584,830 active members*
5,841 visitors online*
Register for free
Login
Results 1 to 10 of 10
  1. #1
    Join Date
    Jun 2011
    Posts
    49

    Plasma voltage drops when cutting

    I built a CNC plasma table several years ago and have been using it as/when required. It's a pretty basic setup, Mach3 on a Win XP Dell PC with an added PCI parallel port, 2 x BOB's, 4 x HaoyuTB6600 drivers, 270oz-in steppers, T5 belt drive on X & Y axis and leadscrew on Z. I have a 50mm deep water tray with steel slats in it. My water tray is sitting on a sheet of MDF and it's electrically isolated from the table frame and the electrical enclosure. Whenever I'm cutting, I always connect the plasma earth lead directly to the steel being cut and everything works well, good quality cuts, no electrical interference, etc.

    Some time later, I decided to install a THC and built an open source Arduino THC from here: https://github.com/regeg/ArdunioTHC

    This THC is designed for use with the Everlast Power Plasma 50 which has a CNC port that gives you full arc voltage via an internal 100K resistor. The THC has a pin you can use to monitor this voltage, and when doing the initial testing, I saw that the voltage drops quite low immediately after cutting begins. It starts at about 130v and drops down to around 30v shortly after the pierce. This causes the THC to lift the torch and obviously the arc goes out.

    I took the plasma cutter off the table and used it manually to cut some steel. I used my multimeter to check the DC volts at the CNC connector while cutting. All was good, it cut perfectly and the voltage remained around 120v during the cut. This tells me there is nothing wrong with the plasma cutter.

    I then installed it back on the table, but completely removed the THC. Again I used my multimeter to check the DC volts at the CNC connector while cutting, and now the voltage dropped just like it did earlier. The cuts were still great but happened at about 30v. This tells me the THC is not causing this problem.

    I then discovered that if I elevate the piece of steel that I'm cutting off the slats by about 40mm, then the voltage does not drop, the THC works perfectly and controls the Z axis to +/- 1 volt. If however a cut piece only partially falls down and makes contact with the slats, the plasma voltage instantly drops.

    The interesting thing is that my slats and water tray are sitting on a sheet of MDF and are electrically isolated from the table frame and the electrical enclosure. My electrical enclosure is metal, and everything inside is earthed to one spot. The power lead coming in is also earthed to the same spot (obviously not the live or neutral wires, just the earth). I have an earth cable that runs from that spot and connects to my gantry, then to my Y car and even my Z axis plate. I have tried removing that earth cable but it made no difference.

    I wondered if the water tray being electrically isolated from the table frame and the electrical enclosure was somehow causing this problem, so I tried hooking up an earth strap from the slats to the earth spot inside the enclosure - this made no difference at all.

    As a last resort, I drove a copper rod deep into the ground beside the table and ran a thick (car battery type cable) to the slats, and also to the electrical enclosure, but this also made no difference.

    Does anyone have any suggestions?
    Eug

  2. #2
    Join Date
    Jun 2011
    Posts
    49

    Re: Plasma voltage drops when cutting

    I thought I covered everything but remembered to add that my compressed air is very dry and using worn or new consumables makes no difference.
    Eug

  3. #3
    Join Date
    May 2015
    Posts
    684

    Re: Plasma voltage drops when cutting

    There is something strange here. Here is a plot (green line) of the Everlast 50 cutting with LinuxCNC at about 95 volts after the initial surge which is about your 130 volts.



    It sounds like there is a short or something acting as a voltage divider as it seems you have eliminated your hardware. But it does make you wonder why its cutting well at 30 volts.

    The thing is that the voltage read is of no interest to us. All we want is a constant height. I would explore seeing if you can sample the volts with the Arduino after say 1.5 seconds and use that reading as your setpoint. I found 1.5secs was enough for any thickness material to settle to a stable height. If thats not long enough for it to settle, extend the time delay. So what if the setpoint is 30 volts?

