Christmas Light Fader

[Jerry Girard]

Okay, so I asked anyone a few months ago for a schematic of a Christmas
light crossfader using discrete components. Instead, I took several
people's advice and built it using a PIC chip. I did it for the capability
of added features. Anyway, I have the circuit functioning except for one
problem. Please see link to schematic:

http://www.lillyct.com/fader.gif

There are two channels that crossfade back-and-forth between each other
evenly. When cross fading, I am getting cross firing of the adjacent triac
which is causing a random flicker. When I disable one triac or pull off its
load, the other fires perfectly at a nice steady fading rate. With both
loads enabled, I get random flicker. They seem to be doing it on the
negative phase cycle.

I am using Teccor L4004L3 sensitive gate triacs. In the code, I either
tri-state of logic-high the port pins to fire. My layout seems sound. Are
there some additional components I should be using to prevent cross talk?
Thanks.

 

[John Popelish]

Okay, so I asked anyone a few months ago for a schematic of a Christmas
light crossfader using discrete components. Instead, I took several
people's advice and built it using a PIC chip. I did it for the capability
of added features. Anyway, I have the circuit functioning except for one
problem. Please see link to schematic:

http://www.lillyct.com/fader.gif

There are two channels that crossfade back-and-forth between each other
evenly. When cross fading, I am getting cross firing of the adjacent triac
which is causing a random flicker. When I disable one triac or pull off its
load, the other fires perfectly at a nice steady fading rate. With both
loads enabled, I get random flicker. They seem to be doing it on the
negative phase cycle.

I am using Teccor L4004L3 sensitive gate triacs. In the code, I either
tri-state of logic-high the port pins to fire. My layout seems sound. Are
there some additional components I should be using to prevent cross talk?
Thanks.

I assume the word after tri-state should have been or. You certainly
do not want to be leaving the TRIAC gates floating if you want to
avoid misfiring. So do not tri-state the outputs. Swing them between
high and low, only.

I would be interested in seeing your layout. I would run the neutral
wire to the TRIACs first and then to the supply capacitors for the PIC
chip. This will keep any inductive ringing in the load circuit from
getting between the TRIACs and the PIC.

 

[Jerry Girard]

Yes, the word is "or". I tri-state then logic-high the outputs to fire.
I've tried logic-low to logic-high with no better results.

The problem is not with the power supply and the PIC. The problems is with
false triggering of the TRIACs. I've been doing some research and this
seems to be a common problem. The cure is a RC snubber network across MT1
and MT2. I've tried 100 ohm and .1uF and that seems to only make it worse.
So I don't know what the deal is.

The L4004L3 is a very sensitive part needing only 3mA to turn on. I think
they are just too sensitive for my application. I might consider ordering
some other versions that aren't as sensitive.

 

[Robert Monsen]

Yes, the word is "or". I tri-state then logic-high the outputs to fire.
I've tried logic-low to logic-high with no better results.

The problem is not with the power supply and the PIC. The problems is with
false triggering of the TRIACs. I've been doing some research and this
seems to be a common problem. The cure is a RC snubber network across MT1
and MT2. I've tried 100 ohm and .1uF and that seems to only make it worse.
So I don't know what the deal is.

The L4004L3 is a very sensitive part needing only 3mA to turn on. I think
they are just too sensitive for my application. I might consider ordering
some other versions that aren't as sensitive.

I've done this kind of thing several times. Get a MOC3012, which is a triac
output optoisolator. You can run it directly from the PIC, and it will
control the triac quite well.

You can get it from www.goldmine-elec.com or locally, I'm guessing.

The other issue is that when the PIC's gate is on, you are subjecting the
protection diodes to quite a bit of voltage. The gate on a triac seems to
follow the line voltage. You'll fry the port if you aren't careful.

Regards,
Bob Monsen

 

[Robbie Banks]

Okay, so I asked anyone a few months ago for a schematic of a Christmas
light crossfader using discrete components. Instead, I took several
people's advice and built it using a PIC chip. I did it for the capability
of added features. Anyway, I have the circuit functioning except for one
problem. Please see link to schematic:

http://www.lillyct.com/fader.gif

There are two channels that crossfade back-and-forth between each other
evenly. When cross fading, I am getting cross firing of the adjacent triac
which is causing a random flicker. When I disable one triac or pull off its
load, the other fires perfectly at a nice steady fading rate. With both
loads enabled, I get random flicker. They seem to be doing it on the
negative phase cycle.

I am using Teccor L4004L3 sensitive gate triacs. In the code, I either
tri-state of logic-high the port pins to fire. My layout seems sound. Are
there some additional components I should be using to prevent cross talk?
Thanks.

Sorry, can't offer more than has been given below. I take it you are using
the ADC to determine the voltage input from the pots? Would you be prepared
to share this code with me, as I'm looking for something like that.

Cheers
Robbie

 

[Jerry Girard]

Yes, I could share the code, though it isn't any of my better work. It's
written in all C. I would include the HEX file, so if you want to flash it
with that, you won't have to worry about recompiling.

How it works is this:

Fade pot - Controls the crossfading speed from 100mS to 5S.
Period pot - Controls the period between crossfades from 100mS to 5S.

When the pots are turned all the way up past their 95% position, they switch
to pseudo-random timing.

It's actually a nice little device.

I'm still refining everything so I will post is when I'm complete in the
next few days.

 

[John Woodgate]

I read in sci.electronics.design that Jerry Girard <[email protected]>

I'm still refining everything so I will post is when I'm complete in the
next few days.

I predict December 26. It always takes longer than you think. (;-)

 

[Fred Bloggs]

Yes, the word is "or". I tri-state then logic-high the outputs to fire.
I've tried logic-low to logic-high with no better results.

The problem is not with the power supply and the PIC. The problems is with
false triggering of the TRIACs. I've been doing some research and this
seems to be a common problem. The cure is a RC snubber network across MT1
and MT2. I've tried 100 ohm and .1uF and that seems to only make it worse.
So I don't know what the deal is.

The L4004L3 is a very sensitive part needing only 3mA to turn on. I think
they are just too sensitive for my application. I might consider ordering
some other versions that aren't as sensitive.

You can make it insensitive by dropping a shunt resistor off that gate,
and you would do well to listen to Popelish- despite what /you/ may
think, your layout is crap. When you fire a TRIAC there is a surge
current on the neutral, and during the negative cycle this will make the
main terminal of the OFF TRIAC negative with respect to the gate- so it
fires. This is fixed by running two separate wires from the VCC filter
capacitor neutral connection to each TRIAC terminal. You also need to
unbundle and keep separate the hot side wires from the TRIACs to the
bulbs to avoid capacitively coupled cross-talk into the OFF gate- if
this is not possible then a small inductance like 10uH in series with
each hot side lead at the TRIAC-in combination with parallel TRIAC
snubber 0.1U + 100R- to kill the slew-rate- will fix the problem.