Low voltage (9V) low noise/THD IC audio power amplifier substitute for LM386

[charles80]

Hello,

I'm currently testing this circuit (not my design) to detect small light changes/vibrations using light.




I'd like to know which IC I could use to replace the LM386 I'm currently using (don't know if it's LM386-N3 or an older one).

Here on this website the author measured his LM386 THD at 7%, while using a goldstar-made GL386 THD was just 0.6%, though it proved impossible for me to source these goldstar ICs.

Reading various opinions on the net, some people would say that NE5532 would be a good pre-amplifier but then I'd have to implement the circuit to use both NE5532 and LM386, which is beyond my current knowledge. Others would swear by the TDA7052A.

Since I already tried to minimize circuit noise by using low tolerance resistors (1%) and polypropylene/polystyrene capacitors whenever possible, my last noise source would be the LM386 which AFAIK is set to its highest gain by placing that 10uF capacitor between pin 1 and 8.

I know that noise (and also gain) could be lessened by removing that capacitor, but since that IC is noisy by itself or at least was never known to be low noise, I'd like to know if there is anything better I'd either replace it with directly (thus using an 8 DIP IC and then wiring pins properly) to try obtaining the best of both worlds or implement the circuit with something else like 2 ICs, though I'd need more help since I don't have the technical knowledge to do so myself.

Of course any suggestion/improvement is welcome.


Thank you very much

 

[Audioguru]

Your circuit is missing the very important 10 ohms in series with 0.05uf (47nF) at the output to ground (they are shown on every circuit in the datasheet) that prevents oscillation at a high frequency.
The datasheet for the American LM386 shows typical distortion of only 0.2% when its gain is low but with your high gain it might produce 2% distortion.

The LM386 is a cheap low power IC made for cheap clock radios. There are a few other cheap low power amplifier ICs with the same low spec's, the TDA7052A is not available in a through-holes package anymore.

 

[charles80]

Your circuit is missing the very important 10 ohms in series with 0.05uf (47nF) at the output to ground (they are shown on every circuit in the datasheet) that prevents oscillation at a high frequency.
The datasheet for the American LM386 shows typical distortion of only 0.2% when its gain is low but with your high gain it might produce 2% distortion.

The LM386 is a cheap low power IC made for cheap clock radios. There are a few other cheap low power amplifier ICs with the same low spec's, the TDA7052A is not available in a through-holes package anymore.

Thanks for your reply. I will try your suggestion to use a 10 ohm and 47nF though I don't know for sure if I should place them before the 470uF or replace it with them.

I've looked on eBay and found many TDA7052A available in DIP 8/same configuration as the Lm386 I'm using, so if it can be placed directly in place of the LM386, I could order some and give it a try.

About price or finding common/generic IC, I don't consider this an issue for me, I'd gladly increase the budget if it's needed to improve the circuit's performance.

That huge 200 gain I guess is needed/very helpful since its aim is to deal with tiny vibrations.


Thank you

 

[betwixt]

Resistor tolerance has nothing to do with noise level and the type of capacitor will make negligible difference. The LM386 isn't that bad but it is noisy and running it at flat out gain gives the worst noise level possible. I would stay with the LM386, remove the gain setting capacitor completely and invest in a low noise pre-amp ahead of it. The NE5532 is an excellent choice but beware of it's power supply requirements, it will not work down to the same low voltages as an LM386.

Brian.

 

[Audioguru]

Thanks for your reply. I will try your suggestion to use a 10 ohm and 47nF though I don't know for sure if I should place them before the 470uF or replace it with them.

Why not do what is shown on the designer's datasheet? The RC is connected with very short wires DIRECTLY to the output pin of the amplifier.

I've looked on eBay and found many TDA7052A available in DIP 8/same configuration as the Lm386 I'm using, so if it can be placed directly in place of the LM386, I could order some and give it a try.

The TDA7052A amplifier is completely different to an LM386. It has a bridged two-amplifiers output so its maximum supply is 6V when it drives an 8 ohm speaker then its output power is about 0.9W at clipping and it will be hot. Its distortion is a little higher than an LM386 at about 0.3% to 1%.

It has 3 times less gain than the LM386 and has a built-in DC volume control.

 

[charles80]

Thank you both for your input.

I thought that resistor and capacitor tolerances would have lessened overall noise.

At this point I don't know what to do since I'm unsure if LM386 at its lowest gain will output anything audible, so if any of you has any improvement suggestions I'd really be grateful.


Thank you

 

[Audioguru]

You obviously need a low noise preamp to drive a little low gain power amp. There are a few low noise opamps available. A very good one is the OPA134 that can work from a single polarity supply from 5V to 36V.

