Unlike a typical pinball controller where the switch to fire a solenoid (i.e. a flipper) is directly connected to the solenoid, my design takes the button press and sends it over USB to the host PC, where my Chameleon Pinball Engine software processes the button press, and sends a command signal back over USB to turn on the solenoid.
As the Chameleon IO Controller has both the gamepad buttons and LED outputs all on a single PCB, effectively the signal makes a round-trip from IO Controller to host PC and back to the IO Controller.
This design introduces latency at several distinct points. There is the latency in the IO Controller to detect a button press and send the new button state info to the host PC. There is latency in my pinball software to recognize and process the button press. There is the latency in sending the new LED state back to the IO Controller. And finally there is latency in the IO Controller to actually turn on the output.
My previous post detailed working on the Chameleon IO Controller's firmware to improve PWM performance, so of those four distinct points affected by latency, I've already addressed the last two. My goal with this new round of testing is to work on the other two points.
Below a certain threshold, latency isn't an actual problem. When you're playing video games, as long as latency from button pressed to pixels updating stays below 40ms, response feels instantaneous and you'll have a great experience. But real world games like pinball are more demanding - physics waits for no-one - so latency needs to be even lower for good ball control.




