ok. i hate to continue to run you guys in circles, as you have explained these things really well in depth, to the point where i actually get it! but, the underlying message i'm getting is that it is user choice, based on what you need. so here is my idea: (yes its made im MSPaint , i'm still trying to learn SketchUp) so, i know what you're thinking. this reminds me of that bad acid trip i had in high-school. well, yes. if all of the lights were constantly lit at the same time, thats exactly what it would be. but, the idea here is that all of the lights strobe in different patterns, to create the visual illusion of the Blue Man Group's live concerts (i use them as an example because they're lighting techniques, and music for that matter are absolutely amazing. www.blueman.com if you arent familarized ). i know almost exactly what i want for the pattern, so this is my question: if each "strip" of same-colored LED's, and each set of same-colored ccfl's are each seen as one item from the controller (basically, each color strip is lit at the same time, not single LED's, or single ccfl's.), what would the best thing to use be? i count 15 different "strings" of lights/ccfl's, and probably 5 30-second to 1-minute programs. so, i need a controller that can be programmed by a computer (dont care what interface i use, just the easiest for whatever device i end up using. i would like the program to be something easy like BASIC, or C++ [though i think C++ is everything but outdated ] but i am willing to learn anything i need to). so, from what everyone has said (thanks emmensily to everyone who has taught me the crash cource on computer controller boards, i really really appreciate it ) i see it as this. if i go the AVR route, i would go like this for my physical connections: Parallel/Serial port--Programmer, with onboard memory that stores my programs--AVR Controller board--signal split (hopefully) 15 ways, 1 each to: 2 rows Blue LED's 2 rows Pink LED's 2 rows Green LED's 2 rows White LED's front panel White LED's front panel Blue LED's front panel Pink LED's front panel Green LED's 4 white ccfl's 2 blue ccfl's 2 green ccfl's 2 yellow ccfl's 2 red ccfl's corner White LED's Blue LED's that surround the inside of the case would this work? is this correct? would a PIC be better, and if so, how would i wire that up? again, thank you so much. when i first thought about this, i didnt think it would be possible to do, and now it seems completely possible. thanks!
Onward to the interface part (incidently the interface schematics are the same no matter what uC you choose). As you are using a 12V item (ccfl) and also clusters of LEDs I would suggest one of the following schematic for turning on/off the 15 items you want to control. The first design is a universal interface with a price tag of about $1 per channel (channel = item/group that can be turn on/off). It can handle as much as 5A per channel (at 12V that is 60W) which is probably overkill for most modding items that are put into a pc. But this way you do not have to worry about ever frying anything because the load got too big. The on/off transistor will eat app. 0.3V of the supply voltage. This is a general purpose on/off module, that can also handle inductive loads, like fans, motors and CCFL light converters. A few design notes: When just controlling LEDs, the protection diode can be omitted. The onboard indicator LED can also be ommitted. Normal transistors are choosen over MOSFETs, for three reasons. a) it is less sensitive to static electricity, b)the TIP125 seems to be a very good value for the price, compared to 5A MOSFETs, c) I have no experience with MOSFETs (yet). The "high flying" PNP with extra on/off transistor (BC547) is used rather than a "low flying" PNP only, as this reduces the voltage drop over the transistor from 0.7V to 0.3V The second design is about $1 per 7 channels. And basically consists of just one chip. It can only handle as much as 500mA per channel, but the chip specs. also says only 500mA total. Meaning each channel could max out the chip, but only one channel at a time. So if the load is only 100mA on an channel, 5 channels can be open at the same time. Anyway you probably get the point, that a little more thought about the loads and patterns of use will have to go into using this cheaper design. The on/off transistor will eat app. 0.7V of the supply voltage. And just because I for some time have been working on a baybus Mod Guide using modular interface boards and uC controller, I also have some PCB design suggestions ready. A 4 channel 5Amp module A 7 channel 500mA module These PCB layouts (and more) are also available in this file The PCB file can be viewed (and printet) using this free Viewer A note about Schematic and PCB layout software: I would like to use the LiveWire and PCB Wizard from www.new-wave-concepts.com as they seem to be just great and perfrect for my use. Unfortunately I did not know of these packages when I bought my Schematic and PCB layout software, so for now I'll stick with what I have. By the way, that will be one wildly light up case you have got there
