-21-2012, 07:55 AM
spindle axis = z-axis = wishfull thinking.

:confused: too criptic for early Sat morning - what do you mean?

Mcgyver
01-21-2012, 08:12 AM
I always check with an indicator the flatness of my clamped piece.

what does that tell you and how do you us it? Clamping easily distorts less than boxy sections. And you really indicate every time after clamping? not sayings its bad, but seems a lot uneccissary effort.




You should tram the mill relative to the table's XY movement, NOT the table's surface. (These two planes are not necessarily parallel).

To do this kind of tram, put the granite block back on the table, or better yet, granite block + sheet of glass. Don't worry if the granite block is not parallel or level.
.

That is an interesting technique, thanks for explaining it. I'm inclined to view it as a good way to survey a machine but of limited value in tramming. ie it might be something you do once to know where you're at regarding wear or a poorly made machine. The reason is the additional effort required compared to a regular sweep method and Jerry's remarks below which are true, what's the point of going through that sort of rigmarole only to machine a wedge? It's milling; the practical limits of accuracy make alignments to tenths moot.


While I may agree with this, I sure would not discount the table surface squareness vs tram....

If the table is a "wedge shaped" thing relative to the ways, then anything you put on top and flatten with any cutter will also become 'wedge shaped" to match unless you pack up one side of the work with shims.

Otherwise it would be like putting shims under one side of the work on a mill which has perfect alignments and tram.

I can see it a combination being useful in working with a very worn machine. Do the survey as BB describes, note the info then take that into account with future trams - ie no need to carry surface plates around, just have the table in the same place and tram for indicator readings to be intentionally out by the noted amounts....then.... wedge the work with shims.

still, I wonder how much the machine has to be out for this to be worth it

Who ever knew tramming was so complex ?:D

Mcgyver
01-21-2012, 08:14 AM
duplicate sorry

dian
01-21-2012, 08:32 AM
well, as i understand it, we have the theree dimensions of table movement (x,y,z), in spindle axis, the movement direction of the quil (if there is one) and the table top. they are all different.

now disregarding the quill, perhaps it makes sence to tram to a compromise between the z axis, the table top and the x-y plane?

Mcgyver
01-21-2012, 08:37 AM
well, as i understand it, we have the theree dimensions of table movement (x,y,z), in spindle axis, the movement direction of the quil (if there is one) and the table top. they are all different.

now disregarding the quill, perhaps it makes sence to tram to a compromise between the z axis, the table top and the x-y plane?

I think I see what you mean, wishful because is the table movement in the Z direction going to be both aligned with the Z and square to x/y at the same time?

beanbag
01-21-2012, 08:57 AM
To answer mcguyver and dian at the same time:

My number one priority is to tram the spindle axis relative to the table x-y movement. The reason is that if you don't, you will NEVER get a flat surface, but will instead get the out-of-tram ridges.

If I am concerned about milling a wedge on accident, I first measure my stock with calipers at 4 corners to make sure it is initially flat. Then I clamp it down somehow and sweep the indicator across the top. If this top surface is level, then I can start milling away. If not, I have to shim it.

And no, I don't always indicate the top surface of my part - only when I care about not making a wedge.

I don't worry too much if the z movement (knee or CNC head) is not exactly aligned with the spindle axis. Usually this error is small. Often you can't do anything about it anyway. IF you use tools of different lengths, your features will be offset by a tiny bit, but who will notice anyway? ,003" is still within spec, right? On the other hand, out of tram marks are really bush league. You can feel even .0003".

I mostly works on mills have have a beat up table surface, so I don't assume that anything clamped on it is flat. I also don't worry about the table surface being a little wedge shaped. Usually this error is small. Often I don't care enough to do anything about it anyway. If I really cared about part flatness, I would be shimming and indicating anyway.

In the end, how you tram will depend on your priorities.

dian
01-21-2012, 09:57 AM
mcgyver, simply because you shouldn assume spindle axis and z-axis are aligned.

beanbag, well, that makes sence.

Lew Hartswick
01-21-2012, 11:49 AM
I've always been under the impression that the spindle axis (the actual rotation axis) IS the Z axis.
NOW! if that axis is not absolutely
perpendicular to the X - Y plane then you just aren't working in true
rectangular coordinates in three dimensions.
...lew...