Slashdot Mirror
Building Your Own Glowing Cyber-Balls?
krezel asks: "So I've been drooling over the Ambient Orb, a cool little gadget 'glowing ball' that you changes colors based the 'health' of things you specify. It can do stuff like fade from red to yellow to green as your stock portfolio improves. However, being a poor college student I can't afford its $200 price tag. I've found lots of sources for super bright multi-color LED's. Cast a couple of them in some translucent resin, hook them up to a power source, and you've got yourself a cheap glowing ball. But I've yet to find any good information on how to build hardware that will let me control relays for devices like this through my serial or parallel port. Basically I'm looking for a cheap way to build a board that will let me control 4-8 relays (for each color) over my serial port, and some info on how to write the software for it. This could be a very cool project, and I plan on making the plans available, and the code Open Source, when I'm done with it. Any ideas?"
Don't you want more than 8 colors? If you use relays, you can only turn on or off each of the red/green/blue colors. But if you vary the current through each led, or vary the duty cycle by pulsing the leds quickly, then you can get more colors (like 24 bit color!).
--Tim
I just wanted to congratulate you for getting the phrase "glowing cyber-balls" on the front page of Slashdot!
When I was looking for computer->analog control chips a few years ago, the best methods I could find were:
Build (or buy) a serial->I2C or parallel->I2C converter; you can get D/A chips with I2C interfaces pretty cheaply.
Use a PIC microcontroller, which gives you serial and analog I/O built in.
the Ambient Orb runs off a wireless network... no computer needed, and you can control it from anywhere in the world (theoretically). To manage that, you'd have to build an 802.11b -> relay interface, at least - if not a cellular one.
Now, assuming you don't want to muck about with that (and who does), your best bet would be to not use relays in the first place - they're loud, slow, and not gradual. Use a Basic Stamp from Parallax and write some code to output a PWM (Pulse Width Modulated) voltage to three different pins - one for each color. (Chances are you'll be using either one 4-pin, 3 color LED or 3 leds (red, green, blue). Infrared or UV leds could be interesting, but aren't recommended...) Then you can either leave the BASIC stamp connected to your serial port and controlled via DEBUG or SERIN (IIRC) commands from your host computer (and write some corresponding code for the host), or you can leave it standing alone and interface to it using any one of the who-knows-how-many add-on boards Parallax sells. (you might want to check out the Communications page - that modem looks like a good thing to try)
This
A solid state relay (S101S05v) is not a mechanical switch.
several plugins for XMMS may get you started. They have plans for the hardware to connect a parallel port to leds, stepper motors, whatever.
I Don't Work Here
Get ahold of a cmos 4066 this chip has plenty of switching power throughput to handle a few leds. Hook the triggers to your parallel port and code a simple pulse width modulator routine to run the triggers. I don't know if the frequency you can achieve on a parallel port would be high enough to prevent flicker, but some capacitors should go a long way to smoothing that out, if not. You might want some current limitin resistors in there as well, so you don't burn out your leds.
BTW, this is rudimentary circuit design in almost any college course. If you want to seroiusly get into building cicuits like this check out "The Art of Electronics" by Horowitz and Hill. Might be a little expensive but will give you a solid foundation in circuit design.
-- The morphemes of your disquisition are ascertainable, but they have eschewed an ambit of transpicuous exposition.
Having done some large castings in casting resin (clear and with opaque or translucent dyes), I can tell you that it's not all that simple to just cast a ball that size either. The casting material is going to be expensive to begin with. And if you don't get the hardner mix ratio just right, that stuff it going to crack and craze like crazy (split a few "paper-weights" in half). It gives off heat (from the chemical reaction as it "cures") which can damage really thick objects, like a 6 inch ball. I'd be willing to bet that what they have is not "hobbiest grade" casting material. It's more likely commercial grade plexiglass type material with a translucent dye added. It might not even be chemically cured like epoxy resins but may be cure thermally or by UV light (former - likely, later - possible but highly unlikely). Plexiglass resins become soft and pliable as you warm them (within reason - moderately high heat burns them easily) but casting resin does not - it cracks and crazes and shatters. The dye would be similar to the casting dyes you would get at a hobby shop. You MIGHT be able to cast a ball that size, if you are lucky, in casting resin but keep it away from large temperature changes and bright sunlight (which damages through both large temperature gradients and UV breakdown damage). You may find that this isn't a cost-effective "do it yourself project" after all.