while the second idea seems like it may be better overall, there are certain parts of the first idea you gave that certainly appeal to me. first of all, its very straightforward: its the same thing, x number of times, going into my controller chip. so, this is the idea that i have now: it looks like the LED's in that first idea at the end of each line simply show whether the line is active or not. so, i could mount the LED's on the top/front of the case and that would give me a manual control of the lights, or i can control that board automatically through the controller board (either an AVR or a PIC) using my program(s). now, each line in that first example actually provides each line with power, so i dont need an outside power source going to each line? what about my ****load of LED's in a line, will it provide enough current to power all of those (current is the factor i need to consider when trying to power lots of things, correct? the voltage will continue to stay the same down the line, if i wire the LED's in parallel?). the ends labelled SL<x> would be the actual lights, and the ends labelled ln<x> would go to the controller board? everything else is pretty self explanatory, i believe...if i understand correctly, the crash course in that board is: the +V line goes in the top (using vague terms here, sorry), and the GND line goes to the bottom right. the transistors and diodes next in each line are the same for each line, however the resistor that i place in the R<9-12> places depend on how much power is drawn from whatever device i have hooked up to the SL<x> end. the LED's in line to the direct left of the SL<x> lines are simply indicator LED's, telling me whether the line has power or not (handy for when i want to control the lights manually). if i wanted to, could i wire a switch into each seperate device line somewhere to allow me to manually control the power to each device (leave them all on when i want the controller board to take over), and put one big switch on the whole thing when i want to manually turn everything off? again i barrage you with newbie questions, but you all have helped a ton. i understand 100x about electronics than i did when i started this thread
yes, that's right. Come again? LEDs on the top/front of the case? Any and all LEDs can be controlled by the uC and can be turned on/off by any deciding source, like (if the uC control program is set up for it) - the PC (through the serial port communication) - or from the uC monitoring a pin from the motherboard - the uC automated program - a user input from a button connected to the uC I think we will revisit these input options again later Yes, the SL(x) symbols on the first illustration are connectors. (on the updated 500mA illustration, the connectors are labeled CNx) One pin on the SL(x) connectors being +12V (controlled by the driver transistor) the other pin on the SL(x) connector is ground (GND). see one example in the updated image for the 500mA driver, where I have illustrated how to run a bunch of LEDs of of one 12V supply. Also I'll elabotate on this with more illustrations. exactly Q: the +V line goes in the top (using vague terms here, sorry), and the GND line goes to the bottom right. A: YEP Q: the transistors and diodes next in each line are the same for each line. A: YEP Q: the resistor that i place in the R<9-12> places are A: ... for the indicator LEDs and the size depends on the type of LED and the supply voltage Q: the on-board LEDs are ...telling me whether the line has power or not (handy for when i want to control the lights manually). A. YEP Yes, you can do that, and we can also visit the option of adding a push button panel to the uC, as a manual input device to the uC. And program the uC so that ie. user operations will override any program setting. You are welcome. And you just keep the questions comming, as I have lots of answers up my sleve. Like I said, I have been wanting to make a Mod Guide, so if you don't mind, i'll use your questions to guide me to make a good Mod Guide. As I'm thinking: If you have that question a lot of fellow modders will probably also have the same question. So during this thread, you'll just have to suffer tailor made answers for your particular questions, but maybe you'll be a gentleman and find a way to live with that. Like I said I have a few illustrations of how to connect many LEDs (and other items) to the controller board, and for calculating the right resistors for use with different types and color LEDs. I will try to gather these illustrations in an upcommming post. But for now it is daytime where I live, and I have e few appointments today, so I'll be back.
