I've the following amplifier circuit:

DC couplet power stage

The two BDX77 are actually a single darlington BDX53C (I do not have the spice model for it), and the same for the two BDX78 (a single BDX54C).

It sounds quite well actually, even if it is not an usual topology for audio amplifiers (it is only didactic and temporary built on a breadboard). But it has a DC offset on the 4ohm load (a 10W speaker) that produces a higher current on the upper darlington transistor than the current of the lower one, with a consequent DC current through the load and a overall unbalancing in power consumption.

I thought that coupling the two emitter followers (the two darlington-connected BD139 before the power stage) in AC mode with the final stage would have solved. In effect, with the circuit below, the current balancing is ok (almost the same current through the BDX53 and BDX54), but the sound quality is largely degraded than with DC coupling: noise, distorsion of any type, clipping, some strange oscillation. I used an electrolitic capacitor (tried 100uF, 470uF, 1000uF). What I'm doing wrong (apart the bad amplifier topology, I know... ) ? it is about the electrolitic nature of the capacitor or that kind of coupling is completely wrong ?

thank you.

AC coupled power stage

  • 2
    \$\begingroup\$ Yo dawg, I heard you like darlingtons, so I put a darlington in your darlington, so you get some gain with your gain. \$\endgroup\$ Commented Jan 29, 2013 at 6:30
  • \$\begingroup\$ After trying to answer and from the resulting responses it is not clear what exactly you are asking. It is difficult to tell what is being asked here. \$\endgroup\$ Commented Jan 29, 2013 at 22:16

2 Answers 2


This seems like a few basically reasonable ideas taken to excess. The individual gain stages may be OK, but three of them simply strung together like that will multiply the non-linearities.

Apparently you discovered darlingtons somewhere, but this is overkill. They are going to add more distortion, but you don't have any feedback to correct things like that except that of the individual stages. You don't need both Q4 and Q5, especially considering the 1 kΩ pulldown on their output. That will be rather assymetric. Even with a single transistor, you have a lot lower impedance pulling up than down. There is also nothing keeping the DC operating point near zero except the open loop bias point of the last gain stage.

What are you trying to do, build a overall amp, a preamp, or a power amp? If you just want a overall amp, I'd replace the whole mess left of the diodes with a opamp. That will be easier to bias with a DC operating point of 0 and will provide lower impedance and more symmetric drive to the power stage. So you don't need a opamp to work on 30 V, give the power stage a small gain, like 2.

Given the impedance reduction you want from the power stage, you don't need three(!) transistors in each leg. With a opamp driving the power stage, you'd start at a lower impedance in the first place, and with even a little global feedback the effective output impedance will go down even more.

Note that this thing will have fairly high and somewhat unpredictable quiescient current. The efficiency will be pretty poor. This topology can't drive the output that close to the power rails, so there will be even less efficiency for the maximum output power. The last two transistors (Q10, Q11) are going to get toasty. They'd better be power transistors on a good heat sink, possibly with forced air cooling.

Since it's not clear what your overall point is, so it's hard to give concrete recommendations. What are you trying to accomplish by not just buying a audio amp?

  • \$\begingroup\$ Hi Olin, thank you for your comment. My goal is just self-didactic: what I had in mind was the possibility to couple one (or more) pure class A pre-amp stages (high voltage gain, low current gain) to a unity-gain stage (push-pull stage with complementary symmetry). Thank you. \$\endgroup\$ Commented Jan 29, 2013 at 8:22
  • \$\begingroup\$ @user: That doesn't explain much. Why is a particular topology the goal? That doesn't make sense. And as I pointed out, you don't have symmetry due to how the power stage is driven. Since you won't say what you are really trying to accomplish, there is little more that anyone can help with. \$\endgroup\$ Commented Jan 29, 2013 at 13:00
  • \$\begingroup\$ Why the topology is the goal is, as expressed already, just didactic: common emitter (voltage gain) coupled to a power stage class AB. Is that possible ? is that not possible ? Probably it doesn't make any sense to an electrical engineer, but my goal is not to build up a HiFi... I'd buy it if I needed it. My goal is just understand. Thank you. \$\endgroup\$ Commented Jan 29, 2013 at 21:15
  • \$\begingroup\$ @user1131951 - I don't think "didactic" means what you think it means. Please try to describe your goal using a different term. \$\endgroup\$ Commented Feb 3, 2013 at 23:47
  • \$\begingroup\$ @Connor: I have no idea what he means by that either, so I just ignored it. Anyway, this question needs to be closed since the OP is no longer cooperating. \$\endgroup\$ Commented Feb 4, 2013 at 14:16

Found the problem and solution: because of my small practical experience I didn't put two large capacitors (1000uF) between +15 and ground and -15 and ground (they are 2 switching power supply). Now my not orthodox amplifier topology sounds good enough.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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