Accurate Edge Finding ??

[Philipintexas]

I have 2 mechanical edge finders, one .500 Diameter and the other with a .200 head diameter. I was told to bring the Edge finder toward the material at high spindle speed, it will form a perfect cylinder, continue until the concentricity of the shaft & head break. At that point you've found the edge. continue in half the diameter of the head and you have the spindle-axis on the edge of the material....

Sometimes I don't seem to be exactly right.
1. How accurate should I expect a mechanical edge finder to be?
2. If I move till the concentricity breaks, haven't I gone slightly beyond the edge?
3. If so, do I compensate a fixed amount?

As with other simple gadgets, technique is everything, am I doing it wrong?

 

[Swifty]

I have used a .200" dia edge finder for years, if I recall correctly, the idea speed is about 900 rpm, but I do tend to go about 1200. Your method is correct, as soon as it throws out zero everything and move over half diameter. If you are doubtful about accuracy, just check a couple of times to see if you get the same reading. There should be no need to compensate for anything, just make sure that you approach slowly.

Paul.

 

[ownthesky2010]

Spot on. I use one of these on my cnc mill and I get 0.01mm repeatability. Good enough for me.
When I need to find the centre of a bore I always compare the y axis and x axis readouts and any difference is always measurable ovality in the bore and not error from the edge finder.

 

[bazmak]

I have never used one but have ordered a cheap one from china
electronic beap type will soon find out if its useful

 

[Metal_slicer]

I have a starrett .2" edge finder and I also have wondered the exact same thing. Sometimes the edge finder will jump offset quite a bit and other times it will jump offset a very tiny amount. I have never tested the accuracy but it seems to work well enough for the work that I have been doing.

 

[nemoc]

Your technique is correct, but I usually advance too quickly which can introduce error. So I back off after it 1st indicates and advance again to within .003 or so and then move very slow until it indicates. If you get inconsistent results check to make sure there are no burs on the edge of your stock. You should get very accurate and consistent results.

 

[toolznthings]

I have a starrett .2" edge finder and I also have wondered the exact same thing. Sometimes the edge finder will jump offset quite a bit and other times it will jump offset a very tiny amount. I have never tested the accuracy but it seems to work well enough for the work that I have been doing.

Take a look at my post in tips & tricks about edge finders getting magnetized.
Could be your problem.

 

[Swifty]

Here's the style that I have, http://www.amazon.com/dp/B002SG7PPC/?tag=skimlinks_replacement-20. Occasionally I lubricate the mating faces with a bit of WD40, I have found that if they get sticky, they don't work as well.

I also have the type with interchangeable tips, but only use it occasionally in awkward situations.

Paul.

 

[deverett]

The cylindrical type seem very popular these days, but GHT back in the 70's described in Model Engineer the manufacture of a similar type of device but using a 'ball on on a stick'.

Works OK for me and can get into places that the cylindrical type can't reach (such as the edges of large diameter cylinders).

Dave
The Emerald Isle

 

[Blogwitch]

To put it bluntly, the best ones of these are the Starret range. I have been using both the imperial and metric ones for many many years. I have tried all of the others, but none come close. This really is a time when spending the money in the right place gives the best results.

Another point, you are using them way too fast to get accurate results. At 100 rpm, I can easily get within a few tenths (double checked with my DRO's and touching on with a cutter). You bring the tip in, and as soon as you detect ANY sideways movement, that is your setting, if it displaces any further, you are way out. The face you are detecting on should also be as smooth as possible, looking like a freshly sawn piece will be no good at all.

Where a lot of people go wrong is that the units do need to have correct lubrication between the sliding faces. It might seem a little gross, but I use a trick from my photographic days when reassembling cameras, I use nose grease. Just rub your finger down the side of your nostril, outside, not on the inside, that is called picking your nose. The grease that is picked up on your fingertip and gently applied to the sliding faces is enough to keep the sliding surfaces perfectly lubricated. Any more and you will have the tips sliding about all over the place, giving wrong readings.

