Class AB stage frequency roll off

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I tried simulating and building a Class AB amplifier to drive a headphone. I have attached a schematic of what I built. I simulated a frequency response plot for the amplifier and the frequency gets rolled off after about 160hz. The 1uF cap and 100k ohm resistor forms a high pass filter at the input right.
Fc = 1/ (2x PI x R x C).

Then from my calculation shouldn't the cutoff frequency be 1.59Hz ? When I built this amp I also get freq roll off after 170Hz. Did I do something wrong ? I also attached the LTspice file that I sim with.
 

Attachments

  • Class ab schematic.jpg
    Class ab schematic.jpg
    38.6 KB · Views: 143
  • class ab freq plot.jpg
    class ab freq plot.jpg
    212.3 KB · Views: 145
  • Class AB Output stage.asc
    2 KB · Views: 31
The 1uF cap and 100k ohm resistor forms a high pass filter at the input right.


Yes, they would if there will be no 2k diode bias.
Since voltage source (placed at supply) have very low impedance you shunt your 100k with something like 1k to groung.
Replace 2k resistors with high-impedance current sources to supply rails and you’ll get your predicted filter response...
 
I don't think so, because the load impedance times the current gain plus one of the transistors is still small compared to 100 kohm.

That is, suppose the transistors had infinite gm and ro but finite current gain beta. The signal voltage across R3 then also occurs across R4 (neglecting signal voltage drop across the bias diodes). The current through R4 is the voltage swing divided by R4, this current has to be supplied by the emitters of the transistors. Their base currents are beta + 1 times smaller than their emitter currents. Hence, after the improvement suggested by BesPav, you effectively see beta + 1 times R4 as input resistance of the complementary emitter follower stage. This is still much less than R3.
 
Last edited:

PRR

Member
Joined 2003
Paid Member
> The 1uF cap and 100k ohm resistor forms a high pass filter at the input right.

Consider the *whole* circuit.

As Marce3lvdG says, you have to drive the Bases. Their impedance is about hFE higher than their load. Say 100*32r is 3,200 Ohms. Much less than the 100K you are looking at!

AND *both* 2K resistors!! That's 1k right there.

100k||3k1||1k is 750 Ohms. I compute 212Hz.

If hFE could be infinite, we have 100k||1k which is 990 Ohms and 161Hz.

hFE=100 is probably low for the parts you show, but they won't be infinite hFE. So the computed bass-cut is somewhere lowere than 212Hz but higher than 161Hz. Your sim suggests hFE close to infinity, which suggests the dead-idle operating point is so near "off" that the transistors are not loading the source. (SPICE freq plot assumes utterly zero signal.)
 
Thank you for the answers. My knowledge of electronics is at a basic level. I simulated and built this circuit for learning. Now I realise there is so much more to this. Now i have to google most of the terms in the answers as I do not understand alot of it. But this is also good as I could learn.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.