I've used something like this as well, except to drive relays that drive reset buttons on a series of machines at uni.
I agree with hotair's use of the ULN2803 chip; however, I think you need a 74HC373 octal latch. Basically when you send a byte to the printer, you need the latch to "catch" the byte and hold it after the signal goes away. The latch is controlled by 1 of the printer control lines that goes low when the data on the bus is valid (I think it is the STROBE pin).
If you want more than 8 leds, you could probably use a 74hc138 (3 to 8 demultiplexer) to control a bank of latches, but that would require alot more thinking on my part, so I'm leaving it with you.
The engineering staff has been good to work with as well.
Also, they seem to sell a product almost exactly like what you describe, with bright LEDs in a diffraction grating, based on the same chipset. I don't know if it has quite the diffiusion you're looking for. (But it does have a buzzer!)
Otherwise, my advice would be to use the parallel port (very easy to program, unless you're a mac user and you don't have one ;-) and don't use relays. In order to drive a relay, you'll need a transistor to switch the coils, and if you've already got the transistor, well, you can see where thats going!
No I don't work for them, but http://www.ibutton.com gets you to devices that can be controlled through a serial (or parallel) port and are cheap. The DS2407 is a switch you can use to control the LEDs. Also check out the TINI links on the page for a Java JVM on a SIMM which can be used to control the LEDs and connect to the net. Nice stuff to work with.
Parallel port devices are fairly easy to program and create (I'd have to say THE easiest), most basic linux programming books go into how to do it.
Considering the simplicity of the circuit to make an LED glow, and the ease of parallel port programming, plus the relative ease of finding information on how to do it I'd have to concur with you.
Plus the low power consumption would probably mean you could forgo any external power source unless you either wanted it brighter or wanted to light up a bunch of LEDs at once. I'm pretty sure you wouldn't need an external power source at all under most circumstances, but I've been up late and shouldn't be wasting time posting on slashdot.
I'd think he should start by looking at some of those overclocker parallel port mini-display information, that'll get him started.
Then just pick up a book on electronics (beginner would do I'd say), the parallel port specification, and some driver knowledge. Boom he's in business. I don't think theres a bit of information you'd need to do this that couldn't be found via google, so let's go for it.
The key to the enjoyment of pop music is to replace any instance of "love" with "C.H.U.D."
Check out Circuit Cellar Magazine -- they are a steady stream of articles and advertisements covering just the thing you want to do.
While you're reading it, also pay attention to PIC Chips and Basic Stamps, which would be a great way to control your orbs without needing a PC (especially the cheaper PIC chips from someone like Microchip Technology)
If you're married to the PC concept, you'll also find advertisements for devices which are controllable via USB. Kinda nice for furure serial-less PCs.
Lastly, though it's a bit out of date at this point, take a look at "Controlling the World With Yor PC" by Paul Bergsmann (ISBN: 1878707159). Great stuff about parallel port interfacing.
Good luck!
One thing to remember though is that you're not allowed access to the ports under Win32 NT-family kernels except through a Ring 0 driver. That can get a little ugly.
Google for "Beyond Logic" and you'll find a site that lays out more info on the legacy ports and on making peripherals for them (and for USB) than you could read in a day.
(Mind, it's pretty easy to make a serial one too. There's a UART called the CDP6402 that's specifically designed to run without a master uC; just add an osc to get 4x the desired baud rate and use an octal latch to maintain the output with an RC circuit to generate the rcv ack pulse, and you're set.)
I was assuming he was talking about doing it under Linux. NT is a bit different. I always seem to assume people who do this type of thing run Linux (I must read too much Slashdot).
That's an amazing site by the way, it's right here since you didn't supply the link.
The key to the enjoyment of pop music is to replace any instance of "love" with "C.H.U.D."
Lotsa links here...
First of all, the 2002 Burning Man project I did that involved a couple hundred RGB LEDs spinning in a persistence-of-vision-based nighttime animated display. Here is the best picture of it. This is the page about the development details.
The LEDs I used were manufactured by Kingbright. The model I used, the LF819EMBGMBC, is big (10mm) and relatively bright for an RGB LED. I couldn't find any U.S. retailers that actually told the truth about whether they stocked them, so I ended up buying 400 directly from Kingbright for I think a little more than $2.50 each. I still have a few left.
