Tramming your mill(/dril) head:A theoretical question-Idea came up

[Chris MateParticipant @chrismate31303]

For those that know all sort of geometries well, I would like to hear your opinion regarding this idea.

I am aware of the current methods used to tram heads with up to two indicators used, and I can do that, but I like how the coaxial indicator works and dont work, really jnteresting.

From that my theory is:
1-You take a block of round bar say 70mm wide by 60 mm in height, put it in lathe and machine the outside diameters, for this case the outside does not matter much.

2-Now you bore it out except you left a bottom of say 5mm not bored out, with a wall thickness of say 15mm. Do not remove from chuck now till finished.

3-Now you drill a very small hole in the bottom centre without removing it from lathe chuck, this is to centre this piece under the spindle using the coaxial indicator with its punch adaptor, this would be like a flat plane start.

4-Now you turn a steep taper of say 15 degrees into this hollowed out bar to near the bottom with concentric hole. The use of the taper would be to "multiply" the effect of it head is out of tram similar to the lenght of bar used with conventional methods for leverage. The taper will give the leverage now, as the coaxcial probe will not run in a perfect circle around the taper if your head is not 100% 90 degrees vertical(Spindle to bed).

5-You can see this as a jig made once to use with the coaxcial indicator to square up the head/spindle to bed(Tram).

I dont want to compare this with the other methods, just opinions in theory regarding this idea. I do see some sort of a simplicity in it. What do you think of the idea in theory-?

[Chris MateParticipant @chrismate31303]

Using the Coaxial Indicator to do it, with a taper.

[JasonBModerator @jasonb]

If you were to work out the effective major and minor axis of the elliptical shape that teh Coax may follow iff the tram were off by say 0.1mm/100mm it would be such a small difference that you would not be able to measure it with the usual dti scales on these coax indicators

You would also need to ensure your piece of bar had the end facing the chuck completely true to the cone, I did not see mention of even facing it.

[JasonBModerator @jasonb]

Just sketched it out in CAD

If we assume the widest point of your cone would be 65mm dia then applied the equivelent change in shape corresponding to 0.1mm over 100 of tram error you would be looking for 0.000032mm on the coax. One is easy to read the other you won't read. Given that most would want something like 0-.01mm or better over 100mm then you will never see th eneedle move on the coax

I drew it as a cylinder, sliced the top off a 0.1:100 (0.06deg) and then measured the major axis of the elliptical end.

You can also see from the Y delta that by measuring over 65mm dia even if done the traditional way your reading will be smaller, half of less of what can usually be swung across a mill table.

[VicParticipant @vic]

Some years back I saw a redneck mill trimming device. It was a steel bar welded to a car disc brake then accurately turned down to fit a collet in the quill. You can guess how it works.
A guy at work told me another way, without the need for an indicator. After turning a point on one end put two short 90° bends in opposite ends of a suitable length of steel bar. The blunt end goes in the quill then you swing the bar left and right until the pointed end just touches the table at each end. He also suggested the use of cigarette papers. I used this method myself before I got a dial gauge.

[Chris MateParticipant @chrismate31303]

Ok, thanks for info. This morning I decided to make the "jig" clamped to mill table. Decided to cut the inside to 18 degrees due to height of round block. At the bottom I left material, then drill a small hole which is concentric to the taper. This hole was drilled with the smallest centre drill with short taper for the punch bit from the coaxial indicator.

I then installed it to the centre of the mill bed. I played for two hours with it. I think I like it for now, must still prove if it has the practical value I want from it, and that is to have something simple enough to tram the head so I can get a smooth cut backwards and forwards in all directions by slight adjustments, and with this I monitor all 360 degree in one sweep, I like this.

1-So I 1st centred the spindle with Coaxcial Indicator installed in 10mm collet/MT4.
…I ligned up the punch bit from indicator so it ran true as possible, this is now at the bottom of "jig" right on bed. I know that depensding on head this may not be 100% yet, but a start.

2-I then fit a longer bit to coaxial indicator and chose a spot around the middle of the taper height.
….Now from here I marked the Y-Axis(Short one) as Y1 and and Y2 on opposite side.
….I then marked the X-axis on "jig" as X1(Left) and X2(Right). So now I have 4 points to concentrate on to make adjustments moving the Bed a bit.

3-I got the Y-Axis which was shimmed by factory and it has "tention" like bolts on top for a slight adjustment which I used after I got the mill. Not all mills has this feature.
…Off the batt I could see I dont have a problem aligning it to the Y-Axis.
…I could also see I have a probem with the X-Axis get it to alighn, I could not initially. I then loosed the 3 head bolts slightly, place the degree box on the quil extented, head adjusted to the top of travel.
……Only after I adjusted the head slightly, the quil now show 90 degree to bed(This of cource is not 100% accurate , but at least an idication),.Only now I could get the coaxial indicator set 100% by adjusting the X-axis more thabn the Y-axis(Very slighty) by moving the bed under the spindle.
….What surprises me the most is that the coaxial indicator stay on zero through the 360 degrees around the taper, with bit of oil on tip. This as a surprise.

