Ohmic sensing, what's best to use ?

Hi all gurus and not so gurus

Now that my THCAD probems with a crap pilot arc machine are solved are solved I'm planning on making a an Ohmic sensor as i'm cutting some 10/10 mm cold rolled steel and it flexes a lot under the weight of the floating head when on small parts.

I got a bunch of Arduinos in various shoe boxes and plan to use one as a sensor.
But ihave some, maybe stupid, questions about the signal to be used for sensing.

I understand the Ohmic sense can have difficulities with rust and paint or any non /bad conductive layer on the metal

Various options do arise, but need testing:

_PWM signal
_DC low voltage
_Capacitive sensing (still uses a counter in the arduino though)
_High frequency (as much as an Arduino can count depending on processor used)
_Combining HF with PWM

It's all a matter of comparing the signal that goes out to the one coming back or comparing it with a base line.

And as i'll be using the wire in the torch used for the HF pilot arc for sensing i can get the arduino to disable the relay piloting the plasma generator ( plasma ON)

Most people are just using some relays as per the circuit in the Plasma Primer.
A few of us are using a second THCAD-5 board with different firmware available for download at mesa that adds 2 extra encoder A inputs and a component I wrote with a lot of guidance from PCW. This improves performance on water tables.

You could follow Hypertherm's example and use proximity sensors that shoot down on an air ram. (they use two). Theirs works underwater.

Yeah i saw those THCAD being used as sensors, great work from your part !

I dont want to use relays as it's slow to react and i want to be able to adjust the sensitivity. Atm i'm not using a water table (it's in the future project) and the system used by Hypertherm is too complicted for my taste, adds unnecessary mechanical parts. If i use an air ram it'll be for an air scribe

I'm going to investigate about sensors and conductivity (used to be my previous job also) and see what's best in terms of reliaibility and selectivity for discriminating false triggers

So don't spend any more than the cost of a THCAD. Recapping, my tests showed we could see the voltage ramp up over 0.04mm of travel so by the time we react to a probe down event we know we are at least 0.04mm past the initial contact. So by monitoring a moving average of the last 10 readings, we don't really have any switch hysteresis. Well we, do but that is actually above the threshold trigger. Its hard to explain but it works!

I could do that but by the time it takes to arrive here i'll have finished my system, i guess.

And there is no thrill in taking a system some1 else devlopped

That depends. You have no idea how hard it was coming up with it! Be sure to learn how to use halcompile and write your own components. A lot of fun and Linuxcnc's best kept secret! I love it...

rodw wrote:

That depends. You have no idea how hard it was coming up with it! Be sure to learn how to use halcompile and write your own components. A lot of fun and Linuxcnc's best kept secret! I love it...

That is where i lack comptency, coding and understanding the codes !
I can do some stuff with an arduino, but C and Py is too much for me, i need to grasp how the HAL system works and so on...

phillc54 wrote:

I was thinking of trying out the capacitive one here: forum.linuxcnc.org/plasmac/39155-alterna...tion?start=10#168859

Capacitance might be ok with some auto compensation for the lenght of wires and cables etc.... it seems the easy way to go, and there is some arduino capacitance libraries already made.