View Full Version : Locating A Point On A Work Piece

05-23-2012, 12:08 PM
Keep in mind, as you read this, that I am not a trained machinist, but a DIY hobbyist type. I do a lot of reading and research on the web, and am always trying to find ways to do things that suite my situations. One of the struggles I have faced many times is finding the starting point, or a center point on a piece of material. Yes, i have several kind of center finders that spin in the mill. Some indicate by snapping the rotating end piece, one has a built in led that lights at the touch of the surface. They all seem to work, and with a bit of math, using the DRO's on the Mach3 screen, I could usually find the point that i was looking for.
Always looking for a better (easier?) way to do things, here is what I have come up with, that works for me.

Reading about laser cross hair position finders, they seemed to be something that would be really nice to have. There are a few in the market place, but not really any that wanted to pay the price for.

I looked on eBay and found some really cheap cross hair laser beam projectors. About an inch long and 3/8 in diameter, they looked exactly like what I needed. I bought several that were advertised with adjustable focus.
After they arrived, I was very disappointed in the width of the projected beam. There was no point where I could achieve a very thin narrow beam. As I had a working distance in mind, for my particular setup, there had to be some way to focus the beam better.

Taking one of the units apart, I found that all of the beam forming elements were made of molded plastic. Really not very accurate, with a very narrow range of focus, never really being able to achieve a well focused beam, at any distance. I started digging around, looking for an old cigar box full of lenses, from my telescope building days, some 40 or so years ago. Needing a sharp cross hair at about 4 inches from the work surface, I started trying different lenses below the little laser pointer. I finally found an old lens that gave me a very find focused pattern 3 1/2 to 5 inches from the work surface. Exactly what I was looking for! The lens was about 3/4" in diameter. Now I needed something to mount it in, about 2 inches down from the little laser unit.

You all probably have several of those Harbor Freight flashlights laying around. You know, the ones that are free just about every month in their flyer. It turns out that is exactly the size I need for the mount! I removed the switch from the end, leaving the rubber push button in place. I then cut off the LED Light end, leaving just enough length to mount the lens and laser unit at the right distance apart. To mount the laser unit, I carefully cut a hole in the rubber push button and pushed the laser unit into the hole. A nice tight fit! I capped it off with some heat shrink tubing around the end of the unit and the wires coming from the unit. The lens was slightly smaller than the opening, so i carefully wrapped it with black plastic tape until it could be pushed tightly into the bottom of the tube.

I hooked the completed unit to a 5 V DC supply, and found that I had excellent focus from about 2 1/2 to 5 1/2 inches. Now I needed a nice firm mounting bracket. I drew up and milled, from .040 aluminum, a small bracket that would mount on my Z Axis frame, and hold the new pointer in place by screwing it's cap through the bracket. Really came out nice!

Here is a picture of the finished unit, mounted on the Z Axis frame.

As I wanted to calibrate the location at a specific eight from the work piece, I used my Z surface finder routine, on my Mach3 screen, to find the surface of the point I wanted to indicate. The routine finds the surface, and then moves the Z Axis to a point 1 inch exactly above the work piece.

Here is a pictures of my Z Height Probe, made from printed circuit board. it really work well and is very accurate at finding surface height.

The next picture shows the laser beam on the work piece. Boy, have you ever tried to photograph a laser beam? What you are looking at is a piece of paper with a hole, placed over a small indent from a center punch. The cross hair is slightly South of the punch mark, as well as slightly West. The bright spot you see, is the South edge of the "crater" created by the center punch. When I centered the cross hair on the "crater," it created a bright glaring light. Made it even more impossible to photograph!

In actual use, I center the cross hair and look for the brightest point of light, that occurs when both X and Y are exactly over the "crater." It has a repeatable placement of +- 3 tenths on the X or Y Axis. Much more accuracy than I need for anything I do!


The next picture shows another addition, a Video Camera Center Finder!
In the picture you can see the Laser Center Finder and to the left the small microscope camera. Yes, another eBay purchase from China! This is a small camera that can be focused from the end of the lens to taking a picture of a room full of people. Have a very good depth of field, and has excellent sharp detail. You can see the aluminum disk below it, with the small punch mark in the center.


As the post is only allowing 4 pictures, I will continue with more information in the next post.

05-23-2012, 12:27 PM
how do you make sure, they are aligned to z?

05-23-2012, 12:33 PM
Here is a picture of my Mach3 screen, showing the video plug-in picture from the Microscope camera shown in the last picture of my previous post. The loaction repeatability is about the same as the Laser Cross Hair. It is easier to use, as you can use the mouse and arrow keys easier, while looking at the screen.


Next is a picture of my Mach3 screen, showing the placement of the video insert from the Plug-In.


The last picture shows a close up of the custom screen area that I use to control the Z Touch Plate, Laser Center Finder, and Video Center Finder.


