Advanced Search

Use halscope to do a plot of encoder.0.velocity pin during a pierce, might be interference, although as you are getting 55 and 75V readings it also seems not right, and this might be due to scaling and/or due to interference.
What divider ratio is Mesa THCAD set to? Usually 1/32.
If this is a new machine, disable auto volts, disable void lock and velocity lock, well disable everything automated until you get the machine working correctly, only then enable the automation.

besriworld wrote:

Well, the reason is very simple Each additional Beckhoff module is a bag of money for the small user.

You need to consider this before selecting a control method.
I've been scouring eBay for various beckhoff modules and the price is not too bad but they are never in Australia so the wait time sucks!
But AUD $132 for above module landed was not bad v's a THCAD given our appalling exchange rate.

A Beckhoff EL3162 has 2 channels of 0-10 volt input with 500v isolation from the Ethercat bus. This could replace 2 THCAD's without worrying about needing encoder inputs.

I don't understand why people want to mix MESA and Ethercat systems even though it can be done.

Hello 
I have managed to get a few motors running with etherCAT, but I still need a simple I/O system with analog out and encoder input, to use a mesa THCAD.

How is it feasible? Do I need a dual Ethernet card? 
The PCs I’m using have a single Ethernet port, so to run both mesa and ethercat I will need another Ethernet system, to have different NICs, right?

thank you.

Installed the THCAD2 and to no ones surprise, it works!
No evidence of the deed though, sorry.

Gents,

Continuing my learning process of LinuxCNC and Mesa cards i've managed (thank you PCW and tommylight!), to start my permanent magnet DC servomotor driven by BTS7960 H-bridges, all hooked up to Mesa 7i95T on the test bench.
forum.linuxcnc.org/configuration-tools/5...-bridges-pmdc-servos

Did some basic PID tuning and motors are not reaching following errors. Ferror is loose however - fine tuning not required at this stage.
On my learning short list now there is two topics:
1) encoder-ratio module
2) Mesa THCAD card. As I'm powering my motors and limit/home switches from old laptop power supply - its voltage sags easily when loading servos so found it good opportunity to practice THCAD and PyVCP by reading power supply voltage and present it on AXIS. 

Ad 1) While reading LinuxCNC documentation i found an interesting function ENCODER_RATIO. First thought was "hey! let's do some experiment and slave Z axis servo to X.". Let's say the final idea is to link spindle encoder with rotary/dividing head servo - i can imaging hobbing gears like that:)
Experiment conditions:
- there is three servos available X, Y, Z so three encoders, all the same ppr = 400
- X will be the master, Z will be slave
- i want Z to turn 4 times slower than X

If i understand the documentation right i did the following:
- detached joint.2.motor-pos-cmd from pid.z.command
- detached hm2_7i95.0.encoder.02.position from pid.z.feedback and joint.2.motor-pos-fb
- connected pid.z.feedback and joint.2.motor-pos-fb with encoder-ratio.0.error
- decided to use one of the switches for slave axis on/off so detached remote e-stop (will not using it now) input hm2_7i95.0.inmux.00.input-06-not and hooked it up to encoder-ratio.0.enable
- connected master axis (X) encoder signals A and B to encoder-ratio.0.master-A and encoder-ratio.0.master-B
- connected slave axis (Z) encoder signals A and B to encoder-ratio.0.slave-A and encoder-ratio.0.slave-B
- set ratio by stating ppr: encoder-ratio.0.master-ppr 400 and encoder-ratio.0.slave-ppr 100

AXIS started with no issues. Checked HAL configuration and what i can see is:
- encoder-ratio.0.master-A and B signals are successfully connected to X encoder
- encoder-ratio.0.enable is ok - turning TRUE when switch is activated by me

And i failed successfully:)
- i can see no change on pid.z.feedback as well as on pid.z.output when i move X axis - no change on encoder-ratio.0.error 

I thing i'm not getting fully what manula says about encoder_ratio. i couldn't find any HAL examples showing correct configuration.
What i found is that unfortunatelly encoder-ratio.0.master-ppr and encoder-ratio.0.slave-ppr are PARAMETERS, not PINS so no chance to make nice 'change ratio' option in PyVCP (imagine you can input different spindle/slave axis ratio and hob whatever spur gear you want!). The only option i see it close LinuxCNC, edit ppr numbers and re-start LinuxCNC to get ratio i want.

My idea was to have motorized dividing head connected and with just one switch make it move in sync with spindle @ required ratio...So it is basically ELS. Ehmmm those dreams.

Anyway, please! Could you guys look at it and poke me in good direction? I did research and tried by myself before i bother you but, as i said before, failed successfully:)
Thank you.