At the moment I like the fact that I can monitor 360 degrees turn with one indicator in one sweep around.
I am pretty sure the head is now 90 degrees as close as I can get it, tighning of 3x bolts included.

Note: Using the coaxial indicator with a boring head, I never thought of the importance of the head to be 100% 90 degrees to bed to get it set up well. At least I learned something from this around the whole plot of using a boring head.

I will take a photo tomorrow to illustrate the setup. Everybody may not like it, but if it allows me to adjust the head slightly to give me a good cut backwards and forwads over the work under load(Not boring head use but normal milling), thats where I am heading, hoping if I have to adjust it slightly off, it will not cut an oval hole with the boring head when that is used, or if it workd for normal milling head slightly off to one side, it may have to be corrected if boring head is used..

I have a good quality shoulder mill with 6 inserts from Walter to proof it for normal milling.
 

Edited By Chris Mate on 30/09/2023 17:44:10

[JasonBModerator @jasonb]

Chris, try putting a 0.1mm or 4 thou feeler gauge under one side of your jig and see just bt moving the X and Y you can get the coax to show zero, I think iy will still show zero as the difference between flat and shimmed is 0.000032mm which is 1/3rd of a micron you you will never see with the naked eye

[Greg Webb 1Participant @gregwebb1]

If the jig was a tube without any angle, would finding the centre at the base and then at the top by moving the quill up give a difference if the head was not trammed.

[PeteParticipant @pete41194]

I have one of the Blake Co-ax indicators, plus a few spare dial indicators and could easily make my own dual indicator tramming tool. I don't use the Blake for setting the head alignment since it's really not quite as accurate as just sweeping the table with a dti. In fact these co-ax tools seem to have at least a bit more hysteresis over what my dti's do. And in hindsight, I've not found much of anything that the Blake can do that in one way or another a dti can't also do, and with in my opinion better accuracy. It just takes slightly longer. With a short length probe, those co-ax tools are a good way to check the lathe tail stock alignment, if it's being used properly, gravity when checking the vertical elevation has little effect compared to doing the same with a dti. And yes I got fooled by that for a few hours until I figured out what was happening. One of the dual indicator tramming tools would get you close a lot faster for the initial alignment if you ever angle the head for certain jobs and then want it back square. But for my mill, then a simple square against the length of the exposed spindle will do that as well. Then it's time to use that dti. A dial indicator and dti has dozens of uses outside of just tramming the head in with a few shop made additions.

And no matter how accurately you tram the head in, you will always get a certain amount of back cutting with larger diameter tools. That head alignment is also a static setting. The machine itself starts to deflect outside of where you set it once those cutting loads come into effect. Those are just physics and simply because there's always various amounts of machine and tool deflections under load. And the harder the material, deeper the cut, and/or duller the tool, the more you'll have. But head tramming is just one item that has to be checked as being correct. With a knee mill or dovetail column the head moves on, then just how straight is it in a true vertical alignment over it's full travel distance. For any machine tool slide, there's 6 possible directions or combinations that a slide can vary or be twisted. And those misalignment's may not be constant or can even reverse direction over the travel length. Then how square is the tables Y axis to the X axis. Most assume all that is correct, actually checking might surprise a few. Again a few simple tools most already have and a dti can also do that, although a boring head can be helpful as well for checking those knee or column alignments. Fortunately most of those only have to be verified once and then many years later.

[JasonBModerator @jasonb]

Posted by Greg Webb 1 on 30/09/2023 18:56:11:

If the jig was a tube without any angle, would finding the centre at the base and then at the top by moving the quill up give a difference if the head was not trammed.

You then start to get into the realms of which do you tram, usually the spindle axis is trammed to th emill table but there is also colum to table and quill movement to table to be considered which may show*

* however it still comes down to showing tiny deviations over a short distance that you would be hard pushed to see.

