G0704 Setup - Tramming/Squaring to Table Issue

[LorenB667]

Dial a 1-2-3 block in true with your X axis, then check it along Y. No need to make any cuts - eliminate that variable from your testing to get started. It may actually be easier to tram your vise in, then check with a 1-2-3 block, but that does add the vise into the equation.

Additionally (and backing up a smidge) When you finish tramming did you run a single indicator across your range of motion? (Actually 6" should do). One downside to the SST kind of tool is that it's easy to get confused and have one indicator reading +0.002" and the other reading -0.002", which can look great, but really indicates a 0.004 offset. Sweeping a single indicator as a final sanity check eliminates this possibility.

GsT

This time I used a parallel to cover the slots. I'm attaching pics of the .0001" indicator at three spots along the front to rear axis. I didn't use any shims. I torqued the column bolts (4 bolts arranged in a rectangle pattern) evenly to 60 ft-lbs (couldn't find a torque spec for the bolts from Grizzly so I found other sources for them).

I did not use a 1-2-3 block.

Non-moving axes were locked down. Quill was locked down.

I ran the gauge on the left to right axis and I'm getting less the .0005" variation from side to side.

 

[GeneT45]

Looks like your head is leaning "back" about 0.0006", which is really not that bad. So your X & Y axis are pretty dialed in. Not perfect, but better than many (most?) people ever try to get. One thing you have not tested is the perpendicularity of your X&Y axis. You can get an idea by indicating the edge of your parallel (if you're careful), but you'll have to dial in one side, clamp it in place, re-check that the side you dialed in wasn't moved by clamping, then indicate one of the "ends" (perpendicular surfaces). If you have anything that's precision ground all over, that'll probably be easier than a parallel.

GsT

 

[LorenB667]

Aren’t the printed numbers .001” each?

 

[GeneT45]

Aren’t the printed numbers .001” each?

Doh! I are dumb. That's 0.006" out, which is not good. Disregard my previous post.

 

[matthewsx]

This time I used a parallel to cover the slots. I'm attaching pics of the .0001" indicator at three spots along the front to rear axis. I didn't use any shims. I torqued the column bolts (4 bolts arranged in a rectangle pattern) evenly to 60 ft-lbs (couldn't find a torque spec for the bolts from Grizzly so I found other sources for them).

I did not use a 1-2-3 block.

Non-moving axes were locked down. Quill was locked down.

I ran the gauge on the left to right axis and I'm getting less the .0005" variation from side to side.View attachment 514495View attachment 514496View attachment 514497

Sounds pretty good to me, how does the machine perform when making parts?

FWIW, many experienced hands recommend using a half tenth indicator since .0001" can be frustrating.

John

 

[LorenB667]

Oh thank God I wasn’t wrong! I was worried o was reading it wrong!

Doh! I are dumb. That's 0.006" out, which is not good. Disregard my previous post.

Now the question is, how do I make adjustments? Correctly…

Do I put the shims at the bottom near the bottom bolts? What thickness should I start with? I’m asking because y’all know more than I do.

 

[GeneT45]

I'd shim the bottom bolts. Washer type shims would be my choice, because then the bolt will keep them from slipping around on you.

Make sure the contact surfaces are fairly smooth and clean. Chips or paint would call for cleaning and re-check before shimming.

To figure out how much, you need the ratio of the amount of Y movement you made, while measuring the tilt in your column to the z-height of the axis of rotation around the X axis, multiplied by your measured deviation. Example

4.3" (travel along the parallel in which you measured 0.0061" deviation) / 4.6" (the height where column will pivot and the indicated surface)

Multiply that total by the measured deviation (0.0061 in this example) and you have the theoretical thickness of your shim(s).

The problem is figuring out where the center of rotation is. And it's likely to change as you adjust. I'd probably assume the point where the column first contacts the base. That should get you pretty close, at which point you'll likely just have to iterate, but use that basic ratio as a kind of sanity check.

GsT

 

[RJSakowski]

As I look at the three photos, the low spot is in the middle @ -.0027", At the far end, you are @ +.0009 and at the near end, you are a+.0028".

A parallel is not necessarily a flat surface on its wide face nor is it necessarily uniform thickness between the wide faces. It is meant to be uniform in thickness between the two narrow faces with a face surface on each.

 

[GeneT45]

As I look at the three photos, the low spot is in the middle @ -.0027", At the far end, you are @ +.0009 and at the near end, you are a+.0028".

A parallel is not necessarily a flat surface on its wide face nor is it necessarily uniform thickness between the wide faces. It is meant to be uniform in thickness between the two narrow faces with a face surface on each.

I interpret the three photos as showing rotation in a continuous, clockwise, direction.
GsT

 

[GeneT45]

**counter-clockwise...

I'm going to bed...

GsT