When the max count of robots recharge the internal buffer gets depleted quite quickly, because the overall recharge rate outscales the buffer rate. I already tried to fix it in my version, but it would be great if you overthink it for other users :) And thank you. I waited for something like this, but never had the time for digging into coding

Mhh i see the problem and fixed it, almost i think :)
Thank you.

Hmm it looks even odder right now. I think the counts of charging ports were just fine. In your new version your Mk2 charges up to 12 bots which means (having infinite bots waiting for recharge) 12x2MW=48MW of energy demand, which isn't met by the internal buffer rate of 8MW, meaning at some point (a quick one) charging will take 6 times longer. If you have a look at the vanilla p.roboport you can see that you can have 2 bots charging at a rate of 1MW while the internal buffer rate is at 3.5MW, so you still have a surplus of 1.5MW, so providing enough energy to the system will recharge the roboport even when the max count of bots is leeching energy.
My suggestion would be to tweak the count of charging ports again (maybe even reducing the count of roboports), but setting the input_flow_limits to chargingports * chargingenergy * 1.75 (depending on how you want to balance it).

Hope that I do not annoy you. Cheers!

Thank you for your suggestion ☺ Sorry my fault...
The problem with infinite comming vanilla bots should now be fixed, with an big power supply of course.

It works fine with:
4x fusion reactor
8x batteries mk2
1x Personal roboport mk3/2
1x Exoskelet
2x Energy shield mk2
And 100 vanilla robots :)

Greetings!

Nice one! Looks great :) Maybe also have them be a tertiary power priority, since shields are important aswell :) My last complaint (for now) I promise :D