Electronic – Is there such a thing as a PWM expander like there is a GPIO expander

Here is some references of MCUs with high PWM outputs (> 12)

• Freescale Kinetis K10 @ 120 MHz series (ARM-Based) : 2x8ch + 2x2ch = 20 PWM channels
• A small number of dsPIC33 have a high number of PWM output:
  • dsPIC33EP256MU814 : up to 30 PWM outputs
  • dsPIC33FJ64GS610 : up to 22 PWM outputs
• Atmel 32-bit AVR MCU have some references with high PWM outputs:
  • ATUC256L3U : up to 36 PWM channels

But as Wouter van Ooijen said in a comment, maybe you should review your architecture and instead of using one big MCU with many PWM outputs, you should use one small MCU for each "component" (SMPS, motor, etc.), and use one big MCU for the general control of all other MCUs. This will lead to a more flexible design: you can add a motor simply by adding a small controlling MCU and linked to the big one in SPI or I2C.

For the CPLD/FPGA approach, you can take a look at Lattice ice40 CPLD and to this reference design : MxN Channel PWM Do some searchs about PWM on Altera and Xilinx website to have other CPLD references.

A PWM with 0% period should be completely off, no spikes. A simple ADC could help with PWM to Digital.

The problem is that while your RC receiver says it's just PWM, it is not. It's a specific protocol that works slightly different. Essentially, it is a PWM signal with 1ms to 2ms pulses at 50hz. A centered pulse is neutral. Even with a push button and not an analog stick, the protocol stays the same.

I highly doubt there is a transistor that does what you want, you would need a custom design or a microcontroller to do what you want. Simple enough to do, read an analog input, determine what the pulse length is, and output to a digital pin the state.

This question is about reverse engineering a product that does just that: Nano Electric Receiver Switch - Circuit Components You may want to read it hint hint