Hypersensing with THCAD - better way to do ohmic sensing
18 Apr 2020 04:46 #164398
by phillc54
Replied by phillc54 on topic Hypersensing with THCAD - better way to do ohmic sensing
You do need to be aware that this will cause an increase in overrun.
Please Log in or Create an account to join the conversation.
18 Apr 2020 04:51 #164400
by rodw
Replied by rodw on topic Hypersensing with THCAD - better way to do ohmic sensing
Yeh I know. Accuracy is determined when we turn off the signal.hence the idea to track the trend. I still have to think it through a bit but if it is trending down, we might be more responsive than say a config that turns on at 22 volts and off at 1 volt which is what I had at one stage. (eg. The hysteresis consumes servo cycles anyway.)
Please Log in or Create an account to join the conversation.
18 Apr 2020 05:03 #164402
by phillc54
Replied by phillc54 on topic Hypersensing with THCAD - better way to do ohmic sensing
I think another variable will be that as you do more cutting there may be more suspended metal particles in the water and that may change the resistance.
Please Log in or Create an account to join the conversation.
18 Apr 2020 05:28 #164406
by rodw
Possibly but the metal should sink to the bottom and the commercial additives should buffer pH etc to a fairly constant value. But it was always acknowledged that different environments would probably need different thresholds. I seem to remember that Islander261 got you to add the pulse function so he could get a preflow of air before his first pierce to combat this. I personally think I would have this problem with a relay config as the 24 volt relays I was using tested out as being 8 volts on and 5 volts off so the probe would stay on anyway. I did notice earlier before I understood the issue that the second probe would often fail unless the air was flowing...
At least now we can compensate for the issue in software... With the right algorithm...
Replied by rodw on topic Hypersensing with THCAD - better way to do ohmic sensing
I think another variable will be that as you do more cutting there may be more suspended metal particles in the water and that may change the resistance.
Possibly but the metal should sink to the bottom and the commercial additives should buffer pH etc to a fairly constant value. But it was always acknowledged that different environments would probably need different thresholds. I seem to remember that Islander261 got you to add the pulse function so he could get a preflow of air before his first pierce to combat this. I personally think I would have this problem with a relay config as the 24 volt relays I was using tested out as being 8 volts on and 5 volts off so the probe would stay on anyway. I did notice earlier before I understood the issue that the second probe would often fail unless the air was flowing...
At least now we can compensate for the issue in software... With the right algorithm...
Please Log in or Create an account to join the conversation.
18 Apr 2020 06:04 #164413
by phillc54
I was just doing some googling and saw this way around the issue:
Replied by phillc54 on topic Hypersensing with THCAD - better way to do ohmic sensing
So then when probing the voltage must exceed 23.8 volts and be trending upwards for 3 servo cycles before turning on.
And then when it fell below that threshold and was trending down for 3 consecutive servo cycles we turn it off.
Well 6 servo cycles (maybe more depending on how many cycles the voltage remains static) increases the overrun 6 times and without knowing how many servo cycles the second config consumes to cover the hysteresis it seems to be a bit of guesswork to determine which way is more responsive.Yeh I know. Accuracy is determined when we turn off the signal.hence the idea to track the trend. I still have to think it through a bit but if it is trending down, we might be more responsive than say a config that turns on at 22 volts and off at 1 volt which is what I had at one stage. (eg. The hysteresis consumes servo cycles anyway.)
I was just doing some googling and saw this way around the issue:
Please Log in or Create an account to join the conversation.
18 Apr 2020 06:47 #164417
by rodw
Replied by rodw on topic Hypersensing with THCAD - better way to do ohmic sensing
I don't like the look of that. There is still a lot of dust in the air compared with a water table so it does not seem to be a downdraft table. The bigger thermal dynamics machines do a water mist cut mode.To be clear, this is purely an issue for water tables and its not THCAD specific. Building a downdraft table would also solve it but I found that the water table gives less dross since I added it.
On a dry table it takes about 0.04mm of movement on contact to gain a full 24 volt signal on my testing so any lag from proposed overtravel is not significant at all. How many servo cycles does it take plasmac to back up say .03mm? (say from 24 volt back to 5 volt to turn off a relay?)
On a dry table it takes about 0.04mm of movement on contact to gain a full 24 volt signal on my testing so any lag from proposed overtravel is not significant at all. How many servo cycles does it take plasmac to back up say .03mm? (say from 24 volt back to 5 volt to turn off a relay?)
Please Log in or Create an account to join the conversation.
18 Apr 2020 06:47 #164418
by rodw
Replied by rodw on topic Hypersensing with THCAD - better way to do ohmic sensing
Oops forgot a link www.plasmaspider.com/viewtopic.php?t=29460
Please Log in or Create an account to join the conversation.
18 Apr 2020 06:54 - 18 Apr 2020 06:55 #164419
by phillc54
On the test component I posted for you it is 0.01 per cycle for Probe Up if sensed before Probe Height then 0.001 per cycle for the second Probe Up.
Replied by phillc54 on topic Hypersensing with THCAD - better way to do ohmic sensing
It wasn't a suggestion, I think it is cute, cnc drag cutting.I don't like the look of that. There is still a lot of dust in the air compared with a water table so it does not seem to be a downdraft table. The bigger thermal dynamics machines do a water mist cut mode.To be clear, this is purely an issue for water tables and its not THCAD specific. Building a downdraft table would also solve it but I found that the water table gives less dross since I added it.
PlasmaC probe up is 0.001mm per servo cycle, so that would be 30 cycles.On a dry table it takes about 0.04mm of movement on contact to gain a full 24 volt signal on my testing so any lag from proposed overtravel is not significant at all. How many servo cycles does it take plasmac to back up say .03mm? (say from 24 volt back to 5 volt to turn off a relay?)
On the test component I posted for you it is 0.01 per cycle for Probe Up if sensed before Probe Height then 0.001 per cycle for the second Probe Up.
Last edit: 18 Apr 2020 06:55 by phillc54.
Please Log in or Create an account to join the conversation.
18 Apr 2020 07:03 #164421
by phillc54
Replied by phillc54 on topic Hypersensing with THCAD - better way to do ohmic sensing
Would it be worthwhile to add something like:
If Ohmic is enabled in the Run Panel then before the first cut in a job pulse the torch before probing commences to get the air flowing.
This may need to be an ini file option as I am not sure that it would be applicable for all machines.
If Ohmic is enabled in the Run Panel then before the first cut in a job pulse the torch before probing commences to get the air flowing.
This may need to be an ini file option as I am not sure that it would be applicable for all machines.
Please Log in or Create an account to join the conversation.
18 Apr 2020 07:12 #164423
by rodw
Replied by rodw on topic Hypersensing with THCAD - better way to do ohmic sensing
PlasmaC probe up is 0.001mm per servo cycle, so that would be 30 cycles.
On the test component I posted for you it is 0.01 per cycle for Probe Up if sensed before Probe Height then 0.001 per cycle for the second Probe Up.
SO we are going to get more efficient at this. Those 30 cycles are currently inherent in most relay configs used by Plasmac that are in use.
When I'm ready to tackle that test, I'll let you know as I'm sure I've lost the link..
Please Log in or Create an account to join the conversation.
Moderators: snowgoer540
Time to create page: 0.166 seconds