Reposting since this has been asked about a few times recently.

Reposting since this has been asked about a few times recently.
https://plus.google.com/102623480969211723334/posts/dUuohJTZkmg

Thanks John. I think I’m going to order one of the plates from triquetra and see if I can get this working.

Can u post a link to the plate? I’d like to get one too.

He regularly posts a coupon code 3corners for $20/off

Here is where I found the original info. https://discuss.inventables.com/t/3-axis-automatic-zero-for-x-carve/22454

Should cost $30 +shipping

I like it!!

I started thinking of whether it would be possible for me to use a touch plate, but all examples I find, on ebay and in this video, seem to use rigid pieces of metal. I seems super bad to run your cutter into a hard metal plate. Since the machine can’t stop instantly, that’s an excellent way to chip an edge on the cutter. It seems the plate would need to have some spring in it to take up excess motion. Does anyone know of such a thing?

If you watch my video on the touch plate widget, I show how you need to run at a slow feed rate so you don’t dig into the metal with your end mill. It is an issue, but by going slow I think it’s not a problem. However, if you use your guage block that is spring loaded you could solve your worry. Just hook the ground up to your plate on your guage block.

Yeah, the problem is that the “calibrated” position on that gauge block is ~1.5mm lower than first touch, and I think the end stop has some damping on it, so don’t know how repeatable that position is. Basically I’d like to find something that has a metal plate pushed up against a hard stop with a spring, The problem with such a design is that you’ll inevitably get some chips in between the plate and that’ll screw up your zeropoint, so it has to be done slightly smarter, I think. Maybe I’ll play with designing something.

I ordered one of the triquetra plates today to try out. I’m not worried about x and y crashes since it’s only held with hand pressure but z could be problematic and will just have to be turned to a really slow search speed.

What about 3D printing a block with conducting filament? Or covering a plastic block with some metal foil? This should work and be also crash proof.

@Simone_Marin I’d have two concerns for repeatability:

First, the top needs to be flat across an area at least as large as the size of your largest endmill, to your required accuracy. My machine repeatability is maybe 0.02mm in z, so I’d want no more than that over 10mm. That seems hard with 3d printing, although you could face the block.

The second is that it needs to be permanent, which means the top surface has to be hard. If it’s not, poking the endmill into it will make depressions and ruin the first requirement above.

You could have a soft material with a piece of metal on top of it, but soft materials tend to not hold their shape very well.

For kicks, I worked out the stopping distance D from an initial velocity using TinyG’s sinusoidal jerk planning. If I did my math right, it’s D = v0^1.5 sqrt(pi/(2*Jmax)), where v0 is the initial velocity and Jmax the configured max jerk setting (including the invisible factor of 1e6 you don’t type in).

The homing jerk setting on my machine is 8000, so Jmax=8e9 m/min^3. The stopping distance from a few select velocities then becomes:
100mm/min: 0.014mm
300mm/min: 0.073mm
600mm/min: 0.205mm
900mm/min: 0.378mm

Only the first distance would be remotely acceptable with a hard touch plate, but in order to cover the range of lengths of my tools it might have to probe for 200mm. That’s 2 minutes of waiting. I can manually measure the offset using my Z-axis setter to within 0.01mm much faster than that, so then it seems pointless.

ok then what about a block featuring both soft and hard surface? The softer surface may be used to stop the head coming at high speed, then it would retract, move on line with the hard surface then probe at 30mm/m

@Simone_Marin You know, that actually sounds like a really workable idea! :slight_smile: Setting up to run two probes will be a bit more complicated, but not by much.

Why 2 probes? It is one probe you just have to add a subroutine to move to a side after the first touch. It could be that simple as to glue a little piece of metal wool to the trinquera block.

Yeah, I meant scripting the two probe commands with requisite moves.

Technically, you can run an axis, stop, change feed rates, and start it again.