good specs for a new(?) design

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hello my friends
The whole story started with me trying to find out the best design for a headphone splitter-amp: an amp that would split the input audio to two headphones. Actually a splitter-buffer but if it could amplify it would be an added bonus.;)

So my interest got to class A output sections.
In the quest I tried to redesign and optimize existing designs and finally this process leaded to a topology with good characteristics. I believe the philosophy is rather unique to the best of my poor knowledge. If that proves to be true and all goes well I intend to publish the circuit in a peer reviewed journal. here comes the first question:
1) where can I publish some preliminary simulation results (without getting deep into circuit analysis) and establishing the parenhood of the idea? What is good and quite respected in this area?

I have not build the amplifier yet. It lives only in LTSPICE and qucs simulations.
In order to have reliable results I used common transistors: bc547b and bc557b
The circuit is biased from a single 8Volt source and current draw is around 10mA. The load is 50Ω capacitively coupled to the output. With this specs i think it is possible to compare to whatever topology you want to simulate at same quiescent current with same load .Here comes the second question:

2) Do these results look bad - good - exceptional? For this reference design the results I get from the simulation are:
-input signal 50mV amplitude - load current 1mA aplitude
Total Harmonic Distortion: 0.000013%(0.000314%)
- 500mV input amplitude, 10mA load current amplitude
Total Harmonic Distortion: 0.000104%(0.000093%)
-1500mV input (this gets us close to available swing without clipping), 30mA (edited typo)load amplitude
Total Harmonic Distortion: 0.000340%(0.000717%)

the spectrum for the three above cases. in each figure the spectrum of the source is shown too.



LTSPICE simulation parameters:
.tran 0 10m 0 100n
.four 1kHz V(out)
tools->control panel->no compression

3) what other measurements do you want me to make in order for you to judge?
noise? frequency response? temperature stability?
please tell me how you want me to set up the simulation in LTspice and I will report further results here.

I hope my post complies with the rules of the forum.
Thank you in advance

Edit: we are talking for the output section only. No VAS no global feedback yet
Last edited:
Joined 2007
Paid Member
The results look very good indeed... but such can be the nature of simulation.

I suspect you will find that most who are interested in this will want you to share more details.

How do you deliver 40ma load swing from a Class A stage (whole amp ?) running at just 10ma ?
mooly thank you for the reply
yes total amp current draw from power supply is around 10mA (excluding some mirrors and current references - 1-2mA additional constant current)

Load current actually is 30mA (it was a typo error) and of course the amplifier has entered class AB.(however it still behaves well which is good for an audio signal with high peak to average intensity ratio)
However the switching seems soft (gradual) to me. Do you know for a 1ms signal period, 10mA power supply draw, +-30mA load draw how much time the active devices are in cut off for a typical push pull?(if you have an optimized push pull for that I would be thankful )

what do you mean by "such can be the nature of simulation"? A reason that I plot the voltage source is to see the floor of the simulation.

I intend to measure output and input impedance and present you the frequency response.

I guess that there is not much interest or the title is not very descriptive.
Thanks anyway for the response.
Joined 2007
Paid Member
Results in simulation can sometimes be a lot better than real world performance for a variety of reasons. The parts are 100% matched in simulation which means good results anywhere where there would need to be close matching... such as your current mirrors for instance.

I'm afraid I don't know how you would derive the conduction overlap for a given bias current. You can observe the currents in the two halves of the push pull stage though.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.