- User Interfaces
- Plasmac
- qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
26 Feb 2023 12:31 #265413
by rodw
Replied by rodw on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Its not as simple as that when the torch is getting contaminated and throwing errors when cutting as the OP described.
Please Log in or Create an account to join the conversation.
- snowgoer540
-
- Offline
- Moderator
-
Less
More
- Posts: 2388
- Thank you received: 779
26 Feb 2023 16:23 #265422
by snowgoer540
Are you sure the slag/dust is causing this? Have you cleaned the torch and verified the voltage goes back to 0v? I am not sure what plasma cutter brand you are using, but in my experience it's either been shorted or it wasnt.
I ask the above question because this is exactly what I saw with my setup, but it had nothing to do with the cleanliness of the tip. The longer I cut, the higher the ohmic voltage remained after each cut was over. Eventually the ohmic voltage would just remain at near max rendering it useless. It's been quite some time, but I seem to remember that either unplugging the torch from the plasma power supply, or removing the ground clamp from the plasma power supply made it go away. I'm not sure why this occurred, but it occurred on more than 1 table. Worth noting that the plasma power sources were Hypertherm in these cases.
This is why the diodes and the relay were critical to the revised THCAD circuit I linked.
Worth nothing that the result of the testing was very similar for both the THCAD and Relay ohmic circuits linked previously. Advantages the relay circuit has are 1. cheaper 2. less complex to implement (no component install required).
Yes, assuming your hal snippets/files are what you are actually running on your table, then you are actually using Mode 0 (the default mode in the component) and not the moving average (Mode 1) Rod has been talking about.
The specific reason is that there is a bug in the code for Mode 0 (line 3 below):
The explanation of the bug is: (ohmic_threshold && is_probing) is evaluated first. Since ohmic_threshold is almost always a number above zero, and is_probing = 1 (per your screen shot), the expression simplifies to ohmic_on = (ohmic_volts > 1 ? 1: 0). So, any voltage above 1 will cause it to trip.
At minimum a fix would be:
Replied by snowgoer540 on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Below is a screenshot of an ohmic test with a slightly shorted torch. The torch is off of the table/material as noted by the 20ish volts of arc.
Are you sure the slag/dust is causing this? Have you cleaned the torch and verified the voltage goes back to 0v? I am not sure what plasma cutter brand you are using, but in my experience it's either been shorted or it wasnt.
I am seeing a ohmic voltage of 13ish...... should be zero. not sure why this is, may play around with resistor values to see if it changes. Turning on or off the plasma doesnt change the values.
I ask the above question because this is exactly what I saw with my setup, but it had nothing to do with the cleanliness of the tip. The longer I cut, the higher the ohmic voltage remained after each cut was over. Eventually the ohmic voltage would just remain at near max rendering it useless. It's been quite some time, but I seem to remember that either unplugging the torch from the plasma power supply, or removing the ground clamp from the plasma power supply made it go away. I'm not sure why this occurred, but it occurred on more than 1 table. Worth noting that the plasma power sources were Hypertherm in these cases.
This is why the diodes and the relay were critical to the revised THCAD circuit I linked.
Worth nothing that the result of the testing was very similar for both the THCAD and Relay ohmic circuits linked previously. Advantages the relay circuit has are 1. cheaper 2. less complex to implement (no component install required).
regardless it shouldnt be tripping the ohmic at 13 volts but it is....... any idea why that could be?
Yes, assuming your hal snippets/files are what you are actually running on your table, then you are actually using Mode 0 (the default mode in the component) and not the moving average (Mode 1) Rod has been talking about.
The specific reason is that there is a bug in the code for Mode 0 (line 3 below):
case 0:
if(is_probing)
ohmic_on = (ohmic_volts > (ohmic_threshold && is_probing) ? 1 : 0);
else
if( ohmic_volts < ohmic_low)
ohmic_on = 0;
break;The explanation of the bug is: (ohmic_threshold && is_probing) is evaluated first. Since ohmic_threshold is almost always a number above zero, and is_probing = 1 (per your screen shot), the expression simplifies to ohmic_on = (ohmic_volts > 1 ? 1: 0). So, any voltage above 1 will cause it to trip.
