I need help with a LED Organ circuit.

Thread Starter

Sphades

Joined Mar 7, 2020
10
Hello guys!
This is my first post here, so I'm sorry if I break any rules or anything. I'd need some help with an LED Organ circuit.

This is the design I came up with:
Circuit1.JPG

It's just a crossover circuit connected to 3 transistors that are supposed to pull the LED's low.
Problem is that it is not working properly. I built the circuit and the only LED's that work, are the ones connected to the Low Pass filter.
My question is:
Is my design wrong in any way, or maybe I simply built it wrong?
How would you approach this?

Thanks for your help!
 

Wolframore

Joined Jan 21, 2019
1,740
I just ran the calculations for your HPF it looks ok although at 3300 kHz not extremely high, your total impedance is about 63 Ohms. That feeding a common emitter and high frequency doesn’t have as much power as low frequency. it would be a good place to use an opamp with positive feedback to saturate it.

I didn’t check your band pass but it may have similar problems

what is your input signal?

You are asking about an actual circuit not the sim right?

sim won’t work for anything other than LP because your AC source is only 250 Hz Sine wave ...

or try using a saw tooth or squarewave which has high frequency harmonics on top for your simulation.
 
Last edited:

Audioguru again

Joined Oct 21, 2019
1,255
You have passive filters that have a huge signal voltage loss and they are so simple (first-order) that they do very little except cut the signal levels.
You should be using second-order or third-order ACTIVE filters (they use an opamp so there is no signal voltage loss).

I simulated your bandpass filter to show its loss and the transistor conducts such a narrow pulse duration that the LEDs will appear very dim. Then you should add a peak detector to light the LEDs for longer durations.
 

Attachments

Wolframore

Joined Jan 21, 2019
1,740
Agreed, I was thinking this would be a good place for a second order active filter. yeah that common emitter input Impedance isn’t high enough to work correctly.
 

Thread Starter

Sphades

Joined Mar 7, 2020
10
Thanks for your answers!

One thing i forgot to mention, is that i amplified the audio signal in with an lm386, before feeding it into the crossover circuit. If that makes any difference. Should i still be using active filters?
 

Wolframore

Joined Jan 21, 2019
1,740
2-4 pole active filters is the best solution. If you want a simple band aid which may not have the effect you want. You can add a BJT after the Lm386 to boost your output. As @Audioguru noted in his simulation they are not great filters and may not have the effect you’re looking for. I would shift the highpass a little higher 5-10k.

You don’t need great fidelity but you do need good separation for it to make sense and good levels for your LEDs to activate.

Try it and let us know how it goes. It’s a simple adjustment.

If you’re interested in something a little better do some research into a state variable filter but set it so the band pass Q isn’t too high. It requires 3 opamps and various components.
 

Audioguru again

Joined Oct 21, 2019
1,255
Hey @Audioguru can you get better response setting the damping factor as butterworth or even Chebyshev? We don't need good fidelity in this situation.
My filters have a gain of one and are not equal-components low phase-shift droopy Bessel, instead they have double the positive feedback so they are Butterworth.
An equal components Bessel filter with a gain of about 1.6 also makes a Butterworth.
 

Wolframore

Joined Jan 21, 2019
1,740
Just wondering if higher roll off from Chebyshev might be appropriate for lighting LED's.

Oh that's cool... it does give you about -20 dB per decade per pole... very interesting!
 
Last edited:

Audioguru again

Joined Oct 21, 2019
1,255
The circuit is an ordinary Sallen-Key (look in Google) second-order filter with parts calculated for a gain of one and a Butterworth response (flat then a sharp corner). It uses the opamp to produce some positive feedback at the corner frequency to create the flat response then a sharp corner.
 
Top