Im currently learning about audio amplifiers and wanted to design a powered crossover circuit to power low wattage desktop speakers just for fun. I tested the sound quality from the configured lm386 and was happy with it. Id like to know if this circuit will work properly to power 3 speakers- each with their own frequency margin (assume all are configured same as the buffer amp). I dont wish to use another chip or create some efficient monster, just have good quality sound within safe operating conditions for the system. Are there any redundant aspects to this, or ways to to improve? The RC values for the crossovers will be calculated with a proper chart, but i thought id throw in variable resistors to tweak frequencies and pots on the gain pins to balance any drops from the crossover. I am expecting a 1amp dc wont do for this circuit x2 for right and left channels. What wattage speakers would work appropriately for right and left channel power draw? I am new so please give feedback. Thanks.

 

I would make the first buffer amp a standard audio quality op amp.
There's no reason to use a power amp for that function as you don't need the power and it will just add some small distortion to the signal.

The quality of the speakers is more important than their power rating.
Just about any speaker you select should have an adequate power rating.

Note that the maximum output power of the LM386 is less than a watt, so don't expect much volume from your system.
For more power look for an audio amp with a bridged output.

 

That's a good point. I didn't realize it would add distortion. Thanks for the advice.

 

just for fun.

That's the important part. When people show us they are not doing something in the usual way, they are usually on a course toward finding out why it's done, "the usual way". Power amps are not low noise signal processors. All this bi-amping and tone control is done before the power stage for a good reason.

 

Ok, so trying to turn that response into something constructive for where I'm at, are you saying to just parallel the line audio to the crossover circuits and power from there, or throw in a unity gain buffer op amp instead of the first 386? I might use the signal from an old wireless receiver board from my headphones so I didn't want to disturb current from it.

 

Ok, so trying to turn that response into something constructive for where I'm at, are you saying to just parallel the line audio to the crossover circuits and power from there, or throw in a unity gain buffer op amp instead of the first 386? I might use the signal from an old wireless receiver board from my headphones so I didn't want to disturb current from it.

Depends upon the output impedance of the wireless receiver board and the value of the resistances you will use in your filter networks.

 

are you saying to just parallel the line audio to the crossover circuits and power from there, or throw in a unity gain buffer op amp instead of the first 386?

A bit over your head?
I'm saying, "Don't use a power amp for a signal processor." Crutschow said the same thing.

There's no reason to use a power amp for that function as you don't need the power and it will just add some small distortion to the signal.

Either do the crossover (or tone control) work using a low noise op-amp before the power amps, or use passive coils and capacitors as a crossover network after the power amp.

 

What do you think the point of these types of forums are? A bit over my head? Crutschow understood what I was asking and worked with it in a clear and concise manner. Even after I gave my second question he didn't respond saying it was ridiculous (even if he thought so), just that it depends on variables. I asked about switching to an op amp with no gain just to isolate a circuit I was unfamiliar with (even if that has ramifications, not the point). He understood that. I know what you're trying to say and of course it is sound, but you coming off like a condescending elitist ass is a waste of time. You're a terrible teacher and I'd appreciate it if you skipped any further questions I may have. I'm sure you were born reciting quantum theory equations. Too bad we all missed it. Take a hike- crutschow, thank you.

 

It is important to remember that an LM386 is not just an opamp with an extra-beefy output stage. It is specifically designed for audio speaker driving with a single power supply rail. It automatically sets the output pin DC level at Vcc/2, and you cannot override this. Also, its inputs are DC-referenced to GND, and you can't change that either.

I mention these things because the signal path from your input stage to the lowpass output stage does not have a series capacitor anywhere, so 6 Vdc is being forced onto the input pin. It might not destroy the part, but it will saturate the output.

Separate from that, the 386 has a minimum gain of 20, so your signal path gain is a minimum of 400, or 52 dB. That's a lot, way too much for a line level input signal. This is another reason to change the input stage to a traditional opamp.

Your filter networks will work as designed only if the network impedances are less than 10% of the LM386 input impedance. More than that and the input stage impedance will alter the filter frequency response. If you want to stay with passive filter sections, a unity gain buffer after each one will prevent loading, because a buffer stage can easily have an input impedance that is 100 times the filter network impedance. A better approach is to use active filter stages; more complex, but much better filter performance.

And finally, you have 11 messages. No matter what your biological age, that makes you a child on this forum. Best to learn more about someone obviously your technical senior before attacking their character.

ak

 

You are also my technical senior, however you didn't throw in unnecessary drivel to bolster your own ego. That was a lot of detailed valuable information which I appreciate and understood. Especially in regards to the impedences and gains. I'll be going back to the data sheet and rethinking the design. Would a 220uf cap to block the dc be acceptable to allow frequencies over to the low pass? Thanks again.