At minimum a fix would be:
case 0:
if(is_probing)
ohmic_on = (ohmic_volts > ohmic_threshold ? 1 : 0);
else
if( ohmic_volts < ohmic_low)
ohmic_on = 0;
break;Please Log in or Create an account to join the conversation.
- snowgoer540
-
- Offline
- Moderator
-
Less
More
- Posts: 2388
- Thank you received: 779
26 Feb 2023 16:26 #265423
by snowgoer540
Replied by snowgoer540 on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Also meant to note: In your screen shot your material setting in the top left corner of the preview has a bunch of period "noise" (. . . . . . . . . . . ).
This was fixed recently, so if you update it should get rid of that
This was fixed recently, so if you update it should get rid of that
Please Log in or Create an account to join the conversation.
26 Feb 2023 20:25 #265437
by rodw
Replied by rodw on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Yeh, thats a bug.
assuming you are using the published ohmic.comp component, download this file
github.com/LinuxCNC/linuxcnc/blob/master...omponents/ohmic.comp
edit line 93 so it says
then
sudo halcompille --install ohmic.comp
drop the sudo if its run in place
you may need to install linuxcnc-uspace-dev to get halcompile
I'll have a look at ohmic3.comp code.
assuming you are using the published ohmic.comp component, download this file
github.com/LinuxCNC/linuxcnc/blob/master...omponents/ohmic.comp
edit line 93 so it says
ohmic_on = (ohmic_volts > ohmic_threshold) ? 1 : 0;then
sudo halcompille --install ohmic.comp
drop the sudo if its run in place
you may need to install linuxcnc-uspace-dev to get halcompile
I'll have a look at ohmic3.comp code.
Please Log in or Create an account to join the conversation.
26 Feb 2023 20:31 #265438
by rodw
Replied by rodw on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
The bug also exists in ohmic3 but if you use the hal file I supplied on this thread it sets mode=1 so it will never be encountered.
Please Log in or Create an account to join the conversation.
- snowgoer540
-
- Offline
- Moderator
-
Less
More
- Posts: 2388
- Thank you received: 779
26 Feb 2023 20:42 - 26 Feb 2023 20:43 #265441
by snowgoer540
To save you the looking, Rod's referring to this line in his config which would be necessary if you want to use mode 1:
Rod's config also has a line that changes the default number of readings from 5 to 10 (unsure if that's necessary or not):
Replied by snowgoer540 on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
The bug also exists in ohmic3 but if you use the hal file I supplied on this thread it sets mode=1 so it will never be encountered.
To save you the looking, Rod's referring to this line in his config which would be necessary if you want to use mode 1:
setp ohmicsense.mode 1Rod's config also has a line that changes the default number of readings from 5 to 10 (unsure if that's necessary or not):
setp ohmicsense.num-readings 10
Last edit: 26 Feb 2023 20:43 by snowgoer540.
Please Log in or Create an account to join the conversation.
26 Feb 2023 21:08 #265443
by jimmyrig
Replied by jimmyrig on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
I've got my ohmic low set to 23 (the screenshot i posted has hal show with all my settings)The issue some had previously was on water tables and water got in the torch. In these circumstances, the voltage variation between on and off was closer to 0.5 to 0.2 volts. At the same time PCW suggested the 24k resistor to extend the voltage hysterisis.
Ohmic3 incldes a moving average algorithm so instead of using a simple low voltage threshold to sense an off condition (ohmic-low), it delares an off codition as soon as it sees a confirmed downward trend in voltage.
If your off is 13 volts, you might like to consider changing ohmic-low to say 18-20 volts.
This is where ohmic sensing gives you flexibility vs a relay. You have flaxibility to alter the on/off hysteresis to cover differing conditions. A relay might turn off at between 5 & 8 volts which would not work for your table.