Short answer:

My Powermax is paired with a THCAD-10.  I changed it to 20:1.  It’s served me well with no issues over the years.

Long answer:

forum.linuxcnc.org/plasmac/39999-thcad-1...powermax45-xp#181612

It depends on your machine amperage. No scaling resistors are rwquired. Most people use 20:1 which gives a 200 volt full scale, 125 amps might be better at 30:1. (300 volts)
The thing is that the THCAD is rated for overvoltage of 500 volts. The voltage will only go over 200 volts on piercing when we do not care about the voltage then. During cutting voltage is between 100-170 volts. Keeping the max voltage range low means the readings are more accurate as more of the ADC's range is used. 50:1 is popular with a lot of non-linuxcnc THC's  (probably with a 0-5 volt ADC). That would be better for a THCAD5.

50:1 worked well you can use the THCAD10 without any modification, 
Some good practices are:
-Place THCAD10 as close as possible to the Plasma power supply.
-Use a shield cable and connect the shield to GND on one side only.
-Use a ferrite bead on the control side to reduce induced noise 
 

Omit the offending line and try again
from
setp hm2_5i25.0.pwmgen.00.pwm_frequency 1000
to
#setp hm2_5i25.0.pwmgen.00.pwm_frequency 1000
---I am pretty sure that is for plasma systems, so the spindle encoder inputs are NOT A/B/Z, but are A0/A1/A2 for 1 to 3 of THCAD.
Anyone, do correct me if that is wrong.
Thank you.

Gents,

I'm slowly preparing items and knowledge to make my plasma cutter based on Hypertherm Powermax 45 XP (just received CPC port upgrade kit). I have a specific conditions - at work for 4 weeks and away from home/garrage and i could only take limited set of toys with me:
- Lenovo T470 with Linuxcnc installed
- Mesa 7i95T
- 2 x THCAD (for THC and ohmic)
- some limit switches
- 3 x H-bridge and 3 x small PM DC servo motors
- bunch of wires
- old laptop power supply to power those servos.

That's all i have to play with LinuxCNC and learn configuration, etc. Ideally i'd like to make those servos live on my bench.
I couldn't take "real" servos and drivers due to weight limitation - my only option was to take those small servos and h-bridge modules. 
If i understand things correctly, out-of-the-box mesa cards are preloaded with firmware containing Step/dir generators and can't be used to control -bridge based servos it this state. So i need to build a specific firmware to replace step gens with pwm generators, right?
I'm stepping here into very dark area for me - creating custom bit file.
I heard that custom bit file are now easier to create with the new FPGA chips on Mesa cards - is it true?
I case of my failure - would it be possible to get some help from you to create custom bitfile? The only source i found is talla83 videos on yt.

Thank you for inputs and have a great weekend.

You could get rid of the short and try adding an isolated 24v power supply (Meanwell hdr-15-24) that is connected to the plasma inputs on the thcad. When cutting, nothing will get damaged. The hypersensing component can see full torch voltage when you are cutting  and it does not care... so this should give you the confidence to do it

You might need to and2 the probe signal so it will only turn on when actually probing.

Here is a quick test where I jumped the shield to the nozzle and then feed voltage from my ohmic into it and then ground the nozzle and the voltage drops like a stone. My thcad-10 is using opto22 and relays to isolate everything so that’s why I have to trigger the ohmic button to send voltage to the nozzle.



youtube.com/shorts/fuJnQ76LXl8?si=pr583jeB8bvqNj0g

Yes, that was one of the original reasons the ohmic test was included. It was intended to check if there was a short like yours before starting a job. It does not happen very often. I only ever experienced it once and I found a bit of rubbish inside the tip,  much like you did. I have thought of setting it up to use one thcad. I'm sure it could be done fairly easily.

Oh ok so you’re not waiting for the voltage to go to zero to trigger the contact. Your looking at any voltage drop and once it does it’s triggering contact. That make perfect sense to use a thcad and watch the frequency output. Then you could set the trip frequency on the encoder input on your mesa board. I wonder if there is a way to double duty a single thcad-10.

You know I stumbled upon something the other day by accident that could work. I was using my fine cut consumables. And I was testing my retraction to make sure my actual retraction was what the display was stating it was. I held the ohmic button down and moved z down till it stopped from the ohmic trip. But I noticed when it touched I seen a +10v on the cut voltage. I did it agian and it was still there. So I started looking around and found a tiny slag bead stuck up in the nozzle bridging the nozzle to the shield. This was back feeding the ohmic voltage into the nozzle and it was showing up on the cut voltage.

Would there be a way to use this to sense ohmic. I don't know the ramifications to running the nozzle bridged to shield but it had been like that for awhile as I had just been cutting.