awesome . so, here is where i stand now: it seems that (for the second, updated design), i need a seperate MC1413 for every 7 channels i have, but the AVR mega8 can only handle up to 14 channels, so i'm limited to 14 different channels. (shouldnt be that big of a deal, i should be able to find a way to group two of the lines into one). sorry, didnt explain that one very well. what i meant was, i can mount the indicator LED's at the front of my case, and they will tell me what channels are on at that point. i think thats all thats on my head right now, will post again if i think if anything more
Q: i need a seperate MC1413 for every 7 channels i have A: yes Q: but the AVR mega8 can only handle up to 14 channels, A: of the 28 pins on a Mega8, 2pins are reserved for the Serial port communication, that leaves 6x PortC pins, 6x PortB pins and 6x portD pins that are easily available for I/O operations (total: 18 pins for I/O) Q:i can mount the indicator LED's at the front of my case, and they will tell me what channels are on at that point. A: that you can
As promised I'll take you through a step-by-step calculation of powering LEDs The basis of electronics calculations is Ohms Law Ohms Law - the connection between Volt, Ampere and Ohm. Ohm's Law: U= R * I Or in other words: Volt = Ohm * Amp U means Volt R means Ohm I means Ampere So if you have 5V and want 20mA through you circuit, it needs to have a total resistance of 250 ohm. because Volt = Ohm * Amp and 5V = 250R * 0.020A or to put it differently 5 / 0.02 = 250 So how did I know that Volt / Amp = Ohm ? Well thanks to some kind of mathematician it turns out that: U = R * I and U / R = I and U / I = R I remember these three rules by only focusing on the first rule: U = R * I, and then putting the three letters into a calculating triangle. So if you know U (Volt) and I (Amp) then you can find R (Ohm). If these math rules are new to you, we may need to look at more examples. Example 1 A Blue LED needs 3.4 Volt to let 20mA run through it. 20 mA = 0.020 A We have a 5V supply, so we need a resistor to eat up the rest of the 1.6V. This is the case in the following picture. In order to find the right Ohm value of the resistor, we know that we want 20mA through it, we also know that there is 1.6V over the resistor. Putting these numbers into the Ohm calculating triangle we get: The current limiting resistor for the Blue LED should thus be 80 Ohm Example 2 Blue LEDs usually need 3.2V to have 20mA running The following illustration is based using the boad described in method 1 (with the TIP125 transistor) 1.5V / 20mA = 1.5 / 0.020 = 75R = 75 ohm If we use the MC1314 chip, we loose 0.7V over the chip, and thus we will have 0.4V less on the resistor, leaving not 1.5V but 1.1V for the resistor. With The MC1413 chip the calculation is 1.1V / 20mA = 1.1 / 0.020 = 55R = 55 ohm Example 3 High Bright Red LEDs usually need 2.2V to have 20mA running The following illustration is based using the boad described in method 1 (with the TIP125 transistor) 0.7V / 20mA = 0.7 / 0.020 = 35R = 35 ohm If we use the MC1314 chip, we loose 0.7V over the chip, and thus we will have 0.4V less on the resistor, leaving not 0.7V but 0.3V for the resistor. With The MC1413 chip the calculation is 0.3V / 20mA = 0.3 / 0.020 = 15R = 15 ohm Example 4 Controlling a LOT of LEDs from the same channel NOTE: It is recommended to use one resistor for each string of LEDs. I do not know why, it is just one of those things I once read, and now I only remember the rule. Shopping for Resistors If you can not buy the exact ohm value you want, then you should choose the closest larger ohm value in the shop. I suppose this will give you a tool for knowing when you hit the 500mA mark in your designs, when you are using the MC1413 chip I do NOT recommend using the MC1413 chip for controlling the CCFL converters. They are inductive loads, and it is usually not specified how many mA they use. They may however just 200mA each, and thus two can be controlled by one MC1413 chip. But who knows? Maybe it would be good advice to get your self a Multimeter. They do not have to expensive to be able to measure simple electronics usually $10 units are just fine. With a multimeter it is easy to find out how much current is being used, and thus no firehazards are created, when builing you own electronic stuff.