    I can't help much further as I've sold my Everlast and replaced it with a 120 amp Thermodynamics machine.
    Rod Webster
    www.vehiclemods.net.au

  4. #4
    Join Date
    Jun 2011
    Posts
    49

    Re: Plasma voltage drops when cutting

    Thanks for the reply Rod.

    It's definitely not the Everlast. I have 2 of them and they both do it. I also have a Hypertherm Powermax 45 and I have been using that lately, but the voltage behaves the same way.

    It's never exactly 30v either, it just drops very low. I have a small program that captures the data during cutting and it shows me the voltage in milliseconds. I'll have to see if I can upload it so you can see. The THC is working perfectly and is issuing up/down commands depending on the actual voltage vs set voltage.

    It's doing my head in as I can't think of anything else to check.
    Eug

  5. #5
    Join Date
    May 2015
    Posts
    684

    Re: Plasma voltage drops when cutting

    I never said it was the Everlast (even if I am pleased to be rid of mine). You seem to have confirmed its not. Does the Arduino sample the torch voltage to establish a voltage setpoint? Does the THC keep the torch height constant even at the lowered volts?

    Here is a plot from today. on 8mm mild steel, 80 amps, 1745mm/min, 120 volt book arc volts, 120.45 volts sampled, Linuxcnc, THCAD10 voltage to frequency interface to encoder input, 30:1 divider native on the Thermodynamics.

    The blue line is the error between the setpoint and the actual volts,starting 1.5sec after ArcOK (red line).

    My point is that if your setpoint starts on the line, you don't really care about the volts. It won't matter if its 3 volts, 30 volts or 3000 volts.





    I have to say that whilst both Everlast and ThermalDynamics have incorrect documentation on the CNC interface port signals, the machine I have now has a much more sophisticated voltage divider board and a much cleaner signal.
    Rod Webster
    www.vehiclemods.net.au

  6. #6
    Join Date
    Jun 2011
    Posts
    49

    Re: Plasma voltage drops when cutting

    Quote Originally Posted by rodw View Post
    Does the Arduino sample the torch voltage to establish a voltage setpoint? Does the THC keep the torch height constant even at the lowered volts?
    I set the voltage using the App that came with the THC (or I can do it on the modified Mach3 screen). The arduino samples the voltage and issues up or down commands as required. Below is a screenshot of the captured data and you can see my setpoint is 100v and the current voltage is 36v so the arduino is issuing THC up command constantly. I can "cheat" and continue to cut this way by setting a THC Max of 2mm in Mach3, so the torch can only go so high, but it's far from ideal if the metal is warped.
    Attachment 411998

    And here's the complete file - if you import it into Excel it is easier to read. This is just one random file of many that I captured when I was testing the unit. In this file, the voltage never got to 120, it got as high as 39 (I'm assuming during the pierce) and then settled on 35-36v for the remainder of the cut.
    Eug

  7. #7
    Join Date
    Jun 2011
    Posts
    49

    Re: Plasma voltage drops when cutting

    Oh, I wanted to mention it looks like you've got your table sorted with Linuxcnc. I wanted to go down that same path and recall following your thread on the Linuxcnc forum a while back. I got as far as setting up a PC with Linuxcnc, but while going through the configuration, iirc I had trouble getting my slave X axis to home separately to the X master to square the gantry up. From memory that was one function Linuxcnc didn't support so I gave up there and then.
    Eug

  8. #8
    Join Date
    May 2015
    Posts
    684

    Re: Plasma voltage drops when cutting

    Quote Originally Posted by xyz-cnc View Post
    I set the voltage using the App that came with the THC (or I can do it on the modified Mach3 screen). The arduino samples the voltage and issues up or down commands as required. Below is a screenshot of the captured data and you can see my setpoint is 100v and the current voltage is 36v so the arduino is issuing THC up command constantly. I can "cheat" and continue to cut this way by setting a THC Max of 2mm in Mach3, so the torch can only go so high, but it's far from ideal if the metal is warped.


