Plasma Current Control on the Fly (synchronous with motion)

More
16 Apr 2021 18:31 #206117 by robertspark
I was wondering is anyone is controlling their air plasma cutter current setting on the fly in sync with motion?

NOT the hypertherm RS485, I am just wondering if anyone is doing it via other means?

I have a Hypertherm Powermax 45 (non XP) that uses a bourns 10K slim pot
www.bourns.com/products/potentiometers/i...ct/PC-Slimline%20Pot

I have thought about using a MESA SPINx1 via PWM to control the amperage on the fly (and within gcode), whilst still retaining the use of the manual pot (knob) on the front of the plasma cutter.

I have had some designs for a while as to how to do it + schematics and PCB's (although no using an SPINx1) but in conjunction with a MAX325 analog relay as I still have and use the hand torch from time to time for different tasks.

The problem has always been doing it in sync with motion before I have started dabbling with LinuxCNC which obviously has available M67
_____________________________________________________________________

Just wondering if anyone is running such a setup, controlling the amperage on the fly?

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 19:03 #206119 by rodw
Grotius shared some photos of a modified Thermal Dynamics 60i machine using an Ethercat pot. Looking at the TD circuit diagrams, I think it would be pretty easy to make a piggyback board to intercept the signals on the 50 pin connector that feeds the front panel. This should work on the TD A120 I have too.

I think for this to be useful, it needs to have an algorithm that is based on what Hypertherm calls the centripetal limit.
For a given cut velocity and machine acceleration, it should be possible to calculate the minimum arc radius that can be negotiated without needing to reduce the velocity. I did modify the state tags introduced in v 2.9 to return the arc radius and if its traversing a full circle or not. See github.com/LinuxCNC/linuxcnc/pull/900 Alas, I have not had the time to finish it off. (or the maths knowledge)

The intention was to use the new pins created in motion that expose the radius and heading (direction of cut) in real time from a custom component that would calculate the centripetal limit and modify cut parameters by some as yet undefined algorithm.

The ultimate would be: When the limit was crossed, lock the torch height and modify current to achieve desired cut voltage. In theory, this should keep the power factor constant at reduced velocity which should keep the kerf width constant...

The other simpler method might be to modify current based on the velocity reduction in force at that point in time.

Another alternative would be to use external offsets based on the X/Y components of the heading (Pythagoras will help) to offset the torch position to account for the wider kerf.

So all in all, reading the arc radius in real time would be a very valuable feature for plasma cutting but the ultimate would be to use it to keep voltage constant by varying the current!
The following user(s) said Thank You: thefabricator03

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 19:11 #206122 by rodw

The problem has always been doing it in sync with motion before I have started dabbling with LinuxCNC which obviously has available M67
_____________________________________________________________________

Just wondering if anyone is running such a setup, controlling the amperage on the fly?


Note that the use of M67 assume you know what current reduction you want to apply when the gcode is written but gcode has no knowledge of the centripetal limit and when the torch slows down.

Linuxcnc is much smarter than that as custom components can read and act on the environment in real time. The enabling step is to modify the core linuxcnc code to provide the info we want to get the project under way...
The following user(s) said Thank You: thefabricator03

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 19:27 #206124 by robertspark
thanks Rod, just looking to use M67 at the moment and get the hardware working.

interested to see if anyone else has any working hardware

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 19:35 #206125 by rodw
Not sure which mesa card you have but if its a 7i76e, you night be better off using spindle speed to control your digital pot using the spindle interface.

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 20:11 #206130 by tommylight
I have thought of doing that several times, but so far i have failed to find a use for it besides reducing the dross on sharp corners and small radii .

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 20:17 #206131 by rodw

I have thought of doing that several times, but so far i have failed to find a use for it besides reducing the dross on sharp corners and small radii .


Thats it, improved cut quality is the goal.

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 20:52 #206133 by robertspark
Yes I have a 7i76E for the plasma, but it is shared with a mill (VFD) and a router (SuperPID), so I have a demand for analog outputs.....

I was going to control the VFD via modbus, but I don't like to control a VFD via comms protocol.... as there is no failsafe (IMO / YMMV)

I was also going to control the superpid via PWM (via another LPT bob on one of he expansion ports).

But I have decided to pick up 2 x Mesa SPINx1 PWM boards....
1) for the SuperPID
2) to have a go at plasma current control.
.... and the internal 7i76e is going to be used for the mill spindle VFD.

the SPINx1 has the advantage of being able to be mounted within the plasma cutter (nice and neat + out of the way + wired directly) and not having long leads with potential induced noise.

{I would have preferred if the SPINx1 used a differential input voltage for the 5v PWM signal.... but maybe I'm being picky..... it does have a relatively low input impedance instead}

I have a few spare LPT bobs knocking about so was going to use a C10S (basically a REALLY compact one of these, no longer sold it appears)
www.cnc4pc.com/c10-bi-directional-parall...-interface-card.html
so was going to use that for all my 5V PWM requirements and interfaces that the 7i76E does not have available.

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 20:56 #206134 by robertspark
I was not fussed about changing the amperage on arcs, I was more interested in changing the amperage on internal tight corners.

I was also keen to be able to change the amperage from the gcode file for dimpling and marking type operations. I like to keep my working settings in the gcode file as I can never remember the optimum voltage settings that work best for me,

Please Log in or Create an account to join the conversation.

More
16 Apr 2021 21:13 - 16 Apr 2021 21:14 #206138 by rodw

I was not fussed about changing the amperage on arcs, I was more interested in changing the amperage on internal tight corners.

I was also keen to be able to change the amperage from the gcode file for dimpling and marking type operations. I like to keep my working settings in the gcode file as I can never remember the optimum voltage settings that work best for me,


An internal tight corner IS AN ARC. Thats what G64 tolerances do.
You will be able to sense this by the corner lock method.

I think the 7i76e already supports PWM's if you do something like this with hostmot2
num_pwmgens=1 num_stepgens=4
Ref: linuxcnc.org/docs/devel/html/man/man9/hostmot2.9.html

Load it with one and review in halshow to see what pins are published/
Last edit: 16 Apr 2021 21:14 by rodw.

Please Log in or Create an account to join the conversation.

Time to create page: 0.085 seconds
Powered by Kunena Forum