My crazy "Keeping the X" solution?

[Tomzo]

I really enjoy my HF Rong Fu clone mill. Got it pretty cheap and with some upgrades (VFD and DRO) it is a pretty capable machine. While I can generally avoid situations where I need to raise the head in the middle of a setup, I have followed with some interest the various solutions folks have posted here. After considering some of the more elaborate solutions I decided to buy a cheap laser pointer and see how that would work. This is where I went into the rabbit hole...

I got the laser and had a spot about 18 feet away that I could draw a line for it to follow, but the spot on the wall was pretty large - at least 5mm across. I felt that would decrease its accuracy and my eyesight would play a factor as the contrast from the bright spot made finding the center pretty challenging. I was pondering this when I realized I had an Arduino Uno laying around, so I created a little circuit board using a light depended resistor (LDR) and five LEDs and wrote some simple code for the Arduino that would change which LED lit up based on the intensity of the light received.

I 3D printed a box for it with a cover that had five holes for the LEDs and another for the LDR. The hole for the LDR is about 1.5mm across and I made adjustments to the sensitivity levels so the lights would not line up unless the center of the light spot was right over the hole. I have a little internal shield in there to prevent internal reflections in the box itself from affecting the readings. I also have some raised lines that I painted white to help me align it from across the room.

Satisfied with myself, I went to put it on the wall and realized that while it would be great if I never moved the mill head, it did nothing unless it could move as well and move accurately up and down. It was like the children's story of the old lady who swallowed a fly - I got a laser but that sent me to design a circuit and write some code and once that was done I realized I needed something to raise and lower the whole device to get it into the new position of the laser. I used a cheap 12VDC motor, a 3D printer leadscrew, and some fairly cheap linear rails to mount the whole thing to a scrap aluminum plate I had laying around. I installed limit switches at each end and ran a toggle switch over to the mill.

The motor was too slow at 12VDC so I opted to supply 18VDC to speed things up. I figured the duty cycle will be very low so this should not be an issue and if it is the motor cost like $13 so I can just replace it. To power the Arduino I happened to have a buck converter left over from another project so it is powered from the same supply that runs the motor.

I set the code to have the first light on whenever the Arduino has power. This is just to tell me it is plugged in as I could see myself running the thing up and down cursing as to why I can't get it to give me all five lights. Aligning it on the wall was an iterative process. I milled a hole in some aluminum with the mill at its lowest point and the quill as far up as possible. I set the rail/detector system to read at all five lights at that point and then moved the mill head up as far as I could for the quill to reach the hole again and adjusted the upper part. I milled some slots in the backer plate to facilitate this but it took at least 30 minutes to get something that I was satisfied with. The images of the device on the wall shows one image with just the one light on and another when the laser is on which shows all five. The toggle switch is just to the right of the laser and the toggle clamp is how I turn it on and off. The laser is rechargeable and has a USB Type A plug under the cap. Leaving the laser on for several minutes kills the battery so I will be soldering up a customer connector to keep 5V on the laser. The way I set the intensity it needs to be at full strength as the hole is small and even when right on the spot if the laser is not putting out full power the detection limits are too low.

Anyway, it "works". After moving the head the detector is pretty sensitive to misalignment. To get all five lights on I have to give small taps to the head in one direction or another to get the lit up. When aligned the tool slides right back into the hole. I also have to tighten the head bolts carefully and in steps as the act of tightening the bolts takes it a bit out of alignment. I sure wish there was a better way to clamp the head around the column, but I guess that is what knee mills are for.

I would be interested in other methods folks use to check the X as I am sure there is a better way to do this than how I did it.

The one thing I know for sure is that putting all of this together, while fun, is probably one of the most ridiculous projects I have done. I have my doubts as to how accurate it will ultimately be, but I am not making stuff for NASA and just getting out the wobbler and resetting my origins with the DRO will probably be just as fast as the process I created here. Once I was deep into the project I had to finish it (I still have some cable management to finish up).

There you have it - my contribution to the long list of "keeping the X" solutions!

Tom

 

[RJSakowski]

I seriously try to avoid having to move the head. This takes some planning in setting up. One thing that helped me was to purchase a set of R8 collets and mount my drill bit in collets rather than using a chuck. This buys me about 2.5" of useful travel. If I have to move the head, I will try to have all the critical work done in the 1st setup. Drilling usually isn't considered a precision operation so if you miss by a few thousandths, it may be acceptable.

