Luxeon blinking fade question

Joshua

Newly Enlightened
Joined
Sep 12, 2002
Messages
53
Location
LA.
I am working on a blinking Luxeon project.
My forward voltage is 9 Volts and I am using a resistor to drop down the voltage for the Luxeon. I want a slow fade effect when the Luxeon turns off. Would the resistor do that for me as well as drop my voltage to safe Red LS/O voltage levels? If not, what would be the best sugestion for getting the red LS/O to fade or slowly die out when turned off?
Thanks!
----Joshua----
 
Might want to add the cap between the resistored LS and the control/supply. Direct-parallel to the LS, with the resistor in the capacitor's charging line would cause the LS to fade on as well. (current-surge into the cap might damage the control-circuit. The Trans/MOSFET should be selected for a high transient current capacity.)
 
Originally posted by Howard Chu:
Add a capacitor in parallel with the Luxeon. Hyperphysics is a good site for the necessary equations.

http://hyperphysics.phy-astr.gsu.edu/hbase/electric/capchg.html
<font size="2" face="Verdana, Arial">I couldn't understand what all the little space were for on that site, so I could not get my info. I need to find out what kind or part number of cap. to get.
My fwd voltage is 9 Volts
My resistor is droping it down to 2.85 Volts.
What kind of cap should I use? Thanks!
 
Originally posted by PsycoBob[Q2]:
Might want to add the cap between the resistored LS and the control/supply. Direct-parallel to the LS, with the resistor in the capacitor's charging line would cause the LS to fade on as well. (current-surge into the cap might damage the control-circuit. The Trans/MOSFET should be selected for a high transient current capacity.)
<font size="2" face="Verdana, Arial">I think a picture might be helpful at this point. You're right about moving the connection, need the cap to discharge thru a resistor to make sure it doesn't dump its energy all at once. Should probably use a 2nd resistor to control charge/discharge time though.
 
Wouldn't it be better to use a cap charge/discharge through a resister to the control voltage on a transistor rather than in the actual circuit with the luxeon? That way you only have to handle the base voltage/current and it can be much smaller. You could add this fairly easily into Odd One's circuit at:

http://www.geocities.com/SiliconValley/Bay/4397/led/led.htm

Hmm, but you couldn't get it to fade on and off without either adding a 555 timer to switch it on and off, or a simple flip flop circuit...

So perhaps it's not all that simple afterall
wink.gif


even doing it this way will only add another half dozen components to the board.

Good Luck,
James
 
It would be better to use a electronic solution instead of just bulk-dump thru a cap; but this will add parts - its a trade off; add big cap or lots of parts

The most eff way to do it and keep the output FET cool, would be to run the MOSFET in Class-C operation and do variable pulse width modulation controlled by a simple CPU, I could do it with just two parts
 
True, the mosfet will heat up and waste power if it's used for dimming instead of just on/off. If you only on/off the things they can carry quite a bit of power with no heat sinking at all, but dimming is another matter.

The PWM is a good idea, but for people that don't have access to the tools and skills necessary to program a PIC, you could alter the other circuit at odd one's site and use a capacitor charge/discharge circuit to alter the frequency of the 555 timer doing the PWM. Hmm, more and more parts. Perhaps I will have to invest in that PIC programmer after all
wink.gif


-James
 
There's really not enough info given to answer this question satisfactorily. You don't know what the "turn-off" mechanism is - if the 9V supply is simply shut off, there's nothing to drive your microcontroller either. Also there's no mention of how slow the fade off should be, whether a fade on is acceptable or whether it must be instant on. There's also no mention of the blink rate, the period of the on/off cycle.

Let's see, for a 9V source, assuming 350mA to drive the Luxeon at 2.85V, you need a resistor to drop 6.15V, which gives 17.57ohms. The closest standard value is 17.6, so let's use that.

Wire up a .1F cap with its own 10ohm resistor going to the +9V supply. So it will charge to 9V when the power is on, with a time constant of about 1 second, and it will take about 4 seconds for the cap to approach full charge. Assuming the power supply is sufficient (max current will be .9A to charge the cap, plus the .35A for the LED) then the LED will be instant-on and the cap will charge.

When the 9V supply is shut off, the circuit will be a loop with the 10ohm and 17.6ohm resistors in series. The cap will discharge more slowly, the time constant will be 2.76 seconds. At a time t=1 second, the cap will have discharged 30% of its energy, pushing 6.26V at 227mA. At time t=2sec it will be down to 4.36V at 158mA. It should provide a nice, gradual fade until it reaches the LED's cutoff voltage.
 
Forget the PIC, check out the 8051F300
complete microprocessor with everything and a NORMAL instruction set

Part comes in DIP and something smaller than a Asprim
 
I Purchased the $99 development kit and 5 DIP package units (hard to solder to the MLP11 package)

I am *VERY* impressed

The CPU has everything built into it, and just requires power and ground to run, you program/debug it in circuit using 3 pins
 
Im sorry, I forgot to give the details. I need the red LS to come on at once and have a 2 to 3 second fade when shut off from the 9 Volts.
 
Hello Josh,

The simplest method i can think of right now is
to build a capacitance amplifier. This allows
you to use a relatively small cap and get large
delays like you are looking for.

One way to build the 'cap amp' is using an
NPN transistor (like maybe the 2N3055A)
and one electrolytic capacitor.
Connect the collector to v+ and the emitter
to the LED through it's dropping resistor
(for an LS at 9v this would be about 15 ohms)
The capacitor (+ lead) connects to the base
and the cap (- lead) to ground.
The size of the cap would be 1000uf, 10v.
To turn on, short the collector to base of
the 2N3055A. To turn off, remove the short.

If you dont mind adding another transistor
to create a 'darlington compound', you can
get away with a cap value of 10uf, 10v.
This second transistor (2N2222 or equiv) would
have it's collector connected to the 2N3055A
transistor's collector, and the emitter of the
2N2222 connected to the base of the 2N3055A.
The cap (+ lead) then connects to the base of
the 2N2222 instead.
To turn on, short the collector to base of
the 2N2222. To turn off, remove the short.

In both cases, the lower base current is supplied
by the cap. Since the base current is lower
then the emitter current, the cap discharges
much slower then a direct connection. This
gives the long delay effect.

If the circuit isnt turned on and off very often,
no heat sink will be required as the transistor
only dissipates heat during the relatively short
delay period (2 to 3 seconds).

total number of parts including the series resistor:
2 parts for the single transistor, 1000 uf circuit, and
3 parts for the darlington, 10uf circuit.

Just in case you find the circuit doesnt turn off all
the way when you remove the short, connect a 1Meg resistor
across the capacitor.

If you need a schematic let me know.

INRETECH:
Are you sure you got the F300 chip in dip
package??

Good luck with your LED circuits,
Al
 
The normal package for the F300 is MLP11 - about 0.150x0.150 but they also have a eval version in a normal 14 pin DIP (0.1 centers, 0.3 wide) for $10.59 each

I purchased 5 of them and I am going to order more, they are so incredible; and yes - I have put one inside a 2AA MAGLITE with a lot of nibbling/cutting
 
Back
Top