You can download my custom screens from: http://www.mrrace.com/SimpleMill/index.htm

The pictures on the site show an earlier version, but the download will have all the additions shown in the above screen shot.

If you would just like my script files for any of the special buttons, mentioned above, let me know and I can email them to you.

As always, your comments and questions are always welcomed.

05-23-2012, 01:33 PM
Hi Dian, good question!

The setting routines are built into the screen buttons for Laser or Video setting.

When you click on the appropriate button, after setting the cross hairs, the X & Y Axis move to a predetermined location, which is programmed into the script that is run behind the button. After making the move, it also sets both the X & Y DRO's to Zero as well.

Kind of neat how I determine the locations. First I take an extremely small ball end bit, usually use a 1mm, and very carefully touch it to the surface of a piece of aluminum clamped to the table. It makes just the smallest of marks on the metal. I then set the X & Y DRO's to Zero. Then I jog the X & Y Axis so the cross hair is EXACTLY on the small mark made by the bit. When I am sure they are exactly centered on the mark, I read the X & Y DRO's and note them down. Those are the two offset figures that get programed into the script that runs behind the button.

Here is the code that is used with either button:

Xmove = -2.5 'adjust this for x move distance
Ymove = -2.5 'adjust this for y move distance

Code "G91 G0 X" &Xmove & "Y" & Ymove 'makes an incremental move the distance that you set
While IsMoving() 'waits while this happens
Code "G90" 'Goes back to absolute moves
DoOEMButton (1008)'Set the X and Y DROs
DoOEMButton (1009)

Really very simple, when you think about it. And once set, it works every time.
Hope that answered your question.

05-23-2012, 10:19 PM
Interesting but I prefer the old fashion way.

Jaakko Fagerlund
05-24-2012, 05:05 AM
Nice, but your accuracy very much depends on interpreting the reflected light or getting the punch mark in a correct place.

Probably easiest is to use an edge finder, always gets me there and is cheap and fast. No need to punch anything, just use a spotting drill :)

John Stevenson
05-24-2012, 05:14 AM
Many thanks for the effort and information, much appreciated.

05-24-2012, 10:13 AM
There are a lot more applications for the Cross Hair Finder than just finding the center of a punch mark. That was just an example to demonstrate how I did the initial setup of the position of the cross hair.

As an example, I can measure the size of a 1 2 3 block +- 2 10ths. I can locate the center of ANY round hole to +- 2 10ths. I can locate the intersection of scribe marks on a blued surface within +- 2 10ths. And all of these are repeatable, time and time again. I do realize that none of this would be possible for me, without have a modern CNC system with accurate Digital Read Outs in front of me on the computer screen. Until a few short years ago, I had no idea that one could accurately move a tool by 10ths!

It really makes it very easy, at least for this fledgeling machinist, to find a given point on any work piece, be it flat or 3 dimensional in shape. I know it is not for everyone, but works great for what I do, until something better comes along. And it is so quick and easy, and always right there to use without have to do any type of setup.

For all the Professional Mechanist out there, who have been using one of the many types of position finders for years, there are always others ways to do things that you may not have even tired. Please don't be to quick to write them off, at least without some exploration time and forethought.

Getting down off the soap box, until next time........

05-24-2012, 10:27 AM
As I see it the beam would have to be perfectly aligned with the axis of the housing OD for this to work and the OD of the light held in perfect alignment with the axis of the spindle. If it is not then the offset would change as the distance from the work increased or decreased. For this to work it would have to be as accurate as a rifle scope is and even they have what is called parallax wherein as the eye moves around the eye piece the impact point will change slightly.

To have it be accurate every time the offset would have to be corrected with each different setup and distance from the work piece. I really don't understand why you want to use a beam to locate a point on the work piece that ultimately will be cut by the program in the CNC. What I mean is, why mark the work piece and then use the CNC program to do the work when the two marks should coincide anyway?

05-24-2012, 11:01 AM
As an example, I can measure the size of a 1 2 3 block +- 2 10ths. I can locate the center of ANY round hole to +- 2 10ths. I can locate the intersection of scribe marks on a blued surface within +- 2 10ths. And all of these are repeatable, time and time again. I do realize that none of this would be possible for me, without have a modern CNC system with accurate Digital Read Outs in front of me on the computer screen. Until a few short years ago, I had no idea that one could accurately move a tool by 10ths!

I am curious what sort of lightly used, extremely rigid, professionally built CNC you have the HF flashlight setup strapped to? Sorry, but you lost all credibility with me the instant you started giving numbers, unless by repeatable 2 tenths you mean +/- 0.2".

If you ever have opportunity, have a good CMM operator with a good machine measure a few parts out to a tenth (.0001), then take a mic and see if you can repeatably measure out to +/- .0002. It takes quite a bit of practice.

