Advanced Search

7 - IO QUANTITY & TYPE

Once the command signal type is determined, the next step is to make a list of all inputs & outputs (IO) required or desired.  Include all sensors, switches, buttons, relays, etc.; all hardware which needs to signal or be triggered by LCNC should be on the list.
      
    • Inputs to LCNC
        ◦ Voltage supplied (or required) by component
        ◦ Current/amps consumed/supplied by component
        ◦ Analog inputs?  i.e. potentiometer, pressure transducer, etc.
    • Outputs from LCNC
        ◦ Does the component sink or source voltage? Or either?
        ◦ What voltage?
        ◦ Current/amps consumed/supplied by component

Are there lots of IO or just a few?  Are most of the IO for an operator control panel, or lots of relays to drive auxiliary functions (coolant pumps, lube pumps, lights, air cylinder valves, etc.)?  Do your drives require an enable signal?

Note that many newer, but not all, Mesa cards can be configured for both sinking and sourcing inputs & outputs.

Answering the questions above will help you form an idea of how many and what type of IO you need.

ADDITIONAL CONSIDERATIONS
Beyond axis drive command signals and general IO, you should consider the following questions:
    • Spindle - how do you plan to control the spindle(s)?
        ◦ On/off relays with manual speed control?
        ◦ 0-10v speed control?
        ◦ PWM?
        ◦ Analog or Step-dir servo?
    • Encoders - do you plan to add linear encoders for closed-loop feedback to LCNC?
        ◦ Do you have 'dumb' analog servo drives and LCNC will close the loop?
        ◦ Do you need/want a spindle encoder?
    • Operator console - do you plan to add/build a feature-rich (i.e. complicated) operator station with physical controls?  These can eat up IO quickly with lots of buttons, switches, and lights
    • Plasma - not specifically detailed in this guide, but the THCAD series of frequency to voltage cards are widely used for plasma arc sensing.  Additional research on which FPGA cards are most recommended for use with the THCAD would be prudent.
    • Expansion - do you think you might want any/all of the above in the future? Some Mesa FPGA cards are more expandable than others, depending on what it is you want to add in the future.

 

tommylight wrote:

From the link you posted:

CNC Plasma Cutting Machine : have 2 pins and 4 pins connecting port at rear of machine. 2 pins connector for The arc voltage output ratio 1:1, 4 pins connector for Signal Control.Easy works with CNC cutting table.The plasma cutter with Non-HF Pilot Arc function (Not HF Pilot Arc) Under this function, Pilot arc will not interference to machine...
You should really take the time to read stuff and not go by the hunch, that is not a blow back plasma, and it does requite THCAD300 as the voltage output is 1:1 noted above.
Still, I would suggest you get the THCAD5 or THCAD10 or THCAD2 and 2 of 1MOhm resistors, just in case you decide to change the plasma later on.

 

The THCAD-5 and THCAD-10 have been replaced with the THCAD-2

JT

For HV/HF pilot arc THCAD5/10 is the correct option
but what resistors must get in case of around 2mhz/10+ kV HF/HV arc ignition ?
any tips to not burn THCAD ?
Any other alternative options for HF/HV pilot arc for reading cutting voltage except THCAD?

From the link you posted:You should really take the time to read stuff and not go by the hunch, that is not a blow back plasma, and it does requite THCAD300 as the voltage output is 1:1 noted above.
Still, I would suggest you get the THCAD5 or THCAD10 or THCAD2 and 2 of 1MOhm resistors, just in case you decide to change the plasma later on.

Hi there,

So I've been going down quite a few rabbit holes the past few days as I'm going to make my own CNC plasma this spring.  I've gotten to the point where I think I know what I need to order - sort of.

I orginally wanted to do a Pi4, but have decided to do a standard computer setup and that will eliminate me having to make the kernel real-time on a desktop (as long as it's powerful enough, which I can get a pretty good one for $200 on Amazon).  

Now, from what I have figured out, with the 7i76e's being out of stock, it appears as though the replacement is the 7i96S (just with a few less outputs I believe).  

If I go with a non-HF plasma cutter that is THC capable such as this one , do I need to get the THCAD controller Mesa card as well?  If so, since this is a blow-back plasma cutter (non-touch) I'm assuming I need to get the THCAD2 not the THCAD-300?