actually a multimeter is the next tool on my list, just because it seems like there are a million things you can do with it, and its pretty much specific to that tool. on a sidenote, what are the basic good options i should look for in a multimeter? your explination of power calculation was exactly what i needed, and the triangle diagram makes it really easy to remember. thanks! my only issue now is time, between work and swimming (i swim mornings from 4:30-6:30, go to work, swim 5:30-8, then get home and have to tend to the wife. course, she isnt particularly fond of me being gone all the time, and you know what that means ) i need about 48 hours in a day just to get things done that i would like to, and get an hour or so of computer dinking in. but once i get this project in gear, i now know exactly what i need and what to look for. thanks! so basically my few questions that are left are: 1. which option should i go with? i was at first leaning towards the first, because it looked simpler, but it looks like the second one may actually end up being simpler. 2. what is the estimated price on all of this? 3. (kind've O/T) what should i look for in a multimeter?
For electronics I have a few really, really cheap Multimeters (<$10 a piece, got them on sale 50% off, so all three was $15), they will do some DC and AC measuring. I use them for DC <20V and <2A measuring ALL the time, and having a handfull of them (well actually only three) means I can monitor Volt AND Amp on some main component, at the same time as I can poke around with the third. I also find myself measuring ohm or "is there a connection" very frequently. So for just electronics, I'd say price is the main issue to look for. What do these cheap instruments NOT offer me, that I am sometimes missing? I suppose this may be a good measure of things to look for. They do not measure AC Amps. Might be interresting some times, but for the last four years I only wanted to measure AC amp like five times. I have an old analog instrument that can measure AC Amp, but I didnt eve bother to get it out those five times, I just did do the guestimated rather than measuring, so AC Amps is a long way down my list. DC Amp has a jump from 200mA and them 10Amp! And I have to move the plus wire. It would be nice with a 2Amp range, but again I manage. I do however blow a few fuses when measuring Amps, so an auto fuse would be a very nice facility. Sometimes I also dream of being able to log some measuring by hooking up the instrument to the computer. But it's probably more of a dream than a real need, as I NEVER sit there with a stopwatch and gater data series. Even my very cheap instruments do Diode test and Transistor test (hFE measuring = amplification ability of a transistor), so any Multimeter should probably offer this. I have only used it like twice, ever, so its not high on my list. What I also often dream of is one of those handheld instruments, as I often find myself wanting to do some oscilloscope style measuring, but again I have managed to find alternative solutions so far. But I do not think that a handy scope will ever leave my wishlist. Glad you liked the Ohm story. Now with the time problem, I AM working on a potion, unfortunately it may be a few years in the pipeline before it's ready for public release So until then; prioritizing, planning and long suffering are the only remedies I know of and yes, I can highly recommend tending to the wife when you are in her precence. I'm supprised she will let you spend a full hour on the compuer, with a schedule like that. Even more reason to give her FULL attention when in her prescence IMHO As for estimated price for such a unit. mega8 $7 rs-232 chip $3 D-sub9 socket $2 1x method 1- 1x 4 channel $8 2x method 2- 2x 7 channel $3 resistors etc. $3 PCB app. $7 Total estimated $33 (+ stamp) And because I like your SIG, i'll offer to make you a board for that price. This might also help your time problem. Let me know if this would be of interrest to you.
