PMI an Hahfran,
What you are both saying is common in many industries. What you learn in school generally teaches toward the end goal, the way we have to use the information. I would have thought that at least on a basic level the physics of how a transistor or diode or other electrical device works would have been shown. But do you think that every EE should also have to have a degree in physics? I worked with polymer chemistry for many years and though I may have a very broad understanding of different polymer chemistry's does that mean that I should have been a chemist to use the polymers that I used. I was not formulating those, but I was doing material applications development. I needed to know how to use the chemistry, how I could apply that knowledge but not how the polymer chains were created. I doubt that many here besides Scott Wercer and a few others will ever have any need to fully understand how the photons are used to etch an integrated circuit, you may be interested on some level but that isn't necessary to apply the finished devices into a working circuit. The same would go for most designing an audio circuit, do you really need to understand the physics to work in this field? The mathematics alone to work on the parasitic level of an electrical circuit is already a fairly extensive knowledge base that you would need to understand to optimize many of these designs.
What you are both saying is common in many industries. What you learn in school generally teaches toward the end goal, the way we have to use the information. I would have thought that at least on a basic level the physics of how a transistor or diode or other electrical device works would have been shown. But do you think that every EE should also have to have a degree in physics? I worked with polymer chemistry for many years and though I may have a very broad understanding of different polymer chemistry's does that mean that I should have been a chemist to use the polymers that I used. I was not formulating those, but I was doing material applications development. I needed to know how to use the chemistry, how I could apply that knowledge but not how the polymer chains were created. I doubt that many here besides Scott Wercer and a few others will ever have any need to fully understand how the photons are used to etch an integrated circuit, you may be interested on some level but that isn't necessary to apply the finished devices into a working circuit. The same would go for most designing an audio circuit, do you really need to understand the physics to work in this field? The mathematics alone to work on the parasitic level of an electrical circuit is already a fairly extensive knowledge base that you would need to understand to optimize many of these designs.
With all due respect, I must disagree. Most colleges I have seen tend to just gloss over that in lecture and maybe a paragraph in a textbook. They jump straight to the HFE explanation. It's simpler (even if not correct) and now that students are seen as "seats" and not students, it's getting worse.
That's exactly how I've seen it taught. Once the wrong way is learned, it's hard to unlearn. Especially because it works most of the time.
That's exactly how my EE classes were. The application is actually a very good thing to teach, but when the rest is glossed over it's not really much of an education.
FoMoCo,
Sounds like you want a class or many classes in solid state physics of semiconductors. Do you really need this much knowledge to work in this field? You can always take additional classes, I have never seen any college or school that would deny you this opportunity. But there comes some point where a school has to set a curriculum that encompasses what they decide is a minimum requirement to graduate with a degree in a subject. Should an EE also have to take music appreciation classes before he is allowed to design audio equipment. Where do you draw the line, what is necessary to do the job. Now if you are given incomplete information leading you to incorrect conclusions and a lifelong misunderstanding of the workings of a device that could be a problem.
Sounds like you want a class or many classes in solid state physics of semiconductors. Do you really need this much knowledge to work in this field? You can always take additional classes, I have never seen any college or school that would deny you this opportunity. But there comes some point where a school has to set a curriculum that encompasses what they decide is a minimum requirement to graduate with a degree in a subject. Should an EE also have to take music appreciation classes before he is allowed to design audio equipment. Where do you draw the line, what is necessary to do the job. Now if you are given incomplete information leading you to incorrect conclusions and a lifelong misunderstanding of the workings of a device that could be a problem.
I just don't think that I should take extra classes for what was common knowledge 20 years ago.
Ah... Now see... That is what is now just glossed over. I had to look it up. It was briefly mentioned as the Shockley equation and never mentioned again.
Hi, off course. Schools are (sorry) for degrees for looking for jobs to get money. They are not being researching that hard.
there in schools are no bjt except models and its function explanation, even there are disagrees about what bjt functions are. For fast system implement indeed the substrate are fragile from anything like stored energy, even interfere when some metal (ground or radiating plate) getting close. That's why chip amps are very limited.
