http://m.youtube.com/watch?v=GYV95Fs6UGE&desktop_uri=/watch?v=GYV95Fs6UGE
Pins 1 & 8 are left open, so a Gain of 20. Most of these "headphone amplifiers" have a gain of 50, or 200. I can't even toss a 10uf betweens 1 & 8 without nasty distortion.
I posted this on another forum. Just was curious to see how people were implenting their single chip lm386 audio amplifiers
*I'm using a 1k & cap in series from pins 1 to 5. This is actually supposed to be a bass boost, which really just cuts the midrange down & lets the low-er frequencies play. I don't care for the bass boost, but this is a GREAT way to cut down on the nasty hiss I'm getting.
It just sounds poor. I'm getting distortion at like 1/8 of a turn.
Thoughts?
This is a pretty common circuit.
Thanks
Pins 1 & 8 are left open, so a Gain of 20. Most of these "headphone amplifiers" have a gain of 50, or 200. I can't even toss a 10uf betweens 1 & 8 without nasty distortion.
I posted this on another forum. Just was curious to see how people were implenting their single chip lm386 audio amplifiers
*I'm using a 1k & cap in series from pins 1 to 5. This is actually supposed to be a bass boost, which really just cuts the midrange down & lets the low-er frequencies play. I don't care for the bass boost, but this is a GREAT way to cut down on the nasty hiss I'm getting.
It just sounds poor. I'm getting distortion at like 1/8 of a turn.
Thoughts?
This is a pretty common circuit.
Thanks
For some reason I can't get the datasheet at TI.com. You'll have to give us more details. What Voltage are you using? You do know it only produces 1.6W? Hiss is very odd though. Tere's no reason for anything modern to hiss badly. Also we need to know what your load is. That will affect where it clips. Min load is 4 ohms.
Have just seen your video and you should replace that 10uF back with a resistor. It probably won't hurt to much but the amp will probably be oscillating since it's driving a 100nF load. Put in 4.7 or somewhere near. 2200 or 4700 in place of the 470u will do better still.
Have just seen your video and you should replace that 10uF back with a resistor. It probably won't hurt to much but the amp will probably be oscillating since it's driving a 100nF load. Put in 4.7 or somewhere near. 2200 or 4700 in place of the 470u will do better still.
"Modern"? The LM386 is at least 25 years old, probably older. I have some in my parts box that are easily 25 years old. The Datasheet is available at Speaker Amplifier and Modulator - Mid/High-Power Amplifier - LM386 - TI.com
You might try a capacitor in series with the input pin. The input is 50 kΩ so you could use say 1 µF or higher. The output capacitor should be big, 470 µF is OK, but you can go higher. Voltage of this cap should be more than 1/2 the supply.
I don't see a bypass capacitor mounted directly at the chip. This is absolutely necessary. Put a 47 to 100 µF cap from pin 6 to 4 directly over top the chip, keeping the capacitor leads trimmed as short as possible. Your capacitor on the breadboard power supply rails can stay there, but you also need the "local" bypass capacitor right at the chip. Normal (aluminum) electrolytic would work, but a tantalum may be better. Keep the output zobel as designed: 0.05 (0.047) µF and 10 ohms. Don't mess with that, it is often necessary for stability. Certainly, taking out the resistor and replacing it with a capacitor creates a pure capacitive load which is a bad thing in general.
Your 4 ohm speaker may ultimately be the cause of all your problems. Try a 8 or 16 ohm speaker. This chip is not designed for difficult 4 ohm loads, and the power output is very low with 4 ohm (only 200 mW max, vs. over 1 W at 16 ohm). Since the little chip will distort at over ~ 200 mW into 4Ω, that's less than 1 V. At a gain of 20, that is less than 0.05 VRMS at the input. So you will never be able to turn the volume pot very high.
You might try a capacitor in series with the input pin. The input is 50 kΩ so you could use say 1 µF or higher. The output capacitor should be big, 470 µF is OK, but you can go higher. Voltage of this cap should be more than 1/2 the supply.