Atmel AVR microcontrollers are just a few bucks each, easily programmable with the STK-500 programmer, also cheap at around $80. I used the ATMega8, which was more than sufficient for my needs. I imagine the original Slashdotter could use one of the ATTiny MCUs, since it really needs only 3 or 4 I/O lines (fewer depending on how many helper circuits you decide to use).
The boards were manufactured by PCBExpress and I was very happy with them. The CAD/CAM software was Eagle, which except for some crashing/redrawing bugs was really amazing. The version I used was free. I tried to buy it but CadSoft has (had?) a fairly crazy pricing scheme that actually left you worse off in terms of acceptable usage if you paid them money than if you used the free version.
The best part of using the Atmel MCU was that GCC can cross-compile for it. So you're basically writing regular old C code but it runs on a little tiny piece of silicon. You'll want to subscribe to the quite active avr-gcc mailing list. Save every message from Marek Michalkiewicz; in my opinion he's the god of GCC-for-AVR development.
1) These are available from Brookstone for $150...
2) Delcom Engineering has a "USB Visual Signal Indicator". This includes RED, GREEN, and BLUE LEDs, Piezo buzzer, 2 meter USB cable and USB powered circuit. The cost? $69.00 each... All you need to add is a globe...
If you want to play around with this stuff, Delcom Engineering also makes USB chips, cables, etc. and they make USB development board for $49.00 that you could you could use to build what ever you wish...
download a set of drivers called "PortTalk", they work very well for allowing programs to access the parallel port under nt/2k/xp and they include quite a bit of sample code for doing it also.
In terms of hardware, you'd need the jack for the port, the power cord, a PIC chip, a transistor for each color, and a bunch of LEDs/resistors of each color. Easy stuff really. Let me know what you come up with. If you use USB, I'd be especially interested, as I have a Mac (no serial port). Good luck!
I hereby place the above post in the public domain.
PortTalk and Port95NT are closed source and produced by a commercial company, whereas there are better libraries such as WinIO that are BSD-license.
And, you'll find that most of the code out there just blindly writes to the data bus at 0x0378. The more complex stuff, such as scanning the BIOS configuration area for addresses and then doing ECP config register writes for each one, and working with the status/control pins, are a bit beyond the average Perl hack.
And then, there's the actual art of getting the info you need.
I've actually done this with eight LEDs (four red, four green) and the parallel port on my firewall machine. I did it mostly because it had a cool smoked perspex cover that slides over the drive bays, and the LEDs shining through it looks pretty cool. Especially when you make them flash and do stuff. Have a look at some pics.
In this case, the LEDs have their anodes connected together, and brought back to the 5v rail, and their cathodes connected to the parallel port pins through 220 ohm resistors. You then bring the appropriate pin low to turn the LED on. There's a good reason for doing it "backwards" - the gates driving the parallel port can sink more current than they can source. That is to say, the transistor pulling the pin to ground is "stronger" than the one pulling it to +5v, so it's more suitable for turning on an LED. You could use two resistors and a small transistor per pin, too. If you like.
Or, if range isn't an issue, how about a DIY bluetooth module for the wireless connectivity?
BTDesigner.com seems to be selling 20m-range modules for about $60.
Anyone know of cheaper sources? By now, Bluetooth modules were meant to cost around $5/module (in large quantities).
Motorola currently has a design contest going with
their MC68HC908QT4 8-bit microcontroller. This
puppy is an 8-pin DIP, FLASH EEPROM programmable,
and the pins appear to be PWR, GND, and 6 I/O
pins. (One of them may be a clock, or it may be
set up to accept a clock if you give it one.)
Start at http://www.circuitcellar.com for details.
www.maxim-ic.com
A data sheet
also check out the related products on that page.
maxim are helpfull with there sample policy. If you were to connect an array of red, green and blue LEDs inplace of the digits, you can change the brightness of each bank of them with 8 brightnesses,
In the data sheet they talk about 127 colours with bi-coloured LEDs, if you had tri-coloured then you would get... ooooh.. 16.8 million.
can be loaded by bit-banging the SPI or I2C interface from a printer port. Im sure someone has made a linux driver for it. Some code to do that was on there site when I looked but I cant find it now..
As far as the PC side goes, just a program to output individual bytes to the serial (or parrallel) port as the backend, and you can do all the "detrermine-what-LEDs-to-light" calculations in a heavier portion of code, updating itself every 60 seconds or whatever from whatever data sources.
As for the microcontroller, it's fairly trivial to write the code to control the lights, unless you want to put in stock-price-fetching/web-page-parsing functionality on the chip. Like I said, i'd recommend doing that on the PC and just have the microcontroller doing the LED-lighting work. If you've done assembly programming on Intel or most Motorola chips before, it's not hard to pick up (very different though from most RISC designs, like MIPS) but if you're not that great with assembly you can build gcc to output code for the HC12.
Regardless of how you generate the bytecode, you'll have to simulate it, and have access to an EPROM writer, and maybe some RAM. The way I implemented mine was to have the initial "boot program" on EEPROM which made a request to the PC for the actual program. The PC-side caught the request, and sent a copy of the 'real' program bytecode which got loaded into RAM, and the booter jumped to the new code. This let me update the HC12 running-code as I found bugs, and let me add more features without re-writing to the EEPROM
As promissed a note about PCM. I know a bunch of know-nothing-kiddies are going to reply to this post saying "it's a LED, it's either on or off, stupid" but as anyone whose done 1 year electical/computer engineering there are 2 ways (that i know of) to control the apparent brightness of LEDs. First, controlling the current flowing though said LED (which was not done in my HC12 LED controller implementation, but could have been with creative use of output ports and resistors), and the second is with pulse code modulation. For PCM, if you want a LED on at 80% brightness, you simply output a '1' 80% of the time, and '0' 20% of the time. The cycle repeats fast enough (remember that we're running on a 16MHz clock which (although it seems puny now) it insanely fast compared to the human eye) so that our eye sees that as 80% the brightness of an "on" LED.
easy, just hook 4 of the data pins rrom a parallel port to one of these
16 line decoders and that will give you control of 16 relays. add a second chip to the next 4 pins, and control 32. Once you have it set up, just write the value you want represented (ie, to turn on the first device, write out a 1, second device 2, third device 4, etc etc) out to the port.
Where's my lobbyist? Right here.
Ok 2 not so easy ways to do it...
serial port method.
Get a pic microcontroler with built in serial port and a DtoA converter. Connect each (1 red, 1 blue, 1 green) to its own individual DtoA output (with a current limit resisor picked so 5V results in full led brightness ((5V-VfLED)/(IledCONT)= R). next hook your serial port to the pic via a RS232 to TTL converter (MAX232 is a good choice). write some simple code to read 3 values from the serial port and dump them to the DtoA reg (too long to post here send me Email for help or look at the ap notes on microchips web sight) and tada you have a serial controled glowing ball....
10baseT ethernet (UDP/IP) method.
go to
http://home.att.net/~wireb/niode/niode/index.html
build a niode and hook the led to that. I am currently working on new commands to allow the DtoA converter to work for a use similar to this (controling fiber optic lights in celing tiles). A lot more work but you would be able to control it from anywhere in the world that can send UDP trafic to it.
Later
Wire
I picked up Designing Embedded Hardware a couple of weeks ago; being a software geek, with no hardware experience (except for assembling PC's, etc), I found this book to be amazing.
I highly recommend it.
If you want a quick fix on how to wire stuff for parallel port, I believe that there are plans in the book "The Robot Builder's Bonanza" to do this. It's great, it's only about $30, and one of the professional societies on your campus probably already has a loaner copy. It's available on amazon.com. Get super-saver shipping if you're on a budget, reply to this with an e-mail addy and I'll even write software for it if I think that your design is sound ;-)
I guess the poster has two options:
Parallel port - Eats his parallel port, and needs lots of wires to be run.
AVR - More complex to design, but easier overall.
I don't see why you reccommended an oscilliscope - For most uC development one isn't needed. (Can be nice to have though!)
Parallel-port AVR programs are simple and easy to build. Atmel did a good job as far as AVR programmability. You can program most AVRs in C with good results, either using GCC, or one of a few other compilers. (As much as I like GCC for most development, Codevision is an EXCELLENT IDE/compiler that is well-maintained. Cornell uses Codevision + AVRs for their microcontroller class, and Prof. Land has had 1-day turnarounds from bug report to bug fix from the author of Codevision.)
