Dimmer

[BMIelektonik]

I use a 16F84 8Mhz.Zero detection with an opto coupler after rectifyer 100Hz and then a schmittrigger to get the levels right. Have a <1ms puls around zero.
It might be a good idea to make the pulse as long as it can be ,and bring it low just before zero. Sharp "s216s01" recomends 1mS min ,
Assembly might be the best way of solving this ,since my goal is to control 5 outputs , I start with one then work my way up

thanks

[George]

I use a resistor as zero detect, 1 meg from phase straight into the port, I then have a 470K resistor on the pin next door to it and swing the bias in an opposite direction to the mains ie:

GPIO.5 is the output pin with the 470K on it - trigger is GPIO.4 which is connected upto the 1M which goes straight to phase. The triac is triggered by a BC327 and i trigger by turning pic from high impedance to low by changing from input to output.

LowSide: 'look for trigger points and trigger triac

IF trigger = 1 Then lowside 'wait for phase to go low
GPIO.5 = 0 'weight input to sense zero crossing
PauseUs dim 'wait for dim period
Output triac 'fire triac
triac = 0
PauseUs 400
Input triac


HighSide:

IF trigger = 0 Then HighSide
GPIO.5 = 1
PauseUs dim
Output triac
triac = 0
PauseUs 400
Input triac

GoTo start

[sougata]

Hi there,

Having the opto-triac driven till the zero cross would take care of inductive loads also.
Provide me with the following :

1. How the outputs would be controlled? (Internal timer /push button etc)
2. Your line frequency

I would try to play with the interrupt part simulate it.

Regards

Sougata

[BMIelektonik]

What i wanted to do is a controller for some decorative lights. I want to make a pattern where light fade in / out.
Powersupply = 50 hz 220V

My first plan was to fade in/out between 2 limits on 1 channel "this works ok".
then i wanted to try 2 channels to push the limit "a bit"
Now I can see this extend to like 5 output creating a wave like motion.
may be it will end up with a lot of assembly code and also lower the demands
on "linear control" and go over to fixed steps and a data table


could even parallell 5 pics to achive my goal this would be the last solution

Thanks

[sougata]

Hi,

Understood. The only problem is that I don't have 16F84 ready. However it should work. Do you have a PIC18F452 ready? This is where I am most comfortable. I also have 16F676, 16F73, 16F619 handy. BTW you don't need five PICs.

Regards

Sougata

[Ioannis]

I have not done this but may be my idea could help.

Using Interrupts, when there is pulse from zero-crossing detector, you could succeed in your house keeping, timing that is.

Entering the ISR, the relevant registers would increment or decrement according the flags you have set in your main program. So there is accurate timing to fire every SCR even 5 or more since the ISR would be triggered every 10ms.

Ioannis

[sougata]

Hi Everybody,

It is possible to trigger n numbers of SCRs within the 10ms(50Hz) time frame. My DMX Dimmer handles 48 of them. There I use a bubble sort algo and actually different interrupt sources. One for syncing the timer with ac, one for firing, listening to USART, fader read and so on. All this handled by two PICs 18F452 @ 40MHZ. The second one is mainly for the download and running of pattern or randomizing them.

I am working on a simplified version using a PIC16F84A (Yes I found one) and trying to acheive a 100% PBP code using the On-Interrupt function. I have no idea about the latency till I get my code running tomorrow and checking on scope. Even the asm interrupt would be quite easy.

You see with a clock frequency of 8MHz--> 2MHz(Fosc/4) / 128 (prescale) / 100 (reset every 10miliseconds), the timer0 would count 156 (50Hz) or 130(60Hz max). Compare this constantly in a loop with preset or increasing / decreasing values and fire the appropriate port.On interrupt

1. Clear all outputs
2. Reset Timer0
3. Return

Thats it. It will be done before the new cycle begins!!!!!!!!!2MIPS

I am working on a two channel version and let you all guys out there juice it up.

BTW most of the PICs have an inbuilt Schmidt Trigger on input ports so the interfacing is pretty easy. I use a couple of resistors a diode bridge and an optocoupler to derive the zero cross sync. It is accurate without any phase latency (as in the case of transformer based).Direct AC feed through high value resistor is also possible and practical but the circuit becomes live and is not recommended for experimenters.(I use it in my remote controlled fan/light dimmer to keep costs under control. It retails for about 18USD!!!!)

The toughest part is possibly randomizing the paterns or using pschychedelic effect with beat detection within the scope of the chosen processor (clock code space, instruction set).

We are having Diwali (the festival of lights) and a holiday to work with a fun project. The entire town is having myriads of lighting displays so why not doing one for the forum. Happy Diwaly to guys out there.

It is midnight now and I can get my hands on the PIC only tomorrow when I get the lunch break. So till then....

Bye!!

Regards

Sougata