I would certainly be interested in your board, PM me with the exact uses if you'd be so inclined . so basically any regular multimeter that measures DC/AC voltage will do, but look for smaller steppings and an auto-fuse? awesome, thanks! EDIT: yeah, Sabrina really hates that i'm never home. this happens every year i swim on the state team, and every year she hates it, but every year she encourages me to go back out for it when i tell her i dont mind not swimming if it bugs her...(the background is ive been a swimmer since like my freshmen year of high-school). so i dunno. ive kinda been in purgatory since i started this year, so theres another decision to add to the list
I didnt see your post until now. I have been busy all day...... designing a board. Maybe you'll like it. Mayb it needs to be changed to fit other physical dimensions. It is the size of a 3.5" harddrive. I have actually designed one end of the board so one can re-use the connector end of a HD (to the right on the board picture) [edit2] The board images will get updated as the project/design progresses, as this saved bandwidth and space on my webh hotel quota. Also I feel no specific need to advertise errors/mistakes that are naturally weede out during a PCB design process [/edit2] Picture without indication of holes, as it makes the white text easier to read. This is the HD connector I'm thinking of using. The board also has a option for a DIL socket (jumper J13 and J14). Below are shown two types of connectors. The high metal pins are the DIL type (with only one line in this picture) The other type is machined chip sockets. One can push one more row of the socket into the one mounted on the board, and a simple and very small connector is made (the machined chip socket is in thr foreground in the following picture) In the DIL sockets (as well as the HD edge socket) 2-pin connectors can be uses, like most motherboards have fornt panel connectors. Or one can take an old HD 40-pin wire, and split it, and solder LEDs etc. to the wires. I have made a few other design choices. (These can be adjusted/changed, but lest see how they fly with you for now) - all outputs can be fitted with on-board resistors. Or a simple jumper can be soldered in, if one prefers to have the resistors mounted with the LEDs - the 500mA driver array, can be either SMD (1.27 pins) or standard 2.54 socket mounted chips. - there is an optino for a third 500mA driver array (IC5), which can replace one or more of the high power channels, if there is no need for power on/off (>500mA) - the High power channels have additional connection jumpers on the board (Out1-4), so no one needs to try to run several Amps through small HD wires. - there was space left over on the board, so I put in an option for an extra uC controller. Who know if that may come in handy one day. - the RS-232 serial port has a 2-pin connectors and not D-sub9 connector, as I supose the controller board will be mounted inside a PC, and thus I see no need for a bulky connector (not much board space left for it either) I have suggestions/ideas, for a flexible serial port connection method. - there is a jumper option (J10 and J11) which allows for ISP programming, without flashing the LEDs/output wildly. - the four indicator LEDs on the power channels, has an optional 4-pin connector (for mounting the LEDs somewhere else than on the controller board) [edit] - The thickness of wires, distance between wires, size of solder pads and distance to solder pads, and also extra via's at DIL jumpers should make this design relatively easy for a "make at home PCB. I use a 0.8 mm drill for most holes and a 1.2 mm drill for larger holes. [/edit] You have summed up my Multimeter advice just fine. Ahh, those priorities, those priorities. Makes for a tough life. But I suppose there is wisdom in the proverb: there is a time for every thing. And sometimes one has to move on, when life moves into a new stage. Sometimes one needs to finish a plan, before a new path can be choosen.
Just a quick note to say you're never going to get strobe or fast flashing effects with CCFL's by controlling them from the low voltage side. The best you can hope for is 1 to 2Hz really, unless your inverters happen to be of the electronic variety with a dedicated driver IC (unlikely). Ideally you want a constant 600-1kV source and have your triacs between the output of the HV source and your tubes.
Hi SteveyG, so what I hear you say, is that truning on/off CCFL's by turning on/off the converters, is supposed to be doing just that (and the max speed should be like one's thumb operating a physical switch). Where as, if one wanted to operate a CCFL (= neon style light) as some kind of fast paced light show then an electronic relay (TRIAC style) should be applied to the AC side of the converter? Good to know. I have a TRIAC design that I have been using for years (for 220V), only I currently have no schematics of it, i'll make that shortly. Maybe you'll have a look at that for us SteveyG, once I make the schematic (I'll probably make it sometime tomorrow). Also, the DC on/off part for the CCFL's will be using PNP's at 12V, and not NPN's at ground. Do you think that makes a difference? Or is the key here, that the CCFL converters are inductive, and it's always problematic to quickly switch on/off coils. Zephyr, I have just been using the on/off modules I hade previously designed, with only 4 on/off channels. And I now see that you plan on using 5 CCFL's . So anyway, I have made a revision 1.1 of the board, so it now has five power on/off's, I'll post an update later. See ya'll later.
I would suggest that you use one of the outputs of the microcontroller to drive the LED of a 'triac driver opto-isolator'. These are opto-isolators which have a triac as the output device and with the LED as the trigger. I'd then use this triac to drive a properly rated triac. Since it's only an LED being driven by the microcontroller, that side of the circuit is simple. I think the inverter would need to be modified slightly though, as the circuit is self resonant, and may not take kindly to no load. In which case it would be better to build a new oscillator circuit to drive the primary side of the inverter.