    And here's the complete file - if you import it into Excel it is easier to read. This is just one random file of many that I captured when I was testing the unit. In this file, the voltage never got to 120, it got as high as 39 (I'm assuming during the pierce) and then settled on 35-36v for the remainder of the cut.
    In my testing and regression analysis of 9000 readings logged every millisecond from 1.5mm to 2.2mm cut height (from memory), the Everlast has a very linear relationship between torch volts and height. In fact 7.53 volts per mm with a 99.5% confidence level. I guess the next question is how do you know the Arduino is accurately sampling the volts? I know you say you do not use the divider but its resistor network was being counted as being part of the divider on my hardware. Unless my hardware was disconnected, the multimeter showed the same incorrect readings as the THCAD10 at the control box end.

    I ended up trouble shooting with 2 x 48 volt power supplies in series (96 volts) attached to the raw arc input on the Everlast divider board so I had a constant known input and connected two multimeters to the system to monitor what was going on.
    Rod Webster
    www.vehiclemods.net.au

  9. #9
    Join Date
    May 2015
    Posts
    684

    Re: Plasma voltage drops when cutting

    Quote Originally Posted by xyz-cnc View Post
    Oh, I wanted to mention it looks like you've got your table sorted with Linuxcnc. I wanted to go down that same path and recall following your thread on the Linuxcnc forum a while back. I got as far as setting up a PC with Linuxcnc, but while going through the configuration, iirc I had trouble getting my slave X axis to home separately to the X master to square the gantry up. From memory that was one function Linuxcnc didn't support so I gave up there and then.
    Linuxcnc's development branch (2.8 master branch) has supported gantry homing since before I started my build. It separates motors (joints) from axes so you can join 2 joints together on the gantry which are aligned during the homing process. To square it, you just need to change one of the homing offsets. I think I have 45mm offset one side and 47mm on the other. All it takes is a few lines in the ini file.

    I will say my journey has been a long one, interspersed with long holidays from the project and many problems due to my lack of knowledge and issues with the Everlast. Trace the wires on the divider board to the ArcOK relay and find its spec sheet. Make sure your hardware draws enough current to exceed its minimum spec or it will become intermittent due to contact oxidisation on the internal relay. If you are too low, add a pull down resistor or something to bring it back into spec.

    I decided to make one more attempt to get the experimental external offsets branch to work for torch height control once I upgraded my inverter and finally got a result. But then had air quality issues so spent another $3k fixing it!

    So now if I get my post processor for sheetcam updated tonight, I think I will be able to walk up to it and cut any part.... without touching any settings.
    Rod Webster
    www.vehiclemods.net.au

  10. #10
    Join Date
    Jun 2011
    Posts
    49

    Re: Plasma voltage drops when cutting

    To be honest, I'm not worried about the Arduino's accuracy sampling the volts. As I said in my original post, if I elevate the steel off the slats, the arduino will reliably keep the torch within +/- 1 volt, which is plenty accurate enough for me.

    It will do it on the Everlast, and since I modified my HP45 to give me raw arc volts, it works equally as well on that also.
    Eug

Similar Threads

  1. Reading Plasma Voltage
    By j5fanclub in forum CNC Plasma / Oxy Fuel Cutting Machines
    Replies: 9
    Last Post: 02-13-2020, 09:09 PM
  2. Plasma torch On OFF automatic solution and voltage
    By eds_miller in forum Waterjet General Topics
    Replies: 1
    Last Post: 01-19-2017, 12:14 AM
  3. plasma THC voltage
    By duraflap in forum Waterjet General Topics
    Replies: 19
    Last Post: 11-28-2016, 04:11 AM
  4. Replies: 7
    Last Post: 06-24-2014, 07:25 AM

Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •