Scanning Z probe support in RRF 3.5

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.

lparnell34

@dc42 Can this be wired to io2 input on the 6hc 1.01 if the resistor is bypassed or can I use the 3hc expansion board?

dc42

@lparnell34 said in Scanning Z probe support in RRF 3.5:

@dc42 Can this be wired to io2 input on the 6hc 1.01 if the resistor is bypassed or can I use the 3hc expansion board?

I2C doesn't travel long distances well. It was designed for connections between ICs on a single PCB. The signal is pulled to ground actively but relies on pullup resistors to pull it high, which makes it susceptible to noise - and I2C has no error correction protocol. Therefore I advise against using I2C with wire lengths greater than a few cm.

RRF for the 6HC doesn't yet support I2C - we haven't yet written a driver for the I2C peripheral in the SAME70 processor.

RRF for the 3HC doesn't currently support I2C, however it should be possible to enable it in the Pins_EXP3HC file because the driver that we already use for the tool board should work.

For now the best ways to connect the LDC1612 are probably to hack a tool board as I did (I will publish details on my blog soon) or to use a SAMMYC21. For the future, we are considering making a dedicated CAN-connected board to support the LDC1612 as well as tool board variants with this chip on board. However, we still need to determine whether this type of inductive sensor is good enough to be worthwhile, given the known temperature sensitivity issues of inductive sensors.

lparnell34

@dc42 thanks. I'm going to use a sammy-c21 board. Can I connect the can-out on the tool distribution board to the sammy-c21? Or do I have to use one of the four can connections for the tool boards?

dc42

@lparnell34 said in Scanning Z probe support in RRF 3.5:

@dc42 thanks. I'm going to use a sammy-c21 board. Can I connect the can-out on the tool distribution board to the sammy-c21? Or do I have to use one of the four can connections for the tool boards?

You can do either.