Want to build an RGB LED controller
that you can program with your own custom sequences and effects? Then read on.
The RGB LED controller has proved to be
very popular project and has been the most frequently downloaded code on
the site since it was made available. I've been contacted by
people who have incorporated this project into all kinds of things
including mood lamps, lighting for a sculpture, accent lighting for
rooms and an illuminated prize trophy.
For 2006 I completely rewrote the
application making it much easier to add, edit and change the
sequence data. I also added a sleep function so a battery
operated version can be built that doesn't need a power switch.
For 2008 I've released version 3 of the
code which now allows you to stop the running sequence at any point so
you can 'freeze' a colour.
All code runs on the 12F629, 12F675 and
the newer 12F683 which, with 2K of program memory has plenty of room for
user sequences.
The original RGB PWM driver application
that I wrote in 2004 had a few shortcomings. Probably the biggest was
that it was not easy to add to or change the sequences. This new
version addresses that problem, is more flexible and now includes the
ability to put the PIC to 'sleep' and 'wake' it again using the sequence
select switch, eliminating the need for an on/off switch in battery
powered applications.
The circuit uses (RGB) Red, Green
and Blue high brightness LEDs that are pulse width modulated (PWM) to
vary the intensity of each colour LED. This allows effectively any
colour to be generated with rapid changing strobe effects, fast and slow
colour fades as well as static colours. The data used to set
and change the colours is held in an easy to edit file so if you don't
like the sequences provided with it, you can modify the sequence data
include file yourself and reprogram with your own sequences.
The code can be assembled for use with
the following PICs: 12F629, 12F675, 12F683. Just select the
correct processor in the MPLAB IDE before assembling.
When the PIC is first powered on after
programming, it should start running the first RGB sequence found. If
you're using the original sequences supplied with the code here it will
run a sequence of red-fade out, green-fade out, blue-fade out
repeatedly.
Press the SW1 sequence select switch to
step through all available sequences. When the last sequence has been
reached it will go back to the first available sequence. Each time
the SW1 switch is pressed the RGB LED PWM values are set back to 0 (LEDs
off)
Press and hold SW1 switch for about 1.2
seconds to put the PIC into sleep mode. Once in sleep mode, press
the SW1 switch for about 2 seconds then release it to wake the PIC from
sleep. If the SW1 button isn't held for two seconds the PIC returns to
sleep - this prevents the circuit from being accidentally turned on.
When operating the RGB controller from batteries (without a 78L05
regulator) a power on/off switch is no longer needed since the PIC uses
only a few microamps when 'sleeping'
About 10 seconds after the SW1 sequence
switch is last pressed the currently selected sequence number is saved to
non-volatile EEPROM memory. When the RGB LED driver is next
powered on, the saved sequence number is read back and will automatically start
running.
Anytime the PIC is put into sleep mode
by holding SW1 switch down, the currently selected sequence is also
saved to EEPROM.
There are two versions of firmware
available for this project. Version 3 adds a run/hold function
that allows the sequence and colour to be 'frozen' but is otherwise
the same as Version 2. Both versions can be downloaded here
and version 2 will remain available indefinitely.
Version 2 firmware
Download source code
rgbsa-inet.zip (V:2.0.2 09/06/2007)
The source code comprises a single .ASM file and four .INC files.
All five files are required and should be extracted to a single
directory on your computer. To reassemble the code, open the
rgbsa-inet.asm file, then select Project - Quick Build from the
MPLAB IDE menu bar.
When you solder the LED's to the PCB, only solder one
lead of each LED. Once all the LED's are fitted, you
can move the LEDs around to get them all neatly aligned.
Once you have them aligned, solder the remaining lead on
each LED. If you try moving them once both leads are
soldered you risk ripping the copper off the PCB.
Where to get parts
I don't
sell complete kits of parts for this project but you can buy
all the components from Rapid Electronics; to that end I
have put together a full
component list here
Turn water clear LEDs
into diffused LEDs the easy way.
Almost all high
brightness 5mm LEDs come in a clear plastic package, but
for many applications what you really want is a diffused
LED.
