Hi everyone,

I'm trying to design my own audio amplifier as an education project. My input is going to be a differential long tail pair with current source and current mirror. I decided to draw up my circuit in LTSPICE and used relatively common 2N5401 and 2N5550 transistors.

Here is a diagram:

Here is the output measured across a 40kohm load (Input in green and output in blue):


As you can see from the measurements, the output is very distorted. Does anyone have any idea how to fix this? I left the LTSPICE schematic if anyone wants to take a closer look.

 

30pF into 40k passes high frequency distortion harmonics but cuts the 1k signal down to almost nothing.
Change the 30pF capacitor to 20nF or more to see the amplified signal.

 

Do you understand that C1 and R5 form a 1-pole a high-pass filter, and what its corner frequency is for 30pF with a 40kΩ load?

Also since I1 is an ideal current source it can be tied to ground.
V2 thus serves no purpose and is not needed.

 

Ah yes, I see the error in my ways. The signal is nice and clean with low distortion. To address crutschow's point, I ended up replacing the ideal current source with a non ideal source (2 pnp transistors in a feedback arrangement).

 

with a non ideal source (2 pnp transistors in a feedback arrangement).

Typically a current source is done with 2 transistors in a current-mirror configuration.
Is that what you have?
A near ideal current-source can be made with a 3-transistor Wilson current-mirror.

 

Typically a current source is done with 2 transistors in a current-mirror configuration.
Is that what you have?
A near ideal current-source can be made with a 3-transistor Wilson current-mirror.

Yes that's what I'm using right now. It seems to work well enough.

Would you use a wilson current mirror in an audio application?

 

Would you use a wilson current mirror in an audio application?

Probably not, since the current mirror doesn't have to be that "stiff".
So it shouldn't make a difference in an audio amp.