Auto Level Sensor Musing

[Ron]

Just a piece of information to add to the printer-scape for an FT-5.
Please don't shoot the messenger.
FWIW......

OK, I just had to try one.
Prior to this I really didn't have much, if any, of a leveling issue to start with.
That is because I run doubled corner bearings, linked Z motors, three point bed leveling adjusters, and a very rigid and stable Z stop.
As a result I basically only had to re-level the bed and adjust the Z stop more or less when I changed nozzles.
But I like to experiment with things just to see what all the fuss is about.

So I purchased the GeeeTech version (yes, I know....) fully figuring that I would purchase a real BL-Touch later if I liked the results and felt that there was promise to it.

First of all, with the controller I am running it was somewhat of a PITA to set up requiring the use of an oscilloscope to diagnose it.
Theoretically it is a simple thing but I had to experiment with several versions of Smoothieware before finding a version that would accept the PWM commands and execute them correctly.
However, finally I did get it to work.

I must admit that it was quite interesting to be able to dial in the first layer thickness merely by setting an offset, and especially so from the slicer (Cura).
But I could do that anyway with the regular sensor.
So far so good.
However-

The sensor rattled like mad from the start when the motions were running (even in 1/32 step mode) with prominent resonance points in the motions.
And it was clearly self-destructing, to the point that the probe is visibly no longer parallel to the sensor body.
Finally, this past weekend I fired up my printer and could not get a decent first layer to print, spending hours trying to tune and fix the problems.
So eventually I commented out the commands to deploy and use the sensor and went back to using just the regular Z stop sensor, and guess what?
The problems then completely went away and I was able to print things again.

Now I know that GeeeTech is a cheapened version of the AntClabs sensor who protest that its competitors are giving it a bad name.
I get that.
Though coming from a company in a region infamous for IP piracy that sounds a bit hollow.
But seriously, I feel that there are two fundamental flaws with this type of sensor, regardless of its source.
First, there is little or nothing to keep the probe tip from moving side to side.
Hence the rattling and wear.
Second, and much more important, the use of a hall-effect sensor and a magnetic armature to precisely sense vertical (axial) position is suspect to begin with.
There are too many things to go wrong with both items.

So, I am not impressed and frankly I really didn't need it to begin with.
Your experiences with these (including AntClabs versions) may be apple pie and sunshine wonderful.
But if the machine is set up rigidly to start with, I really don't see the need.
So into the scrap box it goes.

[S550Stang]

Well written, and thank you for this. I have an i3 2020, and was considering a probe. My reason was that my rig has a crown at temperature. I'm not sure if it is because of the corner constraints and thermal expansion, or thermal stress. I plan on running more tests to figure this out and then develop a solution to minimize.

[Ron]

Thank you, I appreciate that.

Actually, suspected bowing of the mirror tile I am using as a build surface was the only reason I even bothered with it to begin with, despite being curious.
The tile is constrained from moving horizontally but otherwise "floats" over the 1/4" thick heater plate I have fitted.
I understand that with an i3 2020 the bed moves horizontally, so you have a need for more constraints in that plane (mine of course only moves vertically).

You might consider using one of the 25mm rectangular proximity sensors.
Like this- https://www.ebay.com/itm/Inductive-Proximity-NPN-Sensor-SN04-N-DC6-36V-3D-Printer-Heated-Bed-Auto-Level/323171839020?hash=item4b3e8aec2c:g:RDQAAOSwGBdauW-P
This is because they have a bit better sensing range than most cylindrical ones.
An Ender4 printer I built last year uses one of these and it works great as a bed level sensor.
And they don't rattle.

Of course that will only sense the underlying metal surface, not a glass one.

[S550Stang]

I have a piece of thin 1/8 plate glass atop my heat bed. I was thinking of going to a thicker piece to gain more rigidity. I'd like to level that and affix the heated bed to the glass, foil tape for now, but switch to a adhered silicone rubber heater. with quick disconnect to change out beds if need be.

[carefudundae]

I feel that there are two fundamental flaws with this type of sensor, regardless of its source.First, there is little or nothing to keep the probe tip from moving side to side.Hence the rattling and wear.Second, and much more important, the use of a hall-effect sensor and a magnetic armature to precisely sense vertical (axial) position is suspect to begin with.There are too many things to go wrong with both items.

I've had a legit BL-touch on my FT5 for about a year and haven't noticed any wobbling or self-destruction... I have to change the offset a tiny amount whenever i swap nozzles, but that's it.

[Ron]

I've had a legit BL-touch on my FT5 for about a year and haven't noticed any wobbling or self-destruction... I have to change the offset a tiny amount whenever i swap nozzles, but that's it.

That's another great data point. Thanks..!
Still, I didn't really need one to start with, so I don't regret the small investment (<$15) to gather some working knowledge of how these things function.

I do have an idea for an improvement or two to these types of sensors though.
It involves very high sensitivity contact sensor points and a commercial off-the-shelf bi-stable solenoid.
AntClabs didn't invent "their" power saving pulse motion despite their claims BTW, it has been around at least 45 years in commercial form.
And I have some simple ideas to improve the GeeTech one too.

Maybe I'll build one of mine just to see what happens.
But I have a backlog of things to print first......