This is used to feed (and control) an astrophotography telescope. Astrophotography cameras are supper cooled (using Peltier coolers) and draw quite a bit of power. The telescope itself along with the guide scope have dew heaters around the front elements. The circuit measures the ambient temperature and humidity level and computes the dew point, activating the dew heaters (PWM) as necessary. Then you have the mount (2 large motors), focuser, camera rotator, filter wheel, polar scope, etc. All 12V, all controlled using individual USB connections. The loads described are absolute maximum. It’s extremely unlikely all ports will ever be drawing the maximum load all at the same time. However, you don’t want to under engineer it. Where I live, temperatures are usually close to or below 0°C so only during peak summer the high loads come into play (which coincides with not a great deal of night time).
It’s so much about “minimal” space, but the less bulk next to the telescope, the better.
In any case I would separate the high current part from the normal electronics
for 10A your maximum input you need a track size of 10 mm by 35µ copper.
If you move relays and all the high current outputs to a second PCB, this will save size.
On the other side your motors and heater drives will need heatsinks otherwise they will blow up by these currents. Thats an other reason to separate regular electronics from high current part.
I would place the motor and heater drivers near to motors and heaters.
I still do not understand your USB concept.
For modern MOSFETs thanks to their very low on resistance 2…3 Amper currents are not terrible even without heatsink (or a small zone working as it).
There are also fully protected MOSFETs like:
Then I suggest you gain some more experience with working with MOSfets.
These things can easily switch 10A or more without any heat sink, and SMT parts with some extra cooling in the form of copper area on the PCB they can switch much more. Easily 100A, and this makes the copper on the PCB for conducting the current the limiting factor.
A small SOT-23 and according to the datasheet it can handle 8.6A.
With an Rds-on of 20mOhm (max) and 5A of current it will dissipate 0.5W and that would still be within limits without too much copper on the PCB for heat dissipation.
Also, have a look at controllers for BLDC motors. Below motor currents of 20A or so, most do not even have a heat sink. and those FET’s are switching PWM motor currents all the time.
If you want to sleep better. You can also use a “high-side switch”. These are also a mosfet, but they have a built in driver (so they can’t get operated in the linear region where it can get hot) and they also have built in protection for over voltage, over current, temperature, etc.
You should draw a schematic at least. Does not have to be in Kicad. In doing so, you will have to make many engineering decisions about parts to use, pinouts, logic, and so on. I would charge more to design it for you, and would have to pester you with suggestions and questions about “did you mean to do this?” and explain why it was questionable.
When you have a working schematic, post it again.