Flicker Problems - PWM Waveform

[teddpoit]

I hope everyone is having a nice fall. I first want to say that this website is great; it is very helpful on virtually any topic! I recently finished building a FireGod system, which is an awesome design by FireGod, and I am very excited about it! I built the PC interface, one field module, and one SSRNeon design relay (sinking relay). The system works; however, it has a very noticeable flicker.

I have read through the other threads about issues with flicker, and I have tried some of the tips/tricks without success (6V regulator, etc.). I did notice some weird behavior in part of the circuit, and I was hoping that the forum might have some ideas or thoughts.

I put an oscilloscope probe on the output from the PIC to a channel and noticed something weird with the PWM signal. Every time there was a flicker, I noticed that one period of the pulse waveform would drop to zero (no pulse). Then it would go back to the normal PWM mode. This occured even if all channels were stationary. This only happens when the channel is not full on (with discrete periods) and at some intensity between 0% and 100% (or during ramping/fading). Could this be weird behavior out of the PIC? I have tried a couple different ones, and I always get some kind of flicker.

 

[ebrady]

I have seen this behavior on some of my own custom boards.

In my case, it was due to the zero crossing circuit getting "faked out" by noise transients on the AC line.

I was able to correct it by changing the zero crossing point of the circuit to "fire" at a slightly higher line voltage and not quite as close to 0 (somewhere around 8 volts if I remember correctly). Firing at the higher voltage threshold masked a lot of the zero noise off of the line.
Using this voltage level, I then calculated the amount of time it would take mains to cycle to the real zero crossing point and entered this as a software delay in the PIC.

 

[FireGod]

I have posted some new sinking code that may help.