I remember seeing a question similar to this some time ago, but when my
mailboxes grew out of control I deleted old messages. I need to make a
small light bulb (like a grain of wheat type bulb) flicker randomly,
similar to a candle. I would prefer code that doesn't use hardware PWM
(because I have a PIC16C84-10), but I will even go so far as to order a PIC
just for this. Also, does anyone know how long 4 AA batteries will power a
grain of wheat bulb? Many thanks in advance for any help you can provide.
Well, I know that you want to do it with a PIC, because there is an easy
way to do an incandescent lamp flickering using a Triac and a noisy
diode as triggering level...for sure for 110/220 Volts lamps... :)
Wagner
Dennie Lee wrote:
>
> I remember seeing a question similar to this some time ago, but when my
> mailboxes grew out of control I deleted old messages. I need to make a
> small light bulb (like a grain of wheat type bulb) flicker randomly,
> similar to a candle. I would prefer code that doesn't use hardware PWM
> (because I have a PIC16C84-10), but I will even go so far as to order a PIC
> just for this. Also, does anyone know how long 4 AA batteries will power a
> grain of wheat bulb? Many thanks in advance for any help you can provide.
> I need to make a small light bulb (like a grain of wheat type bulb)
> flicker randomly, similar to a candle. I would prefer code that
> doesn't use hardware PWM (because I have a PIC16C84-10), but I will
> even go so far as to order a PIC just for this.
You absolutely, certainly do not need PWM hardware for operating light
bulbs. BTW, I don't think you need a PIC either. The basic PRBNG to
perform this, clocked at 50 Hz, can be readily implemented with a 4024
and 4093 plus 2 resistors, capacitor and one driver transistor.
Of course, you can do it with a SMD 12C509 but you still need the
driver transistor. I suppose you might possibly get away with four or
more outputs in parallel.
Use three AAA cells, not four, if using a PIC. Or use four and a
diode or two in series with the PIC, not the lamp. You don't really
want white light for a candle anyway!
--
Cheers,
Paul B.
I posted some stuff on this a while ago - reference to a magazine
article (or two?).
I'm out of town at the moment - email me in 3 days time if the
following isn't enough and I'll refer you to the original
article(s?).
This all from memory so may be variably accurate:
A "flaming torch" was done with apparently superb reality and it
seems the same result should apply to a candle. I want to try this
soonish myself. They used two monostables with a fast and a slow
frequency - one about 2 Hz AFAIR and the other say 10 Hz. They
combined these (OR?/AND?) to turn a bulb either fully on or off
AFAIR. The time constant of the bulb is such as to never go fully out
on off cycles. They used cellophane around bulb and at a distance
this was apparently indistinguishable from real flaming torches.
If doing with a PIC how you did it would depend on what else it was
doing. If dedicated to this task you simply run two counters
(incremented variables) that count up to 2 different totals and then
reset. Bulb is off except when both counters are in upper half of
their range.
eg pseudo code
Bulb On
Countmax1 = 1000 ; say
Countmax2 = 107 ; say
LOOP
Counter1 = Counter1 + 1
Counter2 = Counter2 + 1
If Counter2 >= Count2max then Counter2 = 0
If Counter1 >= Count1max then Counter1 = 0
If Counter2 > Count2max/2 and Counter1 > Count1max/2 then
Bulb Off
else
Bulb On
Endif
Delay for a while
goto LOOP
Note that the two countmax limits should not be related by an integer
factor to prevent repeated short term flicker patterns.
There are actual light bulbs that simulates candle light,
made of two (just) touching filament shaped like a candle flame (teardrop)
When its conducting, the filament quickly becomes heated and
becomes further apart from the other. It becomes open circuit
and the filament cools down, only to touch again.. repeatedly.
It's very realistic. I've seen them used for decorative purposes
and even on religious altars.
I also saw that kind of lamp for sale, but according to its own
characteristics of physical oscillation and strong burst of current
during the on/off, it *should* has a shorter life... I think.
By the other side, a regular incandescent lamp receiving a variable
current, but with a minimum current (at least) to keep the filament hot,
should last longer...
Something is being used for long time... a simple triac acting as the
regular wall switch, that turns the lamp smoothly from zero to 100% in 2
seconds, can span the lamp life almost the double... or more... since
the lamp never receives a high burst of current (when you turn on the
switch), mainly when the lamp filament is cold, so with a lower
resistance, what sucks more current than when it is hot. The "low
start" triac system just apply some "current cushion" to the lamp's
life... :)
>
> There are actual light bulbs that simulates candle light,
> made of two (just) touching filament shaped like a candle flame (teardrop)
>
> When its conducting, the filament quickly becomes heated and
> becomes further apart from the other. It becomes open circuit
> and the filament cools down, only to touch again.. repeatedly.
>
> It's very realistic. I've seen them used for decorative purposes
> and even on religious altars.
>
> I think a PIC would be and overkill here.
>
> Regards,
> Peter
>
> There are actual light bulbs that simulates candle light,
> made of two (just) touching filament shaped like a candle flame
> (teardrop)
The description of a "filament" shaped like a teardrop sounds a bit
odd to me, and bimetallic switch flashers are generally a bit slow.
