PDF of wxMaxima worksheet for the calculation of power dissipation can be found here: drive.google.com/file/d/1VUT-EAM3LX0n3jq6xZJbed3BmZen3vqt/view?usp=sharing
That's the nicest tutorial for a power amp design for beginners that I've come across. One can't ask for more but if I were to, I'd ask for another part featuring thermal modelling and transistor selection.
Thx for taking the time to create this video. If you hold the Cntrl button down over the component you will get Power as a waveform. In the plot pane hover over that power waveform again press the cntrl button and you will get Avg and Rms values for the selected waveform. LTSpice has a lot of functionality. It is the fastest E simulator on the planet at this time. It is a good read how it achieves this speed. Again, thx for the Maxima and LTSpice exposure.
MAn i have been lookin at this .Symbolic amplifier design..There is only one book i have not been abel to get >and they use mathematica ..You solved all my needs
Hi..Thanks for the detailed video.How can pick an inductance value of speaker for simulation. Is it by sweeping speaker on an impedance analyser and find the inductance of the speaker for the frequency of operation? @25.00 is there any reason why which you didn't plot the power dissipation curve instead of differential of power dissipation
I have a question when it comes to the inductance value of the load. What if you're designing a general purpose audio amplifier rather than pairing it to a specific driver of known parameters? Y'know, what a lot of car audio amplifiers are, what a lot of unpowered bookshelf speaker amplifiers are, what a lot of the DIY enclosures use, etc. Is there a good inductance value to use for this? Do you just pick a generalized range for inductance, such as if I'm looking at an amp for car audio subwoofers where the driver Le (the notation for inductance in T/S parameters) is commonly around 6.5mH at the high end with a handful of outliers being within the 8.5+mH range, just from glancing at some databases, would it be fair to just use Zl=Rl*1e-2 and just have the extra overhead for versatility? In the example of anything but car audio subwoofers, I know drivers are typically
Apologies, it is the Alt button for component power. I had posted hover over the component then press the cntrl button then click, but you use the Alt button, hover, then click.However, on the waveform side you do use hover, Cntrl then click.
Thanks for this video, it is very clear and useful. Only one note: maybe the power computed in Maxima is a little bit different from the one simulated in SPICE because in SPICE you didn't check the quiescent current. I mean, you only set a DC voltage generator for 1.2V but we don't know how much current is flowing through the BJTs. Am I right?
Yes you are correct I didn't check the quiescent current in the SPICE simulation which will definitely introduce mismatch with the calculated power dissipation. If you have time, please find the SPICE quiescent current and re-run the computation in Maxima and let me know by replying to this comment. Yes part-3 will be coming on how to perform typical thermal calculations and heat sink design to keep operating junction temperature of output transistors within thermal limits. Cheers!
Is it really important to calculate peak value of power dissipation? I assume we should integrate that power dissipation over a time period of 1 cycle and use that for heatsink selection. Right?
Why did you not use an ESR meter to include the capacitive reactance and resistance of the speaker? However, I have to admit that I don't know what the frequency is that an ESR meter injects.
Electrical capacitive effects are negligible on loudspeakers because we are dealing with low frequencies in the audio spectrum. But certain loudspeaker mechanical effects are modeled in the electrical domain by capacitance with appropriate values and circuit configuration. Like I said in the video, the model included is a first order approximation which neglects some electrical and mechanical effects.
If you have LTspice download the following simulation and run it. It shows you how to implement a step down XFMR in a real world circuit such as power supply. docs.google.com/document/d/1FsQSZnfkQHuAiSy9G3K7h1u3uwhVJwqDRgmRC1JbG74/edit?usp=sharing You should copy the text in this file and save it with file extension .asc and open it with LTspice. FYI I am using a current source to excite the primary winding.
You considered a lot of things relating the power dissipation of the output transistors, even the contribution of the speaker inductance. But the most important thing is, that the MAXIMUM power dissipation does NOT occur at the MAXIMUM amplitude the amplifier can handle without clipping. It occurs at considerably LOWER amplitudes. I can demonstrate it with LTSpice for my common emitter audio amplifier at maximum amplitude ibb.co/pxyHs0V (0.23 V input amplitude, 2.29 W for Q2) and lower amplitude ibb.co/KsqY4VM (0.17 V input amplitude, 2.57 W for Q2).