Plasti-kote's glass
frosting spray is a quick and easy way to make clear
LEDs diffused.
I've used the frosting
spray on LEDs with this project and it really improves
the colour mixing effect.
Driving large arrays of
LEDs
If you want to drive
large arrays of LEDs using the PIC you need to use a logic level
power MOSFET. The circuit shown below is a generic
driver that will work for most applications. Depending on the
input voltage and the forward voltage of the LEDs used, you may be
able to increase the number of LEDs in each column. You could
also use this type of circuit to control higher wattage LEDs, but
remember the power rating of the LED current limiting resistor needs
to be taken in to account. (power dissipated in the resistor
is = I2R )
Here's a one off
light bar I built using 20 Piranha RGB LEDs and a custom MOSFET
driver board. Assembled into a 25mm x 50mm x 1000mm
aluminium 'U' section. This was fitted under a wall shelf
to illuminate the floor. It's very bright and gives a nice
even illumination without any additional diffuser. Again
this uses the code available to download on this page
I used 0805 SMD resistors, 68R for
the Red and 47R for the Green and Blue LEDs. Running from a 5
volt supply it draws about 1.7 amps max.
A battery operated mini
RGB mood lamp using Piranha RGB LED.
Housed in a four AA battery holder, modified to use three
batteries with the PIC and RGB LED controller PCB located in
the space where the fourth battery would normally fit.
When the mode button is held for over 1.5seconds the PIC
goes to sleep. When the PIC is sleeping pressing and holding
the mode button for 1.5 seconds will wake the it again so
there is no need for an on/off switch.
After I saw these illuminated awards pictured below I
couldn't help but have a go at making one myself.
These were based on my code here on this page.
More photo's of this can be found
here
I bought this light bar from a DIY
store. I've fitted four RGB
lights in it and operate it from a 12volt supply.
Image1,
Image2
The data
used by the application for the RGB sequences is held in the file 'sequenceData.inc'
You can edit this file to add, remove or change the data provided.
You must ensure that it follows the format described. In
particular pay attention to the 'end of sequence' and 'end of all
data' markers and also ensure that each line of sequence data
contains five comma separated entries. (see screen dump below)
In the
screen dump above note the
'end_of_sequence' markers circled in red and the
'end_of_all_data'
marker circled in purple.
You must
have at least one sequence present up to a maximum of 256 individual
sequences, although you're likely to run out of available memory on
the PIC before you reach this limit.
Each line
of data starts with a 'dt' (data table) assembler directive.
All data
is specified using decimal values.
Each data value must be separated
by a comma
The
sequence data on each line has five fields:
Fade
Rate: speed the colours fade from the current values to the new
values. Each step occurs at an interval of 5ms x Fade Rate.
Fade Rate value of 0 indicates the RGB values will be
updated immediately without fading.
Fade Rate value must not be set to 255 except to
indicate end of sequence. (see d. below)
Hold
Time: after fade completes, delay before moving to next line of
data. Interval is 50mS x Hold Time
Hold Time value of 255 following a Fade Rate of 255
indicates end_of_all_sequence data.
Red,
Green and Blue PWM values. 0 = 0% (LED off) through to 255
= 100% (LED fully on)
Typically changes in LED brightness are more noticeable between
0 and 128 than from 128 to 255.
End
of the current sequence data is indicated by the Fade Rate field
being set to '255'. When the application encounters this
it restarts the sequence from the beginning.
At
the end of all available sequence data both the Fade Rate and
Hold Time fields must be set to '255'
After
editing sequenceData.inc the file should be saved and the
rgbsa-inet.asm reassembled. The resulting rgbsa-inet.hex
file can them be programmed into the PIC.
I've recently
bought some
Piranha RGB LEDs from
Rapid Electronics (p/n 72-8998) . This device contains
three LED dies in a clear 4 pin package with 5mm lens. The LEDs are closely mounted and produce good colour mixing when
viewed straight on, If
for the red LED is 50mA
while the blue and green If
is 30mA, however, even driving it directly from a PIC at ~25mA
brightness is very good.
