LM386 low noise alternative?

[neazoi]

LM386 low noise (hiss) alternative?

Hi, I am building a direct conversion (regenerative) receiver that is followed by this amplifier schematic **broken link removed** set at max gain.
The volume is good on my headphones, but there is too much noise. I confirmed that the noise is mainly from the audio amplifier, because I disconnected it from the rest of the receiver and the noise is still there.

Can you please suggest me an alternative with a very low noise and high gain, so that the phones are driven comfortably but with low noise?

Some chips I have available are: NE5532, OP07, AD745, TL071

Will they be of use? I am not sure about their output powers and gains and I am not too good in reading these technical things in datasheets.

An example with the AD745 is attached but I do not know how well (low noise high volume) will it perform compared with the LM386.

Some help would be appreciated.

 

[betwixt]

The gain setting method in the .jp link is not optimal, it should be adjusted by a network between pins 1 and 8 rather than killing the negative feedback like it does.

My suggestion would be to split the stage, use an LM386 (wired properly) at low gain but add a pre-amplifier before it. The NE5532 would give best performance but in this application I doubt you would notice any improvement over the TL071 or similar amplifier. Unless you are using very high impedance headphones I wouldn't rely on a TL071 type of amplifier alone, they are good at low signal, low noise amplification but not very good at driving low impedance loads.

Brian.

 

[KlausST]

Hi,

Neazoi, you are an experienced user here in this forum.
Therefore I'm a bit surprised reading your post here.

You as about low noise Opamps.
My recommendation:
All the manufacturers of Opamps want to sell their devices, therefore they give you a lot of informations to motify you to buy their products.
--> They provide interactive selection guides at their internet sites to find the optimal device for your application.
Use this (Only a couple of years ago we didn't have this help)

I am not too good in reading these technical things in datasheets.

I'm sorry I have to say you need to learn this. There is no way around.

About the Opamp you mentioned.
NE5532 is a dedicated low noise dual audio Opamp. Exactely what you need.
OP07 is a dedicated low noise precision Opamp. A good choice, too.
TL071 is a Fet input Opamp. It has relatively low current noise. This is only good for extremely high input impedance.

Now you have to decide how to go on. You have to decide how much time to spend, how much money you want to spend for the devices and what quality you want to achieve.

From your post I assume just a low noise Opamp will do.
You may use the NE5532. It is cheap and good. But it is a very old device. There are more modern ones. Modern devices may have some benefits like low current consumption, higher gain-bandwidth, less distortion, higher output voltage.

But "noise" also depends on signal wiring, GND signal wiring...here only a good layout design is essential.

Klaus

 

[schmitt trigger]

Please define "noise ".

No, I am not asking that a spectral analysis but rather something of a simple description.
Is it wideband hiss? Hum? Single tone? Popping and cracking? Scratches?

On an amplifier circuit I built many years ago, a scratching sound was related to a defective electrolytic coupling capacitor.

 

[neazoi]

Please define "noise ".

No, I am not asking that a spectral analysis but rather something of a simple description.
Is it wideband hiss? Hum? Single tone? Popping and cracking? Scratches?

On an amplifier circuit I built many years ago, a scratching sound was related to a defective electrolytic coupling capacitor.

It is Hiss, broadband hiss. When I disconnect the lm386 input this hiss is almost the same like when it is connected. so this hiss is due to the audio amplifier.

Note, in this configuration, the lm386 is pushed to 70-74db of gain.
This is a noisy IC, so I wonder if something similar, in terms of volume and gain, but with less hiss, can be achieved with another opamp or a simple audio amplifier like the attached (circuit from another edaboard thread)

- - - Updated - - -

The gain setting method in the .jp link is not optimal, it should be adjusted by a network between pins 1 and 8 rather than killing the negative feedback like it does.

My suggestion would be to split the stage, use an LM386 (wired properly) at low gain but add a pre-amplifier before it. The NE5532 would give best performance but in this application I doubt you would notice any improvement over the TL071 or similar amplifier. Unless you are using very high impedance headphones I wouldn't rely on a TL071 type of amplifier alone, they are good at low signal, low noise amplification but not very good at driving low impedance loads.