This really is a genuine posting, picked up from many years of using these sensors.

I hope it helps.

John

 

[Blogwitch]

Dave,
They are the 'wobbler' type, and as you have said, they are good for getting into tight places, especially down into narrow slots.

Where people go wrong with that type is that they play around with the socket holder and it can soon become unuseable.

I too occasionally use my Starret version to very good effect.

John

 

[Swifty]

John, with all due respect, I know that you are happy with running your edge finder at 100rpm, but that speed is way too slow. A quick google search of edge finder speeds always says around 1000rpm, similar to what I have been using since 1970. With the Wobbler type that uses a ball socket, a far slower speed is needed. Next time I have the mill on, I will certainly try using 100rpm and compare the reading to a faster speed.

Paul.

 

[canadianhorsepower]

I personally dont like using those:hDe:
so I got my self a co-axe centering tool
works great specially when putting something offset
in a four jaw chuck

cheers
Luc

 

[Gedeon Spilett]

Not easy to compare, but the feel is the same type that is experienced with the micrometer.
the frictional force required to move the head of the edge finder should be comparable to that which one feels by measuring a piece between the micrometer rods to expect a similar precision. the optimal speed of the mill may perhaps depends on the stiffness of the spring, I also use it at 1000 rpm, this is really a great tool.
I

 

[mcostello]

I called Starrett years ago to find out what lube to use on their edgefinder. They recommended STP.

 

[purpleknif]

m Not to be a smart alek but... If you need a dead accurate pickup you need to use an indicator and a flat block and swing the indicator both ways. I've been doing this for 30 years and this method will get you within .ooo1" when necessary.

 

[RonGinger]

purpleknif, can you explain this a bit more- I don't see how you use a flat block and indicator.

 

[BaronJ]

Take a look at my post in tips & tricks about edge finders getting magnetized.
Could be your problem.

I made my own and after several months it started acting oddly ! I discovered that mine was sticking due to becoming magnetised.

However a few moments inside the stator of a small motor with the rotor removed soon sorted it.

WARNING: If you try this, the coil gets hot very quickly, so you only need to apply power for 20 or 30 seconds. Which is more than enough time to demagnetise whatever you insert into the stator.

 

[purpleknif]

purpleknif, can you explain this a bit more- I don't see how you use a flat block and indicator.

Hard to explain but I'll try. 1st you need to center your indicator. This is why most machinists use Interapid indicators because the stem is on center. Set the ind. so that when rotated the lowest reading is zero when against the edge of the workpiece, In a vertical situation raise the ind to clear the piece and put a flat piece ( I usually use a TNMG insert ) Against the same edge of the workpiece. Rotate the ind and check the reading and adjust accordingly.
I'm a PC moron but if you need more I'll see if I can get some help to post a few pics.
After 30some years you find there are some really creative ways to make pickups but this is pretty much the standard way to pick up an edge.

 

[Swifty]

I carried out some tests this morning using the edge finder. I used a ground parallel held in the vice as my test piece, first off I used my normal method of running it just under 1200rpm, I tried several times and the largest difference that I got was .0002", more than good enough for me. Next I ran it at 100rpm, straight away I found that I had to approach at a snails pace, as it was slow to react when it found the edge, I had cleaned and lubricated the join beforehand, the largest difference that I got was .0006". Next, using the zero that I had when running it at 1200rpm, I used an indicator to actually check the the centre was on the edge, I couldn't see any difference in the reading of the indicator.

Of course, there are always a few ways of doing things to get the same result, no one is completely right or wrong. In my case, I'm more than happy with the method that I use, I can only suggest that members try different methods until they are happy with the one chosen.

I have to correct myself in a comment I made in an earlier post, although I started work in 1970, edge finders of this type were not used, I assume that they were commonplace later in the 70's.

Paul.