Hypersensing questions

rodw wrote:

Uthayne wrote:

Here's some plots of a "wet torch" condition. Both touching the plate and not. The data was taken right after a pulse command, which had no real effect on lowering the voltage.

My low voltage is 23.8v and high is set at 23.9v

great. It would be good to see a plot of the transition from the on to off state to see the voltage fall off. That is something I will be doing.

I have written a new version but need to make sure it compiles. With a bit of luck I will test today.

So 2 new pins. mode 0 to select current method and mode 1 to follow the trend. The other pin is the number of readings the trend must be seen for (eg up or down for x cycles).

I was a bit worried about the noise you show might prevent a simple trend with x successive readings being lower than the last. I do have some code that computes a moving average which I was hoping to avoid incorporating.

But the plan is that if there is a downward trend, the probe will turn off the moment the voltage falls below the high volt threshold.

Forgive me if I'm noticing something everyone else is already aware of. So I have my shield grounded to earth. The interesting thing is that when i dip the tip into water, I can make the false ohmic trip go away by removing the powermax ground from the table. The ohmic goes back to working like normal then (I can trip it with my fingers as it makes about 5 volts).

I havent had enough time to really wrap my head around the whole thing, but for some reason, I feel like a well placed diode would solve the problem. Obviously it's feeding back through the torch, to the ground, and then through the table to complete the circuit.

Am I on to something, or do i have my head up my ...

Also, when I dip the torch in water with the ground on the table my voltage spikes to 25+ volts, even though my PS is only set to 24.5.

Just more data points.

One other thing, can someone explain the 390K resistor. I know it's a "scaling resistor", but if you can put that into laymens terms for me as to what it's doing?

rodw wrote:

snowgoer540 wrote:

One other thing, can someone explain the 390K resistor. I know it's a "scaling resistor", but if you can put that into laymens terms for me as to what it's doing?

Its described in the THCAD manual. Internally the thcad-5 has a 0-5 volt range with a voltage directly connected but that added resistor changes it to read 24.5 volts per the maths in the manual.

So I changed my mind and kept coding. I figured noise would likely be a problem. I decided that if we did a moving average of (ohmic-volts - ohmic-threshold), a positive average would be an upward trend and a negative average would be a downward trend. Whilst this involved some cut and paste of some (complex) earlier code I wrote, it actually simplifies the calculations to something like:

          avgvolts =	avgarcvolts(ohmic_volts - ohmic_threshold, (int)is_probing, (int)reading_count);
          if(is_probing){
            if(!ohmic_on)
              if(ohmic_volts > ohmic_threshold)
                ohmic_on = 1;
            else{
              if((ohmic_volts <= ohmic_threshold) && (avgvolts < 0.0)){
                ohmic_on = 0;
              }              
            }
          }          

where avgvolts is the moving average of the last N readings (defined by a hal pin)
I just hope we can see the voltage drop away as I did in my testing. Halscope plots needed to confirm this.

Very politely told me to RFM, much appreciated . Thank you for explaining.

So in my head this isn’t an issue that can be solved with code, but it also occurs to me that we may be trying to solve two different issues. I’m not an electrical engineer (mechanical), but to me the problem is that the power source completes the circuit when the nozzle is wet. If I measure the resistance from the center of the nozzle, to the ground I get 16.54K ohm. That will obviously be a bit higher with water. Until we find a way to account for that somehow, water in the torch will always trip the ohmic sense. Perhaps tommy is right, it needs to be more about current than resistance? I can make the hal show shown5+ volts just by grabbing the table and the nozzle with my bare hands. It makes sense that a drop of water in the nozzle easily overcomes 16.54K ohm.

Am I making any sense? Lol

rodw wrote:

We were measuring resistance in mega ohms so maybe there is a wiring issue.

On the other thread, Phill just posted

phillc54 wrote:

An interesting quote from John in this post

I know there has been a lot of discussion at times about "false" triggering of ohmic sensing. I can say with certainty that I have never had a problem with this that was a true false trigger. Yes, I have probing errors all the time but they are always caused by a flooded torch end or dross stuck in the torch end. So the probing errors are always cause by a true shorted torch which is why you check for that condition before the start of probing..

So it was John who insisted that the re was a probe test button on plasmac so he could check this before running a job.

SO what we are trying to do is to come up with a solution to a known problem... Maybe we won't solve it yet! I can only work on what I am seeing on my table...

Understood, I am just trying to wrap my head around what problem you are shooting to solve exactly?

As for resistance, at least for hypertherms, it seems like the resistance is much much lower. BUT it makes perfect sense as to why I am seeing voltage spikes of 25.3 or so when there is water shorting the torch tip. I read the manual, and the higher the ohms, the higher the scaling. SO, my resistance is effectively 390K + 16.5K which results in a raise of the perceived voltage (I think)... but it makes sense to me because having a series resistor of 406.5K means my VFS becomes 25.325V...which is what I am seeing it spike to.

So my shower thought for this, is what if we could isolate that case? Meaning this, the threashold starts at whatever the low is, but anything higher than the high threshold is disregarded. So, you're cutting, and go to make the next cut, but theres water in the torch... When it goes to ohmic sense, the voltage is much higher than 24 volts. But when the torch contacts the material, the voltage should actually go DOWN. And then when it hits 24, it trips the ohmic sense in plasmac. I can confirm the voltage drops with a direct short from torch to table, rather than a short through the plasma cutter, but I suspect it will.

In that case, the problem could be solved with code afterall

Or am I wayyy off base?

So, you're cutting, and go to make the next cut, but theres water in the torch...

From Johns description this doesn't happen during a cut which kinda makes sense as you would normally have airflow through the torch.

Understood, I am just trying to wrap my head around what problem you are shooting to solve exactly?

Me too, originally I thought that it was water on the surface of the material on the first cut. I would think subsequent cuts would be ok because of the airflow. Having the two closely related threads just confused matters I think, as you guys know I am easily confused...

rodw wrote:

Water ponding on the material poses a different problem based on my observation. Surface tension makes the water stick to the torch as it retracts and when it finally breaks free, the torch is significantly higher than the desired pierce height... Often too high to actually cut through the material.

Again mostly confined to that first cut becasue once the air is flowing it blows the water away from the contact point. The solution to a lot of issues is to trigger air flow before probing the first time.

Hence my earlier suggestion of an automatic quick torch pulse prior to probing commencing to get the air flowing

The real issue is water entering the torch and bridging between the two electrodes used for torch ignition. I did find that a 120 amp nozzle has a much bigger orifice so is much easier for water to enter.

Is this during probing or between cuts?

But in my environment I manages over 1500 probe event as last week and it all worked ort allowed me to recover from any fault. In one instance, over 500 pierces unattended...

So there is no issue.