Is the PWR_FLAG Still in use in V7?

KiCad V6 already has the ability to assign colors and wire thickness to a net class. You can do this in: Schematic Editor / File / Schematic Setup / Project / Net Classes. You can also give the Default Net class some weird color, with the result that any net that is not in another defined netclass will get this color.

1 Like

I think adding a netclass property that any net in the class is assumed-power-driven is a reasonable feature request.

1 Like

I’m getting the impression that KiCad is moving int the other direction. Net class assignment used to be some text only thing, and now it’s moving to a combination of the net class directive labels and the filters.

I intended to say that it could be a property of a battery pin or connector pin, which can be set in case that is where the power comes into the circuit.
I could imagine, analoguously to an output of a voltage regulator, which is defined as a power output.

But perhaps it can be a net property instead; I haven’t decided whether one or the other is better.

You can already set a pin to type power_output, if you want to do this for a regulator or something. But doing so for a connector symbol is somewhat dangerous (if you care about the value of the check) because the connector symbol could be used elsewhere in a situation where it no longer provides power.

It would need to be a property that is not inherent to the symbol, but set separately for each use of the connector in the schematic. So there is a difference compared to the voltage regulator.
I don’t know if this is the best solution, I just thought it would be logical to put the source of power at the point in the schematic where it best matches the situation in real life…

This is what is intended with the visual directives.

Run the ERC, right click on the warning and then you get a choice of various actions including ignoring the inpuit-output error. You can also edit the ERC rules.

Does this undermine the benefit of the ERC?

As I understand it, all lines connected to, say, the +3V3 power pin are the same net because power pins are like global labels. The ERCheck works because power pins are type “Power Input”, so there must be exactly one “Power Output” connected to the net to satisfy all the “Power Input” power pins. For scenarios where no “Power Output” pin is connected (eg. when power comes via a connector pin or passes through a passive), the PWR_FLAG emulates a “Power Output” pin and says to the ERC, “I warrant that power is sourced here”.

Therefore, wouldn’t a netclass solution be equivalent to turning off the error, since it’s effectively saying "this net doesn’t need a “Power Output” connected to it?

Yes, it would be turning off the error, but for only one netclass rather than the whole schematic.

1 Like