At 20mA outputs
are: Red - 3,000mcd, Green - 4,200mcd, Blue - 2,180mcd.
While this might not sound very bright compared to 10,000mcd and
better 5mm LEDs it actually compares very well; it's much
brighter than the figures suggest. With all three colours in a
single package it performs much better than discrete 5mm LEDs,
especially where good colour mixing is concerned.
Although my PCB
layout can't accommodate this LED directly, my code will drive
the LED quite happily. You'll need to connect GP4 on the
PIC to Vss so the outputs are active low. Currently (March
2007) the price is only £0.66 for one-off quantities which
compares well with buying discrete red, green and blue 5mm LEDs.
April 2007
I've been working on a constant current
source for 350mA Luxeon type LEDs that can be used with RGB LED PWM
controller on this page and the simple serial controller also on this
web site.
Saving
the Internal Oscillator calibration instruction.
For
details on the internal oscillator calibration word, download the
Microchip datasheet for the 12F629 / 675 and read section 9.2 and in
particular sub-section 9.2.5.1
This
only applies to the 12F629 & 12F675. The 12F683 uses a
different method for calibration
Multiple RGB lights not staying in sync This note
follows a query I
had from someone who built these RGB lights. Because the
timing is generated from the internal 4Mhz oscillator and there is
no way to synchronize each device with others, even if they are
running the same sequence, the colours quickly get out of sync with
each other.
If you're building a light bar or any
application where you need more than one unit and they need to stay
in sync you should build the 'serial
controlled RGB driver'.
Oscillator Calibration Instruction
I've had a number of enquiries from people who can't get the code to
work and it has transpired that they / their programmer has erased
the OSCCAL value. Without the calibration RETLW instruction the code
crashes so you must ensure that it is present or the code will not
run.
How
can I recalibrate the internal oscillator on a 12F629 or 12F675?
I also understand that Microchip's PICkit2 Development
Programmer/Debugger has a feature that will calculate and replace a
missing OSCCAL value.
Verify
Errors after programming The RGB light programme code uses the PICs EEPROM to store the last
used mode. When the PIC is powered up for the first time after
it has been programmed it checks to see if the last sequence value
saved in EEPROM is less than or equal to the number of available
sequences. If it is not, the code set the value in the EEPROM to 0.
I had a series of e-mails from a guy in Greece who had lots of
problems with "Verify 0000!" errors. I eventually built the
programmer hardware he was using and ran the software to try and
find out what was happening. It turned out that this
particular combination of programmer/software was powering up the
PIC between writing the code into the PIC and verifying it.
This caused the EEPROM to get initialised to "00" but the
software was still expecting "FF" so the verify failed.
The programmer hardware was
PLMS OziPIC'er
and the software was
IC-Prog 1.05D. I must
stress that apart from this idiosyncrasy this is a perfectly capable
PIC programming setup and once aware of what is happening it is easy
to account for and work around.
(A special thank you to
Panagiotis from Athens - thanks to the Internet, his excellent
English and much patience we got to the bottom of this problem
despite the language barrier and the distance)
Editing, adding changing the RGB sequence data.
All the sequence data is now held in the sequenceData.inc file so it
is really easy to add your own sequences. The application
works out how many sequences are present, where they start and
finish and takes care of page boundary crossing. All you have to do
is put the correctly formatted data into the file, save it then
reassemble.
Please take your time when editing the sequenceData.inc file since
errors are likely to cause the RGB Driver to do unexpected things or
crash.
If you see an 'Warning [220]' error during assembly like that shown
below, you've added more sequence data than the PIC has available
memory.
Clean: Deleting intermediary and output files. Clean: Done. Executing: "C:\Program Files\Microchip\MPASM Suite\MPAsmWin.exe" /q
/p12F675 "rgbsa-inet.asm" /l"rgbsa-inet.lst" /e"rgbsa-inet.err" Warning[220] C:\CODE\RGBSA-INET.ASM 158 : Address exceeds maximum
range for this processor.
If you see an Error [112]
message during assembly check for missing comma separators in the
sequenceData.inc file
