SPIS - EP, How to calculate electric currents by HET in circuit

steevy · 21 April 2025 18:15

Hello everyone!
I work in a laboratory with Hall thrusters, and it just so happens that I need to understand what current of ions and electrons (especially electrons) flows to the surface of a spacecraft under a potential comparable to the accelerating potential of the thruster. We have already conducted an experiment in the laboratory and obtained linear volt-ampere characteristics of probes in plasma, i.e. the dependence of current on voltage on a flat probe located in the thruster plasma. And we obtained linear characteristics, where, logically, the electron current goes to the positive potential, and the ion current to the negative. But when modeling the same situation in SPIS, I encountered the fact that the net electric current to the nodes corresponding to the flat probes does not coincide at all with the experimental results. Moreover, it is not clear whether the electric current in the circuit is closed correctly. The electric circuit is as simple as possible: between nodes 3 and 5 (these are electrodes - probes) there is a potential difference, for example 50, 100, 200, and 300 Volts. How can I understand what current flows between nodes 3 and 5? If you look at the currents on the surface, it turns out very strange: When the potential of the electric field (positive) on the probe decreases, the ion current for some reason decreases, although it should, on the contrary, better get to the surface and increase accordingly. If anyone has encountered such a situation, then please explain how electric circuits work, how to see what current flows in them?

I am attaching the calculation files of the model at this link: SPIS-EP Current – Google Drive

And some pictures:

Exel file inside g-link.

Thanks in advance for your reply, advice or any help!

cyberpsycho · 22 April 2025 04:57

I ask you to help a colleague (Please). We both encountered the same thing.
We would appreciate any clarification.

frejon · 24 April 2025 16:53

Hi Ivan, Karolina.
A few words of warning,

Don’t use unlimited ‘UnlimitedGlobalMaxwellBoltzmannVolDistrib’ for anything. I guess I could have said "if you don’t know what you’re doing’, but it’s such an unrealistic approximation of densities and currents at attractive potential that I think it should be fully deprecated. KineticMaxwellBoltzmannVolDistrib is there if you really need speed and performance without unreasonable approximations, but why not default to PICVolDistrib.

Also, GlobalMaxwellBoltzmannVolDistrib distribution of ions is not recommended for ions in the manual. Your temperatures are pretty high in both of these cases, in comparison to the potentials, so it hasn’t affected your results too much so far, but it’s asking for trouble. Just use PIC.

I also think the input plasma densities are strange, it’s not a quasineutral plasma at start, but maybe that is as intended.

As I’m sure you’ve noticed, SPIS is telling you that your mesh is of bad quality. Optimize the .msh in GMSH using both or any optimization method until you get your quality under control. you may need to increase resolution where necessary.
You are extracting results before the simulation has converged. As you probably noticed, the 300V simulation potentials are still evolving for node 3 and 5, and a comparison to lab measurements will be useless

And onto your actual question. I couldn’t reproduce your google sheet currents (using the emitted and collected currents, the maximum I could see was -0.3 Amps in node 5 in the 300V case, from the cathodeelectrons. I think there is some issue with how you or SPIS have extracted these log10 currents. I always look at the collectedcurrents.txt, emittedcurrents.txt and potentials.txt in the output folder of the numkernel, and they make sense to me.

log10(abs(collectedcurrents for node ==5)

steevy · 1 June 2025 16:54

Hi Fredrik

Thank you for your reply!
And sorry for my late reply, it took some time to do the calculations.

Following your advice, I fixed the distributions for the electrons and ions on PICVolDistrib and increased the mesh quality by optimizing. I also increased the calculation time so that the surface potentials had time to equalize. However, this did not allow me to get an acceptable result.

As before, the balances of the currents at the input and output from the electrodes on electrical nodes 3 and 5 are not equal to zero in sum. Which means that the charge conservation law is not observed. This time, as you recommended, I used txt files (though for the currents I looked at individual nodes, since the total collected current is difficult to analyze due to the large number of parameters for monitoring).

So I ask again, “what to do with this?”)))
Why is the current coming to electrical node 3, connected to electrical node 5, different from the current that goes to this electrical node 5?

Here is a link to my new calculations and also an Excel file:
https://drive.google.com/drive/folders/19IgEFzxUWMIbyPl_7ykUgRFlmdFeKZHF?usp=sharing

And a few more questions and clarifications:

How to check the convergence of the calculation in SPIS? What criterion to use? How to check compliance with the laws of conservation of energy, residuals, energy balance, etc.?

And where exactly can I find information about distributions and what exactly is meant by such things as, for example, Total_current_on_node_3? (In the Detailed Design Document I did not find a transcript of what UnlimitedGlobalMaxwellBoltzmannVolDistrib is, although it is present in the validated model propulsion-6.1.0-2021-06-23_09.44.07.182)

Plus I clarify that I used the default SPIS version, and not my own build, so I do not think that these are errors in the SPIS data conversion.