05-24-2012, 11:09 AM
In actual use, I center the cross hair and look for the brightest point of light, that occurs when both X and Y are exactly over the "crater." It has a repeatable placement of +- 3 tenths on the X or Y Axis. Much more accuracy than I need for anything I do!

can you elaborate on that? how you get to .0003"? I've not used a laser centreing devise but always wondered at their accuracy because the laser seems so wide....have you somehow got the bean focused so its only a tenth wide and then view it through a usb microscope or something? :confused: apologies if you've covered this, just wasn't clear to me

05-24-2012, 11:29 AM
Ok, the camera thing I can get. With magnification they can pretty accurately calibrated and I've seen several other people get under 0.001" positioning accuracy with it by making one that mounts into a collet in the spindle and taking care to make sure there is no axial deviation.

The laser.... That is also tough for me to believe that it is holding that accuracy in a process that depends on the hieght of the head, and your veiwing position, and the calibrated eyeball. It sounds like perhaps you have it going to a known height above the workpiece before the laser is used maybe? If so then I guess it's a matter of calibration to get it project to the point the spindle centerline is sitting over but have a hard time getting on board that it can decern tenths.

I like the collet mounted cameras best. They are right in the spindle centerline and the magnification and cross-hairs being imposed into the image seems to make them easily as accurate as the mechanical edge finder and not dependent on the material to be conductive like the simple touch probes. No offsets to deal with either. X/Y zero is right where you see it.

As I read that it seems to be kind of edgy so let me add.... Thanks for post up your project!

Weston Bye
05-24-2012, 02:15 PM
Good information, thank you.

My first experience with a laser was working for my dad, laying sewer pipe. the laser was shined down the center of the pipe, and a plexiglass target fitted into the end of the next section of pipe to be installed. With the pipe plugged into the previous section we shifted the free end around until the crosshairs aligned with the laser and then filled sand around the pipe.

I wonder about using a uniform "target", say a disk with a tapered projection on the bottom that would drop into a center punch hole and center up there. Then, on top of the disk, crosshairs or a target or something on a surface that allows a crisper view of the laser than shiny metal.

edit to add: a more direct treatment might be a little squirt of flat white spray paint directly onto the metal in the region of the center punch.

05-24-2012, 04:10 PM
OK guys, sorry that I probably don't speak "Machinist," and I don't claim to be even close to being one either. I guess I need to be much clearer when talking about what I am trying to measure and how close I am getting. And, to the Gentleman who claims to be sometimes a bit gruff, no problem! In my working career I dealt with a lot of folks that were a lot more gruff, some on a daily basis, then I expect anyone in this group will ever be. I am here because I enjoy learning how to do things, and hopefully just because I may be way behind the abilities of those in this group, you will understand and help me understand how to do more, even better. If that kind of mentality does not exist, within this group, then I get the pictures and will simply go to any of the many other forum's that I participate in, where newbies and their projects, thoughts, and ideas are welcomed and guided by those who take part.

OK, down off the soapbox again!........

So let me get back to basics as to what I do and would like to do. I enjoy working with metal. I am currently building a lot of sheet aluminum stuff for builders of experimental airplanes. This stuff, for the most part, is well within tolerance when being built to specifications +- 1/32 of an inch.

When I built my little CNC/Router, my goal was to be able to build small items, like little air driven engines, and hold tolerances within a thousandth of an inch, .001. Using my little CNC and Harbor Freight lathe, I have been very successful and have built several that really run well.

What I guess I have not been clear on, and it is clearly my omission, is that I am trying to keep within one thousandth, .001, in any project that I do. That was my goal from the outset of learning about CNC milling with this particular combination of compute hardware, CNC, and software. I do believe that I have been able to meet that goal. And what I have not made clear, here on this forum, is that when I am talking about tenths, (10ths) I am talking about how close to one thousandth +- I am able to get with my setup. The only place in my setup that is capable of measuring 10ths and moving by 10ths (.0001) is seen on the DRO's of my Mach3 screen. And yes, I can jog by .0001 increments if necessary.

I think that I recently demonstrated the ability to work within those limits, .001 +- a few 10ths, in my building of Weston's magnetic clock. I was really pleased to be able to cut his aluminum 3.2 inch Rotor Spoke Plate, and his steel Rotor Rim with the 3.196 center opening, and be able to cold shrink the Spoke Plate inside of the Rotor Rim. Absolutely nothing was necessary, as far as fitting goes, except to carefully file the four holding tabs from the Spoke Plate. And all of this on a cobbled together system that I have less than $2000 invested in. I know to you who are "Professional Machinists" that is not much of an accomplishment. But to someone like me, with absolutely no "Machine Shop" training, it is a big accomplishment.

