Yet another amp - simple inverted with 10x gain, mosfet output stage

[WuYit]

A more radical change would be necessary, otherwise you lose performance.

[juma]

Let's wait and see what Sonny thinks about changing anything - this is his amp and his thread ...

[sonnya]

Input impedance.

It's correct if you going to feed it from an passive preamp. Then it's to low.

If you are going to feed it from an preamp or attenuator with an bufferstage. If this buffer is designed the wrong way around then it will be degraded.

As for the construction of using inverted design as this is: The feedback network as is it should be. I would not use higher impedance for LM3886 as well.

National states:

"With the increase in gain, there is
also a reduction of the power bandwidth which results in a
decrease in feedback thus not allowing the amplifier to respond
quickly enough to nonlinearities. This decreased ability
to respond to nonlinearities increases the THD + N specification.
The desired input impedance is set by RIN. Very high values
can cause board layout problems and DC offsets at the
output. The value for the feedback resistance, Rf1, should be
chosen to be a relatively large value (10 kΩ–100 kΩ), andthe other feedback resistance, Ri, is calculated using standard
op amp configuration gain equations. Most audio amplifiers
are designed from the non-inverting amplifier configuration"

This will give a low inputimpedance as well.

But i can follow your point that you want a higher input impedance.

Juma: I have just seen your post.

I did test it yesterday. But because of difference in current gain, you will have with 10k resistor to ground around 500mW on the "in+"

I have drawn an alternative version i will post a bit later.

[sonnya]

The reason to post this circuit was that it is symmetrical with low part count to relatively low price.

I have a highperformance circuit but that is not cheap at all.

But i got your points.

[WuYit]

We'll get nowhere with that kind of explanations. What is the signal source?

[lineup]

Sonnya
Where do you think DC-offset adjust should be done?
In a mirror perhaps?

I get -126 mV offset in my simulator.

[WuYit]

Without increasing costs...

[sonnya]

WuYit:

that is okay. Can you get the JFET's as SMT and complementary?

The parts i have choosen besides of the capacitors and mosfet are SMT.

Lineup:

You can get balance in current that is not a big problem. Its more the voltage drop VGS and VBE that will make up the big problem.

My Mirand A1 v11 design uses a servo as i has not added dc blocking capacitors. This one is not cheap and i have a bit more sofisticated circuit on the drawing board.

I wanted to make a cheap symmetrical amp. One way is to use mosfet in the output.

In this one i wanted to remove the DC servo to make i cheap, and i do not think we are getting away without a DC blocking Cap.

In an noninverting design i am down to 33mV at the output with a 4K feedback resistor. But then the step response is not good enough. That i ahve to solve.

[lineup]

You say bias 500 mA. At least.
With 100-200 mA there is higher THD.
This circuit works best in Class A or near full Class A,
where the THD is much lower.
I get a value of THD 0.005% with bias 500 mA.

[lineup]

I tried the JFET buffer, but got too much THD (0.037%)

So I added one OPA134 preamp at the input.
Now THD is only 0.004%.
See diagram.

To adjust DC-offset to zero I use one 50k-trimmer (R4)

I used transistors BC550C, BC560C and one 2N5401, 2N5551