Advanced Search

Hi Andy,

So I did tie up all the unused inputs to zero, and there was definitely an improvement. Another guy I talked too also suggested that I install a first order low pass filter on the arc voltage input from the plasma cutter, so I will try that as well to see if I can get and improvement.

---

Thanks to all for helping me.

glad you made it work.
Tom

groblerjf@gmail.com wrote:

the analog inputs is susceptible to interference from each other.

The problem you are seeing might be due to the capacitance of the multiplexer. It might help to tie all unused analogue inputs to 0V.

Hi Tom ; Andy,

I finally got my build to work properly. I do not know whether my solution is the correct method, but it works. I debounced the Arc OK signal with a 100nF cap and that seems to do the trick.

However, I will definitely be ordering a Proma THC or other, because the little Arduino seems to be too problematic and the analog inputs is susceptible to interference from each other.

Regards
Jacob

Ok, so the THC that I am using is my own that I built with an Arduino. Locally I cannot buy a THC so the next best thing was a micro-controller.

According to the Hypertherm manual, I need an arc voltage of approx 112V for a 1.6 mm plate. As far as I can tell, the THC is working correctly, because I can see that LinuxCNC receives the move up or move down signals.

O, so I sort of understand why I should rather cut thicker metals, but I want/need to cut the thinner material.

Regards
Jacob

The slow motion is a result of the ARCOK signal going on and off . Set the voltage on the THC correctly and try cutting a thicker metal part so the voltage at the torch is stable for longer periods of time.
What type of THC are you using?
The normal voltage on most of the machines i have and use while cutting 3-6mm thick plates is 113-128 V.
Tom

groblerjf@gmail.com wrote:

There is a new problem however. For some reason now, as soon as the arc establishes, and motion begins, the machine moves extremely slow and I have no idea why. Everything works fine, even my own implementation of THC, and normal motion is at the correct speed, but when I issue the M3 command, the motion is extremely slow.

It is slightly hard to work out what is happening without the HAL file, but I notice that you enable adaptive feed in the INI, so my assumption would be that you need to calibrate the input to the adaptive feed pin.

Hi, so I did not wrap the wire in Al foil, but I did add a shielded cable, not sure whether it helped, may have installed it wrong. What I did do is debounce the float switch electrically with a 100nF across the switch, I went one step further however and installed and opto-coupler. I also rewired the switch to be normally open.

There is a new problem however. For some reason now, as soon as the arc establishes, and motion begins, the machine moves extremely slow and I have no idea why. Everything works fine, even my own implementation of THC, and normal motion is at the correct speed, but when I issue the M3 command, the motion is extremely slow.

Attached is my thc_toma.ini file

Any ideas???

Regards
Jacob

Another thing that can help is debouncing the input signals with the hal debounce component

Thanks, will do it in the morning.

Regards
Jacob

Do a test an wrap the limit switch with alu foil properly, afetr you change the wire. If the switch is close to the torch that will help a lot.

It is absolutely necesary to use shielded cables in your case, you can use even RG 58 or RG59 tv antena or video camera or audio cable with one wire in the midle and one as a shield. Just connect the shield to the ground on your bob and the midle wire to the input.

I also forgot to mention one of the common errors i have seen around : stop grounding everything at the same point. That is o.k. for other machines, but it can be fatal on plasma cutters. I have explained that in this forum and some others so i won't repeat myself, just to prove a point : almost all power supplyes have ground connected to minus terminal internaly as it is supoesed to be but the plasma cutter clamp that is usualy clamped to the material or the ribs on the cnc is the positive.
Tom

Ok, so the system I am running is a Powermax 45 with machine torch.

So I implemented the solution you suggested Tom.

- I rewired the limit switch so that it is normally closed. I installed a pull up resistor on the 5 inputs.
- The plasma cutter is as far as possible fom the PC.
- The cutting torch cable and the ground cable of the plasma cutter is on the opposite side of the control cables as well.
- The only cable that runs next to the other cables is the pc interface cable that is connected to the plasma power supply.


I have not yet tried the online diagnostics tool as Andy suggested as my PC is not connected to the internet. I did however discover that it is indeed the float switch that is causing the problem. I used a standard mill config and just arced the torch. If a z limit is defined during configuration then an Estop is triggered, but when that same pin is set as unused, then the system arc's and I can cut further.

Any suggestion on how to fix this.

My z float is just two wires dangling next to the machine torch, should I perhaps just install shielded wire and see if that does not fix the problem.




Thanks
Jacob

Hi

The problem you describe is exlusively related to High Voltage/High Frequency start systems in plasma cutters, so Andy was right regarding the electrical problem.

There are solutions for this as noted below:

Cabling

---

the torch and ground cable from plasma should not be parallel with other cabling -NEVER EVER- exept when using high quaity RF shielded cabling.

The limit switch used for finding the work piece

---

it has to be configured and connected as NC or normaly closed. This for a simple reason: it stays always connected to ground exept when triggered by touching the work piece so when the torch starts it is at ground potencial and way less prone to RF interference.

Plasma cutter

---

as far as posible from everything else. This is limited by the length of the torch cable, but having the computer and electronics on one side of the machine and the plasma cutter on the other helps a lot.

BOB

---

This is more complicated to fix and requires some soldering skills and 5 220 - 330 ohm resistors. The resistors have to be soldered to 5 inputs on the BOB and to +5V so it keps them pulled up all the time until there is a signal on the inputs.

That is it.

I have done this several times ( over 15 ) on a wide range of CNC plasma cutters and this is the exact procedure that i use and it always works, always.
There are also High Voltage start plasma cutters and they cause much less troubles, and scratch start, but the latter is not usable for cnc.

regards
Tom

OK., that is quite a config isn't it?

I would guess there is something funny going on with the inputs to estop-latch.

Perhaps it would help to get a graphical view of the HAL wiring?

wiki.linuxcnc.org/cgi-bin/wiki.pl?Rockhopper_Web_Server

(I have never used that, but I have heard it can help)