Spis 6.1.0 Bug in Langmuir Probe instrument sweep, current collection

SPIS
1

Hello all.

My students are getting some ridiculous low currents on our LP sweeps, for any population. It is consistently at least 10^6 lower that what we expect.

The SpisNum log is interesting, and points to some issue with the summation of currents on each surface, here’s a snippet for the last surface (the count starts at 0, so there’s no surface 285) :

Density in volume seen = 1456079.6433853859 m-3
Collected Current I = -1.9013031090224007E-11 A
Density in volume seen = 0.0 m-3
Collected Current I = -0.0 A
Density in volume seen = 54768.30407562608 m-3
Collected Current I = -1.2546814071436615E-12 A
Density in volume seen = 0.0 m-3
Collected Current I = -0.0 A
Density in volume seen = 1880750.2063190837 m-3
Collected Current I = -1.889095200286357E-11 A
Density in volume seen = 0.0 m-3
Collected Current I = -0.0 A
Density in volume seen = 1926533.2395256464 m-3
Collected Current I = -2.3361649991081503E-11 A
Density in volume seen = 0.0 m-3
Collected Current I = -0.0 A
Surface number 284 over 285 surfaces on the detector | curr dens =-8.5207107E-13
Surface backtracking done at 100

Langmuir_probe_id24 current equals: -4.4239304E-14 A

The number in bold is also what appears in the Langmuir_probe_id24_IVcurve_Seq1.txt file with potentials.
I don’t understand why there are 8(?) different current collections, half of which are zero. the non-zero current values appear… more reasonable. If 1 out of 285 surfaces collects 2E-11 Amps, then I would expect the sum of all currents should in nanoamps, not femtoamps.

It’s also remarkable consistent. Although noisy, and extremely small, we can recover e.g. the electron temperature by analysing the slope of the sweep for these sweeps

I couldn’t figure out how to attach a file, so here’s a dropbox link to the instrument xml file

https://www.dropbox.com/s/rp7yprkve6tlkty/Langmuir_probe_id24.xml?dl=0

and the entire project

https://www.dropbox.com/s/k0095s7pjq476qf/PopDurationLP_Electrons.spis5.zip?dl=0

2

Just to add to this, if I manually add all the ≈ 50000 listed currents in the SpisNum.log file, and collect them by bias potential, I recover a very nice sweep with an error of only 2% from what I expected. I think that confirms that there’s nothing wrong with this simulation, just that SPIS isn’t collecting and summing the currents properly (or overwriting it, or whatever)

3

Thanks Fredrik for investigating the problem. We will try to correct the bug for the next version