Trying to Square a Fully Disassembled X3, I have too many moving pieces...Help please

Hello All,

Thank you for your time to read this and help.

Due to moving I had to disassemble my CNC X3 into pretty much the smallest pieces I could. So I removed the Table, Column, Head, and Y mount.

So now trying to reassemble it, I have too many pieces and I can't seem to figure out how to measure it to make it all square again.
Here is what I tried to do.
I set the tension in the X and Y axis gibs, and adjusted the backlash down to the minimum I could get it, (0.0002")
I pulled out my tramming gauge, then I squared the left right (X) table to the head.
I adjusted the left and right by loosening and adjusting the 4 main bolts that attach the head to the Z drive. (This part went ok getting it to within .0001.
There is where it goes down hill, I can't seem to figure out how to square the head to the column, and the column to the table.

If I take my tramming gauge (May be an issue here too, more below) place it in the spindle and adjust the head front to back (nod) to be square to the table, it doesn't tell me if the column is square, meaning it may not square making it move up at an angle. If I adjust the column to make the head square, to the table, there is no way to tell if the head is square to the column, leading back to the problem of it not moving directly vertical, instead moving at an angle. I can't figure out how to measure the head to the column square.

Also the tramming gauge I have was given to me from an old machinist, it is a single dial gauge that is on an arm and goes to a 1/2" rod. I place it in a 1/2 collet and measure one side, then spin it to measure the other side. So this leaves me with three floating variables. 1. If there is any runout in the spindle it will translate to the holder, 2. Any problem with the flatness of the table will show up here and 3. What ever variable I am trying to measure. Yet both #1 and #2 are only adding errors to my #3 square measurement.

I just feel like I have way to many moving variables to start squaring from. If anyone could help me out it would be greatly appreciated.

Thank you for your time.

Hoss may jump in. He has some videos on his site I believe showing this in detail.

I just started my first cup of coffee but here goes anyways.
If it were me, I would make sure the bed is nice and level, no twist, then I would square the Head/Spindle on the column using Rollie's method.
Once that is done, you know the spindle is parallel to the Z ways in both planes and shouldn't have to adjust it anymore.
Mount the column to the mill base and indicate it in so it's parallel to the X and Y using a precision square or cylinder. You do this by by shimming the column base and not adjusting the head (spindle is already aligned).

The order you do this in is pretty important. If you do it in the wrong order, you will never have consistent readings and will fight it.

You need large right angle plate on the bed. And then raise and lower the z up-and-down. The bigger the plate the better if you do this first and then what you did afterwards it will come out okay

Hi, you can make your own precision square that will be dead square by getting a 100mm (4") diam ROUND steel tube, 600mm (2 feet) long and weld on a round base, same diam as the tube, and another piece of steel across the bore of the other end, and machine the OD and face the base flat, undercutting the centre of the base or it won't sit flat.

This will be the cheapest but most accurate square you can make/get.....no matter how much money you would otherwise spend on a shop bought square.
Ian.

BTW, a very quick and easy way to detect if the column is leaning any which way is to assemble the mill completely and adjust all gibs as normal.

Then fasten a round disc at least 75mm (3") diam X 12mm (1/2") thick to the table approx. in the middle of the table travel.

Bring the head down close to the table and mount an indicator in a holder in a chuck in the spindle

Rotate the indicator around the disc (by hand) to centralise it 0 - 0 both sides of the X and Y axis by adjusting the X and Y travel.

Lock both X and Y slides.

Then raise the head up the column about 300mm (12"), and with a long rod ( needs to be the biggest diam you can hold) held in the chuck in the spindle mount the indicator again and do the rotation again.

If you get zero and zero once more around the disc in two planes the column is vertical and square to the table surface and any run out will be the lean of the column in either plane, for and aft and left to right.

No square is needed for this exercise, and it works for two planes in the vertical axis.

The column must be bolted down firmly before any testing is started, and if it runs out, shimming or scraping the base of the column will correct it......if you know what you are doing.

The higher up the column you can go for the second test the more accurate you will get.
Ian.