For those who are lazy and have some extra cash, the STK500 devel board is $89-109 or so depending on where you look and is wonderful for prototyping circtuits. (Has built-in switches and LEDs, TTL-to-RS232 converter for AVRs with a UART, and a serial port based programmer.)
Some AVRs have built-in internal RC oscillators - Not as stable of a frequency reference, but for many applications (such as this one), it doesn't matter, and it's much easier to use.
Most AVRs have only one or two PWM outputs, and VERY few if any have three, which isn't as serious of a limitation as it may seem - Since the PWM rate only has to exceed 100 Hz or so, he can easily do 8-bit PWM with a software loop. (Essentially, having 8 or 16 software-driven PWM channels.) 8-bit PWM would probably be overkill for this application, even 4-bit might be more than sufficient.
AVRs are also excellent for driving parallel LCD displays, although the software is much more difficult. (It's easy with Codevision, except that only the commercial versions of CV can use the LCD library.)
retrorocket.o not found, launch anyway?
Hmm.
$299 list.
$199 Thinkgeek
$150 Brookstone Ships week of 5/15/2003
Not sure if building your own worth the effort.
My 2
Actually you can get 12 "output" channels out of a parallel interface if you really want to... the 8 data channels, plus the "strobe", "autofeed", "init", and "select-in" channels (pins 1, 14, 16, 17, but 1, 14, and 17 are "inverted" watch out).
c uit.jpg
I use the parallel port to drive three stepper motors. Pretty easy to do, and your circuit would be just as easy. You could actually use the same circuit, but simply change the +12V input to the correct voltage, and hook the LED's in where the motor phases would be.
See my crude circuit diagram here:
www.lenticularshareware.com/downloads/stepper_cir
FYI, you can also get 4 "input" channels from the parallel port at the same time, I use them for limit switches to set the position of the steppers to "home". (see diagram)
MadCow.
I used to have a sig, but I set it free and it never came back.
Not sure if anyone covered this yet, but wouldn't you want to duty cycle the various colored LED's so that you can blend lots of different colors. You don't need actual relays to switch this level of power do you?
I'm not an electrical engineer, but you might find what you're looking for at AAG Electronica. They have serial port adapters, sensors, switches, etc that work on a 1-wire network. I have a weather station hooked up with their gear. I think you'd need the adapter (~$15) and their switch module (~$30). http://www.aagelectronica.com/aag/index.html
This is interesting. I have a basement full of old ICs including drivers, microprocessors, digital/analog devices and thousands of LEDS and other things I've bought surplus over the years for various projects. I've thought of packaging them up into little "hobbyist paks" and selling on EBay, but never could think of what kind of project to sell them for. Perhaps a kit with parts like a UDN2987 octal driver, 8 leds, a connector and a schematic showing how to build a parallel port I/O interface along with some Linux C driver code for $15-20 shipping included? I think I even have artwork for a PCB I made to do this years ago!
Think anyone would be interested?
Then again, do I really want to deal with the support emails from people who can't hold the right end of a soldering iron...
http://www.cygnal.com The C8051F300 is an 11-pin micro with 8 digital/analog peripherals. It can provide serial access and then an additional 6 outputs. Run the leds through a ULN2003 darlington array (7 transistors in a 16 pin package) and hook them up to the micro. Cost $15.
I have a project that you can look at to get some Linux source for this and some simple hardware to look at for hooking up 8 LEDs to a parallel port. This is very simple and safe to do. I have it running on my desktop at home giving me the CPU usage in bright blue LEDs across the front bezel of the machine. Looks awesome. Very easy to adapt the code to utilize other values to control it. Just replace the function that gets the CPU usage with whatever you want. Diskspace on a server, network utilization, flowrate in the sewage system, whatever.
My name fits again.
Actually a PIC microcontroller is VERY easy to program as you can do it in BASIC. Getting started with them is also VERY inexpensive. You can get the PCBs for a parallel programmer for free off the 'Net and the parts for it from any reputable electronics parts supplier. They also can be powered from the serial port if you want to keep it simple. A PIC can power a number of LEDs with nothing other than a current limiting resister on the ground side of the LED. The PIC microcontroller is about $5-$8 from Reynolds Electronics.
Nothing is impossible. It just hasn't been figured out yet.