I just made simplest possible full bridge class d amp where i have triangle oscilator,two comparators and full bridge output....also i made negative feedback loop using op amp.Feedback loop is made in basic mode as diferential amplifier.All is made as basic school example class d model.Please dont ask for schematic because it is as simple as i said.Amp works great and output noise is ultra low.When i connect negative feedback loop ( i can easy change op amp ) it works....gain is going down and all works fine.Issue happens at saturation and clipping.If i use LM358 as feedback op amp i have high pitch cracling at clipping (low power but noticeable).... if i use NE5532 i have serious crackling (high power).....If i use nothing and i remove negative feedback loop i have excelent sound and no cracling on clipping.I know that op amps i mentioned are two worlds....but among all i tested these two works fine.
I tried to lower input levels on both inputs of diferential amplifier design-didnt help.I tried to do various compensations - no luck.Main difference between these two op amps is that LM358 can swing output to ground and NE5532 cant....this pushed me to decrese input levels even more.....didnt help.Any idea is welcome.I will test and publish results.
I tried to lower input levels on both inputs of diferential amplifier design-didnt help.I tried to do various compensations - no luck.Main difference between these two op amps is that LM358 can swing output to ground and NE5532 cant....this pushed me to decrese input levels even more.....didnt help.Any idea is welcome.I will test and publish results.
The LM358 hasn't got the power bandwidth for audiol, it clips at 5kHz or so with large output swing and slews at about 0.5V/µs (*). It may well be increasing the distortion at higher frequencies.
The main differences between the NE5532 and LM358 is the far larger gain-bandwidth product and good symmetric output drive capability of the audio opamp. LM358 is not for audio, I say it again!
(*) NE5532 slews at 9V/µs
The main differences between the NE5532 and LM358 is the far larger gain-bandwidth product and good symmetric output drive capability of the audio opamp. LM358 is not for audio, I say it again!
(*) NE5532 slews at 9V/µs
I agree .... LM358 will not work fine as well as NE5532.....but what i did is that i made both ICs to work with very small output signals in relation to power supply voltage (at clipping i have 2Vpp and power supply voltage is 12V).So i didnt have clipping at all.I can exclude negative feedback completely and measure only its output which i did.What i noticed is that both amps behave the same at clipping level.....kind of selfoscilation signal appears on output.....one ceramic cap and problem is solved.Thank you for your explanation.
The National Semiconductor (now part of TI) LM358 is the dual version of the BJT LM324 (released in 1972), which was a very old design that could handle ground referenced inputs (uses darlington lateral PNP input differential pair), mostly designed for automotive applications. It was an improvement on the LM741 which could not handle ground referenced inputs.
It is a general purpose cheap op amp with very low slew rate, poor input offset and and not low noise input.
There are many op amps that that operate way beyond the capabilities on the LM358 (dual).
Application Design Guidelines for LM324/LM358 Device
TI has 578 dual op amp variations to choose from, many are much better than the LM358: https://www.ti.com/amplifier-circuit/op-amps/products.html#p480=2;2
It is a general purpose cheap op amp with very low slew rate, poor input offset and and not low noise input.
There are many op amps that that operate way beyond the capabilities on the LM358 (dual).
Application Design Guidelines for LM324/LM358 Device
TI has 578 dual op amp variations to choose from, many are much better than the LM358: https://www.ti.com/amplifier-circuit/op-amps/products.html#p480=2;2
Last edited:
So if i have constant voltage outputwith zero THD .... absolutely perfect amp with voltage negative feedback loop......and i connect the speaker at for example 100W output.....and output voltage is still perfect with zero THD.....will current be perfect and with zero THD and also constant too?I am talking about single frequency input of lets say 1KHz. 
The AC version of Ohm's law says that the amplifier's output current i = v / z
so the output current of a perfect amplifier depends on the inverse z of the speaker.
If z has nonlinearities, so will the i of the amplifier. If z has a nonlinear phase, so will i.
This does not take into account transients due to stored energy in the L and C in the crossover.
Load current transients for certain nonperiodic inputs can be several times larger in peak value
than the steady state current for the same peak input.
AES E-Library >> Input Current Requirements of High-Quality Loudspeaker Systems
so the output current of a perfect amplifier depends on the inverse z of the speaker.
If z has nonlinearities, so will the i of the amplifier. If z has a nonlinear phase, so will i.
This does not take into account transients due to stored energy in the L and C in the crossover.
Load current transients for certain nonperiodic inputs can be several times larger in peak value
than the steady state current for the same peak input.
AES E-Library >> Input Current Requirements of High-Quality Loudspeaker Systems
Last edited:
So if i maintain perfect absolute zero thd voltage on output , current wont be perfect because speaker impedance and inductance are not constant at all at any given path of the move of the speaker cone......which means that what is transmited out from speaker , even if the output voltage of the amp is superior and perfectly looking , will not bee with zero THD but with distorsion.What is the point to have negative feedback for voltage then?Why we are making singing sounding SMPs instead of power amplifiers which audio amplifier is?From what i described it is very simple to see that audio AMPLIFIER amplifies power and not just voltage.....and it should be made as output power-voltage controled source and not as voltage controled voltage source.....right?If we go deeper.....lets say that we have negative feedback loop for current and not for voltage.... in that case voltage will be mirror of all non linearities of the impedance of the speaker...so that wont work too....On other side if both are existing in one amp will that work ....i will say NO ..... so why we use negative feedback?Is amp without any feedback some more right way to go even it doesnt look nice on scope?
in that case voltage will be mirror of all non linearities of the impedance of the speaker...so that wont work too....On other side if both are existing in one amp will that work ....i will say NO ..... so why we use negative feedback?Is amp without any feedback some more right way to go even it doesnt look nice on scope?
Most loudspeaker systems are intended to be used with an amplifier that has a low source resistance,
perhaps 0.5R to account for the speaker wires. It could be done differently, but would need to be
an integrated system. This has become much more possible recently with active speakers becoming
more common.
Current-Drive - The Natural Way of Loudspeaker Operation
perhaps 0.5R to account for the speaker wires. It could be done differently, but would need to be
an integrated system. This has become much more possible recently with active speakers becoming
more common.
Current-Drive - The Natural Way of Loudspeaker Operation
Did you read my last post and to what is this what you wrote here related to in to my previous post?If i turn off negative feedback loop in my amp it has perfect sound...if i turn it on sound becomes more controled (in terms of what you set on EQ that stays constant) and flatter.....On other side with no feedback i have same control working (from EQ) but sound is more dynamic and natural.My personal thought about this is....no feedback loop and freedom to speakers !......and ...freedom to amp-speaker relationship!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.