Z Axis Touch Plate Recommendations

So I will agree with Peter Van Der Walt. I fumbled around untill they released the XYZ Probe. I have not looked back! Works good.

 

A high-precision momentary switch with a carbide face is what you need for faster probing and overtravel protection. Apparently the VMC-style Z-probes with the air knife that are about $60-80 are surprisingly good, and given that the body height at activation is fixed and repeatable, you can literally just set it on your workpiece to probe Z (and remember to adjust for its thickness, or you'll be cutting air). Ultra precision limit switches by themselves (without overtravel protection, don't bottom them out) from Metrol are $100-200 for a basic one, I have one, they're nice (and super low profile, which is handy for low-Z gantries). They're probably what is used in the high-end toolsetters, and will give sub-micron repeatability. Definitely overkill for this kind of machine, but you'd never need to question it, that's for sure.

Of course this is only in Z for "fake" Z-heights. To properly XY[Z] probe material sides, an electronic 3-point-contact "Renishaw-style" probe mounted in the spindle would be the way to go. Personally for the time being I plan on just using a Tschorn 3D taster for XYZ, but it would be nice if it were more automated. Maybe I'll make an electronic touch probe at some point.

If you're wondering what I mean by "fake" z-heights, setting WCS Z0 by probing the material with the tool is technically wrong, because Z0 should be part of the machine- typically the spindle face or a fixed offset from it. Now because grbl doesn't support toolchanges, it really doesn't matter that much in practice, but anyone using multiple tools with optional stops for manual changes in a single g-code file will need to properly understand the difference between setting z-zero (typically at the top of the workpiece, sometimes on a repeatable fixture) and probing their Tool Length Offset (and grbl does support TLO, for whatever reason).

 

As long as you set Z0 to the bottom of the material in CAM, that's a perfectly valid option too. Some people only ever zero to the table and then work from there (if they have vises/pallets/fixtures of whatever kind, those go into CAD as well!)

And like I say, the difference between "technically wrong" and "practically wrong" in machining can be a vast gulf sometimes. Just pays to be aware of the rules so you know why you're breaking them. Other than that... Whatever it takes.

 

Hahaha. To be fair, as we've seen repeatedly with things like CONTROL's lack of G53 view, "simple" doesn't always equal "beginner friendly", depending on the person's type of learning. Sometimes the lack of information can be just as bad as too much- but at least with it they have control over how much they attempt to absorb at once.

This way anyone who wants to look up "proper" probing can do so with some terminology to get started, and anyone who doesn't and simply wants to keep tapping a conductive tool on a piece of foil can also happily do so, and everyone will go on their merry way.

The OB XYZ Probe is a great idea, but it's also important to bear in mind that probing is inherently a controlled crash. The "controlled" part is important, in that the force has to be minimized, and since the mass is consistent, that occurs two ways: speed of impact (to reduce the velocity element of deceleration) and elastic yield (to increase the time component of deceleration). The reason the OBXYZP can work as a Z-height setter as a rigid material is that it's almost always used with soft materials that can elastically deform under pressure. That's a tiny amount of extra time vs a sprung switch's huge travel, but it's something. To compensate, the probing feed rate has to be quite slow. It also depends on the material elastically deforming- if the workpiece stays partially squished (say XPS foam), your probing is inaccurate.

With a proper switch-based toolsetter or three-point probe, the yield is in the spring, which provides far more deceleration distance than a semi-squishy material like plywood, so you can run probe cycles faster. It also allows the contact body (the button top for a setter, the ball tip for a probe) to be made from material as hard or harder than the tool/workpiece itself (carbide or ruby, typically) which provides tool longevity for high cycle counts as well as probing precision by reducing material deformation. If you used a solid carbide disc to probe the top of an aluminum workpiece with a tool, you'd probably chip a cutting edge.

The OBXYZP will, no matter how slowly you probe, eventually succumb to the hard carbide edges constantly hitting it and over time lose flatness and precision. Will that matter for the relatively low cycle counts and mdoerate precision operations typical to extrusion machines? Highly unlikely, I'm sure everyone who bought one will use it happily for years. But I also wouldn't want to treat it as a reason not to give beginners information that may influence their decision, or become useful later on, or help solidify the theory of machine operation and allow them to operate their machines more effectively. I also don't want to assume that every beginner reading this is entirely new to not only CNC, but also machining, engineering, or mechanical pursuits in general. Many things may make complete sense to them immediately once introduced.

Information wants to be free!

 

I would like to switch to a mechanical z-touch probe because I just had a failure with the OB XYZ probe where it didn't stop and the bit crashed into the metal probe, even though the LEDs turned green indicating that it had touched it. I have no idea what happened. I had to hit the emergency stop, restart, and then home to get the machine back to functional.

Any suggestions which z-touch probe to get and how to make it work with the blackbox?

I am thinking to have the z-probe permanently mounted in one corner so that the tool can automatically go there and zero z for a new tool. Is that possible with the blackbox/Interface?

 

The wiring is fine and hasn't changed. I have even tied the cable to the blackbox so that there is no possibility that the connector gets wiggled or pulled out using the probe. I haven't extensively tested this fail mode but it looks to me that perhaps there was some interference on the cable. The problem with this is that it is a really bad thing to happen. It destroys the probe, the bit, and the calibration of the z-axis. It is not something I want to have happen again. Thus a fundamental change is necessary. A fixed-mounted z-probe seems a more reliable and less fail-prone setup.

 

How does one wire in a tool length sensor to the blackbox? Is this simply a signal going from low to high?

Some of these tool-length sensors have a second switch, an overtravel protection switch that one could potentially combine with the z-limit switch to prevent the tool from crashing into the sensor. Has anybody done that? Would I wire this overtravel switch in parallel to the z-limit switch?