You're not confusing this with a standard "flicker flame" neon are
you? These have been used for most of my lifetime(!) and contain a pair
of parallel "teardrop"-shaped electrodes excited by the AC via a
resistor in the base. I presume the facing surfaces of the electrodes
are covered by an insulating coating.
Since the neon glow discharge area is directly proportional to the
current flowing, in these lamps the current is such as to light only
half the area of the electrodes. With the electrodes flat and parallel,
and I suspect dependent on heating effects and the electrode coating,
the glow is unstable and rapidly moves to occupy different areas on the
electrodes, causing the "flicker" effect.
If due to damage or manufacturing error the electrodes become other
than parallel, the "flicker" effect is compromised or lost.
Now you've reminded me. They used to be sold in the radio parts shops
and I've never owned one, always wanted to. An associated device was a
glow lamp with flower-shaped electrodes in which the leaves glowed green
and the flower, orange or red.
--
Cheers,
Paul B.
There is a lamp 110Vac tear drop style, with a funny shape filament,
that keeps bouncing and flickering... it is common to find it in special
arts and hobby stores. It bounces from one side to another and
flickers, trying to simulate a candle flame in the wind... well, it does
not a good job, except if from distance, since the filament rithm is
almost predictable. The interesting point to see a lamp in action is
that the filament really bounces almost one inch from one side to
another inside the lamp... it is curious at least.
>
> Peter Tiang wrote:
>
> > There are actual light bulbs that simulates candle light,
> > made of two (just) touching filament shaped like a candle flame
> > (teardrop)
>
> The description of a "filament" shaped like a teardrop sounds a bit
> odd to me, and bimetallic switch flashers are generally a bit slow.
>
> You're not confusing this with a standard "flicker flame" neon are
> you? These have been used for most of my lifetime(!) and contain a pair
> of parallel "teardrop"-shaped electrodes excited by the AC via a
> resistor in the base. I presume the facing surfaces of the electrodes
> are covered by an insulating coating.
>
> Since the neon glow discharge area is directly proportional to the
> current flowing, in these lamps the current is such as to light only
> half the area of the electrodes. With the electrodes flat and parallel,
> and I suspect dependent on heating effects and the electrode coating,
> the glow is unstable and rapidly moves to occupy different areas on the
> electrodes, causing the "flicker" effect.
>
> If due to damage or manufacturing error the electrodes become other
> than parallel, the "flicker" effect is compromised or lost.
>
> Now you've reminded me. They used to be sold in the radio parts shops
> and I've never owned one, always wanted to. An associated device was a
> glow lamp with flower-shaped electrodes in which the leaves glowed green
> and the flower, orange or red.
> --
> Cheers,
> Paul B.
I've seen such a light, where the filament is attached to rather long wires
inside the lamp. I remember seeing a small ring magnet glued to the top of this
bulb. Apparently the filament support wires had a resonant frequency of around
60Hz, making it dance wildly back and forth below the magnet. It is a neat
idea, but I wonder how durable the filament is, the lamps were large and selling
for about $20 US each. I imagine one could hide a more powerful magnet in the
base so it wouldn't be seen.
>
> No Paul, he is not confusing it.
>
> There is a lamp 110Vac tear drop style, with a funny shape filament,
> that keeps bouncing and flickering... it is common to find it in special
> arts and hobby stores. It bounces from one side to another and
> flickers, trying to simulate a candle flame in the wind... well, it does
> not a good job, except if from distance, since the filament rithm is
> almost predictable. The interesting point to see a lamp in action is
> that the filament really bounces almost one inch from one side to
> another inside the lamp... it is curious at least.
>
> Wagner
>
> "Paul B. Webster VK2BZC" wrote:
> >
> > Peter Tiang wrote:
> >
> > > There are actual light bulbs that simulates candle light,
> > > made of two (just) touching filament shaped like a candle flame
> > > (teardrop)
> >
> > The description of a "filament" shaped like a teardrop sounds a bit
> > odd to me, and bimetallic switch flashers are generally a bit slow.
> >
> > You're not confusing this with a standard "flicker flame" neon are
> > you? These have been used for most of my lifetime(!) and contain a pair
> > of parallel "teardrop"-shaped electrodes excited by the AC via a
> > resistor in the base. I presume the facing surfaces of the electrodes
> > are covered by an insulating coating.
> >
> > Since the neon glow discharge area is directly proportional to the
> > current flowing, in these lamps the current is such as to light only
> > half the area of the electrodes. With the electrodes flat and parallel,
> > and I suspect dependent on heating effects and the electrode coating,
> > the glow is unstable and rapidly moves to occupy different areas on the
> > electrodes, causing the "flicker" effect.
> >
> > If due to damage or manufacturing error the electrodes become other
> > than parallel, the "flicker" effect is compromised or lost.
> >
> > Now you've reminded me. They used to be sold in the radio parts shops
> > and I've never owned one, always wanted to. An associated device was a
> > glow lamp with flower-shaped electrodes in which the leaves glowed green
> > and the flower, orange or red.
> > --
> > Cheers,
> > Paul B.
You can get PWM "for free" if the code is complex enough that
there's somplace in it that gets visited at seemingly random
intervals. Just toggle the output bit at that point.
Also, does anyone know how long 4 AA batteries will power a
> grain of wheat bulb? Many thanks in advance for any help you can provide.
No, but this can be determined in about 48 hours with 4 AA batteries
and a grain of wheat bulb.