Brian.

I have also seen this page. https://ludens.cl/Electron/audioamps/AudioAmps.html The second circuit, is not too complex yet he says it acheves much better performance than the lm386. What do you think?

 

[betwixt]

It really depends on the headphone impedance. Low impedance ones (8 Ohms for example) need relatively high currents to drive them, they are essentially ordinary loudspeakers on a headband. High impedance ones (2,000 Ohms for example) need lower current but higher voltage. The transistor output stage can never deliver more than 9/1000 Amps (9mA) to the load because of the 1K resistor so it isn't suitable for driving a low impedance. You will note it says "1K" headphones on the schematic, indicating it is designed to work into a high impedance load only. The LM386 is the opposite, it is optimized to deliver high current to its load so it will work better driving a low impedance.

You can 'under load' the LM386, it will happily drive a higher impedance although maybe not quite as loud. As you noted though, it is a noisy device, especially when driven well beyond it's intended gain by shunting the negative feedback path to ground. If you do the signal amplifying in a low-noise amp and keep the LM386 only as the power driver it will perform far better and should work with headphones or a loudspeaker, you get the best of both Worlds!

Brian.

 

[neazoi]

It really depends on the headphone impedance. Low impedance ones (8 Ohms for example) need relatively high currents to drive them, they are essentially ordinary loudspeakers on a headband. High impedance ones (2,000 Ohms for example) need lower current but higher voltage. The transistor output stage can never deliver more than 9/1000 Amps (9mA) to the load because of the 1K resistor so it isn't suitable for driving a low impedance. You will note it says "1K" headphones on the schematic, indicating it is designed to work into a high impedance load only. The LM386 is the opposite, it is optimized to deliver high current to its load so it will work better driving a low impedance.

You can 'under load' the LM386, it will happily drive a higher impedance although maybe not quite as loud. As you noted though, it is a noisy device, especially when driven well beyond it's intended gain by shunting the negative feedback path to ground. If you do the signal amplifying in a low-noise amp and keep the LM386 only as the power driver it will perform far better and should work with headphones or a loudspeaker, you get the best of both Worlds!

Brian.

Thanks, I will do as you suggest.
So, even in 2016 it seems there isn't yet a single IC that can combine these features! I am curious...

I wonder if this would work for the purpose https://ludens.cl/Electron/audioamps/AudioAmps.html (the second schematic). It isn't much more complex, yet it uses one IC not two. But he does not mention the output power or the gain. Any estimation?

 

[chemelec]

If you built the Correct Circuit, using Pins 1 & 8 for gain, it would probably be better.
And Anytime your Powering HEADPHONES from ANY POWER AMP, You should Include a 100 Ohm Series Resistor going to the Headphones.
Both these Will reduce that Noise.
The Resistor will protect the headphones from getting too much Power and protect your Ears from Extreme Loudness.

 

[Audioguru]

The problem is the very poor performance of the Regenerative Receiver, not the amplifier. The LM386 works very well with a real radio circuit.
The regenerative receiver is AM anyway and AM produces lots of noise. You can fiddle with a Regenerative Receiver to make it "slope detect" an FM radio station but it still picks up noise.

You said the LM386 produces hiss (high audio frequencies). If you add a lowpass filter then the hiss (and high audio frequencies) will be reduced, just like an old AM radio.

 

[neazoi]

The problem is the very poor performance of the Regenerative Receiver, not the amplifier. The LM386 works very well with a real radio circuit.
The regenerative receiver is AM anyway and AM produces lots of noise. You can fiddle with a Regenerative Receiver to make it "slope detect" an FM radio station but it still picks up noise.

You said the LM386 produces hiss (high audio frequencies). If you add a lowpass filter then the hiss (and high audio frequencies) will be reduced, just like an old AM radio.

The Regenerative stages are actually impressively quiet, compared to my IC-728! It is oscillating so it is a DSB detector. The hiss is due to the audio amplifier, I have verified it by disconnecting the input to it.
The LPF you propose could cure the problem, whereas not kill the desired gain. I have not connected the amplifier to the shound blaster input to see the hiss spectrum, but it sounds exactly like the hiss from old audio tapes.

