Tell Me What's Going On With This Edge Finder

[JoeLee]

I've had this edge finder since about 2008. I've always had problems with it's accuracy. I've done several tests to prove it. Every time I would indicate the edge of a part or the fixed jaw of the vise when I drilled or milled a part I found that I was always off about .004.
Last test I did was yesterday. I ground and squared up a block of steel exactly 1" in width. Touched off on the back of the block and zeroed the DRO using the edge finder. came in exactly .500
I came down with a center drill and then drilled working my way up to 7/16" finished the hole with a 1/2" dia. end mill to eliminate any drill walk.
I measured the hole, it was off center by .0004.

So, I called Acu Rite. Talked to a tech and explained the situation. He asked me if I checked the run out on the probe tip. I hadn't so I did. The run out on the probe shank in my mill was about .0001, the run out of the probe tip was .0035. I thought OK, there's the problem.
Called the tech back and he said to loosen the 4 set screws on the top of the probe and tap the body around until I get zero run out at the tip.

I set the probe up in my grinder work head with a 4 jaw chuck so I could dial the shank in to zero run out. Once I did that I tapped the body around until I had zero run out on the probe tip.
I locked the set screws, check the tip as I went along and then checked the shank again. Dead zero at both ends.

I put the edge finder back in the mill, checked the shank, no more than .0001 run out, checked the tip of the probe and found about .002 run out. What's going on??

Any ideas??

Indicated the shank to zero run out.

JL..................



Indicated and adjusted teh body until I got zero run out on the tip. Locked the set screw and double checked both ends.



Put it back in the mill, checked the shank, less than .0001 run out. Checked the tip and got about .0025 run out.

[Doozer]

Wonky collet??
Wonly spindle taper?
Why not clock in the adjustment screws while it is installed in the mill?
Why are you doing it in the grinder?
Clock it for center in the mill, then mark with magic marker,
turn it 180° and read runout again. turn it 90° and read runout again.
I think you will find your issue.

-D

[chipmaker4130]

I'm sure Doozer nailed it. Like a chuck, everything needs to be dialed in on the machine it will be used on. If you're one of those who've removed the index pin for the collets, you'll be toast without the original marks mentioned above.

[redlee]

And using it horizontally , gravity will cause errors.
Definitely dial it in where your using it.

https://www.alignmentknowledge.com/dial-indicator-alignment-basics/

[JoeLee]

Wonky collet??
Wonly spindle taper?
Why not clock in the adjustment screws while it is installed in the mill?
Why are you doing it in the grinder?
Clock it for center in the mill, then mark with magic marker,
turn it 180° and read runout again. turn it 90° and read runout again.
I think you will find your issue.

-D

Collet is good. Spindle run out on the R8 taper is about .0002. With the 1/2" collet in it and a ground gauge pin the run out on the pin is .0001. The collet, it repeats too. So good there.

The only reason I set it up in the grinder with the 4 jaw was so I could get the shank run out to zero.

Gravity from horizontally mounting it isn't an issue because of the spring tension on the probe tip.

So I did set it up in the mill as suggested.

JL.............

[JoeLee]

I'm sure Doozer nailed it. Like a chuck, everything needs to be dialed in on the machine it will be used on. If you're one of those who've removed the index pin for the collets, you'll be toast without the original marks mentioned above.

We all think alike........... I never removed the index pin. I used that as my alignment point for the next time I use the probe.
The plug port is in line with the key slot on the collet.



JL.................

[JoeLee]

After zeroing it in the mill to less than .0001 run out on the probe tip I zeroed the fixed jaw of the vise. Drilled a test hole in a piece of steel. There is now a difference of .002 of the hole being off.
That's better than the previous .004.
May be due to slight head nod or knee sag being amplified by the 4 1/2" length of the probe. I think they should have made that probe a little shorter. That may have increased it's accuracy.

JL.................

[Dan Dubeau]

After zeroing it in the mill to less than .0001 run out on the probe tip I zeroed the fixed jaw of the vise. Drilled a test hole in a piece of steel. There is now a difference of .002 of the hole being off.
That's better than the previous .004.
May be due to slight head nod or knee sag being amplified by the 4 1/2" length of the probe. I think they should have made that probe a little shorter. That may have increased it's accuracy.