I see what you're both saying about the CCFL's. Ideally, they would provide the most consistant, bright strobing affect, but I do worry that I will cause some sort of a voltage spike, or extremely shorten the life of the CCFL's by doing it. The other idea was the possibility of actually getting a strobe light, but how much power do they take? And, would the controller board be able to control the power they take? Alvin, I really appreciate the work you're putting into this, designing boards and racking your brain just so I can create the perfect lighting affect. It's very cool of you . Would it be easier to use a different lighting type, because CCFL's require such a high voltage (current? whichever one) that they are hard to work with? I can't think of any other lighting type that will achieve the same affect right now...possibly things like this or this? Here is a schematic to build a strobe light, and it doesn't look all that complicated. Well, I'm off to a movie with the wife (this is slightly a thread on balancing life and relationships ), trying to show an attempt to spend time with her. Thanks!
Huh. A PIC16F874A with 8X 74HC573 and 8X ULN2803 could do the job much easier with 64 outputs without the uses of optos or those high-end transistors. You dont need 5A output for this project. But thats just my opinion. I'd go with the PIC since you can then control the whole mess through the internal USART using simple VB or even hyperterminal from the PC. OT: In no way is an AVR better than a PIC. They have the same speeds, they have the same hardware peripherals, and come in enough different flavors to find the right package for the job. Dont bother listing 'this or that' since microchip has already done it too. What Microchip has done is make it easier to learn, easier to build, easier to buy, easier to program, and they keep giving away the ghost. Hence the reason why theres so much more support found on the web. Are you gonna tell me that "AVRs have the ATMEGA this and has that". Then you havent bothered reading about PIC30F series, and the PIC10F series, and hundreds of other projects they have coming in the next few months. If you like AVR, so be it. They are good chips. My personal opinion is not to use a bulldozer to dig a small hole. You dont use a 3GHz dual processor machine just to read txt files. PICs can do the simple jobs at hand, for less cash and less developing time. Projects Ive done: Create a standalone graphics processor. On the fly manipulation of analog output signals for 64 independant channels. Real-time evaluation of analog signals with 24bit resolution. RS-485 bus monitoring 245 slaves with bidirectional control/data acquisition and PC data logging. And Ive yet to get into the 18F family. All of these things were done using 35 instructionset. But this is just my personal opinion.
ummmmmmmm... AVRs also have internal UARTs. An AVR project I'm just finishing up work on is entirely serial controlled. Hell AVRs can even bootload through the serial port... So I don't see why you think AVRs don't have this capability. You say PICs have the advantage of being easier to learn and whatever. If this were not the case than I'd say Microchip would have a serious problem! I mean PICs have been around since the 70s - OF COURSE they're going to be easier to learn, have better support, much wider user base, and a much larger community. AVRs have been around for a couple years, but already they are starting to take up a significant amount of space in the 8 bit embedded market. Things are a'changin'. Anyways - you say that in no way the AVR is better than the PIC. That I disagree on, but instead of arguing about it, I can just point you to this topic: http://www.avrfreaks.net/phpBB2/viewtopic.php?t=25212 which sums up my views pretty nicely.
SteveyG: - about the Triac design; using an opto-isolator, and then a power Triac, is exactly what will be used. - "it would be better to build a new oscillator circuit to drive the primary side of the inverter" Yes, I think you are right :gulp: However, this is where my hobby skills definately no longer qualify. Do you have any good designs, links, hints, suggestions that may lead to a workable oscilator design. Hazer: - " 8X 74HC573 and 8X ULN2803 ". Great, this is the chip suggestions I've been wanting (but obviously been too lazy to dig out my self). I'll definately make a driver module based on those chips. Just keep you comments comming! I have been avoiding the ULNxxxx chips as at RS-Components they are more than $1 a piece. Fortunately I have now found a danish supplier who will sell them for less than 50 cents so long as you buy more than 10 pcs. - We have already covered my reasons for using the AVR chips, see this post #18 And I agree that PIC's are just as good or for many a better choice than AVRs as you are quite right about the support and availability issues. - You bringning up the PIC/AVR point again does however remind me of the benefit of designing interface modules that are uC independant. And thus it would be wise to design systems that use two sections of PCB, one for a uC and one for the interface/driver chips. I'll definately keep that in mind for future reference. - 5Amp power for this project? No, that's not needed. It's just that the TIP125 transistors are way cheep, and thus I find it easier to just stick with a known design. - I'll just keep this board design available, as it IS a generally useable driver board, with low power (500mA) and high power (5Amp) options. And the 3.5" HD size is probably a good size for PC add-on boards that are not meant to use a PCI slot space. hmm... maybe it would be wise/possible to add holes/edge connector that actually would fit into a PCI slot (just without any PCI wire connections), as that would make for a quick alternative mount in a PC. I like flexibility and built-in easy options. Zephyr: - sorry that I did not pick up on the need to quickly flicker the CCFL's. I should have picked up on that sooner. Good thing SteveyG picked up on that, and commented. About the alernatives: yes, Lazer LEDs may be able to create some good lighting effects. I think that maybe the Hyper Lights are just Luxeon Super LEDs with some kind of optic spreadder/diffussion box attached. Wether you choose to move to power/super power LEDs rather than CCFL's I still think it's a worthy cause to come up with a good design for controlling CCFL's at higher speeds. These changes in design during a project is why I like the idea of modular drivers/controllers. Designed right it'll be almost like working with electronic LEGOs. - I'll have a look at that strobe light thing. - So what movie did you guys see?
I wondered what the Hyper Lights were. I may simply decide to move to more powerful LED's, if not for a more controllable, longer-lasting lighting affect, but I do agree it would still be nice to find a way to control CCFL's toggling quickly. I'm off to bed, but we saw Oceans 12. Good movie, both of us loved the first one, and this one outdid it
Alvin: Good to hear you found a good source for those ULNs. I know what you mean about using discretes instead of packaged ICs, but its nice when the IC is cheap enough and makes wiring that much easier. I wish digital pots didnt cost $2-5 each. Nleahcim: I didnt really say PICs are better than AVRs, except that they are easier. I know the AVR has internal USART. It also has internal oscillation. 25mA/pin driving ability. SPI/I2C MSSP module. PWM module. Built-in comparators. Slave parallel module. External memory interface. Internal EEPROM. Settable internal weak-pullups. Every option that AVR has, so does the PIC. So it really comes down to personal style. Last but not least, I dont see the point of the link you gave. Nowhere in that 120post forum thread was there anything that shows an AVR is better than a PIC. The only thing I gathered from it was a slight bias for AVRs, since the forum is about AVRfreaks...... It all comes back around to the same thing: In no way is the AVR better than a PIC. Every peripheral found in an AVR, you can find in the PIC. The MIPS speed for each are pretty much the same (lets not get into the 4:1 arguement, it always comes out to be the same MIPS). The price is always about the same (one thing not mentioned in that thread is that there are 4 times more sources for PICs than AVRs so you cant compare price based on one supplier). The advantage (arguably) is that PICs use RISC. 35 instructions is easier than 70+. But to play devils advocate, those 35 instruction are really only benficial if you use assembly. Otherwise, using a higher level language will cut your efficiency to the point where it wont matter how many instructions you can choose from. And here comes the arguement this thread has lacked: assembly vs high-level language. If you use a 8pin uC to light a couple of LEDs, why would you need a higher-level language? Its really not that hard to learn the 35 IS RISC. Its easier than learning a higher-level language. But if you already know that higher-level language, then that arguement turns around. I can see where taking the easy route is beneficial in that area. My personal opinion is that I can make the most efficient code using assembly to the point where i hardly ever have to troubleshoot my code, and runs twice as fast as anything done in a higher level compiler. But thats just me. Others have no use for efficeincy and preffer to stick with what they know. Final conclusion: The PIC and the AVR is basically neck to neck, so saying one is better than the other is garbage. The slight differences are not even enough to say that personal prefference makes the decision.
