Hello. I am getting into IC and Solid State amps and have completed many successful projects. Recently I have considered building a stereo desktop amp with built-in speakers using the simple LM386N-1. That is very easy, however, I would like to build in a mini sub-woofer as well, with a separate linear pot. Will I need a third LM386 chip to do this or can I make a sister board and bridge the woofer to each of the other two amps? Also, I was curious as to how I can filter out all the the high frequency signals to get a more pure bass sound from it. Any suggestions and/or instructions would be greatly appreciated! Thank You!!
Adding a capacitor across your feedback resistor (high side) will reduce high frequency. Otherwise series a choke on the output. Capacitive reactance (effective resistance) is figured by Xc= 1 / (2pi x Freq-in-megahertz x Capacitance-in-Farads) For an inductor it is
Xl= 2pi x Freq x inductance-in-Henries.
Basically a capacitor couples more signal the higher the frequency and an inductor couples more the lower the frequency. Putting a cap around your feedback resistor will couple more higher frequency back for comparison, thus reducing that component of output.
Sorry if it is confusing, but it is.
Doc
Xl= 2pi x Freq x inductance-in-Henries.
Basically a capacitor couples more signal the higher the frequency and an inductor couples more the lower the frequency. Putting a cap around your feedback resistor will couple more higher frequency back for comparison, thus reducing that component of output.
Sorry if it is confusing, but it is.
Doc
I don't mean to be judgemental but many aspects of the project don't make much sense.
USB 5V/500mA isn't very good at all for class A/B amps, and barely suitable for class D. I mean sure you will get sound but with little to no extra parts and minor expense for higher current chips, it is as GloBug wrote, just a matter of a different transformer.
But, supposing you want it portable for use on a laptop, I could see why to do it but think there isn't a good reason to put a sub in. Output power too low, just split the difference and use a little larger L & R channel drivers.
Linear pot? Why not a log pot? It seems like a lot of complexity and added time and size relative to the result of a LM386 amp. Main point is, within the limits of USB power you will have higher output including more current available for a sub (or L/R channel bass) by going class D.
USB 5V/500mA isn't very good at all for class A/B amps, and barely suitable for class D. I mean sure you will get sound but with little to no extra parts and minor expense for higher current chips, it is as GloBug wrote, just a matter of a different transformer.
But, supposing you want it portable for use on a laptop, I could see why to do it but think there isn't a good reason to put a sub in. Output power too low, just split the difference and use a little larger L & R channel drivers.
Linear pot? Why not a log pot? It seems like a lot of complexity and added time and size relative to the result of a LM386 amp. Main point is, within the limits of USB power you will have higher output including more current available for a sub (or L/R channel bass) by going class D.
That amount of power is really small for a subwoofer. On the other hand class D usually means surface mounted, which does not suit everybody, and you don't get that much more power out of it either. 500 mA are 500 mA, whether they pass through class B or D.
Talking about desktop speakers however, the subwoofer will not be so sub. Such a miniature woofer will probably play its lowest note at a higher frequency than the usual crossing point to satellites, so not much power is needed either.
To answer post #1, either way you will need a crossover.
When you use a separate amp for the woofer, the crossover will be before the amps. The best solution would of course be active, but that means you need several op amp channels that would also feed from the 5 V 500 mA supply and reduce the available output power accordingly.
A passive crossover before the amps could be an option. It is not as big, and doesn't need extra power. Drawbacks are that it is not easy to get more than 12 dB/octave steepness, and the channel separation will suffer a bit from creating the mono signal for the subwoofer. For small desktop speakers however that is probably an acceptable compromise.
You can use two smaller woofers instead of one or a double voice-coil woofer. Then you can use two separate amps and skip the mixing part.
For your other idea to bridge the two channels into the woofer, you need a passive crossover, i.e. big inductors and capacitors between amp and speakers. Then you need to invert the signal of one of the channels to get the inverted signal for one leg of the woofer, and then invert the polarity of the corresponding satellite, so that it is in phase with the other again.
Or again, use two smaller woofers or double voice-coil instead of bridging one.
A third approach would be to use a band-pass woofer and skip the crossover. That requires 8 Ohm speakers, because you just connect the woofer in parallel to the satellites. The satellites work unfiltered and contribute what bass they can. The woofer is low-pass filtered by the band-pass cabinet. It also require to use two woofers or a double voice-coil model. And the band-pass cabinet will be slightly bigger than a closed or ported box.
Talking about desktop speakers however, the subwoofer will not be so sub. Such a miniature woofer will probably play its lowest note at a higher frequency than the usual crossing point to satellites, so not much power is needed either.
To answer post #1, either way you will need a crossover.
When you use a separate amp for the woofer, the crossover will be before the amps. The best solution would of course be active, but that means you need several op amp channels that would also feed from the 5 V 500 mA supply and reduce the available output power accordingly.
A passive crossover before the amps could be an option. It is not as big, and doesn't need extra power. Drawbacks are that it is not easy to get more than 12 dB/octave steepness, and the channel separation will suffer a bit from creating the mono signal for the subwoofer. For small desktop speakers however that is probably an acceptable compromise.
You can use two smaller woofers instead of one or a double voice-coil woofer. Then you can use two separate amps and skip the mixing part.
For your other idea to bridge the two channels into the woofer, you need a passive crossover, i.e. big inductors and capacitors between amp and speakers. Then you need to invert the signal of one of the channels to get the inverted signal for one leg of the woofer, and then invert the polarity of the corresponding satellite, so that it is in phase with the other again.