Thank you for any help you can provide.  I'm going to do this in stages and would like to get the Mesa cards first so I can play around with them a bit before my next paycheque to get the cutter.

This post describes a process for integrating a BCL-AMP capacitive sensor to LinuxCNC. These capacitive sensors are typically used for height sensing in fiber laser cutting systems. The sensors come potted in an aluminum case with a small coax connector to connect to the laser head and a GX16 4-pin circular aviation connector to drive the device. We've been trying to get this working for a while at our hackerspace (sector67.org), and were finally able to get it working.

The amps look like this:

 

There are a number of capacitive sensors with this same basic form factor. This post covers the “BCL-AMP Amplifier Preamplifier Sensor For Friendess BCS100 Height Controller FSCUT1000 FSCUT2000 System” amp. As of the time this article was written, this amps is available for $56 US at www.aliexpress.us/item/3256804585924423.html.

Electrical interface to the amp
The functionality of the GX16 4-pin circular aviation connector on the amp seems to be one of the great mysteries of the Internet. As far as we have been able to find, no data sheet or manual for the amp exists, and the circuit diagrams provided by machine integrators do not provide enough detail to determine how to interface with the amp. As a result, we’ve had to reverse engineer the device to determine how to properly interface with it. The results of that experimentation have led us to the following pin functions, with the pins numbered per the GX16 connector spec:PinFunctionNotes1+5V DC5V DC supplied to the amp2Signal outA ~5V variable frequency pulse train output3GND0V DC supplied to the amp.4SHIELDThis is the one pin that you can find conclusively documented on the Internet. This pin should be tied to the cable shield. Internal to the amp, pins 3 and 4 are tied together to the ground plane, with pin 3 going through an RF choke inside the device.
With this hookup and some capacitive input on the small coax connector you can observe a variable-frequency output on pin 2 of the amp.

Isolating and cleaning up the amp output
The variable frequency output of the amp is unfortunately not a very clean signal. To clean up this signal we used a SN74HC14N schmitt trigger:

 

which creates a much cleaner square wave. VCC is wired to +5V and per the data sheet all of the inputs except for one are tied to ground so they don’t float. The output of pin 2 on the amp is wired to one of the trigger inputs, and the output of the trigger is wired to our Mesa 5i25 hostmot2 encoder A-phase input. The SN74HC14N also provides a modest level of signal isolation.

Capacitive frequency response
The following values were measured in a bench setup using a variable capacitor:CapacitanceFrequency85nF2.50MHz123nF2.00MHz166pF1.67MHz210pF1.54MHz233pF1.44MHz
The response is not linear. The closer the metal is to the laser head (the higher the capacitance), the lower the frequency output is. When there was a short the output frequency drops significantly.

Interfacing to LinuxCNC
Because the amplifier outputs a variable frequency signal, the output of the amp can be treated as an encoder input with the velocity representing the value of the signal. This makes it conceptually similar to a THCAD Mesa device. We are using a Mesa 5i25 board for high speed I/O, and so we can configure a hostmot2 encoder (linuxcnc.org/docs/html/man/man9/hostmot2.9.html#Encoder) via pncconf to read the amp output. Since we are only reading a pulse train and not a quadrature encoder this encoder needs to be in “counter-mode” so that each rising edge of the phase-A input is counted, and the value on phase-B is ignored. 

The relatively high frequencies output from the sensor (up to ~3MHz) create some special considerations for interfacing. The Mesa card’s FPGA is fast enough to read the signal, but for the fastest signals of the amps, the “filter” setting of the hostmot2 encoder should be set to “False” since 15 sample clocks at 25MHz is too long to successfully sample a ~3MHz signal. The non-default (filter = false) 3 clocks should be fast enough, and with the schmitt trigger the signal is clean enough to work well with a three clock transition.

(bit r/w) filter

If set to True (the default), the quadrature counter needs 15 sample clocks to register a change on any of the three input lines (any pulse shorter than this is rejected as noise). If set to False, the quadrature counter needs only 3 clocks to register a change. The default encoder sample clock runs at approximately 25 to 33 MHz but can be changed globally with the sample-frequency or muxed-sample-frequency pin.