 

[betwixt]

The first thing to understand is that the LM386, or whichever type you use, does not give less output if you reduce it's gain. The output can remain the same, you just have to put a bit more in to get the same output.
In fact by using them at lower gain settings, you can actually improve the quality of the output because the negative feedback level is greater.

Designs of amplifiers that produce low levels of noise are difficult when high powered stages are also present on the same silicon, that's why it's better to utilize high power stages and low level pre-amplifiers in different devices. It also makes power supply filtering and good signal grounding far easier to achieve. As with most amplifiers, if you can run them off a higher supply voltage, their design and performance improves but if you have to use 9V, the pre-amplifier suggested by Audioguru sounds a good choice. Avoid standard op-amps, especially the common 741 type as they are probably noisier than the LM386! They were designed over 40 years ago and never intended for quality audio work.

Brian.

 

[charles80]

Why not do what is shown on the designer's datasheet? The RC is connected with very short wires DIRECTLY to the output pin of the amplifier.

I meant that I don't know if I'd still keep in place that 470uF capacitor and just add the 10 ohm and 47nF or place them (10 ohm & 47nF) and remove the 470uF, though I guess the former is the correct answer given this example circuit on texas instruments' datasheet:
**broken link removed**

There are a few low noise opamps available. A very good one is the OPA134 that can work from a single polarity supply from 5V to 36V.

Thank you very much for your suggestion.
I don't know what would be the right pinout wiring between OPA134 and LM386, so unless somebody would lend a hand I'm stuck at this point.

As of now, I think I understand that your suggestions about things to fix are the following:
- remove that 10uF capacitor between LM386's pin 1 and 8 to set it at its lowest gain setting (20)
- place a 10 ohm resistor and a 47nF capacitor connected directly to LM386's pin 5 as shown in texas instruments' datasheet and either keep that 470uF capacitor in place or remove it (unsure about this, see above question)
- use OPA134 as preamp and put it ahead of LM386

However, even if above statements may have reached correctly my mind, switching to action on the breadboard and later on the soldering station is another thing.

For example, would I need/would it be better to leave that 50K trimmer in its place (connected to LM386's pin 3 and then grounding one end and connecting to that capacitor the other one) or would I have to put it between OPA134's pin 1 & 8 (offset trim)?

Also, how would I connect OPA134 and LM386 and hence how would the circuit be upgraded? Should I have to put any capacitors to set OPA134 values or would I just have to place it ahead of phototransistor's signal?

I know that last thing you'd want to read is me asking for an update circuit drawing or hand-drawing since it may look like spoonfeeding, but I am otherwise far from catching up how to apply what you all kindly suggested.


Thank you

 

[Audioguru]

I meant that I don't know if I'd still keep in place that 470uF capacitor and just add the 10 ohm and 47nF or place them (10 ohm & 47nF) and remove the 470uF

A speaker works from AC. The 470uF capacitor passes AC and blocks the DC (half the supply voltage) at the output of the amplifier from going into the speaker and offsetting its cone.

I don't know what would be the right pinout wiring between OPA134 and LM386, so unless somebody would lend a hand I'm stuck at this point.

With a single polarity supply, the output of the preamp opamp is at half the supply voltage but the input voltage of the LM386 must be at 0V because it has internal biasing. So use a coupling capacitor between them to pass the AC but block the DC.

1) remove that 10uF capacitor between LM386's pin 1 and 8 to set it at its lowest gain setting (20)

Correct.

2) place a 10 ohm resistor and a 47nF capacitor connected directly to LM386's pin 5 as shown in texas instruments' datasheet

Correct.

3) use OPA134 as preamp and put it ahead of LM386

Correct.

However, even if above statements may have reached correctly my mind, switching to action on the breadboard and later on the soldering station is another thing.

The OPA134 opamp and LM386 power amp operate at high frequencies. A breadboard with all its strips of contacts and wires all over the place with capacitance between them will probably cause oscillation. Also, the wires will probably pickup hum and other interference. Instead of a breadboard use a pcb or a compact stripboard soldered circuit.

For example, would I need/would it be better to leave that 50K trimmer in its place (connected to LM386's pin 3 and then grounding one end and connecting to that capacitor the other one) or would I have to put it between OPA134's pin 1 & 8 (offset trim)?

The 50k pot is the volume control. Offset trim is needed on a DC amplifier but your preamp is an AC amplifier so offset trim is not needed.

