...
We will likely be using parallel led strings (this is pretty common in lights now days) so we need a good way to deal with this. We want to balance led strings. One example would be if there are leds in parallel and they have a forward voltage of 3.3V, since there is up to 20% variation it could be a huge difference in the voltage needed. Additionally, the lower voltage draws more current than expected compared to the higher voltage string.
In general, if temp goes up, the forward voltage will drop and it will draw a lot more current → we should mount on a common heat sink (I believe this is being done in the current boards)
We can use a current mirror (this is being done in the current boards) for regulating current
if using dimmable lights, pwm is the way to go
Potential upgrades
Use independent opamps for each string instead of a current mirror, might have extra voltage but the current level should be good. More effective, less efficient, other considerations include needing a FET in a power package and that the switching frequency will be lower
We could add balsast resistors though this is probably not needed since we are using the current mirrors
...
Sounds expensive but using a mag-amp with one dc driver for dimmable leds
Increase the FOV of the lights
increase power efficiency
Vary threshold voltages (use low voltage thresholds)
stack transistors in series
if we are using linear regulators switch to switching regulators
we can use fast switching mosfets
add more thermal vias if possible.
Make sure we keep the ground and power planes for low impedance paths
Adding decoupling and bypass capacitors (for filtering) near power pins of components
Center high power components for better thermal efficiency
even thought the driver boards are small, try to keep things away from the edge of the board
if necessary add heat sinks
Test the current boards with a PDN analyzer?
https://youtu.be/oFdsiRMZZWk?t=145
We could have led chasing LOL
Another example of a sequential led circuit
...
https://www.homemade-circuits.com/sequential-bar-graph-turn-light/
Rough Sample Nita made
There are some RC filters which delay the "ON" signal sent from power distribution, making the voltage on each of the nodes increase one after the other (you can see on the probes). The NMOS should be on each node but is just on one for illustration. the diode models an LED (which would be technically on another board - space might be constrained so we have separate boards for driver + LEDs)
...