JL.................

Tram of the mill? Difference in tool lengths with make that error magnified.

How are you drilling the hole in the mill and determining it's out 0.002"?

Are you moving the knee at all during the drilling process? Or just quill only?

[JoeLee]

Tram of the mill? Difference in tool lengths with make that error magnified.

How are you drilling the hole in the mill and determining it's out 0.002"?

Are you moving the knee at all during the drilling process? Or just quill only?

Yes, that is what I said.

I'm not moving the knee during drilling. Just the quill. Just a small 1/8" dia. hole with a center drill in the chuck. Then I measure each side of the hole to the edge of the test part.

It should be even if it's centered.

JL....................

[Mcostello]

Indicate both sides and split the difference.

[Paul Alciatore]

Yes to not moving the knee, BUT did you lock it down? Or is it just floating in the dovetails? If it isn't locked down, then 0.002" or even 0.004" is no surprise.

And then there is the X and Y movements. Are those also locked down? A thousandth here and a thousandth there and sooner or later we are talking about a serious accumulated error.

Yes, that is what I said.

I'm not moving the knee during drilling. Just the quill. Just a small 1/8" dia. hole with a center drill in the chuck. Then I measure each side of the hole to the edge of the test part.

It should be even if it's centered.

JL....................

[genea]

Is it possible that gravity is playing a role? Setting horizontally, testing vertically.

[Paul Alciatore]

Unless you have some kind of super accurate four jaw chuck, I would think that the run-out measurement made in your first and second photos is totally worthless.

What follows is conjecture as I have no experience with that model edge finder. But I have done some work on possible designs for better electric or electronic edge finders and have seen many of the problems that this presents.

I tried to look up that edge finder and what I found was one that looked similar, but with a nice, heavy electronic cable running from it, assumedly to some form of electronic readout. And the description on the Acurite web site says, "The 3-D, spring-loaded probe records the located edge, even if over-travel should occur." That says to me that the probe is intended to swing from side to side (the over-travel) as it contacts the work piece. Also I do not see any cable on yours so I have to assume that it is not connected to any control box or DRO or CNC system that can make any corrections and you are only working with the LED that is built into the edge finder to show when the contact with the edge occurs. They also state that it works with both conductive and non-conductive materials so it is not relying on electrical contact by the probe tip. Therefore, the probe tip MUST deflect by some amount before that LED can come on. They also state that it's accuracy depends on the software so, lacking any corrective software or hard wired circuitry to compensate for that deflection, it can not turn on without that small deflection.

The probe looks like it extends about two inches below the body of the edge finder. The diameter of the body of the finder is about one inch or less so the lever arm for that switch is probably less than a half inch. That is at least a four to one mechanical advantage but it is working in the WRONG direction. The probe tip must move 0.002" for the switch end of the lever to move 0.0005" and that is probably a best case assumption because it is hard to imagine a "switch" with only a 0.0005" gap between the contacts in the OFF position.

I would not be at all surprised that you get several thousandths of error using this edge finder without any electronic correction. And even then, the OEM's instructions seem to be saying that you should find a center line on a part by splitting the difference between the detected location of two opposite sides. This would make it more of a center finder than an edge finder.

All in all, if you do not have the electronics to get the best performance from this edge finder, you would probably be better off with a $15 snap type. Or use this one but with a shorter probe. A MUCH shorter probe. The shorter the probe is, the more accurate it will be.

[Paul Alciatore]

Gravity may play a role with some DTIs, but I highly doubt that it is causing several thousandths error with this edge finder.

Is it possible that gravity is playing a role? Setting horizontally, testing vertically.

[JoeLee]

Indicate both sides and split the difference.

I've done that, only because I've run into this issue, but shouldn't have to.
Zeroing the back edge of the vise should be all that's needed. Providing..... the vise is inline with the table travel, the movable jaw doesn't tip the part and the work is sitting square in the vise and was measured accurately. I'm good with all that.
I believe that edge finder is supposed to be accurate to .0005.

JL.................