I want to address some of the comments/questions ask in the above posts. I will do that in the next post. I wanted to clear the air about the .001 and 10ths issue before going on any further. Hopefully this gets us all back on the same page and what I am trying to do with my "strapped on HF flashlight" may now be a bit clearer than I made it previous to this.

05-24-2012, 06:25 PM
Here is how I initially setup the cross hair projector position and height. I used a very small ball mill and jogged down to the surface by .001 increments. I watched through a magnifying glass to detect exactly when the tool touched the metal plate. I then retracted the tool and set both the X & Y DRO's to Zero on the Mach3 screen.

Next I used my Z Axis touch plate and routine to find the surface height of the material. The routine compensates for the thickness of the touch plate, and then moves up one inch (1") from the surface of the material.

Now the tip of the tool is exactly one inch above the material, and the cross hair projector is about 4 inches above the surface of the material. I now jog the X & Y Axis so that the cross hair is exactly over the center of the very small mark made by the tool. I now read the distance that the X & Y Axis moved to reach the center of the mark. Those values are then programmed into the script that will be used to move the tool back over the center mark on the material. And yes, I look through a magnifying glass to position the cross hair as close as possible to the center of the mark.

The axis of the cross hair projector and the axis of the tool must be quite close to being parallel as my measurements of the X & Y distances only change by a couple of 10ths (.0001) if I do the same routine as above and only retract the tool height to .125 above the work surface.

When you project light through a lens, it get wider as it gets further from the projector. Take a good quality laser pointer. Measure the width of the dot projected a few inches in front of the lens. Then shine it on an object 10 feet away and measure the width of the beam there. You will find that in that 10 feet of distance, it may have only increased by a factor of 2 or 3. Nothing like the light from a projector spreads.

With the setup that I have, the "HF flashlight strapped" to my Z Axis frame, moving it 7/8 closer to the work surface does not change the width of the cross hair enough to be perceptible to the eye.

A couple of you have commented about using a center punch mark to mark the work to be milled, not making any sense. I guess I did not make it clear that the reason for the center punch mark in the metal plate in the pictures, was to demonstrate how easy it is to position the cross hair over a given point, the punch mark making a very small round bottom reflector that helps to demonstrate how easy it is to do the initial setup for the X Y Axis after touching the metal plate with the tip of the ball mill.

In reality I use the cross hair projector much like you would use an edge finder. Working off of a known point on the piece being machined, and using the DRO's, I can easily indicate the starting point for a hole, as an example, and with the click of the mouse, move the bit into its cutting position. Kind of like having an electronic rule that lets me measure in different directions and find the wanted starting point.

Hopefully some of this makes sense to you.

Jaakko Fagerlund
05-24-2012, 06:44 PM
The only place in my setup that is capable of measuring 10ths and moving by 10ths (.0001) is seen on the DRO's of my Mach3 screen. And yes, I can jog by .0001 increments if necessary.
Your "DRO" (just a counter, not a measuring device) may show 10ths, the Mach3 an issue 10th moves, but how have you verified that your machine actually moves that much (or little)?

05-24-2012, 09:56 PM
Hi Jaakko, I think I can answer that.

I have zero backlash ball screws, lead screws of known pitch, and the rated resolution is given as .0001. I am using 6 wire stepper motors, size NEMA 23 being driven by a Micro-stepping solid state Controller. There were a lot of test that had to be made, in the initial setup, to guarantee proper stepping for a given length. I remember there was a formula for using the pitch of the ball screw and lead screw to pick the proper ratio for stepper movement. Then setting a bunch of jumpers to give the proper instructions to the controller to drive the stepper/lead screw combination.

The final test, after everything was setup and verified was to use a 1 2 3 block to check all three axis for proper jogging distance. Once that all checked out, a final test was to set up an indicator, on one of the stepper motors. It was simply a coupling on the top of the Z axis stepper, a 3 inch piece of wire held sticking horizontally out of the coupler, and a white piece of paper under the wire on one side of the stepper frame. Started with the wire in a position over the paper, and carefully marked a line on the paper below it. For my setup stepper ratio, jog .1 inches was to be 1 revolution. Did that 50 times to make sure it was exactly over the line drawn on the paper after 50 presses. Then jog .01 and press it 10 times for one revolution, then jog .001 and press 100 times for 1 revolution, and last, leaving jog at .001 press 3 times. Then jog .0001 and pres jog back the other direction 30 times. The end result, the pointer was over the starting mark on the paper.

So, I can honestly say that when I jog .0001 it actually moves .0001 on any axis.

The alternate way was to use a dial indicator to simply check the movement. Back when I built the unit, I did not have a dial indicator, so used the simple steps recommended above. Some time later I did buy a dial indicator, and indeed my CNC will jog .0001 on all axis. That was the first thing I checked after I got the dial indicator.