First of all, before answering your questions, let me make some suggestions about the layout. I have seen your previous post and this has been a good attempt, but:
1.- The "Switching Loog" is too large. (Switching loop is Vin, Trafo, Q1 and R8, you call it,"high current loop", maybe it is no quite precise) And moreover, this loop include the control area. This is really EMI problem!! I suggest something like this.
(Think in current loops always!)
Make smaller the switching loop
- Place R8 near from C1. Really near.
- Rotate Q1 90ยบ clockwise
- Place R6 and C6 as close as possible (acap) from Trafo pins.
After that,
- Think about the "Gate net", "Sense net" and "pin 15 from controller net", It must be route separatly! The "gate net" is a radiant net and the others ones are sensitive nets. Route them.
- Route Vin and Enable controller nets.
- Complete the route of primary
(The controller has not a Decoupling Capacitor ? Are you sure of this?)
2.- A connector for trasformer sounds a problems maker. Could you solder your Custom Transforme in the PCB? It would be better. If you cann't, I you sugggest a 2 row 2 columns conectors. In this way you can gain more space between primary and secondary, and more space between trafo pins.
OK and now. Your questions.
1.- Yes, try to keep separate switching ground from control ground, but connect to the same ground, in your case Bottom layer. For this, try to keep the "switching loop" components together according the placement I suggest.
Connecting R8 to Vin through one track and then one via to GND is not a good idea in your current layout.
2.- Ummm... I would try another layout before consider this question.
3.- I cann't undestand this question. What do you mean?
4.- As a rule, It better If you fill the empty area with hatched copper. It improve the etching process in PCB manufactureing.
5.- As a starting point, fill as much as you can. Do you need consider any electrical isolation between primary and secondary.
Good luck!
There are pros and cons for either.
A signal full ground plane has the advantage that signal 0V is a high integrity 0V and can be relied upon but, not when there are currents flowing of any significance. Despite a ground plane being very low impedance, volt drops can still occur when significant currents flow. How much volt drop being a problem depends entirely on the smallest signal that you wish to amplify.
A star point system avoids those "significant currents" mentioned in the previous paragraph by making sure that tracks carry signals (0v return tracks) do not share with these significant currents. The down side is that you end up with magnetic loops that can have voltages induced from one another and the "significant current" flowing in a different track can be coupled.
Best Answer
I'm thinking tie both sides of the level shifter to the input ground and have the output current take a long path through the single point ground?
I would do the opposite, so keep the loop at the output short because that's where the current flows.
The inputs are high impedance so almost no current flows. For noise reasons the argument could be made that the inputs should have the better ground connection but since we're dealing with digital input signals, I would not be too concerned about noise at the inputs.
I think in the end there isn't that much difference between ground at input side or ground at output side. This isn't RF, not very sensitive (using low voltages) and also not high current.
If you still have doubts you could consider to connect the levelshifter's grounds to a local ground plane and then connect that plane via a link / zero ohm resistor / blob of solder to the output side or the input side.