The advantage of comman emitter amplifiers is that you have nearly rail-to-rail output voltage swing. So the power dissipation at peek power is considerably lower than in the case of emitter follower amplifiers. Some people don't like common emitter amplifiers because they are suspected to have higher THD, especially odd harmonics, compared to emitter follower amplifiers. But this is the question what you want to get. ibb.co/brmM8tZ shows the FFT with a third harmonic of -53dB compared to the signal. This is absolutely ok for me. ibb.co/ZdGrNqW shows the linear FFT plot. The output voltage at the 3rd harmonic is 22 mV compared to a 10 V signal. It is impossible to hear this contribution.
When you use a MOSFET output stage ibb.co/TKcDc0N you can make the circuit very linear. ibb.co/zFb30Jt shows the FFT. The THD is considerably lower. The 9th harmonic comes with the highest level leading to a distance of -59 dB.
Hi thank you for your comment and sorry for the late replay. Indeed if the load you are driving is purely resistive then yes the maximum power dissipation on the output transistors does occur at an output voltage that is lower than the maximum output amplitude the amplifier is capable. But as can be seen in the video, the amplifier is driving a complex impedance that has an inductive component and as such, the maximum power dissipation occurs at the maximum output amplitude. I will demonstrate this behavior in a video shortly. Stay tuned.
@@gkdresden Although common emitter output stages can give nearly rail to rail outputs, they are highly unstable (there is whole host of reasons why this is the case and I might discuss it in future videos). Therefore it is best to stay away from such circuits.
Why does the inductance-resistor combination of the speaker increase the Power dissipation of the output transistor? Doesn't the increase impedance due to the added inductor of the speaker increase the total impedance of the speaker load and thereby should decrease the power dissipation?
Go back and re-watch the video...the answer is there ;) I will give you a clue... It has something to do with zero crossing of the voltage across collector and emitter Vce(t) and collector current Ic(t). (Also think about what will happen to the power dissipation in the transistor if the load is only an inductor...)
Cant you just take the peak output voltage and divide it by a worst case load to get the maximum current then multiply it by the peak voltage to get the peak amount of power the output device would need to handle?
What happened to the day's when people just built an amplifier just going by the look and feel of the components and worried about the math later... Truth be told, if I had to go through these calculations just to build something I would seriously look for a new hobby LoL... I think that it's great that you did it, I know that there are people who love mathematics but I'm not one of them LoL..
"What happened to the day's when people just built an amplifier just going by the look and feel of the components and worried about the math later...." I'd imagine you only get that good having done the calculations many times, having built those circuits so many ways, and having had a deep grasp of the underlying mathematics. As in software engineering, any other trade or performant art, you can only sensibly abstract away the technical details to think at the architectural level after you've mastered those technical details.
If you take a step back and consider that what you begin with are mathematical descriptions of something both physical (electronics hardware) and intangible (i.e., frequency) and what you end up with is something that works as you've predicted (i.e., a bomb ass high fidelity speaker -- or an airplane that doesn't fall out of the sky because some controls engineer somewhere did his math homework), it's quite miraculous. The system of symbols and abstract models, on paper, mimic reality! The science of it is much more interesting of a strategy than some guy who picks up a random capacitor and tells you to put it in your circuit and observe whether it blows up.
Minh Tran you can throw a capacitor in your circuit anywhere and if it blows up, then don't put it there again unless you found out that you actually enjoy blowing things up LoL... Maths on the other hand just takes the fun part out of electronics... All that you need is some basic numbers that are available from datasheets and basically just have fun experimenting, Your test equipment will tell you what's happening and also give you the math that you may need... I think that if I started with mathematics first before I did anything else, then that's clearly saying that maybe I'm in the wrong industry... Whatever you are into, you should have already have plenty experience because you already know what isn't going to work because you have already thrown that capacitor into the same circuit before and you definitely know that it won't work without any mathematics.... And yeah I get it that some people develop an erection while doing maths and I'm okay with that, I'm just not one of those people.... I can just look at the circuitry and see what is wrong with it and where I haven't thrown a capacitor into it yet! But the real beauty of it is that I can do it all in my head. If what you are doing is fun, then you must be passionate about what it is that you are doing but if it isn't fun, then it's just a job and you don't have the passion to make a difference in the world. As an employer, I'm not really interested in your schooling qualifications... I really just want to see the passion in your eyes and hear it in your voice... That's the difference right there!
Minh Tran Oh and as for coming up with something as predicted! You are better off reading some service manals and fixing some tv's... hopefully you will achieve something as predicted!... But please don't tell me that you are going to do some mathematics and come up with a prediction that you are going to come up with something that no one else has done before! So basically the Wright Brothers went and did some mathematics and came back with the answer that they are going to be able to make a contraption that will definitely be able to fly! Or maybe you might think that they just made a frame and threw some capacitors at it to see if something comes out of it? if something did come out of it then I'm sure that they would have busted out the mathematics to work out why it worked... But if it didn't work, I'm sure that they would have tried something else based on their simple observations of what happened when they threw the capacitors at it... No point in doing the math for something that clearly didn't work... Because while you're doing mathematics, someone else is on the verge of inventing the next world changing idea, he can worry about the math later because everyone else is going to do it for him.. and that could potentially be you!
