Z-axis carriage-mounted finger guage

Browse around for inspiration, or share your work to inspire others!

Moderators: al wolford, sbk

Post Reply
Posts: 679
Joined: Tue Jan 06, 2015 3:24 pm
Location: Hoquiam, WA

Z-axis carriage-mounted finger guage

Post by Rando »

Hey, y'all:

At our local (Tukwila, WA) CNC Club meeting, held at the ever-hospitable Rockler Woodworking store (sorry, shameless plug, but whatEVER...), the topic came up of squaring the bed to the gantry, and I mentioned I'd mounted a dial indicator on my z-axis carriage.

Here's the rationale behind what is clearly a questionable choice: On the one hand, it did lead me to drilling a 1/2" hole in the side plate of the z-axis carriage, and then mounting a delicate measuring device on the end of a non-positively retained swing arm, tightened only enough for moderate resistance,...on a moving, vibrating spindle carriage. Right! What could possibly go wrong?

(Sharkcutup, this is where your signature mantra would...err, should be applied :D ).

But on the plus side, any time I want to verify that the stock or part is true in the vise, or that the vises themselves are aligned, or where really is the high-spot on a piece of stock, and so on and so forth, I just raise the z-axis up, and flip it down. The length is set so the measuring finger can be below most bits I use, so I can typically safely do most top-surface measurements with the bit still in place...though not rotating! Then when the measurement is done, I just flip it back up, move the carriage back to my traditional starting place, and then proceed with whatever I'm doing.
Indicator arm in the down position
Indicator arm in the down position
At the top, and working down, there's a 1/2" hole drilled into the side plate of the z-axis spindle-carrying carriage, roughly inline with the spindle centerline. It doesn't have to be exact, because the mechanism is intended to be adjustable but rigid during normal measurements. Anyway, from the bolt, you've got a collection of the normal stuff you'd get with a cheap magnetic base, such as this one from Harbor Freight: http://www.harborfreight.com/multiposit ... -5645.html
Indicator arm in the up position
Indicator arm in the up position
I didn't use all the parts from that base in this arm, so there are additional parts that I use with the other magnetic bases I have/use for more-traditional measurements. Spindle-runout and tramming, that kind-of thing.
Closeup of the cheap dial indicator.
Closeup of the cheap dial indicator.
The finger indicator is one like in this search (it should return a listing from MSC's selection, hopefully something will be on sale!)

https://www.mscdirect.com/browse/tn/Mea ... y&dir=desc

The finger can be rotated into position, and it has several dovetail pillows to attach the mounting stud to, so it's very flexible in positioning. The fact that the finger can be rotated is a huge blessing should you go to the end of it's range (we've neeeeever done that, have we?), because instead of things bending, in most cases the needle will just be rotated and you have to start the measurement over. Better to do over with a tool that works than with one that's now broken.

Just remember "sine error". When the direction of the surface displacement being measured is significantly out of parallel with the arc segment swept out by the probe finger, the displayed distance can have significant under-reporting of the true displacement distance. The amount of error is proportional to the sine of the angle between the two, thus it name, sine error. :D

So, it's an idea. It works for me. I do mostly metal-working, trying to get the most out of the machine accuracy-wise, so these kinds of things are necessary, if maybe not "normal" ;-).

One thing I've found it moderately useful at is deciding whether an issue I was seeing was truly backlash, or an actual toolpath error, since you can put the probe tip in contact with a surface normal (perpendicular in all other axes) to the axis you're testing, and then just jog back and forth at a step size of 0.001, or whatever it will let you actually step in. On the SharkHD, the "step size" from these steppers is 360 degrees / 200 steps/revolution with a lead screw pitch of 8 TPI (but two threads, so effectively 1/4" per turn), or 0.00125" per stepper motor step. So, the actualy machine resolution is that value, not the claimed 0.001", but again...whatEVER. With an accurate dial indicator, you can clearly see how when the machine tries to incrementally position itself one-thousandths of an inch at a time, you can see the "extra step" that occurs in the aliasing (of 1.0 mil against a measurement window of 1.25 mil). Over longer distances it doesn't matter, but when you're trying to get accurate physical part sizing, well, these are things you want to know :D. As Dirty Harry said "A man's Got To Know His Limitations...."

Another of the jobs I helped a guy with this last summer involved engraving on the bottoms of bullet cartridges. Most production bullets have the stamp on the bottom of the brass, but he wanted to have a special engraving done for a present. We ended up trying to carve 0.040" lettering, in a circle, with cutting depths of about 0.002-0.005". We built a jig to handle 25 at a time. But, the engraving never went quite perfectly, even though the jig, made on the same machine, we knew had to be flat to the machine. It wasn't until I wrote some custom GCode to take a plunge-style digital indicator attached to that arm, and probe three places around the circumference of the 25 individual to-be-engraved surfaces, did we figure out what was going on. I could also have used VCarve to create modified "drill" cycles for the probes, since it already "knows" the true centers of the fixture bores. But, I didn't :P.

Instead, I just took the 25 center coordinates and manually manipulated them to get the probing locations, since in a square matrix of 25 holes, there are really only 10 numbers to calculate for the centers, and then the three probe locations are just offsets from those center locations. Easy Peasy. I used good-old G04 Pn.nnn "dwell" commands to stop the probe long enough for me to capture the value. From that we were able to determine that the cartridges weren't actually being held in place. And, having that GCode also let us test and re-test as we modified how we held those cartridges in place, until we finally got a fairly reliable process...then we ran out of cartridges...hahaha....and the press-stamp finally came in, and we called it good. But, without that attached probe, I don't think we ever would have been able to find and fix the problem.

Anyway, enjoy!
ThomR.com Creative tools and photographic art
A proud member of the Pacific Northwest CNC Club (now on Facebook)

Post Reply