I don't see a bypass capacitor mounted directly at the chip. This is absolutely necessary. Put a 47 to 100 µF cap from pin 6 to 4 directly over top the chip, keeping the capacitor leads trimmed as short as possible. Your capacitor on the breadboard power supply rails can stay there, but you also need the "local" bypass capacitor right at the chip. Normal (aluminum) electrolytic would work, but a tantalum may be better. Keep the output zobel as designed: 0.05 (0.047) µF and 10 ohms. Don't mess with that, it is often necessary for stability. Certainly, taking out the resistor and replacing it with a capacitor creates a pure capacitive load which is a bad thing in general.
Your 4 ohm speaker may ultimately be the cause of all your problems. Try a 8 or 16 ohm speaker. This chip is not designed for difficult 4 ohm loads, and the power output is very low with 4 ohm (only 200 mW max, vs. over 1 W at 16 ohm). Since the little chip will distort at over ~ 200 mW into 4Ω, that's less than 1 V. At a gain of 20, that is less than 0.05 VRMS at the input. So you will never be able to turn the volume pot very high.
Put 5.6K in series with 10u from pin 8 to pin 1. This will give you 6dB of bass boost, but gain will remain 20 1.2kHz on and up.
Do put a 1uF from bypass to ground.
You may be over-driving the chip -- back off the input voltage.
Last edited:
In the datasheet (under application hints), it states that when you do not bypass pin 1 to 8 for gain boost, then the minimum value of R between pin 1 and 5 is 10 kΩ, otherwise the amp will become unstable. You are exceeding this minimum.
I remember as a youngster having no real success with various power amp chips of the day. They always seemed to be oscillators first and amplifiers second and easily provoked into failure. I've learnt a lot since then but discrete designs were always more successful in my early experiments 😛
The LM386 was a favorite for ham radio low power operation since it's so simple to implement. It's generally worry free.
From your video, I see some issues. Big problem is your grounds. If that is the grounds on the right, you have signal, power and output all together and that will cause oscillations and crackle with music. You must isolate the signal grounds from the output. No wonder it won't work at higher gains! The cap on the supply, and the output RC filter must be close to the IC. You should not spread the components out so far like you have. Don't use the LM386 with 4ohm loads. Use 8 ohm minimum with 9v supply.
that sounds like ... dunno. hard to belive.
i never managed a discrete amp on the first attempt, in fact i had waaay too many problems with them.
chips on the otherhand are very easy. if done with least minimum care...
I remember using the LM386 on a project in elementary school, something like 4th grade. it came in an 8 pin DIP package back then. That was 36 years ago. This chip has been around forever and has always been a piece of cake to implement.
Dude, I was 9 and I was able to make it work well. What's up with that?
Dude, I was 9 and I was able to make it work well. What's up with that?
Yes, those are my grounds on the right.
I need to isolate the signal grounds from my output?
How?
I do not understand how I can isolate a ground when all the grounds go to a ground rail, in which they are all connected.
Yes it doesn't make sense at first. Remember that in circuits with a varying signal, there are resistance, capacitance and inductance in a piece wire that can affect the circuit. What you want to do is to keep the high currents from the output to power ground from flowing in the input signal ground otherwise some of the signal is picked up at the input and can cause problems.
To accomplish this on your breadboard is to connect the power ground right to the IC's pin 4 row. Connect the input wire to the same row as well as the lead for the speaker return and output RC network. This forms a "star" ground pattern.
Also, twist the input and its ground wires together to form a twisted pair type cable that reduces their ability to pickup external signals. Doing the same with the speaker leads helps also.
Last edited:
Thanks for the twisted advice! Power GND is ran straight to Pin 4 on the IC.
Connect the input wire to the same row as well? You mean, connect the GND from the input wire directly to Pin 4 on the IC? & connect the GND from the loudspeaker & GND from the zobel network directly to Pin 4 on the IC?
Now the source GND, power GND, & loudspeaker GND are all at Pin 4 on the IC VS them all being on the GND rail.