Or again, use two smaller woofers or double voice-coil instead of bridging one.
A third approach would be to use a band-pass woofer and skip the crossover. That requires 8 Ohm speakers, because you just connect the woofer in parallel to the satellites. The satellites work unfiltered and contribute what bass they can. The woofer is low-pass filtered by the band-pass cabinet. It also require to use two woofers or a double voice-coil model. And the band-pass cabinet will be slightly bigger than a closed or ported box.
The USB should be capable of powering 4 LM386's, so the easiest solution may be to passively high-pass filter the satellites (via the output capacitor, for instance), and use the bass boost circuitry found in the LM386 datasheet for dual woofers instead of just one. There's probably milliamps to spare for an active crossover, but it hardly seems worth it here, as this is a very basic, casual-listening, near-field system.
I can highly recommend start using dual polarity power supplies Vee, Vcc. There is a serious background behind the matter.
Consider ordering a trafo and think about LM3886 for satellites and LME49830 for a sub. That would bring you to another level of sound reproduction since you have accomplished several simpler projects already.
Computer PSUs are no good for audio actually.
Maybe just so for nearfield listening.
If USB is 5 V DC, 500 mA DC, and each LM has a quiescent current of 24 mA, you are left with ~400 mA DC. You get 1,77 V AC, 141 mA AC -> 25 mW of clean output power, if you ignore losses, and only at an ideal load impedance of 12,5 Ohm. Lower impedance means less voltage, higher impedance means less current, in both cases less power. So, if you use four 50 Ohm speakers to achieve those 12,5 Ohm in total, you get less than six clean mW per speaker. At any other impedance you get less.
Mine too. Freudian slip after reading it draws 24 mW of quiescent power.
The question is not, whether an LM386 can do that, but what happens when four have to do that from one USB power supply. You might end up with less output power than before due the current limitation, if you don't reserve another USB port for the additional amps.
Thank you for the advice. I have recently built a dual mono system into a altoids tin and housed the speakers, which are 1" x 2", 2W, 8ohm speakers, into two altoids smalls tins. Used a A100K dual ganged audio taper, and 1uF capacitors between pins 1 and 8 for gain control. It sounds shockingly loud and clear for what is is, and have had requests from people to build many more. The USB powers this little rig effortlessly and can be powered by a 9v battery as well. The circuit, USB cord, battery clip, and a battery for back-up, all fit within the main tin, and the speaker tins are tethered by 6 inch lengths of wire wrapped in heat shrink. Again, the 5v/500mA that the USB gives for power is more than enough for 2 LM386's and I have never had any issues with it. As for using the LM3886, I am working up to this as I feel more comfortable, as I am pretty much a total noob. The total size of the unit I am looking to build next will be a 4 inch cube made from cherry wood stained a mahogany color. I have a 3 inch sub-woofer that I powered with a breadboarded LM386, and after playing with various cap. values, got some pretty impressive results. Again, total noob here, and considering all advice, some of which I don't quite understand. Just looking for the simplest, cleanest way to bridge the sub, preferably on a sister board with it's own pot. As for the complicated algorithms; thank you for sharing as I was curious as to how that is figured, and that seems as though is may be the cleanest setup. As for a passive crossover, that seems to be the simplest setup, as does the band pass woofer, although I am not sure of a good place to obtain small speakers for projects like this. Luckily I live near a place called SkyCraft in Orlando, and they have the tiny speakers I used, along with damn near any damn thing else related to or dealing with electronics. Cool place and they have a website. Thank you all and I will try my best.
You might try some other ICs that can deliver a higher output swing into the load at the given supply voltage. Here are some examples. These were tested on a regulated 9 volt supply. Power was measured just before clipping into non inductive 4 and 8 ohm resistors. Some of these ICs won't work on 5v, but the point is that there are many modern ICs that do much better than the LM386 and are just as easy to use.
As you can see, the LM386 does not perform as well, especially on 4 ohm loads. Bridging the chip into an 8 ohm load is like running it on a 4 ohm load single ended. The 7267A nearly doubles output power on 4 Ohm loads.
An externally hosted image should be here but it was not working when we last tested it.
As you can see, the LM386 does not perform as well, especially on 4 ohm loads. Bridging the chip into an 8 ohm load is like running it on a 4 ohm load single ended. The 7267A nearly doubles output power on 4 Ohm loads.
Last edited:
My position is basically that the distortion will go through the roof above 200mW for the LM386, but it would need to get up above that level to draw 1/2 amp from the supply.
Running from a powered USB hub might be a good idea; AFAIK those just supply 5V to the ports, with current limit set by the power supply itself. There are also low voltage op amps that could be used to make an active mixer for a sub out channel. Just sayin'.
Running from a powered USB hub might be a good idea; AFAIK those just supply 5V to the ports, with current limit set by the power supply itself. There are also low voltage op amps that could be used to make an active mixer for a sub out channel. Just sayin'.
I just don't see the point which seems to be having a surplus of LM386 chips... when so much time and construction and end benefit contradicts using LM386.
Point is, in a good design you spec the parameters and pick the best chip, not just deciding that you'll use LM386 any way possible. I might be unreasonable, it's just one opinion. I don't see any situation where LM386 has merits except if someone happens to have a bunch of them, but relative to total time and cost, other low power ICs are not very expensive.
Point is, in a good design you spec the parameters and pick the best chip, not just deciding that you'll use LM386 any way possible. I might be unreasonable, it's just one opinion. I don't see any situation where LM386 has merits except if someone happens to have a bunch of them, but relative to total time and cost, other low power ICs are not very expensive.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.