- #1
In summary: It uses a differential pair of transistors to amplify a single input. The output is balanced to reduce offset voltage. It is optimised, and was designed to test alternative subcircuits to the Advanced Amplifier.In summary, the component values are blurry and the current shapes of Q4 and Q5 don't look the best, but are beautifully righted in Q6 and Q7 in the alternative schematic.
- #2
Baluncore
Science Advisor
2023 Award
- 15,316
- 9,138
Can you please rename a copy of your .asc file to .asc.txt and attach it to your next post.
Then I can run your original file with all the same parameters.
Then I can run your original file with all the same parameters.
- #3
ntetlow
- 21
- 2
ok. thanks
- #4
Baluncore
Science Advisor
2023 Award
- 15,316
- 9,138
@ntetlow
The component values are too blurry to read, but when I follow the schematic in my head with typical values, the topology seems OK. My guess would be that you are seeing a dead-band and step during each crossover of the output stage, that is being fed back and around the loop. I suspect your selection of component values is the problem, rather than the simulation parameters.
The component values are too blurry to read, but when I follow the schematic in my head with typical values, the topology seems OK. My guess would be that you are seeing a dead-band and step during each crossover of the output stage, that is being fed back and around the loop. I suspect your selection of component values is the problem, rather than the simulation parameters.
- #5
ntetlow
- 21
- 2
Thank you for looking and for your info though I'm going to have to look up 'dead band' and 'step' to try and understand what you mean.
I got the schematic from http://www.ecircuitcenter.com/Circuits_Audio_Amp/Advanced_Amplifier/Advanced_Audio_Amplifier.htm which I would expect would work without error. I have run both the schematic and the accompanying netlist with the same result (i'm fed up with finding errors in published documents!)
I got the schematic from http://www.ecircuitcenter.com/Circuits_Audio_Amp/Advanced_Amplifier/Advanced_Audio_Amplifier.htm which I would expect would work without error. I have run both the schematic and the accompanying netlist with the same result (i'm fed up with finding errors in published documents!)
- #6
Jony130
- 627
- 138
Crossover distortion are the cause of this spikes. Too low output stage quiescent current.
Try this circuit:
Try this circuit:
- #7
Baluncore
Science Advisor
2023 Award
- 15,316
- 9,138
Attached is a copy of the original, but for LTspice.
I have not labelled the nodes.
Let me know what problems you find.
I have not labelled the nodes.
Let me know what problems you find.
- #8
Baluncore
Science Advisor
2023 Award
- 15,316
- 9,138
Errata:Jony130 said:
Q10, a PNP type QMPSA56 is cross labelled with itself.
Replace PNP text with QMPSA56
Replace text QMPSA56 with Q10
- #9
Jony130
- 627
- 138
Yes, I notice this. And I reupload the corrected file.Baluncore said:
- #10
ntetlow
- 21
- 2
Audi_amp_bjt_1 seems to have got rid of the spikes (thank you) but why are the base, collector and emitter currents of Q3 so ugly, is this normal? Also, the current shapes of Q4 and Q5 don't look the best either but are beautifully righted in Q6 and Q7, is this too normal?
May I also draw your attention to http://www.ecircuitcenter.com/Circuits_Audio_Amp/Diff_Amp_w_Mirror/Diff_Amp_w_Mirror.htm. this is a preamble to the advanced amp circuit and is used to double the current gain from Q1, Q2 using a current mirror, this seems to work OK. However, if the same is supposed to happen in the Advanced amp, it doesn't. In the mirror circuit the collector currents of Q1 and Q2 are 180 degrees out of phase whereas in the advanced amp circuit they are not. Any ideas why?
May I also draw your attention to http://www.ecircuitcenter.com/Circuits_Audio_Amp/Diff_Amp_w_Mirror/Diff_Amp_w_Mirror.htm. this is a preamble to the advanced amp circuit and is used to double the current gain from Q1, Q2 using a current mirror, this seems to work OK. However, if the same is supposed to happen in the Advanced amp, it doesn't. In the mirror circuit the collector currents of Q1 and Q2 are 180 degrees out of phase whereas in the advanced amp circuit they are not. Any ideas why?
- #11
Baluncore
Science Advisor
2023 Award
- 15,316
- 9,138
@ntetlow
I might be an idea to find the source of oscillation in your OP model. Is the same oscillation present in the Advanced_Amplifier_0.asc that I attached to post 7 ?
The Advanced Amplifier is a learning exercise, not a practical optimised audio circuit.
One problem with Audi_amp_bjt-1.asc is that the feedback is taken from the second last stage, not from the output. Another is that, to reduce crossover distortion, the output stage has a quiescent current of 800 mA, without any current limit to prevent the catastrophic thermal runaway. You need a checklist of parameters to verify, so when you optimise a design, you do not end up with something like 800 mA quiescent in the output stage.
Here is an alternative schematic that operates in class-A up to 5 mA output, then switches on one half of the class-B circuit for higher loads. The base current is balanced to reduce offset voltage. It is not optimised, and was designed to test alternative subcircuits to the Advanced Amplifier.
I might be an idea to find the source of oscillation in your OP model. Is the same oscillation present in the Advanced_Amplifier_0.asc that I attached to post 7 ?
The Advanced Amplifier is a learning exercise, not a practical optimised audio circuit.
One problem with Audi_amp_bjt-1.asc is that the feedback is taken from the second last stage, not from the output. Another is that, to reduce crossover distortion, the output stage has a quiescent current of 800 mA, without any current limit to prevent the catastrophic thermal runaway. You need a checklist of parameters to verify, so when you optimise a design, you do not end up with something like 800 mA quiescent in the output stage.
The collector currents in the mirror should be in phase. The collector currents in the differential pair should be out of phase. You must be careful how you specify and measure currents because currents can be reversed by reversing a component.ntetlow said:
Here is an alternative schematic that operates in class-A up to 5 mA output, then switches on one half of the class-B circuit for higher loads. The base current is balanced to reduce offset voltage. It is not optimised, and was designed to test alternative subcircuits to the Advanced Amplifier.
- #12
FAQ: Spikes in LTspice voltages and currents
What causes spikes in LTspice voltages and currents?
Spikes in LTspice voltages and currents are typically caused by sudden changes in the circuit, such as switching on or off a component or a sudden change in the input signal. These changes can create high-frequency oscillations that result in spikes in the voltage and current.
How can I reduce spikes in LTspice simulations?
To reduce spikes in LTspice simulations, you can try increasing the simulation time step, using more accurate models for components, or adding damping to the circuit. Additionally, you can try using different simulation methods, such as Gear integration or Bilinear Transform, which can handle high-frequency oscillations more effectively.
Are spikes in LTspice simulations a cause for concern?
In most cases, spikes in LTspice simulations are not a cause for concern. They are a natural result of the simulation process and do not necessarily reflect real-world behavior. However, if the spikes are significantly larger than expected or occur in unexpected places, it may indicate a problem with the circuit design.
Can I eliminate spikes in LTspice simulations completely?
No, it is not possible to completely eliminate spikes in LTspice simulations. The simulation process is based on mathematical models and approximations, which can result in small errors and high-frequency oscillations. However, by following best practices and using appropriate simulation settings, you can minimize the impact of spikes on your results.
Do spikes in LTspice simulations affect the accuracy of my results?
In most cases, spikes in LTspice simulations do not significantly affect the accuracy of your results. However, if your circuit is sensitive to high-frequency oscillations, the spikes may affect the overall behavior of the circuit. It is important to carefully analyze the results and consider the impact of spikes on your specific circuit design.