Scanning Z probe support in RRF 3.5

[dc42]

@matej1006 it's an inductive probe so the usual caveats apply: bed surface must be metallic and uniform, and the sensor is likely to be sensitive to temperature. In fact I am using a flex steel bed with magnets holding it down, but the magnets don't appear to be causing a problem because they are not obvious in the height map. The sensing IC (LDC1612) is the same one that is used in the Beacon.

[appjaws]

@dc42 that's fantastic.
Will we be able to have something similar using your original ir probe?

[samlogan87]

Very cool. Would this still work with a garolite sheet on top of aluminium?

[dc42]

@appjaws said in Scanning Z probe support in RRF 3.5:

@dc42 that's fantastic.
Will we be able to have something similar using your original ir probe?

Yes, in principle. It will need modified firmware on the IR sensor to reduce the amount of filtering and provide a true analog output.

@samlogan87 said in Scanning Z probe support in RRF 3.5:

Very cool. Would this still work with a garolite sheet on top of aluminium?

If the garolite sheet is thin enough, then yes. Currently I am using the 20mm diameter coil that came with the breakout board and I have the coil about 8mm above the bed; so plenty of room for a sheet. It will measure the distance to the aluminium, so obviously the garolite sheet must be of uniform thickness and in contact with the aluminium over the whole bed, in order that the top surface of the garolite matches what the sensor is measuring.

[CNCModeller]

@dc42 said in Scanning Z probe support in RRF 3.5:

@appjaws said in Scanning Z probe support in RRF 3.5:

@dc42 that's fantastic.
Will we be able to have something similar using your original ir probe?

Yes, in principle. It will need modified firmware on the IR sensor to reduce the amount of filtering and provide a true analog output.

So can we use any analogue sensor input, for instance a linear hall effect sensor measuring the location of a ball rolling over the bed surface?

[samlogan87]

@dc42 ok cool. My sheet is 3.5 mm thick and seems pretty flat

[gringo]

Looks like the Beacon Probe. Good Job
When will it be available to the public and what is neededÜ

[oliof]

@gringo new firmware, an LPC1612 sensor, and a (modified for now) toolboard.

[TLAS]

@dc42
Wow, kudos. It’s been a while since I’ve been like “that’s awesome” when reading about a firmware update.

I have a laser sensor I’m upgrading to that has an analog output and trigger - shouldn’t have a problem in a similar setup I think.

Now, this begs the question… when can I use this to audit a print? It would be awesome to scan the print during any travel moves (or a deliberate linear pass) to make sure the print hasn’t pulled off or anything.

[o_lampe]

Good bye BLtouch

My favourite sensor would be the IR-probe, too.
Maybe I'm seeing problems, but inductive or capacitve sensors are reacting slower than light sensors. Because they store energy and may have a filtering/smoothing behaviour?
Plus, the IR sensor is much smaller and lighter.

[o_lampe]

@dc42 Is there a way to store probe points only when the probed height has changed? That would compress the heightmap a lot and we can use a finer probe grid. Or do the probed polygons have to have the same size/shape?

[JoergS5]

That's impressive speed and I can think of multiple use cases, like non-planar. I'd wish there is a comparable solution for non-metalic objects, to 3D scan the already printed object.

[brunofporto]

@dc42 One can always use a mall single beam LIDAR as those small ones used by drones..... Those are not tailored for small distances but something similar for short distances may exist.

[JoergS5]

@brunofporto maybe it's possible to gather the different types in a future RRF version.

[dc42]

@o_lampe said in Scanning Z probe support in RRF 3.5:

My favourite sensor would be the IR-probe, too.
Maybe I'm seeing problems, but inductive or capacitve sensors are reacting slower than light sensors. Because they store energy and may have a filtering/smoothing behaviour?

The prototype that I put together reacts very quickly, even faster than the IR sensor does because the IR sensor has some filtering. Currently I use the 20mm diameter x 1.5mm thick sense coil supplied with the breakout board, but I intend to try a smaller one.

@brunofporto said in Scanning Z probe support in RRF 3.5:

@dc42 One can always use a mall single beam LIDAR as those small ones used by drones..... Those are not tailored for small distances but something similar for short distances may exist.

I look at LIDAR sensors occasionally but so far they don't have sufficient precision or reproducibility to be good enough to use as a Z probe. Even if we accept a precision and reproducibility as large as 10um (so 20um round trip for the laser), that would require measuring the laser round trip time to within an accuracy of about 0.07 picosecond.

[CNCModeller]

@dc42 could we use the analogue output from a linear hall effect sensor measuring the location of a ball rolling across the bed surface as an input for this?

I assume there is already a need to for some form of analogue input to distance calibration routine?

Many thanks
Barry M

[brunofporto]

@dc42 plus the issue with different types of surfaces...... Still have my laser filament sensor

I like @CNCModeller idea of the indirect measurement with a ball but then there is the dynamic problem of a ball "jumping" on small defects or residual first layers.

[dc42]

@CNCModeller said in Scanning Z probe support in RRF 3.5:

could we use the analogue output from a linear hall effect sensor measuring the location of a ball rolling across the bed surface as an input for this?

Yes, in principle.

[o_lampe]

@brunofporto
The ball could be held down by a (plastic) spring. Like a big ballpen.
That was at least my idea for building an inductive probe for glass plates and other non-metallic surfaces. The 8mm pinda probe would've been ideal.
But I find non-touching probes better suited for fast scanning moves.

[o_lampe]

@dc42 said in Scanning Z probe support in RRF 3.5:

The prototype that I put together reacts very quickly, even faster than the IR sensor does because the IR sensor has some filtering

But you can easily change the IR-probe circuit, because it's your design?
These inductive/capacitive probes are a black box.

@CNCModeller I like your idea too, but I'm concerned about the probing range. Should be no problem for mesh probing, but levelling often starts with two or more passes bed-tramming. That would better be done the classic up/down way.