I designed a basic audio amp for learning purposes and now I'm trying to upgrade the differential stage with current mirror. However, I encountering a problem - the currents on both branches of the differential amp are different beacuse of the additional current from T1's base (the mirror itself works as expected tough)

In fact, I achieved better symmetry with simple resistor rather than a current mirror. Am I doing something wrong?

I've marked the problematic currents in yellow squares on picture below. As You can see, the difference is about \$23µA\$ which is quite high comparing to \$3µA\$ difference I achieved using a simple feeding resistor.

Agreed - this is just a simulation (and not the best one for sure) but still...

The circuit can be found here!

enter image description here

  • 1
    \$\begingroup\$ Zobacz co sam Douglasa Selfa ma do powiedzenie na ten temat: "Note must be taken of the base current the VAS transistor will draw from the input stage; it must not be allowed to unbalance the input transistor collector currents significantly. A high VAS beta helps, but will not be found in company with a high Vce(max). The EF-VAS is in my experience wholly free of this problem because of the extra stage of current amplification." \$\endgroup\$ – G36 Apr 4 at 6:18
  • \$\begingroup\$ Also, notice that the input stage tail current values are 4mA to 6mA in Douglas designs. \$\endgroup\$ – G36 Apr 4 at 6:28
  • \$\begingroup\$ @G36 yep, I missed that part in Selfs book. Emitter follower is the answer. Thank You! \$\endgroup\$ – Dawid W Apr 6 at 9:02

If the current mirror was perfect then the 300uA current and the left-hand 294.1uA would be equal to each other. As things stand the mirror is pretty good with only 5.9uA difference between the two sides. A mirror can be improved with the addition of an extra two transistors to cancel out the effect of the extra current in the right hand side of the mirror due to the base current and also to make the collector voltages equal to each other to cancel out Early effect. If an extra transistor is added either side of the mirror then you will need to turn T1 into a Darlington (two transistors) to give enough head room for the extra left-hand mirror transistor.

At dc (quiescent conditions) there will always be a mismatch in the collector currents of the differential pair because T1 base needs some base current to hold the amplifier's output at the mid-rail voltage. However if the mirror is designed to be almost perfect with the left-hand mirror transistor collector current virtually equal to the right-hand diff-pair transistor collector current then you will find that the diff pair collector currents are very well balanced when an ac signal is being amplified.

This is the advantage of using an accurate current mirror, the diff pair collector currents will be well balanced with an ac signal. If the current mirror is not accurate or you use a resistor then, as you have found, the diff pair collector currents may be well balanced for dc but when ac is applied there will be significant 2nd harmonic distortion due to imbalance.

Another advantage of using a mirror over a resistor - It doubles the amplifier's slew rate.

  • \$\begingroup\$ Uhm thanks but it's not 300\$uA\$ and 294\$uA\$ that bother me but rather 300\$uA\$ and 323\$uA\$ - the mirror works fine but the way it's connected to the differential amp makes it pretty useless since there is still lack of symmetry present at the differential amplifier collectors. I'm not sure if it supposed to be connected this way and if such difference is acceptable (I think not). \$\endgroup\$ – Dawid W Apr 1 at 17:51
  • 1
    \$\begingroup\$ @Dawid W To summarize my answer if it was unclear. - You usually get diff amp collector currents asymmetry with a dc signal, as you are experiencing, this is to be expected and is unimportant. When you apply an ac signal, if the mirror is near to perfect, the diff amp collector currents will be almost symmetrical and it is only when amplifying ac that it is important to have as near to perfect balance in the diff amp in order to minimize 2nd harmonic distortion. \$\endgroup\$ – James Apr 1 at 21:26

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.