One trick that I use is to locate a hole with a dowel pin prior to moving the head and when I finish moving the head, I relocate that hole. If the work is in a vise, I will loosen the vise slightly and allow the hole to move to accept the dowel pin and then tighten the vise. Finally, if I need precise positioning, I use my edge finder to re-reference the work.

 

[jmkasunich]

I've always thought that if I ever had a round column mill I'd do the laser thing. But my plan would be to mount a long skinny mirror (I have a couple from old laser printers) on the opposite wall. That would bounce the beam back to the front of the mill head so I could easily see it with my MkI eyeball while standing at the mill tapping it into alignment and tightening it down. As a bonus it would double the travel distance and thus the sensitivity. The long skinny mirror would be mounted vertically so that as the mill head goes up and down the beam still hits the mirror.

 

[pontiac428]

I cut a hole in the roof of my shop so I can bounce a laser off the moon. It doesn't matter what phase the moon is in, but the time of day is critical.

 

[Tomzo]

I've always thought that if I ever had a round column mill I'd do the laser thing. But my plan would be to mount a long skinny mirror (I have a couple from old laser printers) on the opposite wall. That would bounce the beam back to the front of the mill head so I could easily see it with my MkI eyeball while standing at the mill tapping it into alignment and tightening it down. As a bonus it would double the travel distance and thus the sensitivity. The long skinny mirror would be mounted vertically so that as the mill head goes up and down the beam still hits the mirror.

Hmm - that may actually work with my solution. I could mount my detector to the mill instead. Getting the mirror aligned correctly would take some doing but would be oodles simpler than what I built! Oh well, what's done is done!

 

[RJSakowski]

As you have discovered, aligning the round column head via a distant mark has its complications. However, there was a thread a few tears back about using a laser to align a mill vise.

As you discovered, the width of the laser dot leads to a lack of precision but your solution should improve that considerably. To align an RF30 mill to +/- .001 requires resolution of 20 arc seconds of rotation. The resolution on a wall 20 ft. away has to be .024". Tramming a 5" vise to +/- .0005" would require the same resolution.

 

[Tomzo]

As you have discovered, aligning the round column head via a distant mark has its complications. However, there was a thread a few tears back about using a laser to align a mill vise.

As you discovered, the width of the laser dot leads to a lack of precision but your solution should improve that considerably. To align an RF30 mill to +/- .001 requires resolution of 20 arc seconds of rotation. The resolution on a wall 20 ft. away has to be .024". Tramming a 5" vise to +/- .0005" would require the same resolution.

Yes - precision with this method is a challenge. That 0.024" is about half a millimeter. My hole for the light is 1.5 mm but the light needs to be right on the center to get all five to light up. I am not sure if I am within half a millimeter but my guess is that I am within 1 mm.

I am thinking of using my DRO to measure how well I get back to the same spot, but again it requires the quill to move. I would set an indicator against the shaft of an end mill (in an ER32 collet), set the indicator to zero and then zero the X on my DRO. I can then raise the head, move the table an inch or so, lower the quill and then come back to where the indicator reads zero and see whether the DRO goes back to zero.

Again, perhaps more trouble than just getting the wobbler out and resetting my origins on the DRO. It was a fun project anyway....

 

[markba633csi]

Rube Goldberg would be proud! LOL

 

[extropic]

Here is a much simpler idea that would allow you to raise the head. The method assumes the column is precisely cylindrical, precisely perpendicular to the table travels and that the quill travel and the spindle axis are precisely parallel to the column. Additionally, your table has to have some spaced exposed to securely (temporarily) mount a reference plate (described below).

Don't reset your DRO during this procedure so you can return to your desired location after the head is raised.
The reference plate is simply a plate with a precisely round hole in it. I'm thinking a 2" hole would be comfortable, but any size could work.
You retract the quill completely and mount the reference plate securely above the table such that you can center the reference plate (move X & Y) under your spindle using a DTI.
The reference plate needs to be stable enough to allow reliable DTI readings. Create a structure as required.
After you have the spindle in the center of the hole, lock the table X & Y and DO NOT MOVE THE TABLE AGAIN until the head is at you desired height.
Raise the head to your desired height and pivot the head until the spindle axis is re-centered in the reference hole.
If the spindle axis will not re-center (with the head clamped), you have just discovered that one or more of the assumptions are false.
If the DTI shows an error you can live with, Bingo, you're golden.
If the indicated error is unacceptable, the method should give you some idea of the magnitude of any orthogonality or parallelism errors that exist.

 

[pontiac428]

^ and that, my friends, is using a datum.

The difference between using a datum and making it really really hard is like the difference between getting yourself to school and riding the short bus.