As Pete says even with perfect tram you may get slight back cutting as the front pass may deflect the spindle slightly upwards and also cause minute flex in the machine particularly if your facemill has the blunter inserts which means that as the deflection and movement is removed once the front edge has cleared the work the tool drops down and takes a skim on the trailing edge

[NealebParticipant @nealeb]

Biggest issue with this tramming game I have found, on both a Warco VMC and a Wabeco 1400 which both have variations on swivellling heads, is less about how to measure any errors and more about how to move the head under control to remove them. Usually ends up with soft-faced hammer on a non-critical part of the head, but it's not something I like doing to the machine, especially when there is a DTI in place to show the movement. And then re-clamping without the head moving…

[Chris MateParticipant @chrismate31303]

Hi Neabb, I did bolted a long thick aliminium bar to the top of my mill head to assist in moving the head, and it does help a lot making it easier, the leaverage is such I can bump it wih my fist, however I see in future I will make something "fexable" to adjust it accurately with two opposing bolts, like I do with vice and zeroing the X-Axis swivel function, no bumping.

Here are two photos added:

 

Edited By Chris Mate on 01/10/2023 12:24:52

[Chris MateParticipant @chrismate31303]

30mm x 90cm aliminium bar bolted to head to make it easier to move, tap with fist.

[Graham MeekParticipant @grahammeek88282]

Whilst the theory may be good, introducing the tapered component also introduces other errors. If the taper is not dead square to the face in contact with the table. This will be a source of error as an elliptical path will be shown straight away. Adjusting the Head to suit this condition will give an error on a true vertical.

I am also not keen on the amount of extension being shown on the Quill. The Quill might well show two different conditions as regards vertical due to the amount of Quill engagement in the Head casting.

I don't think you can beat the clock on the table, or a known parallel, but again with the minimum of Quill extension.

Regards

Gray,

[Greg Webb 1Participant @gregwebb1]

I don't think you can beat the clock on the table, or a known parallel, but again with the minimum of Quill extension.

I was always taught to have the quill at full extension for better accuracy. Does anyone have a definitive answer with an explanation or reference to the reason for extended or retracted?

Thanks

[Tony Pratt 1Participant @tonypratt1]

Posted by Greg Webb 1 on 02/10/2023 10:43:28:

I don't think you can beat the clock on the table, or a known parallel, but again with the minimum of Quill extension.

I totally agree that you can't beat the 'clock on the table ,or a known parallel' method of tramming but I was always taught to have the quill fully extended , thinking about it I don't know why so would be best to have it fully retracted. As an aside our Swiss made Genevoise jig borers were serviced once a year and they checked the quill squareness with a long arm at 90 degrees plus DTI straight onto the table, enough said I think!

Tony

[Anonymous]

For a fixed rotation point of the head and a given angular error with an extended quill the point of rotation of the Dti will be further off true centre, but I am not convinced it makes any difference to the vertical readings fro the DTI. I always have the quill retracted and locked when milling, so it seems reasonable to tram in that position.

Andrew

[JasonBModerator @jasonb]

My thinking is that I use the mill with minimal quill extension when cutting so that is the best way to check the tram, as Graham says it may well be different when extended so why test it in that position.

Even having the quill free or locked is likely to affect readings, so test it as you are likely to use it would be my way of doing things.

[Greg Webb 1Participant @gregwebb1]

As an aside our Swiss made Genevoise jig borers were serviced once a year and they checked the quill squareness with a long arm at 90 degrees plus DTI straight onto the table, enough said I think!

Tony

I agree, they see the advantage of measuring far away from the pivot point.

Looking at the difference between extended & retracted perhaps reason for extended is that it is a little easier to hit the number you want on the dti. The dti will move the same distance over a longer arc allowing the reading to be less sensitive to linear movement.

[PeteParticipant @pete41194]

With a knee mill or one with a dovetail column the head moves on, it's much better to have that quill fully retracted for some simple reasons. Extending it out transmits much more leverage on the head due to the machining loads trying to move the head back out of tram. Add the vibrations from cutting and the head can still move even when solidly bolted in place. All 4 of the head bolts on mine are torqued to 45 lb. ft. and I've still had the head move on me over some time. A fully retracted quill is also much more rigid the closer you can get the tool tip doing the work to the spindle bearings. So more accurate dimensions on your parts with less taper due to distortion in the head and spindle, plus a bit better surface finishes. An extended quill and the side loads from something like a larger end mill, then that extra leverage increases the loads on the spindle bearings. And the rear dovetail column mills aren't that rigid. Extend the quill and that leverage problem is going start flexing and bending that column.

That's one reason something like an ER collet chuck is less capable taking very heavy cuts over a Morse Taper or R8 collet. With those the tool itself is a lot closer to those bearings. Industry almost universally uses ER collet chucks because of those tool changers, Or for real rigidity, hydraulic and shrink fit tool holders. The only time I use the quill on my knee mill is when drilling, tapping, reaming or single point boring. Anything else and the quill is retracted and locked and the knee used for any Z axis moves. That's also imo more accurate for precise elevation changes since gravity is always forcing all that weight of the table and knee fully down against the knees feed screw thread flanks so there's no backlash. The only down side is how many times you have to turn the crank handle.