Your hal file will need something like the following:

With this configuration in place, you can use Halscope or Halmeter to see the “velocity” response of the encoder change with the distance to the metal. The effective response range is around 0.1 inches.

As of LinuxCNC 2.9.2, pncconf does not enable setting the filter of the encoder, so I’ve created a shell script with the Linux sed tool to fix the generated configuration:

The output of the encoder velocity from a 0.1” Z move can be seen below:

 

Our next steps are to set up plasmaC and use the capacitive sensor encoder's "velocity" value to hold height over the work piece.

the pin you are looking for should be,    arc_voltage_in however not sure if modbus will be fast enought for the plasmac THC control loop,
if you are working in a "conventional"  plasma machine, the recomendation is to use MESA THCAD card to read the voltage. 

Try using the MesaCT wizard to make a config for it:
forum.linuxcnc.org/27-driver-boards/51384-mesa-ct-2-1-1
Download here
github.com/jethornton/mesact/releases
-
Here is a config for 7i95 (not 7i95T although it should work by just adding T in the ini file) for a plasma machine with step/dir and encoder feedback, so if you do not have encoders you might need to omit the corresponding lines.
4 joints, 3 axis, floating switch, Mesa THCAD, so you might also need to change the mode if that is a THC in the picture.

Howdy folks. I've been working to create a workable fiber laser control system using LinuxCNC and eventually plan to use PlasmaC. I have a base machine configuration working, and am at the point where I am ready to integrate a capacitive sensor for torch height control (the on to PlasmaC). The sensor I am integrating is a BCL-AMP style connector that is outputting a variable frequency signal at around 3MHz and around 5V. There seems to be a complete lack of documentation for this capacitive amp, but I've managed to reverse engineer it enough so that it is basically working. But I don't trust it that much unfortunately.

Which leads me to want to ensure I don't inadvertently send more that 5V to the Mesa card. Opto isolators like the PC817 or even H11L1 are too slow. Would a zener diode voltage clamp be fast enough? Any other ideas?

I believe that at 3MHz I'll probably need to set the filer attribute of the hostmot2 encoder to false. I'll also need the encoder in counter mode to provide a velocity value that can basically act like a THCAD for PlasmaC. Any other advice on integrating a signal like this?

Thanks,

Scott

I'm back!
So everything has been going well so far, I have all my motors and end stops wired at the table. Drag chains are installed, slats installed but no water pan yet. I will probably do water pan this spring. I need to now mount my cabinet and my electronics inside and then I will then start to wire up my ohmic sensing THCAD and voltage THCAD. 
My question today is. Can I run my wires in steel conduit for a short run or will this cause any issues?

fmj435 wrote:

By don't ground the HV side do you mean to just leave the star ground off the thcad-2 that I see on some of the wiring diagrams? Thanks for the reply I will give that a try.

I don't know about the star ground but I was thinking frame ground of your control box (which I guess will also be attached to the star ground).
I did not bother running a sepreate lead from the enclosure to the star ground as both were bolted to the table.
I had copper ground points in my enclosure so I used them. Also the mains ground and motor cable shields were attached to these.
Also run seperate ground wires back from the motors to yuor star. This helped to quieten noise. Every bit helps.

I will say the Everlast voltage is quite noisy compared with the Thermal Dynamics I had after it. But Linuxcnc handled it.
 

By don't ground the HV side do you mean to just leave the star ground off the thcad-2 that I see on some of the wiring diagrams? Thanks for the reply I will give that a try.

The Everlast has a voltage divider builtin. Why reinvent the wheel?
The only issue is that with a THCAD, the internal resistors are seen by the thcad so what the THCAD sees is distorted.
Set yout THCAD to use a 10 volt range, no resistor and use the 16:1 range on the everlast.
The divider seen will be 24:1 (not 16:1). This will give you a full scale reading of 240 volts.
Just calibrate your software to 24:1 instead of 16:1
Do not ground the HV side of the thcad as that will allow a mains 50/60 Hz ripple show up.
You must use a RFI filter on the control panel mains power entry. 

Ah, that makes a lot more sense on why I was confused. Thanks for clearing that up.

The 2.4M resistor should be on the negative input (the plasma voltage)
Ground is positive on plasma machines.

The division ratio can be specified in the setup (/32 suggested)