tagging people who might be interested
@StollD @Galileo88 @Poodmund

Test environment not as controlled but still relevant. Using KSP 1.3.0, Kopernicus 1.3.0-6, MFI 1.2.4, MM 2.8.1 and latest GPP dev build.

New save, launched directly up into orbit. On panel deployment half the panels tracked Grannus as exposure was 1.0 and the other half tracked Ciro.

http://imgur.com/a/zkMxV

Logs

I did some more tests and it turns out that, if you rotate the craft so that the solar panels face the brightest star, then they will start tracking that one

So I guess the issue is that the code only accounts for the current orientation of the solar panels, and not the whole sky

@StollD I had some time to think and I think Kopernicus should automatically focus on the best star regardless of AOA

so, even if a star is behind the solar panels, the solar panels should still try to follow that one, instead of facing the best option in front of them

this is because in 99% of the cases you will have a close star and one or more distant stars, whenever you turn away from the closest star you will risk getting "stuck" on one of the far away stars that will provide infinitesimal amounts of Ec/s

the only issue of this would be if there are two stars that are similar in Ec/s production, but the one with highest flux is in a position that doesn't allow for 100% AOA

in that case you might want the solar panels to lock on the star with the lower flux if they can maximize AOA

this situation is much more unlikely than the other, so until we get the manual tracking to work I'd say that this change would help lessen the amount of bugreports for solar panels

I agree with Sigma's sentiments above entirely. In my reference case, the secondary star was/is pretty much never the better star to orient the panels to as it gives almost nil EC even when they have 1.00 exposure to that star.

If a check is already made to list available light sources and decide on which can provide most EC in its front 180 degrees field of view, is there anything stopping Kopernicus from determining the same from a 360 degree FOV?

Just tested the functionality on the new Manual Tracking and it seems I get runaway exposure in the values of thousands that can suddenly ramp up quickly then randomly start to drop back to the correct level. This can happen when coming out of time warp or just by simply spinning the craft around changing the angle of incidence between the panels and the tracked star.

This was occurring on an install of KSP 1.3, Kopernicus latest build, MFI 1.2.4, MM 2.8.1 and a barebones version of GPP.

Logs: https://www.dropbox.com/s/dbciiat645yv5cp/Kopernicus%20Solar%20Panel%20Logs.zip?dl=0

Screenshots: https://i.imgur.com/tmiOlAI.png https://i.imgur.com/eQTh82G.png

Apart from the above, the implementation of the tracking was working fine, as intended. 😄

Im also currently experiencing this issue, with only GPP and Kopernicus, planetshine, scatterer. Panels face the wrong way, or face the right way but dont have any energy flow even with exposure at .99

Seems totally random for most craft but my stations have now become unusable due to being unable to charge.

I tried the new features out of curiosity and I have a couple of suggestions:

1- make the button for "use relative exposure" stay down when the feature is enabled and up when it is disabled, or alternatively change the text to "use relative exposure" / "use absolute exposure"

2- the auto tracking seems a bit weird, in GN while in orbit around kerbin the star chosen is K-stor Cb even if the Sun is in clear view screenshot

Logs.SolarPanels.zip

1. should be fixed
2. can you please try it again? issue 1. might have caused this.

still targetting kstorcb

Logs.SolarPanels2.zip

note: I generated this logs with SB and GN installed

to test them use the development branch of GN and the test branch of SB (I just updated the dlls so you can test)

(you need to use the textures from the latest releases of those mods, since I don't include texture files in the github repo)

honestly, I think we can release kopernicus as it is now, and try to figure out what's going on later

the manual selection of targets is just enough

maybe we could add a button to "set all solar panels to the same target" to make it more convenient for big stations with a lot of panels

closing this

solved in v1.3.0-8