[andy_con]

hi everyone

currently i use a wobbler but its not accurate enough.

what is the most accurate way to find an edge?

[neilw20]

How accurate do you want? 0.0005" is OK?
Mine works like a charm.
http://www.youtube.com/watch?feature...d-cp_9dg#t=56s

[fizzissist]

hi everyone
currently i use a wobbler but its not accurate enough.
what is the most accurate way to find an edge?

Well, you know it isn't accurate enough, then you must know by how much. If you know by how much, there has to be a pattern, or trend, and that you can compensate for. It's a perfect opportunity to take advantage of the old adage about doing the same thing over and over but expecting different results.

Do the same thing over and over to get a statistical average, and compensate for that amount. If the value is inconsistent in TWO directions, then you're not using the tool consistently.

I rotate the edgefinder at 750rpm (Haas max rpm with the doors open... ) and touch the locating dia to cause it to wobble.

Slowly bring the dia up to a burr-free edge, dropping your feed to .001 or .0001"/step. The dia will stop wobbling, then suddenly jump to one side or the other. At the instant it does, the outer diameter edge of the edgefinder is coincident with the edge of the part.
Since God won't let two things occupy the same spot in space and time, that dia is going to give because it's the weakest element in the system.

When you've hit that sweet spot, either turn off your spindle and retract in Z, or retract QUICKLY ...(the edgefinder's end can scar the top of the part if you go slowly. trust me.)

Note the location in X or Y that you're locating, move the table towards the center of the spindle by exactly 1/2 of the edgefinder's dia (most of us use a .200"dia so the math doesn't confuse us) and there's your edge.

I'd used an edgefinder the other day and had it in the control. A Haas tech came by and calibrated the probes, and for fun we probed the edge I'd done. It was within about .0007" of the probe's value, as Neil suggests.
Since I had .05 of rough stock on that side, I hadn't spent a lot of time on it, and I'd been jogging in .001" steps.

[txcncman]

Machinists were making parts with +/-0.0001" precision and accuracy for decades before electronic edge finding came along. How do you think they accomplished this and you can't? Were they really that much smarter than machine operators now?

[andy_con]

so its a case of " a bad workman blames his tools"??

[fizzissist]

so its a case of " a bad workman blames his tools"??

It's always worked for me!!

Only trouble is, when the customer's QC sends the part(s) back, excuses don't matter. Actually, I don't like to deal with QC from other companies. If I can't make it right the first time, then I don't deserve the job.

[fizzissist]

Machinists were making parts with +/-0.0001" precision and accuracy for decades before electronic edge finding came along. How do you think they accomplished this and you can't? Were they really that much smarter than machine operators now?

You may have pointed out the problem.
Machinists as compared to machine operators.

[neilw20]

How accurate do you need?
[ame="http://www.youtube.com/watch?v=DSo81MmB31w"]CMM or CNC positioning error inspection using a FEANOR LP30 laser interferometer - YouTube[/ame]

http://www.renishaw.com/en/interfero...xplained--7854

[andy_con]

How accurate do you need?
CMM or CNC positioning error inspection using a FEANOR LP30 laser interferometer - YouTube

*** just seen your link to renishaw sorry.

not that its in that video but i have a reinshaw probe set here

[RICHARD ZASTROW]

There are some attachments for edges and square corners for locating same. They have a jig ground hole directly above the edge / corner to precisely position your spindle over that point with an indicator

I believe they are magnetically held in place.

Dick Z

add: Flexbar has the edge type, I didn't see the corner type.