Please Log in or Create an account to join the conversation.
26 Feb 2023 21:12 #265444
by rodw
Replied by rodw on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
num-readings sets how many readings are used for the moving average. It could be 100 if you like (0.1 second) but that might be too long.
I settled on 10 because I wanted the system to still be responsive. It seemed to work well.
Its a bit hard to explain but the thcad is sensitive enough to see a change in voltage as contact is made or broken. A moving average will smooth data and lag behind the raw value. When contact is first made, there should be 9 readings near 0 volts and one at 24 volts so the moving average is quite low (eg. 2.4 volts). If superimposed over the raw signal in a graph, it will look a bit like a diving board sticking out when contact is made and actual voltage rises above the board.
When QTplasmac notices contact, it slowly starts to back away from the material. Once the raw voltage crosses below the moving average line, an impending break in contact is confirmed. This could be well above the ohmic-low threshold which acts as a fail safe level. The intent was to sense probing activity with a wet torch on a water table to see the 0.2 volt difference between on and off voltage some people were experiencing in a water table due to conductivity through water in the torch. the 24 k resistor was designed (by PCW) to widen this votage hysteresis.
It has been bulletproof on my table for several years.
I settled on 10 because I wanted the system to still be responsive. It seemed to work well.
Its a bit hard to explain but the thcad is sensitive enough to see a change in voltage as contact is made or broken. A moving average will smooth data and lag behind the raw value. When contact is first made, there should be 9 readings near 0 volts and one at 24 volts so the moving average is quite low (eg. 2.4 volts). If superimposed over the raw signal in a graph, it will look a bit like a diving board sticking out when contact is made and actual voltage rises above the board.
When QTplasmac notices contact, it slowly starts to back away from the material. Once the raw voltage crosses below the moving average line, an impending break in contact is confirmed. This could be well above the ohmic-low threshold which acts as a fail safe level. The intent was to sense probing activity with a wet torch on a water table to see the 0.2 volt difference between on and off voltage some people were experiencing in a water table due to conductivity through water in the torch. the 24 k resistor was designed (by PCW) to widen this votage hysteresis.
It has been bulletproof on my table for several years.
Please Log in or Create an account to join the conversation.
26 Feb 2023 21:24 #265445
by jimmyrig
Attached are screenshots of hal show before and after cleaning when ohmic test is clicked. I was going to take a picture of the nozzle but the chunk of slag fell out.... will try again in a few.
This was after first powerup with the plasma, pc and electronics box all shut down for 12 hours. the values were the same as before shutdown last night. So this particular problem isnt a hysterisis problem.
Thanks for the code working on that now and will give it a shot!
Replied by jimmyrig on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Yes, 100% sure. That doesnt mean I wont run into the issue you were having as they might be independent. ...so far i have been only able to run for about 5 min before a failure, each time the tip gets a small chunk of slag in it which causes the shutdown.
Are you sure the slag/dust is causing this? Have you cleaned the torch and verified the voltage goes back to 0v? I am not sure what plasma cutter brand you are using, but in my experience it's either been shorted or it wasnt.
Attached are screenshots of hal show before and after cleaning when ohmic test is clicked. I was going to take a picture of the nozzle but the chunk of slag fell out.... will try again in a few.
This was after first powerup with the plasma, pc and electronics box all shut down for 12 hours. the values were the same as before shutdown last night. So this particular problem isnt a hysterisis problem.
Thanks for the code working on that now and will give it a shot!
Please Log in or Create an account to join the conversation.
- tommylight
-
- Away
- Moderator
-
Less
More
- Posts: 19188
- Thank you received: 6430
26 Feb 2023 23:04 #265450
by tommylight
Replied by tommylight on topic qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Any particular reason for not using the float switch?
Please Log in or Create an account to join the conversation.
Moderators: snowgoer540
- User Interfaces
- Plasmac
- qtplasmac 2.10 arc ok, thcad thc, and Thcad Ohmic troubleshooting help needed.
Time to create page: 0.260 seconds