How should I add an LPF? A simple 0.1uF capacitor shunted at the output of the amplifier maybe?

- - - Updated - - -

If you built the Correct Circuit, using Pins 1 & 8 for gain, it would probably be better.
And Anytime your Powering HEADPHONES from ANY POWER AMP, You should Include a 100 Ohm Series Resistor going to the Headphones.
Both these Will reduce that Noise.
The Resistor will protect the headphones from getting too much Power and protect your Ears from Extreme Loudness.

I have not tried it, but it won't have so much gain, it will be about 40dB or so.
Thanks for the resistor tip, I almost burnt my ears when I flipped a switch yesterday...

 

[FvM]

Problem seems to be that you are using old LM386 with very sensitive headphones. LM386 gain according to datasheet is only 46 rather than 76 dB, by the way. The device doesn't even have a noise specification.

Obvious solution with easily available components is to cut back LM386 gain to G=20 and place a G=10 low noise preamp, e.g. NE5532.

 

[neazoi]

Problem seems to be that you are using old LM386 with very sensitive headphones. LM386 gain according to datasheet is only 46 rather than 76 dB, by the way. The device doesn't even have a noise specification.

Obvious solution with easily available components is to cut back LM386 gain to G=20 and place a G=10 low noise preamp, e.g. NE5532.

Elsewhere another circuit has been posted, claimed to have 30db of gain and plenty of power for driving headphones. It is not too complex.
Do you think it will do the job as far as concern volume and lower noise than the 386?

 

[neazoi]

You said the LM386 produces hiss (high audio frequencies). If you add a lowpass filter then the hiss (and high audio frequencies) will be reduced, just like an old AM radio.

To describe it better, the "hiss" is like the noise of old audio tapes, without dolby/noise reduction. A constant background noise.

I have also found this anti-hiss circuit https://www.wentztech.com/radio/Technical/Misc/lm386.html I want to try it today, but in my circuit pin 8 is left unconnected, so I do not know if connecting to this pin will ruin things.

 

[chemelec]

With a 10uF Cap placed between Pins 1 & 6 you can get a Gain of 200. (Positive of this Cap goes to pin 1) Putting a 1K2 Resistor in series at pin 1 to the cap, will give a gain of 50, and Other resistor values will give other Gain values.
Check out the Data Sheet.

This is a VERY USEFUL, Low Power Amp!

 

[Audioguru]

With a 10uF Cap placed between Pins 1 & 6 you can get a Gain of 200.

Correction, the capacitor goes between pin 1 and pin 8.

 

[chemelec]

YES, Sorry for the TYPO.

 

[Audioguru]

The anti-hiss circuit is too simple and it will simply cut all high audio frequencies including hiss. Dolby noise reduction was not simple, it was smart. So it boosted (pre-emphasis) low level high audio frequencies during recording then cut (de-emphasis) all high audio frequencies when levels were low during playback which reduced the hiss.

 

[neazoi]

The anti-hiss circuit is too simple and it will simply cut all high audio frequencies including hiss. Dolby noise reduction was not simple, it was smart. So it boosted (pre-emphasis) low level high audio frequencies during recording then cut (de-emphasis) all high audio frequencies when levels were low during playback which reduced the hiss.

The anti-hiss circuit had absolutely no effect, I tried it.

 

[schmitt trigger]

The 10pF capacitor in the anti-hiss circuit seems to be waaaaaaaaaaaaaay too low to be an effective. Its turnover frequency appears to be around 1.5 Mhz.

Substitute it for 1,000pF and then 10,000pF and let's see what happens.

 

[neazoi]

The 10pF capacitor in the anti-hiss circuit seems to be waaaaaaaaaaaaaay too low to be an effective. Its turnover frequency appears to be around 1.5 Mhz.

Substitute it for 1,000pF and then 10,000pF and let's see what happens.

With a 2.2nf and 10k it oscillated.