amrvis3d for rho_electrons

[chiehjen]

I tried to visualize a 3D simulation using amrvis (the default tool). It works fine for positive charge, but when I select "rho_electron", it shows nothing (except for the box), like below:

For positive charge, i.e., aluminium in this case, it works fine:

The thing is if I define the charge of e- to positive, i.e., instead of

ele.charge = -q_e
, use
ele.charge = q_e

then amrvis3d works again for rho_electron (at least the shape is correct, but the charge is opposite), as below:

So it appears that the problem is amrvis3d cannot plot materials with negative charge. For 2D cases, amrvis2d has no problem in showing negative charge materials.

Does anyone encounter problem in 3D visualization of electron density? Please tell me if there's any solution to that.

[RemiLehe]

@chiehjen Thanks for reporting this.

@WeiqunZhang @atmyers Have you seen this error before with amrvis? Is this a known issue?

[WeiqunZhang]

This is not an amrvis issue. By default it shows three slices at i = 0, j = 0 and k = 0. You can click (including right, middle and left clicks) on the pictures to move the slices to different positions. You can also click on the "arrows" in lower right part.

[chiehjen]

This is not an amrvis issue. By default it shows three slices at i = 0, j = 0 and k = 0. You can click (including right, middle and left clicks) on the pictures to move the slices to different positions. You can also click on the "arrows" in lower right part.

I could click and move the slices. However, the problem is it always shows nothing but only the box for "rho_electron". This occurs only if the charge is negative and for 3d simulation plotted with amrvis3d.

For 2d simulations plotting with amrvis2d, there is no such problem, it can display negative value for "rho_electron".

[WeiqunZhang]

Could you show a picture with slices not at (0,0,0)? Maybe something at (64,64,128)?

You can use mouse to click on the top-left, top-right, and lower-left panels or the arrows shown below.

You can also provide the values from command line. amrvis..... -xslice 64 -yslice 64 -zslice 128.

[WeiqunZhang]

If you are talking about the volume rendering in the lower-right panel. You need to adjust opacity. Click on "Render", "Lighting", and then modify "minRayOpacity" and "maxRayOpacity". By default high values have higher opacity. So the outer layer could block the view of the inner part.

[chiehjen]

If you are talking about the volume rendering in the lower-right panel. You need to adjust opacity. Click on "Render", "Lighting", and then modify "minRayOpacity" and "maxRayOpacity". By default high values have higher opacity. So the outer layer could block the view of the inner part.

Thanks a lot for the info. Moving to (64,64,128) or other non-zero value helps in the panels showing the 2D slides. For the lower-right panel, however, it doesn't work to me. I tried adjust "minRayOpacity" and "maxRayOpacity" for values from 0 to 1, but the 3D structure of the target is still not there. Below is an example:

Maybe because "rho_electron" is negative, so any Opacity value from 0~1 block it? I tried negative value also, but it doesn't work either.

[WeiqunZhang]

You might try minRayOpacity = 0.95 and maxRayOpacity = 0.05. Here, min and max refer to rho_ele. It's not that minRayOpacity must be less than maxRayOpacity. You want the higher density region to be more transparent.

[chiehjen]

You might try minRayOpacity = 0.95 and maxRayOpacity = 0.05. Here, min and max refer to rho_ele. It's not that minRayOpacity must be less than maxRayOpacity. You want the higher density region to be more transparent.

Thanks again. I tried minRayOpacity = 0.95 and maxRayOpacity = 0.05, but still cannot get it to work:

So it seems it's a bug in displaying density of negative charge particles.

[WeiqunZhang]

It does seem to have a problem. The original author of Amrvis has retired. So I am not sure whether there is a hidden option somewhere. Nevertheless, I have a PR that seems to fix the issue. @chiehjen Could you please give it a test?

AMReX-Codes/Amrvis#35

Click on "Invert Opacity" and then click on "Draw".

[chiehjen]

@WeiqunZhang It works after adding the lines in /Amrvis files and recompile. Thanks a lot !

[WeiqunZhang]

This is resolved. By the way, another approach is to load a different palette. The default palette has higher opacity for higher value. But one could use a different palette.