[iGG1e]

Been chasing my tail a bit recently trying to work out the cause for steps in material when face milling, as this is what first brought my attention to the issue.

Judging from the step it appeared that the column was tilted forward slightly. So using my largest 642 block as a reference (okay, Chinese so accuracy can be called into question) I did some travel tests to establish how the column was tracking in the Z axis with respect to X and Y with a DTI. A bit of tweaking and I got it to what I'd call acceptable. So a face mill test with a block of aluminium he'd in the vice I normally have on the table and it came out pretty well spot on where the two passes overlapped.

Zip forwards a few weeks, vice had been on and off a few times and now I want to face a block again. Now the step is back. Not good!

I put the 642 block in the vice this time and checked the travel as before. Showed tilt in both directions. Took the vice off and tested the block mounted on the table and found it was pretty much as I thought it was when I first tweaked it. Vice back on, test, lean is present but to a different degree.

I got various parallels (thin, thick etc.) out and spanned the table width and shone a light source along the edge and you can see that the front edge of the table is lower than the middle, with the back edge also lower than the middle, but higher than the front edge. You can feel that the parallels don't seem to plant themselves solidly, you can detect the a slight rocking.

So I've come to the conclusion that the table is warping things mounted to it due to the crown in it and the slightly wavy XY surface it seems to have. The table finish wasn't good from new, with a nice factory fitted dent due to them cranking down on the 4th axis for shipping. Not sure whether I see if I can get it surface ground (I doubt the dent wouldn't disappear due to its depth) or try my luck with fitting a new extended version of the table that ArcEuro are selling. That adds its own complications.

Has anyone ever tried to correct or refinish their table?

[neilw20]

How big is the cutter?
How sharp is the cutter?
How deep is the cut?
What is the material?
Is the machine bolted down?

I have found that a new 10mm carbide cutter on stainless steel taking a skim cut has a vertical load around 150 grams.
A worn out cutter needs 5KG.

Put some scales under the head, and a DTI from the table, then increase the pressure slowly.
5KG gives me over 0.001" -- 30 microns measured, and that is with the column bolted to a brick wall.
Without the bolt it is more than 20 times that.

Some is column flex, some from the vertical gib, and most from the base flexing.

Don't blame the Chinese. That's as good as it gets for that much metal.

With some tweaking I am maintaining 2 microns in Z all day, and I am using balance weights CNC lowered on to the head to balance the cutting forces.

1. Use a brick wall.
2. Get the gibs PERFECT.
3. Use sharp cutters.
4. Know the limits of the machine. It is most physics limits.:idea:
Seeing nice criss-cross cuts, all directions, (is it square?) requires better than 0.0001"
:cheers:

[iGG1e]

Thanks for the reply Neil.

The face mill is an Iscar 50MM eight insert body using APKT10 aluminium inserts. The depth of cut for these passes was 0.25 and 0.5MM on a block of 6082. This resulted in a step/ridge in the overlap zone of the U shaped path needed to cover the whole surface. I did reverse the path just to be sure and got the same result.

The machine is bolted down, level and table/head travel is nice and smooth. Gibb adjustment is good and everything is well oiled.

Actual tram of the mill is good and after the first time I saw the steps and corrected the travel, I was happy with the way it was running. But subsequent removal and refitting of the vice shows that it can be reintroduced. I use a dedicated pre-set torque wrench when mounting the vice and it is this type of vice. I use four clamps to hold it down. Probably depending on the tightening sequence used and the T slot it is engaged in, is influencing just how the cut will come out.

I did a few more tests with the block in various positions and was getting varying results depending which T slot was being spanned by the block. I'm going to investigate the surface grinding route as, even with the crudest of measurements, I can see that the table isn't flat.

[jeffrey001]

I don't know the thickness of the table but could be related to how much torque you put on the screw holding the vise.

Just my 2 cents.


Jeff

[iGG1e]

Agreed Jeff, cranking down on the bolts could influence things. The torque wrench is set to 20NM/15ft lbs which for me equates to a moderate hand tight with an allen key. As four clamps are used, I don't feel the need to crank down hard. That of course, being my perception of not a hard cranking.

The approximate T slot dimensions are 12MM wide x 9.5MM depth gap with a 21MM wide x 9MM deep flat of the T section. The central table thickness (the thinnest section between the dovetails) is roughly 26.5MM, with it going to 38MM in the area of the two outer T slots. These are the T slots I use for the vice.

[Ed from NY]

I don't know the thickness of the table but could be related to how much torque you put on the screw holding the vise.

+1 on the vise+torquing being the variable there. If you put a DTI as you torque you will see that you can move the vise quite a bit. I use a 6" Kurt which weights over 70 lbs and it does move. With a smaller vise... well.

Then there is always the chance of some chips or a ding on the table or vise.

The other thing is because the head is not very rigid try climb milling, and do it in a way that puts the head+column under compression, not under tension. That also means that you would be doing zigs, and not zig-zags.

Zig-zag Zig-zag Oi oi oi :cheers:

[iGG1e]

I'm going to take the table down to my friendly local precision engineers and put it on their measuring table etc. just to see if I'm totally losing my marbles or if there is an actual issue, whether it is worth correcting...or viable.

[Ed from NY]

Oh nevermind

[iGG1e]

Hehe Ed, to answer your thought about the Z gibb moving, previously I spent a fair bit of time getting a replacement (as the first was so bad it was out of adjustment range before it did anything meaningful in the dovetail) to fit and run well.

If the test next week just ends up with the conclusion that this is how it is going to be, I'll have to take it into account when planning machining operations to avoid tool paths that will likely end up forming the step.

[Ed from NY]

Then of course, the smaller the facemill the smaller the step. Or otherwise get a facemill that can cover the worpiece in one pass, but you knew that already

IMHE, a lot of the "precision" problems are toolpaths, gibs, and work holding.

I just machined 10 stainless steel pieces 7" long which are now under 50' of water holding the gate mechanism of one of our large water reservoirs here in NJ. Yes, it was a very simple part, but they came out flawless. I had adjusted the gibs just before that

[neilw20]

My machine runs 4 hours per day, and I regrind my gibs every 2 weeks.
That is the penalty for having a low cost machine.

[iGG1e]

Are you making your gibbs out of anything in particular?

Not entirely sure who in the UK stocks replacements for these machines and the one I did get from Hugh was out of his parts donor machine if I recall correctly.

[neilw20]

No, using existing ones. Just take the tiniest skim off them to remove the low wear marks.

I use a small surface grinder which I now bought, but can be hand done on a ground surface plate with some 800 grade paper,
then 1200 grade to finish, checking flatness with a straight edge.

I originally used the SX3 to grind the Y axis ones, while temporary locking strips are used in place of the Y gibs.
Y axis does not need to move while grinding.