‘But please don't tell me that you are going to do some mathematics and come up with a prediction that you are going to come up with something that no one else has done before!‘ Math is a vast and terribly interesting subject on its own but engineers use it as a tool of analysis. Specifically, it allows us to describe, model, and simulate the behavior of physical components and systems (i.e., dynamics of sound, flight, fluids, electricity). It is a design aid, like a saw to a carpenter. Mathematical models help us design and understand complex circuits. Edison, Tesla, Franklin, Joseph Fourier, Bernoulli, Newton, Einstein, Currie, Feynman, Maxwell, Alan Turring, Marconi, Von Neuman ... to name a few, were pioneers of novel technologies who were renown for their mathematical prowess. Do you think they had a manual for what they were doing? How do you think we describe and observe things beyond human perception - Black holes, gravity waves, light, atoms? We we express their properties in mathematical equations and refine them. To reiterate, math is a tool.
PDF of wxMaxima worksheet for the calculation of power dissipation can be found here: drive.google.com/file/d/1VUT-EAM3LX0n3jq6xZJbed3BmZen3vqt/view?usp=sharing
Nicely done. I feel like I just went through a EE lesson on class AB amp circuits! :)
Will there be a part 3 to this series?
That's the nicest tutorial for a power amp design for beginners that I've come across. One can't ask for more but if I were to, I'd ask for another part featuring thermal modelling and transistor selection.
Thx for taking the time to create this video.
If you hold the Cntrl button down over the component you will get Power as a waveform. In the plot pane hover over that power waveform again press the cntrl button and you will get Avg and Rms values for the selected waveform. LTSpice has a lot of functionality. It is the fastest E simulator on the planet at this time. It is a good read how it achieves this speed.
Again, thx for the Maxima and LTSpice exposure.
This is excellent! Where is the rest of the videos?
You’re awesome. I’ve been learning about bjts and mosfets and this video really puts the lessons together 🙏🏾
Thanks for this! Been looking forward to the continuation of this series :-)
Excellent AB amplifier design tutorial and simulation of output power. Many thanks!
Great pair of videos. Looking forward to the next one!
Thanks for the tutorial. I think I can take it from here to design my own.
Looking forward to part 3
You won a new subscriber, thanks for very detailed and educative video.
amazing
Awesome video! I just can't wait for the next part
Man this is unbelievable, thanks for the video
MAn i have been lookin at this .Symbolic amplifier design..There is only one book i have not been abel to get >and they use mathematica ..You solved all my needs
Thanks for the education, the vas stage has always been my problem
Alvrg soy mexican. En estados unidos están a otro nivel. CHido el amplificador.
please upload the input stage!! your videos are great
I see the mount in the background for the sdr antenna, I have the same one!
Hi..Thanks for the detailed video.How can pick an inductance value of speaker for simulation. Is it by sweeping speaker on an impedance analyser and find the inductance of the speaker for the frequency of operation? @25.00 is there any reason why which you didn't plot the power dissipation curve instead of differential of power dissipation
Nice video!!!
I have a question when it comes to the inductance value of the load. What if you're designing a general purpose audio amplifier rather than pairing it to a specific driver of known parameters? Y'know, what a lot of car audio amplifiers are, what a lot of unpowered bookshelf speaker amplifiers are, what a lot of the DIY enclosures use, etc. Is there a good inductance value to use for this? Do you just pick a generalized range for inductance, such as if I'm looking at an amp for car audio subwoofers where the driver Le (the notation for inductance in T/S parameters) is commonly around 6.5mH at the high end with a handful of outliers being within the 8.5+mH range, just from glancing at some databases, would it be fair to just use Zl=Rl*1e-2 and just have the extra overhead for versatility? In the example of anything but car audio subwoofers, I know drivers are typically
In LTspice you can draw graph of any component power dissipation, just click on the component with Alt key pressed.
This is great. Are you going to do the input stage?
Apologies, it is the Alt button for component power. I had posted hover over the component then press the cntrl button then click, but you use the Alt button, hover, then click.However, on the waveform side you do use hover, Cntrl then click.
Thanks ........
plz upload input stage design also
Awesome video! Just subscribed and looking forward to learn more. Btw, still using those dollar store pens! Haha, those are the best.