[Greg Webb 1Participant @gregwebb1]

I use a solid stop system to give a repeatable stop & rigidity to my Bridgeport quill. It eliminates the play in both the nut & the adjusting screw circlip. The videos show the system & the issues with the original setup.

Solid stop system.

Issues with existing factory setup.

[Michael GilliganParticipant @michaelgilligan61133]

Nicely done, Greg … and very nicely demonstrated

MichaelG.

[Graham MeekParticipant @grahammeek88282]

Having the quill extended moves the clocking point radius further away from the point of rotation of the Milling head. With a constant swing radius of the clock you end up with a pair of triangles. The "Opposite" sides (clocking radius), are common. The "Adjacent" is longer with the Quill extended as well as the "Hypotenuse" Thus when compared to the Quill "fully in" the angle between the Adjacent and Hypotenuse will be greater in this instance. Than when compared to fully, or extended.

Experience tells me the fully extended position is going to be more finicky to set because of the smaller angle.

The other thing is, and it again boils down to triangles. The top of the Quill touching the Quill housing bore top right while the bottom of the Quill touches bottom left due to the Quill Lock.

The clearance between the quill and the housing will have less effect when the Quill is fully home than when the Quill is extended. If 0.025 mm is spread over 150 mm the angular difference is going to be less than 0.025 mm spread over 50 mm. (please note this is an example and not a particular case).

It is however less than I would expect on something "like repeat like" a Bridgeport quill. The "Limits and Fits" for a "Precision Location" H7/g6 gives a total clearance for an 80 mm diameter quill to be 0.059 mm. One "fit" better at H7/h6 will only reduce this total clearance to 0.049 mm. Transition fits would reduce the clearance still further, but the spring required to return the quill would need to be stronger, making the machine less sensitive. Plus as the accuracy of "fits" gets better, so the costs go out of the window. No doubt this is why Jig borers have exceptionally long Quills.

Regards

Gray,

[Chris MateParticipant @chrismate31303]

Thanks for all the info and opinions, I appreciate it, and chewing through it.

If proper mills has issues this thing can only be worse.

So in short its safe to say if you want to tram the mill, its best to tram it in the vicinity to the vertical height you going to mill with as little quil stickout as possible. I always mill with no stickout of quil to start with, however I never trammed it specifically to the vicinity of the vertical height of operation.

Well, the stickout of the quil with head advanced to the top, I was thinking more leverage to pick up inaccuracy in tramming effort, and all of this giving me more perspective over the range of possibilities, and as I learn to know my mill, I will settle somewhere with knowledge thats best to use, understanding where its coming from.

Just for fun I will tram it this way with the head on top, quil most advanced and everything locked, then lower it right down with quil retracted and head as near bottom as possible, and not the variation, I am sure it may not be pretty….I know understand why the precission grinding machines were invented.

As soon as I am happy with the method of tramming, and actually moving the head by fractions controlabl(I will have to make something to accomplish this so it not bumping), I want to experiment with the cut quality finish of a good quality facemill, to see how slightly on/off tramming effects it under some load of cutting(not as much as a heavy proper mill) X/Y choice of movement/locked, and back over the work, versus all around conventional milling around the work…etc.

[Chris MateParticipant @chrismate31303]

I re-indicate it with coaxial indicator today at the top with quil extended everything locked-Ok

I then moved quil up, and lowered head to bed, check indication and it was off X-axis.

I then installed vice, re-indicate it in vice and it was spot on all directions. I took a cut ar flat piece of steel with shouldermill with etal inserts. I can see the 3x cut lines like in CNC work, but could not feel or measure the lines with -.01mm dial indicator. I also had the level box at 90 degree on left side of head, and it did not change from 90 degrees the whole time, not 100% accurate, but at least it did not move withing its limits of 1 degree either side of 90. I am happy with the front and back cutting as it cut forward(Conventional) and backwards(Climb milling).
This crap mild steel block, so it was not a mirror finish like I get with aliminium on same cutter.

So my next step is to buy/make a short arm for attaching a dial test indicator in spindle to the taper as well and see if I can determine if the coaxial indicator is more sensitve in visible movement per line as the test dial indicator .01 will show. I can see if I wind the quil down, the tip of lever touching the taper seems to move half the distance as to one line on coaxial indicator to next line. Will have to see how that compares to the test dial indicator.

What I do like is dialing it in in X1/X2 & Y1/Y2 points is actually quick and easy after chosing a position on taper, and you get all directions in one go while the indicator stay facing you. So it looks like the verical area around the vice jaws are ok till/if proven wrong by dial test indicator
.

[Author]

Posts