Edge Finding

[krwada]

OK ... This is absolutely driving me nuts here. Hopefully, any of you skilled machinists can tell me what is going on.

I set up my milling vise and tram it to less than 0.001". I am using a Kurtz clone here.

I then find the edge using an edge finder on the stationary part of the vise. and find the edge and hit the ZERO on my DRO.

I then put a precision parallel in the vise so I can find the very same edge ... but going in the exact opposite direction.

I get a a very consistent offset of 0.005"!

Now ... I have fully checked out the DRO using a dial indicator, and I get a hysteresis of 0.001". This is to be expected because the DRO (Accuremote), is specified to 0.0005". Knowing about encoders, I fully expect a forward and back hysteresis of 0.001"

What is driving me nuts is the 0.005" hysteresis using an edge finder. I am not using one of those wigglers. I am using one of those edge finders that is chucked up using a 3/8" collet and has a 0.200" diameter pin.

Can anyone tell me what is going on? Now, I have no idea about where the true edge is when I do this operation.

Thanks;
Ken

[kcstott]

One what speed are you running the spindle? Edgefinders are not to be run faster then 1200 RPM.

Is it a Starrett Edge finder?? or some cheep knock off?

and are you working into the edgefinder gently to where it just runs true then jumps? or are you dialing in a little hard??

it's all feel brother.

[krwada]

One what speed are you running the spindle? Edgefinders are not to be run faster then 1200 RPM.

Is it a Starrett Edge finder?? or some cheep knock off?

and are you working into the edgefinder gently to where it just runs true then jumps? or are you dialing in a little hard??

it's all feel brother.

I am running the edge finder at 600 RPM. I did a bunch of tests and found that between 600 but no greater than 1000 RPM works best.

Indeed, I am using a cheap knock off. I just ordered a Mitutoyo edge finder. I also placed an order for a Starrett wiggler too.

I have no idea which brand is best however. I know that you cannot go wrong with Mitutoyo or Starrett. However, I have heard that Starett is re-branding cheap Chinese knock-offs in their name.

And yes ... I am dialing it in very slow and careful. That last 0.002 to 0.001" and how you dial it in makes a huge difference.

Thanks!
Ken

[kcstott]

Well damn near twenty five years ago when I bought mine Starrett was still making them in the US.

Next question. In the collet or in a chuck or collet holder extension?? I try to mount my edge finder as close to the spindle as I can for precision stuff. meaning 3/8" collet in the spindle and the edge finder noting else.

Now your parallel did you stone the edges to be sure no burrs?? One little burr will screw up a reading.

I'd be willing to bet you are dealing with back lash. i.e. lead screw error.

if you're not dialing the same way you will pick up numbers I don't care what DRO you have, I don't care what mill it is. It will pick up numbers. and that is why we always worked one direction and one direction only per axis.

What kind of machine?

[krwada]

Well damn near twenty five years ago when I bought mine Starrett was still making them in the US.

Next question. In the collet or in a chuck or collet holder extension?? I try to mount my edge finder as close to the spindle as I can for precision stuff. meaning 3/8" collet in the spindle and the edge finder noting else.

Now your parallel did you stone the edges to be sure no burrs?? One little burr will screw up a reading.

I'd be willing to bet you are dealing with back lash. i.e. lead screw error.

if you're not dialing the same way you will pick up numbers I don't care what DRO you have, I don't care what mill it is. It will pick up numbers. and that is why we always worked one direction and one direction only per axis.

What kind of machine?

Direct into the collet. I also mount the edge finder as close to the spindle as possible. I also move the main Z axis such that I only drop the quill by .2"

I am new to the DRO. When I learned machining 30 years or so ago. I was taught to ALWAYS measure and do the layout + cuts ALWAYS in the same direction.

I thought the DRO changed all of that and we did not need to use the same discipline as the old school machinists.