Great to hear from you! And yes those pens are awesome.
Thanks for this video, it is very clear and useful. Only one note: maybe the power computed in Maxima is a little bit different from the one simulated in SPICE because in SPICE you didn't check the quiescent current. I mean, you only set a DC voltage generator for 1.2V but we don't know how much current is flowing through the BJTs. Am I right?
I'm looking forward for part 3!!! :-D
Yes you are correct I didn't check the quiescent current in the SPICE simulation which will definitely introduce mismatch with the calculated power dissipation. If you have time, please find the SPICE quiescent current and re-run the computation in Maxima and let me know by replying to this comment. Yes part-3 will be coming on how to perform typical thermal calculations and heat sink design to keep operating junction temperature of output transistors within thermal limits.
Cheers!
@@SmithKerona Did you ever make Part 3?
Is it really important to calculate peak value of power dissipation? I assume we should integrate that power dissipation over a time period of 1 cycle and use that for heatsink selection. Right?
Hİ, which are you using books ?
Why did you not use an ESR meter to include the capacitive reactance and resistance of the speaker? However, I have to admit that I don't know what the frequency is that an ESR meter injects.
Electrical capacitive effects are negligible on loudspeakers because we are dealing with low frequencies in the audio spectrum. But certain loudspeaker mechanical effects are modeled in the electrical domain by capacitance with appropriate values and circuit configuration. Like I said in the video, the model included is a first order approximation which neglects some electrical and mechanical effects.
Why did you choose BJTs over MOSFETs? They should be more linear .. no?
Did you do the follow up videos to this video
Would you please dervid the power dissipation( Pd ) formula at the beginning of the video?
Nice video on how to audio amplifier design. Real designers use their calculators and pen.
I have SPICE. I have tried to input a step-down power transformer but cannot figure out how to do it. Can you help me? Thanks
If you have LTspice download the following simulation and run it. It shows you how to implement a step down XFMR in a real world circuit such as power supply. docs.google.com/document/d/1FsQSZnfkQHuAiSy9G3K7h1u3uwhVJwqDRgmRC1JbG74/edit?usp=sharing You should copy the text in this file and save it with file extension .asc and open it with LTspice. FYI I am using a current source to excite the primary winding.
You considered a lot of things relating the power dissipation of the output transistors, even the contribution of the speaker inductance. But the most important thing is, that the MAXIMUM power dissipation does NOT occur at the MAXIMUM amplitude the amplifier can handle without clipping. It occurs at considerably LOWER amplitudes. I can demonstrate it with LTSpice for my common emitter audio amplifier at maximum amplitude ibb.co/pxyHs0V (0.23 V input amplitude, 2.29 W for Q2) and lower amplitude ibb.co/KsqY4VM (0.17 V input amplitude, 2.57 W for Q2).
The advantage of comman emitter amplifiers is that you have nearly rail-to-rail output voltage swing. So the power dissipation at peek power is considerably lower than in the case of emitter follower amplifiers. Some people don't like common emitter amplifiers because they are suspected to have higher THD, especially odd harmonics, compared to emitter follower amplifiers. But this is the question what you want to get. ibb.co/brmM8tZ shows the FFT with a third harmonic of -53dB compared to the signal. This is absolutely ok for me. ibb.co/ZdGrNqW shows the linear FFT plot. The output voltage at the 3rd harmonic is 22 mV compared to a 10 V signal. It is impossible to hear this contribution.
When you use a MOSFET output stage ibb.co/TKcDc0N you can make the circuit very linear. ibb.co/zFb30Jt shows the FFT. The THD is considerably lower. The 9th harmonic comes with the highest level leading to a distance of -59 dB.
Hi thank you for your comment and sorry for the late replay. Indeed if the load you are driving is purely resistive then yes the maximum power dissipation on the output transistors does occur at an output voltage that is lower than the maximum output amplitude the amplifier is capable. But as can be seen in the video, the amplifier is driving a complex impedance that has an inductive component and as such, the maximum power dissipation occurs at the maximum output amplitude. I will demonstrate this behavior in a video shortly. Stay tuned.
@@gkdresden Although common emitter output stages can give nearly rail to rail outputs, they are highly unstable (there is whole host of reasons why this is the case and I might discuss it in future videos). Therefore it is best to stay away from such circuits.
Why does the inductance-resistor combination of the speaker increase the Power dissipation of the output transistor? Doesn't the increase impedance due to the added inductor of the speaker increase the total impedance of the speaker load and thereby should decrease the power dissipation?
