http://www.microchip.com/Download/Appnote/Category/rDesigns/40171a.pdf is an app note for a Lamp Dimmer for the PIC12C508

Fansler, David says:

1. The schematic does not totally match the article
2. The load would go between Hot Out and [Return].
3. You would supply a Neutral (white wire) to Return.
4. You would supply Hot (balck wirefrom power company) to Hot In
5. Ground (green) is a good idea. It can be tied to Neutral (which is done in US breaker boxes).
   
   Dave Bell says (Paul B. Webster and Donald Budette aggree)

6. Don't! This is done in US breaker boxes, *but that's the only legal place to do it*! The Green/Grounding conductor should go to case ground if there is a metal case, and not be tied to the circuit anywhere.
   

7. Vcc is generated by the circuit above the PIC (RV1, R1,2, C1-3, D1-3)
8. I would take JP1 to be an external input to allow control of the circuit by an outside source (another PIC, automation system, etc.)
9. While simple, a fair warning is given in the article that this circuit has no isolation from the AC main and therefore can bite you!
10. If used for lighting purposes, this circuit will produce filament hum at low light levels (larger filaments hum more) A large choke in line with the load will kill the hum.

Donald Burdette says:

• C3 absolutely MUST be specifically rated as an Across-the-Line capacitor. C3 is where the line voltage gets dropped down to the Vcc level. If this cap fails shorted, something will blow up (quite literally). Line rated caps will not fail shorted. Don't make the mistake of assuming that R1 is to protect you from this. It will dissipate about 300 watts if the cap shorts, which will probably cause it to explode. R1 is there to limit the surge current when you plug this thing in.
  
  Paul B. Webster says:

• It should really therefore, be a component rated for the purpose. Called a "fuse resistor", commonly used in TVs and so marked on the schematic. A slo-blo fuse of about 200mA rating may be a good addition.
  

• If you want to connect this circuit to any other devices, be very careful of grounds. JP3 pin 2 is connected to the line voltage neutral. This can be several volts or more different from ground. That's enough to damage your equipment or make it act wierd. Opto isolators would be a very good idea.
• This circuit switches the neutral line, which is normally not the correct thing to do. It means that when your lamp is off or partially dimmed, both sides of it are HOT, which I think is a violation of NEC rules and is an electrocution hazard.
• Your switches should be carefully selected to ensure that the contacts cannot be touched, even if the switch breaks, as many inexpensive switches will do when there are kids, pets, or clumsy adults around. If your circuit is built and functioning correctly, there's little danger, but if you plug it in backwards or if something fails, those contacts become HOT.
• I agree with David Fansler's wiring recommendations, except for one - if you are in the US, or probably anywhere else, NEVER connect ground to neutral. Ground is a separate circuit designed for your safety, and should never be used to carry current. If you connect it to neutral, will reduce its effectiveness. Also, you may reduce or defeat the effectiveness of GFCI breakers, which are now required for certain outlets in most US home construction.

Speaking of GFCI breakers, I would recommend that this circuit be plugged in through one. It will reduce some of the dangers mentioned above.

M. Admin Davis says:

...note that the triac should ALWAYS be considered to be on. If the only load is a neon bulb, for instance, it won't turn off since it draws so little current.

You might want to:

• use 0 and -5V as the uP supply because triacs trigger better with a negative voltage pulse.
• use a diode in series with the gate resistor to prevent the AC that comes out of the gate, when the triac is conducting, getting back into the uP.
• use RTCC as a zero-crossing detector, set it to FFh. The next positive transition on RTCC will cause it to rollover to 00h. Check it with mov RTCC,1 and test Z flag. (My take on this is that RTCC can't be used to detect both positive- and negative-going crossings)
• Detect both the rising and falling approach to the zero voltage point. This way, integral numbers of both positive and negative half cycles can be used. For example, loads with high inertia such as heating elements would be better suited being supplied with complete half cycles. With motors and lights this would cause jerking or flickering and they are best supplied with a part of every cycle.
• Put a choke in series with the triac to minimise EMI and suppressor cap(s) across the load.
• If you assume the threshold for the CMOS inputs is -2.5V (with this PSU 0V is high and -5V is low) then the positive- and negative- going zero crossing points are slightly assymetrical, which should be accounted for if very fine control is needed.
• Detect the zero crossings - When reading the input pin a change from a stream of 1's to 0 is a negative-going zero crossing ie the mains is heading into its negative half cycle. A change from a stream of 0's to 1 is a positive-going crossing.
• Trigger early in the cycle when using whole half cycles to minimise current rush and electrical noise. However with inductive loads, the current and voltage through the triac may not be enough to cause it to latch properly and the trigger pulse should be lengthened or a series of micropulses sent throughout the cycle.
• For dimming, divide a mains half cycle into a finite number of equal time intervals, say 50 or whatever is convenient for the frequency you have the PIC running at. The triac will continue to conduct after being triggered so the power to the light will come from the moment the triac is triggered until the next zero voltage point. For 100% power send a gate pulse out as soon as a rising or falling voltage is detected after a zero point. For other brightnesses, start counting the time intervals after the zero point and then send a gate pulse. As this means the power will be applied at a non-zero voltage, steps will have to be taken to cut noise, such as chokes or suppressor caps. This may be difficult with high wattage applications.

Questions:

• Richard Hamilton Le Mesurier asks: " I am looking for digital dimmers to be used ultimately for home automation - but I want to do it all myself - anyone able to help with ideas?"

Interested: