Quote:
Originally Posted by carmatic /img/forum/go_quote.gif
i guess i'll be looking at low q values for what im doing
|
Well, you can approach it that way, but if you need a shallow roll-off, you usually want to keep the Q high enough that it doesn't add extra ripple -- the value for this depends on the filter type -- and just use fewer poles in the filter. I would only lower the Q to get shallower fall-off if I got down to 2 poles and it still wasn't shallow enough.
(One can argue that a 1-pole active filter doesn't exist: it's just a passive filter followed by an op-amp configured as a buffer. It's not until you ask for the second pole that you have something different than you'd get with passives only.)
Quote:
| how much complexity do i need for a non-audio application like what im trying to do? |
That's up to you, but I would think an active filter, a rectifier and a power driver stage per band isn't too much to ask.
If you need something simpler, you can probably find what you need pre-made in catalogs catering to live music performance. You know you're reinventing the wheel here, right?
Quote:
| how many watts should the resistors be? |
It depends on where in the circuit they are and what their values are, which affects how many volts and amps go across them, which tells you power dissipation. You figure this out with the several formulations of Ohm's Law:
V = IR
I = V/R
R = V/I
P = VI
P = I^2R (that's current squared times resistance)
P = V^2/R
You may be able to get away with generic 1/4W or 1/8W resistors for everything. It's not clear to me that you will need a resistor in the power path, which would have to be multiple watts. You might need something small, like a 1 ohm 5 W resistor in series with the LEDs, for safety or something like that.
Quote:
| im not an electronics guy... |
You're probably going to have to become one if you want to see this project through. The only other way is to find someone who will design the circuit for you. That won't be me, and it doesn't seem like anyone else wants to comment in this thread.
Quote:
| the op amp symbol always has the negative and positive power supply terminals implied , even when they are not shown... |
Not always. They're usually shown when there's something that needs to connect to the supply pins besides just power, like bypass capacitors.
But yes, I take your point: you don't know how to read schematics very well, yet. That's a fixable problem.
Quote:
| i am guessing...i can use the bandpass filters as the power source of the lights |
Not unless you use studly power op-amps. Your garden variety op-amp can only put out about 1/125 the current you need for this application. There are op-amps that can source 5 A, but it'll be cheaper and easier to just hang a single bipolar off each filter's output to handle the load drive.
Thinking more about this, though, there are several complications:
1. You'll need a heat sink on each drive transistor. I suspect the voltage drop across the transistor won't be very large, but the full LED drive current will go through it. The heat sink may need to be rather large.
2. You need a voltage to current conversion in here somewhere: LEDs vary brightness as a function of current, not voltage. So, you want the voltage across the string to be roughly constant, with the current varying as a function of the music. A bipolar transistor being driven though a base resistor is a crude sort of current amplifier: you adjust the base resistor so the signal varies the collector-emitter current appropriately.
3. The power supply requirement for this is looking to be at least 450 W. That's a nontrivial amount of power. You're basically building a specialized power amplifier here.
Quote:
| i dont need a separate transformer anymore if i can somehow get the op amps to output 24v and enough current without any damage |
Power has to come from somewhere; there is no free lunch. If you need 24V and 15A (three strings in parallel), then you need a power supply capable of putting out no less than 360 W, and that assumes the circuit is 100% efficient, which it won't be. My 450 W number above assumes there is only about 20% loss, which is awfully good for a linear circuit like this. It's achievable, but you might want to shoot for more like 500 W, just to be safe.
Quote:
| vcbo, vceo, vbeo and all that |
I recommend reading
The Art of Electronics by Horowitz and Hill. The chapter on bipolar transistors will cover this. The next chapter over is on op-amps, and there's also a chapter on power supplies. Most of what you need for this project is in there. It might even have the filter stuff.