Also, I am using a Grizzly G0704. I had access to a Rockwell Delta. The problem with the bigger machine is that it would have cost 4x the cost of the machine just to move the dang thing 10 city blocks!

So ... I got the Grizzly G0704 instead. I evaluated the minimills, and I was not impressed with what I saw. I saw that the G0704 is barely adequate for what I need it for.

I need to deburr my parallels? I had not thought to do this.

[kcstott]

Where there's a will there's a way. Not knowing your situation though you may have had other reason not to attempt a move. but here's how I did mine




Ok New machine, New DRO, New edgefinder, and you have no clue which is correct.

First thing is to check the tram on the mill head. If that is less then .001" over six inches then go to the following.

Get a mag base 1" travel dial indicator and place it so it can read your saddle or table movement which ever way you are edge finding. Be careful of cosine error. Set up your edge finder and use the edge finder to locate your surface. dial in slowly until it just jumps. Zero your indicator and zero your DRO. Zero your hand wheel dial as well if you can.

Back off the surface in the opposite direction say .200" or so and come back into your surface again until your dial indicator reads zero. What does your DRO say? How close to zero did your edge finder jump? that will give you a good indication of what your repeatability is.

If your dial indicator says .000" and you DRO says .000X" you can trust the DRO to a degree.

Now you can check back lash. Set up your dial indicator and dial into it with the saddle to say .500" now set you DRO to zero and dial off of or into your indicator .400" each way and note the difference between the DRO the hand wheel dial and the indicator. That is your back lash error for your DRO and the lead screw if you are also able to read a graduated dial. Some of these smaller machine don't have decent dials on them.

What you are doing is eliminating where your error is coming from. Then knowing what it is working around it.

Now this assumes a few things. Like I said the tram of the head to the mill vise, the gibbs being correctly adjusted and so forth.

Also I'd really run that edge finder at 1200 rpm they seem to be a little more responsive at that speed. but That's just my personal preference.

[ar15barrels]

Any chance the edge finder has the sticky preservative oil on it still?
Is it dry or oiled?

[kcstott]

Any chance the edge finder has the sticky preservative oil on it still?
Is it dry or oiled?

Good point

[krwada]

Where there's a will there's a way. Not knowing your situation though you may have had other reason not to attempt a move. but here's how I did mine




Ok New machine, New DRO, New edgefinder, and you have no clue which is correct.

First thing is to check the tram on the mill head. If that is less then .001" over six inches then go to the following.

Get a mag base 1" travel dial indicator and place it so it can read your saddle or table movement which ever way you are edge finding. Be careful of cosine error. Set up your edge finder and use the edge finder to locate your surface. dial in slowly until it just jumps. Zero your indicator and zero your DRO. Zero your hand wheel dial as well if you can.

Back off the surface in the opposite direction say .200" or so and come back into your surface again until your dial indicator reads zero. What does your DRO say? How close to zero did your edge finder jump? that will give you a good indication of what your repeatability is.

If your dial indicator says .000" and you DRO says .000X" you can trust the DRO to a degree.

Now you can check back lash. Set up your dial indicator and dial into it with the saddle to say .500" now set you DRO to zero and dial off of or into your indicator .400" each way and note the difference between the DRO the hand wheel dial and the indicator. That is your back lash error for your DRO and the lead screw if you are also able to read a graduated dial. Some of these smaller machine don't have decent dials on them.

What you are doing is eliminating where your error is coming from. Then knowing what it is working around it.

Now this assumes a few things. Like I said the tram of the head to the mill vise, the gibbs being correctly adjusted and so forth.

Also I'd really run that edge finder at 1200 rpm they seem to be a little more responsive at that speed. but That's just my personal preference.

Yes, I have trammed the mill to 0.001"

Also, I have tested the backlash on the DRO with the dial indicator. As I have stated, I measured a 0.001" backlash between the DRO and the dial indicator.

