Oh Ok. I just remember that Bob said a few months back that the book was going to the publishers in April or maybe that was his manuscript or something.
Maybe Bob could update us as to the status of the second edition.
As I'm sure that everyone here is interested in an update.
The manuscript is nearly complete, and should go to the publisher around the end of this month. I apologize for it taking so long to get the second edition out. I'm hoping it will be published and available for sale before the end of September, but it is often difficult to predict publisher turn-around time.
The second edition will be about 750 pages (up from about 600) and will have 36 Chapters, up from 31. Some of the new chapters are on Noise, Switch-mode power supplies, professional power amplifiers, a second chapter on output stages, and a chapter on building an example amplifier. The chapter on advanced forms of feedback compensation has been substantially expanded.
I deeply appreciate your interest in the book and your patience.
Cheers,
Bob
Thanks for the update Bob! This is great news
You're nearly there!When is the third edition coming out?
Yes, to the greatest extent possible. Thanks for pointing this out. I found an example of this in Figure 19.4. I checked the original manuscript PDF of the Figure that was submitted and it was correct. Somehow that change crept into the figure at the composition step, and it was not caught. I apologize for that.
A great many of you here at DIYaudio have given me very helpful and valuable feedback on errors and suggestions for improvement, and to all of you I owe a great debt of gratitude.
Cheers,
Bob
Mr Cordell, your competency, your curiosity that goes with a rare open mind, your modesty and the perfect balance between pure technology and this little pinche of art and feeling that give to audio a spécial place in the world of technology makes your books unequaled and precious. A reference, a must have.I too can't wait for you to hold it in your hands
How to thank-you ?
Thank you for your very kind words. I love to learn and share what I have learned.
Cheers,
Bob
Bob or maybe someone else, will be so kind and explain in more detail on how speedup capacitor "speed-up" the switch-off time in the output stage. I ask about it because W. Marshall Leach on his page about Leach Amp says that this capacitor serves no purpose.
The Leach Amp - Output Stage (fig. 8)
The Leach Amp - Output Stage (fig. 8)
Leach is correct insofar as quiescent and small-signal conditions. However, under conditions of fast signal transitions, the bias spreader may not be able to keep the voltage across the emitters constant, because the bias current in the drivers may not be able to provide sufficient turn-off current for the output transistor that is turning off. Depending on the ft of the output transistor and the rate-of-change of signal current flowing, it takes a certain minimum amount of turn-off current to pull stored charge out of the base of the output transistor to support the rate of change of signal current. If that required turn-off current exceeds the driver bias current, that transistor's driver transistor will turn off, and no longer be able to control that output transistor. Under that condition, the emitter-to-emitter voltage at the drivers will actually increase.
The idea of the capacitor is to allow the "other" driver transistor to help contribute turn-off current in a push-pull manner. Ideally, the capacitor prevents the emitter-emitter voltage between the drivers from increasing.
I am not a fan of that capacitor because it is actually not very effective, and may create time-dependent hangover biasing effects. This is a nonlinear situation, and whenever you introduce a capacitor into a nonlinear situation you may get un-intended consequences. I prefer to employ adequately high driver bias current in combination with adequately fast output transistors to solve the problem.
When simulating your amplifier under large-signal high-frequency conditions (e.g., full power, 20 kHz, 4 ohm load), be sure to do a simulation where you plot the collector current of the driver transistor to make sure it is not completely turning off.
Cheers,
Bob
Yes.
It is always a surprise, when some discover that MHz FT transistors are not fast enough for audio, when they were only supposed, at first sight, to reproduce a 20-20 000Hz bandwidth.
If we want an amplifier to be linear enough in phase, we first need 10 time (at least) this bandwidth). And, if we don't want TIM, this is only the bandwidth of the requisite input filter. Not to talk about the phase margin for stability. The whole amp has to be a lot faster
As an other example, the parasitic gate capacitances of mosfets power devices require huge currents from the drivers at HF.
Well, it is happy that, nowadays, the things are so much better than in the good old time, before the first "planar" BJT power devices ;-)
Bob's book presents, and all of the example amplifiers in the book obey, this guideline:
See Figure 2.15 for an early example.
Bob's readers are quite well aware that audio power amplifier design work involves frequencies hundreds of times greater than 20 kHz.
Aim for an open loop unity-gain frequency of 10 Megahertz or higher. With the recommended 20X closed loop gain, this gives a closed loop bandwidth of 0.5 Megahertz.
See Figure 2.15 for an early example.
Bob's readers are quite well aware that audio power amplifier design work involves frequencies hundreds of times greater than 20 kHz.
I hope so.
It is just that I remember to had been flamed on a forum because I said that the actual devices on the shelf begin just to be fast enough for our audio purposes.
Hi Tryphon,
I thought that TIM was proved to be a myth. Certainly at audio frequencies it is as the emf is almost an instantaneous thing. The feedback happens practically in phase with the input signal. It's not like you have to wait for electron drift.
If TIM did exist, it would be plainly visible with today's oscilloscopes above 100 MHz bandwidth. Any spike that needs that kind of rise time wouldn't make it down the speaker wire due to inductance and capacitance, and for sure in a crossover. I can't see any reasonable explanation that could explain how this would be audible, yet invisible even to a crude instrument like an oscilloscope. Even if I feed in sharp pulses from my oscilloscope timing calibrator (Tek TG-501) into an amplifier, I can't see anything even with a spectrum analyser good to 500 MHz, or my 40 MHz spec-an. Tried both.
I know insisting that TIM exists is romantic, something like "I know something you don't know" kind of thing. But that bug is dead and buried.
-Chris
I thought that TIM was proved to be a myth. Certainly at audio frequencies it is as the emf is almost an instantaneous thing. The feedback happens practically in phase with the input signal. It's not like you have to wait for electron drift.
If TIM did exist, it would be plainly visible with today's oscilloscopes above 100 MHz bandwidth. Any spike that needs that kind of rise time wouldn't make it down the speaker wire due to inductance and capacitance, and for sure in a crossover. I can't see any reasonable explanation that could explain how this would be audible, yet invisible even to a crude instrument like an oscilloscope. Even if I feed in sharp pulses from my oscilloscope timing calibrator (Tek TG-501) into an amplifier, I can't see anything even with a spectrum analyser good to 500 MHz, or my 40 MHz spec-an. Tried both.
I know insisting that TIM exists is romantic, something like "I know something you don't know" kind of thing. But that bug is dead and buried.
-Chris