Most of the audio amplifier schematics for this chip require a 250uF on the output. I'm currently running a 480uF. Why such a high value? 2200-4200uF ??? I can wire up a few 470uF's in parallel, but I couldn't make out a difference. I do not have an o-scope.
What is the difference between running a 100uF capacitor from Pin 6 to GND rail VS running a 100uF overhead the chip from 6 to 4. Pin 4 is GND.
**Note to self: do not hot glue anything on the breadboard. I ended up destroying my breadboard. Anyways, I broke it into pieces & threw the bottom plate (which is conductive btw) into the garbage. I think this was my problem. The chip sounds a lot better on my new "broken" breadboard.
I am not using the 10k POT. So at full tilt, the POT is not resisting at all & at least tilt the pot is resisting 10k?
Idk. I can barely get to 50% volume on my lap top before it distorts. The 33nF on the input helped a lot. The lower frequencies distort much quicker than the higher ones. Honestly, I can't tell whether my lap top is louder or my single chip IC is louder. Wouldn't exactly call it an amplifier, more like a passive component, lol
Connect the input wire to the same row as well? You mean, connect the GND from the input wire directly to Pin 4 on the IC? & connect the GND from the loudspeaker & GND from the zobel network directly to Pin 4 on the IC?
Now the source GND, power GND, & loudspeaker GND are all at Pin 4 on the IC VS them all being on the GND rail.
Most of the audio amplifier schematics for this chip require a 250uF on the output. I'm currently running a 480uF. Why such a high value? 2200-4200uF ??? I can wire up a few 470uF's in parallel, but I couldn't make out a difference. I do not have an o-scope.
What is the difference between running a 100uF capacitor from Pin 6 to GND rail VS running a 100uF overhead the chip from 6 to 4. Pin 4 is GND.
**Note to self: do not hot glue anything on the breadboard. I ended up destroying my breadboard. Anyways, I broke it into pieces & threw the bottom plate (which is conductive btw) into the garbage. I think this was my problem. The chip sounds a lot better on my new "broken" breadboard.
I am not using the 10k POT. So at full tilt, the POT is not resisting at all & at least tilt the pot is resisting 10k?
Idk. I can barely get to 50% volume on my lap top before it distorts. The 33nF on the input helped a lot. The lower frequencies distort much quicker than the higher ones. Honestly, I can't tell whether my lap top is louder or my single chip IC is louder. Wouldn't exactly call it an amplifier, more like a passive component, lol
Last edited:
IMO, this was more true of the 386 than any other part. They were famous for oscillating and general misbehavior. We didn't know as much about grounding and construction and there was no helpful Internet community to help out the newbie. Properly implemented and bypassed, they work fine in their limited little way, but there are way better choices now.
I breadboarded the lm386. It sounds fine and does not oscillate. I get plenty of sound without the gain boost cap across pin 1 and 8. I used a .27uf (.1uf should be fine too) film cap you see going over the chip from pin 6 to 4. This is supply decoupling and must be close to the IC pins.
I made a jumper from pin 4 over to the right for the output RC filter (zobel), power ground and speaker ground.
470uf is fine for the output coupling cap. Larger values pass lower bass. If you can't tell the difference, you don't need the larger cap. Change the output zobel back to datasheet recommended value. .1uf is fine for the cap.
The smaller 'lytic cap is Vs bypass. I'm using a 9v wallwart for power and there is a low hum in the speaker without it. The white jumper wire returns it to the signal ground side.
I made a jumper from pin 4 over to the right for the output RC filter (zobel), power ground and speaker ground.
470uf is fine for the output coupling cap. Larger values pass lower bass. If you can't tell the difference, you don't need the larger cap. Change the output zobel back to datasheet recommended value. .1uf is fine for the cap.
The smaller 'lytic cap is Vs bypass. I'm using a 9v wallwart for power and there is a low hum in the speaker without it. The white jumper wire returns it to the signal ground side.
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
- Status
- Not open for further replies.