Quoted from Edenecho:
What is visually the difference between chained leds and led inserts through the pd-led? Are chained basicalle one single chain going through every led under the playfield while with the pd-led its one wiring out to each led insert and back to the board?
"Chained" means that the RGB signals are all driven from a single, common source. The LED is enabled by a bit that shifts through the chain. As this bit shifts down the chain, the RGB levels change for the next LED in the chain. Once you've shifted through the entire chain, the whole process repeats itself. This bit is driven by a single ended driver. It's a clock signal more or less and is the source of all of the headaches Gerry is RIGHTFULLY warning you about. Quite frankly, Gerry could modify his FPGA in about an hour to support chained LEDs and announce support for them using PROC (hell, I'll write the damn verilog myself! ). He won't because he is 100% correct ... they are problematic under a pinball machine playfield. Yes, you would only need 4 wires for signaling, but you will have problems with that clock wire if it is single ended.
"Parallel" means that you will have three wires driving each LED for a red, a green, and a blue level. If you want to drive 8 RGB LEDs, you will need 24 wires. It's more painful to wire up and it costs more, but it is the proper solution (unless the chained LEDs use a differential signal ... then then and only then will chained LEDs be acceptable for pinball use).
In terms of visual appearance while playing, you won't see any difference if you're driving serial LEDs properly vs their parallel brethren so long as you're using a reasonable timing on the LED chain and there is no noise. It basically comes down to lighting an LED for X uS then shifting the "enable" bit downstream to light up the next LED for a few uS ... rinse, lather, repeat.
I work in the robotics industry, develop FPGAs for a living, and design a PCB here and there ... I deal with noise all of the time on high speed serial links for vision applications.
Trust me, use the parallel solution. You are going to save yourself a LOT of grief in the long run. You WILL have intermittent light failures using single ended chained anything under a pinball machine playfield. Shielding won't help, routing won't help, nothing will help.
You have a shot at getting lucky and things may work at first. However, you might bump a solenoid a few mm closer to a single ended wire and suddenly cause intermittent bit flipping every X number of times that solenoid fires. Hell, I've wired up piss poor examples of single ended connections that would cause a clock on a wire to go crazy if you move your HAND past said wire just as an example as to why long, single ended traces/wires are silly (I'm being a bit unfair as there was zero ground wire near the clock wire, but you get my point).
While I am no device physics expert (I know people that understand this stuff as if it were "normal" to understand this stuff ), I do know enough that differential wiring and differential routing do plenty to control effects from noise. A differential receiver is looking for a difference between two voltages to signal a one or a zero. A differential driver drives a signal at a + level on one line and a - level on another like. They are intertwined ... the intertwining has the effect of canceling out some noise. More importantly is how the receiver works. Again, the receiver looks for a difference between the two voltages. If some kind of noise is induced on the twisted pair and causes the lines to move to a higher or lower potential, the receiver will still interpret the one or zero properly because it is merely focused on a difference and the wires will tend to drift in the same direction.
Now, single ended connections require a voltage threshold to indicate a zero or one. Let's say you have a single ended wire driving the clock input of a flip flop (or the lamp select input on your serial LEDs) with a threshold voltage of 1.8V that is rising up to 3.3V. Now, let's say a solenoid fires, induces a spike on the single ended line that causes a fast ramp rate, spikes the signal up to 4V, causes a swing down to 1.5V for about, I dunno, 60 us, and swings back up and levels off at 3.3V. See the problem? You've caused two clock edges to occur ... you might get lucky and 60us would be ignored by one LED for a bunch of different reasons. You might have an LED that is sensitive to the 60us time and it might flip state. If you were using a differential input clock, the + line would jump up to 4V, but the negative line would also follow suite ... no glitch occurs .
There's also the issue of noise causing stress on the inputs when the voltages swing really high or low, but the amount of time the voltage is out of spec usually causes no damage.
You want to think of all of those single ended connections as mini antennas too. You will induce electromagnetic currents on that antenna and will get intermittent bit flips thanks to pinball solenoids. A bit flip in a one hot shift register (which is all an LED chain is) is pretty much fatal . These kinds of errors are a complete pain in the ass to debug ... especially when you are writing brand new software and have no idea if you light show issues are your fault or some hardware failure.
Moreover, these kinds of problems make game coding not much fun if you ask me . I'd rather be solving gameplay problems vs. hardware failures.
Do yourself a favor, just get the system that offers the parallel LEDs ... you will never thank me for this suggestion as you will NOT have any issues at all, thus no need for thanks .
Good luck on your game!!! You home brew authors do some amazing stuff!!!