Go back and re-watch the video...the answer is there ;)
I will give you a clue... It has something to do with zero crossing of the voltage across collector and emitter Vce(t) and collector current Ic(t). (Also think about what will happen to the power dissipation in the transistor if the load is only an inductor...)
@@SmithKerona Thanks. Much appreciated.
Cant you just take the peak output voltage and divide it by a worst case load to get the maximum current then multiply it by the peak voltage to get the peak amount of power the output device would need to handle?
no
@@AndrewTa530 explain
Sir part-3 sir explain full circuit please sir I need to understand please
What happened to the day's when people just built an amplifier just going by the look and feel of the components and worried about the math later... Truth be told, if I had to go through these calculations just to build something I would seriously look for a new hobby LoL... I think that it's great that you did it, I know that there are people who love mathematics but I'm not one of them LoL..
"What happened to the day's when people just built an amplifier just going by the look and feel of the components and worried about the math later...." I'd imagine you only get that good having done the calculations many times, having built those circuits so many ways, and having had a deep grasp of the underlying mathematics. As in software engineering, any other trade or performant art, you can only sensibly abstract away the technical details to think at the architectural level after you've mastered those technical details.
If you take a step back and consider that what you begin with are mathematical descriptions of something both physical (electronics hardware) and intangible (i.e., frequency) and what you end up with is something that works as you've predicted (i.e., a bomb ass high fidelity speaker -- or an airplane that doesn't fall out of the sky because some controls engineer somewhere did his math homework), it's quite miraculous. The system of symbols and abstract models, on paper, mimic reality!
The science of it is much more interesting of a strategy than some guy who picks up a random capacitor and tells you to put it in your circuit and observe whether it blows up.
Minh Tran you can throw a capacitor in your circuit anywhere and if it blows up, then don't put it there again unless you found out that you actually enjoy blowing things up LoL... Maths on the other hand just takes the fun part out of electronics... All that you need is some basic numbers that are available from datasheets and basically just have fun experimenting, Your test equipment will tell you what's happening and also give you the math that you may need... I think that if I started with mathematics first before I did anything else, then that's clearly saying that maybe I'm in the wrong industry... Whatever you are into, you should have already have plenty experience because you already know what isn't going to work because you have already thrown that capacitor into the same circuit before and you definitely know that it won't work without any mathematics.... And yeah I get it that some people develop an erection while doing maths and I'm okay with that, I'm just not one of those people.... I can just look at the circuitry and see what is wrong with it and where I haven't thrown a capacitor into it yet! But the real beauty of it is that I can do it all in my head. If what you are doing is fun, then you must be passionate about what it is that you are doing but if it isn't fun, then it's just a job and you don't have the passion to make a difference in the world. As an employer, I'm not really interested in your schooling qualifications... I really just want to see the passion in your eyes and hear it in your voice... That's the difference right there!
Minh Tran Oh and as for coming up with something as predicted! You are better off reading some service manals and fixing some tv's... hopefully you will achieve something as predicted!...
But please don't tell me that you are going to do some mathematics and come up with a prediction that you are going to come up with something that no one else has done before! So basically the Wright Brothers went and did some mathematics and came back with the answer that they are going to be able to make a contraption that will definitely be able to fly!
Or maybe you might think that they just made a frame and threw some capacitors at it to see if something comes out of it?
if something did come out of it then I'm sure that they would have busted out the mathematics to work out why it worked... But if it didn't work, I'm sure that they would have tried something else based on their simple observations of what happened when they threw the capacitors at it... No point in doing the math for something that clearly didn't work... Because while you're doing mathematics, someone else is on the verge of inventing the next world changing idea, he can worry about the math later because everyone else is going to do it for him.. and that could potentially be you!
‘But please don't tell me that you are going to do some mathematics and come up with a prediction that you are going to come up with something that no one else has done before!‘
Math is a vast and terribly interesting subject on its own but engineers use it as a tool of analysis. Specifically, it allows us to describe, model, and simulate the behavior of physical components and systems (i.e., dynamics of sound, flight, fluids, electricity). It is a design aid, like a saw to a carpenter. Mathematical models help us design and understand complex circuits.
Edison, Tesla, Franklin, Joseph Fourier, Bernoulli, Newton, Einstein, Currie, Feynman, Maxwell, Alan Turring, Marconi, Von Neuman ... to name a few, were pioneers of novel technologies who were renown for their mathematical prowess. Do you think they had a manual for what they were doing?
How do you think we describe and observe things beyond human perception - Black holes, gravity waves, light, atoms? We we express their properties in mathematical equations and refine them.
To reiterate, math is a tool.