Gantry squaring - purely manual or CNC12-assisted?

[kb58]

First off, I realize that regardless of method, it's on the user to ensure the gantry is square before enabling anything.

I'm currently using a C86 board and signal-sharing commands to the two gantry servos. I know that CNC12 (Pro and up) offers the option of sending different commands to each gantry servo, enabling automatic squaring using the limit switches. For those of you who switched from signal-sharing to having CNC12 handle the final squaring, are you happy with the operation? Any regrets? I'm asking because sometime when things are made more automated, there are unintended consequences. Thanks.

[cnckeith]

Several methods for gantry squaring, they all work. they all have pros and cons.

1.) manually pull the gantry against hard stops that are square and home out at that location. (simple, easy, quick to get working)

2.) square gantry using switches with SOFTWARE paired gantry axis motor, (either use the stock centroid wizard generated auto square macro or write you own)

3.) square gantry using switches with HARDWARE paired gantry axis motors (sometimes also called "relay autosquaring")

4.) square gantry using axis motor drive home to hard stop or home to switch mode(s) (not recommended for newbies)

[ShawnM]

I have setup a 6x12 router (Chinese made with steppers) with auto squaring and it works flawless. The router can cut 72" in the X axis and the gantry itself is 8-1/2 feet wide and heavy so it can easily come out of square when the motors aren't locked. At start up and once Acorn squares it during homing with the axis pairing and auto squaring it stays rock solid all day. A great option for large format machines.

I also have a 5x10 plasma table that is built very solid (US made) with no chance of it ever coming out of square and it has hardware paired Clearpath servos on it and simply homes to the home switches with no chance of it coming out of square.

Both work as they should and as Keith stated there's an option for everyone. Really depends on the machine what method you use. And as you stated you must build/buy a square machine, auto squaring is not meant to fix a machine that is not square.

[kb58]

The reason I'm considering changing the homing operation is because when the ClearPath servos power up, I've seen them sometimes shift position. The consequence of that didn't sink in until after I'd squared the gantry and cycled power, and it was slightly out of square again. Having the squaring function actively re-square on power-up solves any issue of the gantry gradually walking out of square over multiple power cycles.

[ShawnM]

I have not seen this issue on the multiple Clearpath setups I've done. They simply power up with the orange led rapid blink waiting to be enabled by Acorn. Since they are not enabled any stray step or direction signal would do nothing. You can freely move the motors once powered up before they are actually enabled. Why not manually sqaure the machine after you power it up?

After any power cycle you'd have to re-square and/or re-home the machine anyway so I confused by your comment about "multiple power cycles".

I also have never used the C86 boards for many reasons. It's anyone guess what that board is doing in line between Acorn and the servos.

[suntravel]

Hardeware parinig, and squaring fixed with a belt, ok for small router

Uwe

[kb58]

I have not seen this issue on the multiple Clearpath setups I've done. They simply power up with the orange led rapid blink waiting to be enabled by Acorn. Since they are not enabled any stray step or direction signal would do nothing. You can freely move the motors once powered up before they are actually enabled....

I misspoke by assuming that the servos come up before the Acorn. Both the servos and Acorn are powered on at the same time, but on separate supplies, so maybe they don't. Since the servos power up disabled, that suggests that the Acorn/C86 may be bumping their positions slightly.

...I confused by your comment about "multiple power cycles"...

I mean that if the router is repeatedly power-cycled, and if each time the servos shift position slightly, it means that over time, those slight offsets (if in different directions) could add up and become a problem. It's just a theory and is untested.

[kb58]

I made a six-second video showing the issue, last entry in my build album, https://photos.app.goo.gl/BDVkuuegUtyfm8v47
I watched both sides several times and while the servos bump the same direction, it's unclear whether it's the same angular amount.

[Gary Campbell]

Stepper motors and by extension step & direction DC servos may, in fact will, "jump" into a nearby magnetic detent on startup. Sometimes clockwise, sometimes counterclockwise. Sometimes a few encoder counts, sometimes more. I don't even know if stopping the motor at exactly the same shaft location makes it predictable.

This is why we home, and in some cases why we need independent auto square. That said, there is no real substitute for a machine that is rigid and is build with all the axis travels orthogonal to each other

[kb58]

Agree, but even assuming an infinitely stiff assembly, the servos may try to move to slightly different positions at power-up. Being perfectly stiff, the gantry will remain aligned, but because of that, there will be higher holding currents constantly trying to push the servos to their differing power-up positions. This isn't a reflection on ClearPath servos or any steppers for that matter, it's just a nature of the design. Without absolute linear encoders on the axes, the Acorn has no idea where anything is until it's homed. I now think that some sort of auto squaring is necessary. As my setup is now, it drives until it sees either limit switch, so the lagging servo will remain offset, and any holding currents will persist.

Just now I power-cycled the router several times (but not homed). The first two times, the delivered holding torque in the left-hand servo was +0.5 divisions on Teknic's scope app. The third time however, delivered holding torque was -0.5 divisions. This indicates that it or its twin on the other end of the gantry moved the opposite direction. All this does is underscore that dynamic squaring is pretty much a requirement for an accurate setup.