I used a mag base test stand and a dial indicator and placed a 3/4" precision ground drill rod in a 3/4" collet. As best as I could, I found the dead center of the drill rod as it was chucked into the quill. I then moved the Y axis back and forth and compared the DRO readings with the dial indicator readings.

Other than the 0.001" backlash ... both the DRO and the dial indicator agreed very well.

Next, I will do the tests you recommended to find out exactly how the edge finder compares to the DRO and dial indicator. I had not thought of indicating on the saddle while finding the edge in both directions. I shall do this next.

As to moving the Delta Rockwell. I rented a truck that had a lift on it. The problem I had was the lift was too small and the mill was really top heavy, and it was attached to its original skid that must have been well over 40 years old!

Moving that thing around darn near killed me and my younger son. So ... I decided to take the 'safer' route and ordered up the Grizzly G0704. Up here, where I live, (Silicon Valley), getting manual operated mills seem to be getting easier every year. All the shops here have converted to the professional grade CNC stuff here. The machine brokers here are telling me that one can get a decent used manual mill for a very reasonable cost... However, it will cost a lot more to move it than the cost of the machine.

Thanks KC;
Ken

[krwada]

Any chance the edge finder has the sticky preservative oil on it still?
Is it dry or oiled?

It is dry. Does the edge finder need oil? It was chock full of that preservative stuff when I got it. I pretty much hosed it down with WD40.

Perhaps I did not clean it very well. The tip does feel sticky. I was told that the tip needs to move around like it is on a bearing. It does not do this.

How does one clean an edge finder?

[ar15barrels]

I was told that the tip needs to move around like it is on a bearing. It does not do this.

How does one clean an edge finder?

Degrease it completely with brake cleaner and then blow it out with air.
WD40 dries up and leaves it sticky.
I get good results completely dry, but I have also seen recommendations to oil it.
Either way, it needs to be completely FREE moving to work correctly.

[krwada]

Degrease it completely with brake cleaner and then blow it out with air.
WD40 dries up and leaves it sticky.
I get good results completely dry, but I have also seen recommendations to oil it.
Either way, it needs to be completely FREE moving to work correctly.

I will now get some brake cleaner and do what you recommend.

OK ... so I can assume that the edge finder is NOT supposed to have a directional hysteresis?

Thanks!
Ken

[ar15barrels]

OK ... so I can assume that the edge finder is NOT supposed to have a directional hysteresis?

It could have some, but not more than a couple thousandths.

[kcstott]

It could have some, but not more than a couple thousandths.

Yeah at the most.

usually the hysteresis comes down to operator feel.

Bear in mind that the edge finder is also registering all the run out your spindle and collets have too. I don't think that will add up to more then .0005" total run out but you never know.

The reason for indicating in the saddle is it's just easier to get to.

As for your edge finder. it should have a cap on one end, and the stylus on the other and a spring in between.
the surface that the stylus rides on (the face of the tube) needs to be clean along with the face on the stylus. My edge finders have no more oil on them then what is needed to prevent rust and that is just a drop of machine oil spindle oil etc.

[krwada]

A couple thousandths

Yeah at the most.

usually the hysteresis comes down to operator feel.

Bear in mind that the edge finder is also registering all the run out your spindle and collets have too. I don't think that will add up to more then .0005" total run out but you never know.

The reason for indicating in the saddle is it's just easier to get to.

As for your edge finder. it should have a cap on one end, and the stylus on the other and a spring in between.
the surface that the stylus rides on (the face of the tube) needs to be clean along with the face on the stylus. My edge finders have no more oil on them then what is needed to prevent rust and that is just a drop of machine oil spindle oil etc.

A couple thousandths is exactly what I am observing with my edge finding. A 0.005" hysteresis means there is a registration error of 0.0025 in both directions.

I will also look into deburring the parallels. I had not thought that this could be the problem. If the parallel has a burr in it ... this can easily account for the 0.005" offset. 0.005" is not much. However, for the work I am about to do, it is alot!