Audio Amplification Problem

Audioguru again

Joined Oct 21, 2019
6,673
The load has nothing to do with crossover distortion. The distortion is much worse at lower levels because the output transistors need +0.7V and -0.7V to turn on since they are not biased for class-AB like most other opamps and audio amplifiers.

You can eliminate the crossover distortion if you add a resistor from the output to ground or to the positive supply so that one of the two output transistors smokes when operating in class-A. but you cannot reduce its hisss and you cannot improve its lousy frequency response.

The datasheet of an LM380 shows that its input resistance is 150k ohms. An LM386 is 50k ohms.
 

Wolframore

Joined Jan 21, 2019
2,609
Hey @Audioguru again here's the schematic for the LM380 - are you saying this is not class AB? I agree there's a distortion issue but how would you classify this output stage? Modified AB?
1605896746651.png
 
Last edited:

Wolframore

Joined Jan 21, 2019
2,609
This is the screenshot posted... it states LM358... too many LM3xx numbers floating around. hard to keep them straight. The LM358 does have a strange output for an opamp. What were they thinking? Ugh I've used these before and perhaps it wasn't a great idea... even if it wasn't used for audio. Thanks for pointing this out AG.

1605904324904.png

this is an ugly output for an opamp: LM358 opamp
1605904716896.png
 

bloguetronica

Joined Apr 27, 2007
1,541
Indeed, the LM358 is a class B. I think I've used it as a comparator of sorts, years ago. At the time, I didn't wanted the typical comparator, or to deal with an open collector. It is a mess.
 

Audioguru again

Joined Oct 21, 2019
6,673
My post #100:
" The lousy old LM358 is never used for audio because it has severe crossover distortion (its distortion is not spec'd)".
It was designed to use very low idle current so they did not bias the output transistors properly.
 

Attachments

Last edited:

Wolframore

Joined Jan 21, 2019
2,609
Sam sorry for this tangent, there are other issues with the LM358, slew rate is 0.3v/uS and it sources way more current than sink, the output is not symmetric at all.
 

Audioguru again

Joined Oct 21, 2019
6,673
It does not matter if an LM358 can source more current than it can sink. If its output is not overloaded with current higher than the weakest transistor can produce, then the emitter-follower outputs are perfectly symmetrical.
My post showing the crossover distortion is perfectly symmetrical.
 

Wolframore

Joined Jan 21, 2019
2,609
That’s with a 2k load generally they’re used with 8 Ohm speaker. It look heavier on the top to me, it looks like the crossover wasn’t centered correctly on the scope.
 

Audioguru again

Joined Oct 21, 2019
6,673
That’s with a 2k load generally they’re used with 8 Ohm speaker. It look heavier on the top to me, it looks like the crossover wasn’t centered correctly on the scope.
Your numbers are mixed up.
A lousy old LM358 is a very low current opamp with crossover distortion.
LM380 or LM386 are power amps.
 

Wolframore

Joined Jan 21, 2019
2,609
Thanks for the correction - I'm doing it too.

Take a close look at this older CRT screen... it does not have the auto zero we're used to. Notice the top is higher than the 1V division while the bottom is on the 1V line. Now see that the crossover is below the center line. It does look pretty good besides the crossover which the opamp is trying to fix internally but it's there.

1605928424911.png
 

bloguetronica

Joined Apr 27, 2007
1,541
The LMV611 noise is far worse than the LM358. It has worse THD and PSRR, making it lousy in comparison, despite the input noise being significantly lower. I've managed to use it successfully to control the amplitude of an AD9834, despite the fact the LDOs had no margin for an effective PSRR, and despite the fact that USB Vbus is very noisy.

The output, with the frequency set at 8MHz, shows no noise at all. The LM358 would be even better. If a bad op-amp can be used for DDS generation, could be used for other things, even audio-related DC biasing stuff, since the human ear is rat level lousy compared to an oscilloscope.
 

Attachments

Wolframore

Joined Jan 21, 2019
2,609
the LMV611 is not worse, in fact it has a Class AB output. The slew rate is slow but faster then the LM358 and claims THD of 0.022% which I would trust. The PSRR is 100 dB for both.

We have many opamps for many applications. It seems that the LM358 were designed for audio for Walkie talkies when they were operating off 9V batteries to prolong battery life. Walkie talkies never sounded that great and was more Important to save power.

if you have to imitate that sound like in a recording studio this might be a great use for one
 
Last edited:

bloguetronica

Joined Apr 27, 2007
1,541
It is worse in terms of noise, in my particular case, mainly because of PSRR. The objective of my comment is that one can't say a particular component is useless or lousy, because it always depends on the application (except for the LM741 that, frankly, had its time but no longer finds use in modern devices - it is still used to teach electronics, though).
 
Top