Also, how would I connect OPA134 and LM386 and hence how would the circuit be upgraded? Should I have to put any capacitors to set OPA134 values or would I just have to place it ahead of phototransistor's signal?

The preamp opamp needs an input bias voltage at half the supply voltage. Add another coupling capacitor at the of the output of the volume control that feeds the biased input of the preamp opamp. The preamp opamp needs two feedback resistors to set its AC gain.

I know that last thing you'd want to read is me asking for an update circuit drawing or hand-drawing since it may look like spoonfeeding, but I am otherwise far from catching up how to apply what you all kindly suggested.

Sorry, but this simple project is far too complicated for your inexperience.

 

[betwixt]

Give the guy a chance!

Sorry for the quality, it's from my mobile phone while on a train journey!



It isn't complicated.

Brian.

 

[Audioguru]

It isn't complicated.

That is how it is done.
I would use a 470uF capacitor from +9V to ground to stop the circuit from motorboating (putt, putt, putt etc.) when the battery internal resistance increases as it runs down.

 

[charles80]

Give the guy a chance!

Sorry for the quality, it's from my mobile phone while on a train journey!



It isn't complicated.

Brian.

Thank you so much for your kindness

Buying you a couple beers would be the least I'd do.

I've got one more doubt: which value should be the capacitor you draw at OPA134's pin 3 (signal in) between the two 100K resistors and would it be ok to connect phototransistor's collector/gather at capacitor's IN and phototransistor's emitter at GND or should I apply a 10K resistor to phototransistor's emitter?

Thank you

 

[Audioguru]

I've got one more doubt: which value should be the capacitor you draw at OPA134's pin 3 (signal in) between the two 100K resistors

The value of the coupling capacitor is calculated with a simple formula involving the source impedance (the output of the photo-transistor), load resistance (the two Why does the software on this site do this to my text?
100k resistors in parallel) and the frequency that you want the low response to roll off. We do not know your desired roll off frequency. Try a 10uF film capacitor.

would it be ok to connect phototransistor's collector/gather at capacitor's IN and phototransistor's emitter at GND or should I apply a 10K resistor to phototransistor's emitter?

Use the original photo-transistor circuit.

 

[betwixt]

If you told us what the expected frequencies are it would help us advise on the values. As it is, the whole of the audio spectum would be handled quite well but the use of coupling capacitors anywhere in a circuit introduces a frequency dependent element because their reactance (like resistance but to signals rather than DC) increases as the frequency drops. In essence, they are less effective at passing low frequencies than high ones and this applies to the input coupling capacitor, the loudspeaker coupling capacitor and the one in the pre-amp feedback path. Don't lose sleep over it though, the change over frequency is gradual and in an amplifier of this simplicity, I wouldn't expect you to be striving for sound excellence.

What is your photo-transistor actually picking up?

Brian.

 

[charles80]

If you told us what the expected frequencies are it would help us advise on the values. As it is, the whole of the audio spectum would be handled quite well but the use of coupling capacitors anywhere in a circuit introduces a frequency dependent element because their reactance (like resistance but to signals rather than DC) increases as the frequency drops. In essence, they are less effective at passing low frequencies than high ones and this applies to the input coupling capacitor, the loudspeaker coupling capacitor and the one in the pre-amp feedback path. Don't lose sleep over it though, the change over frequency is gradual and in an amplifier of this simplicity, I wouldn't expect you to be striving for sound excellence.

What is your photo-transistor actually picking up?

Brian.

Sorry for the exceedingly late reply. That photo-transistor is picking up a laser beam reflected off a surface.

Thanks again

 

[kam1787]

https://ludens.cl/Electron/audioamps/AudioAmps.html

 

[Audioguru]

That photo-transistor is picking up a laser beam reflected off a surface.

It smells like a circuit for snooping (spying) on somebody.
Voices in a room vibrate the glass on a picture hanging on the wall. The laser beam shines into the room through the window and reflects off the picture glass with the vibrations modulating it.

 

[charles80]

It smells like a circuit for snooping (spying) on somebody.
Voices in a room vibrate the glass on a picture hanging on the wall. The laser beam shines into the room through the window and reflects off the picture glass with the vibrations modulating it.

I posted that exact phrase knowing that you or somebody else could draw this conclusion, so keep in mind it was no breaking intuition/idea.

If you've got any tips I'm more than willing to read it.

Thanks

 

[Audioguru]

The LM386 article with the high distortion is missing the important 10 ohm resistor and 0.047uF capacitor in series to ground at its output.

The photo-transistor is an emitter-follower with NO voltage gain. If it is made into a common-emitter circuit then it will have plenty of gain.