In audio there are more complex function of bjt. Some bjt are sound better than other and linearity figures could not help
.When I did my EE degree I was taught the basics of semiconductor physics and bjt operation using the standard models talked about so far. But most of it never sunk in - I learnt enough to pass the exams, but frankly at that stage of life I was more interested in girls and drinking beer and enjoying life
Later on after thinking about it more and reading some excellent textbooks that I wish were standard in my courses such as those by your good selves Mr Self and Mr Cordell, I came to understand transistor operation more deeply, but there's still plenty to learn.
My point is, what you are taught at uni is only a grounding in these areas and while it would be nice to have these things taught more clearly and correctly, most of the students are not really going to "get" it at the time and the expectation is that they will study more deeply later on if and when required. If a particular individual chooses to believe all that he or she was taught in those early stages throughout their life/career without questioning it or trying to understand things more deeply, that is their choice and there's nothing a university can do about it. Besides, these days universities seem to be very much run as corporations, and as each year government funding is battled for, the accountants who "manage" them like to reduce costs by sacking admin staff and expecting the lecturers to do their work as well as their normal teaching and research. So give the academics a break I say!
Later on after thinking about it more and reading some excellent textbooks that I wish were standard in my courses such as those by your good selves Mr Self and Mr Cordell, I came to understand transistor operation more deeply, but there's still plenty to learn.
My point is, what you are taught at uni is only a grounding in these areas and while it would be nice to have these things taught more clearly and correctly, most of the students are not really going to "get" it at the time and the expectation is that they will study more deeply later on if and when required. If a particular individual chooses to believe all that he or she was taught in those early stages throughout their life/career without questioning it or trying to understand things more deeply, that is their choice and there's nothing a university can do about it. Besides, these days universities seem to be very much run as corporations, and as each year government funding is battled for, the accountants who "manage" them like to reduce costs by sacking admin staff and expecting the lecturers to do their work as well as their normal teaching and research. So give the academics a break I say!
Mr Self, to get back to your question (and I only have the 4th ed. to go on here): I would very much like to see your insightful and superbly clear analysis applied to some of the alternative topologies to the generic Lin. I also would prefer a larger format/ page size that would allow clearer graphs etc. But it sounds as though you've already addressed both so thank you and I look forward to reading it.
Can I also just say that I agree with Carlos that a little more respect is due. Whether or not we all agree with your philosophy regarding the value of subjective evaluation or think the sound quality (for those of us who believe such a thing exists ) of your designs is pleasing to our ears is irrelevant - the fact is you have taught most if not all of us plenty about our favourite subject and we are lucky that you've been generous enough to share it by publishing such excellent texts (as they all are). Indeed I suggest you have given many of us a solid foundation of knowledge to build on so that even if we do decide to start tweaking or designing our own gear, we at least have some idea what we're doing!
I still find it mind-boggling that many of my favourite recordings must have been produced using mixing consoles you designed or at least of similar design, yet the last step in the chain seems so critical to my listening enjoyment and I have found other approaches (eg discrete preamp designs) more satisfying. It makes no sense rationally, but there you are. I admire your stance against all the BS in this industry. So thank you, and best wishes with the new edition.
Can I also just say that I agree with Carlos that a little more respect is due. Whether or not we all agree with your philosophy regarding the value of subjective evaluation or think the sound quality (for those of us who believe such a thing exists
) of your designs is pleasing to our ears is irrelevant - the fact is you have taught most if not all of us plenty about our favourite subject and we are lucky that you've been generous enough to share it by publishing such excellent texts (as they all are). Indeed I suggest you have given many of us a solid foundation of knowledge to build on so that even if we do decide to start tweaking or designing our own gear, we at least have some idea what we're doing!I still find it mind-boggling that many of my favourite recordings must have been produced using mixing consoles you designed or at least of similar design, yet the last step in the chain seems so critical to my listening enjoyment and I have found other approaches (eg discrete preamp designs) more satisfying. It makes no sense rationally, but there you are. I admire your stance against all the BS in this industry. So thank you, and best wishes with the new edition.
Obvious this is spoken by someone who is not paying $400 a credit for "education." At that price I expect a professor to do more than half-*** read me a poorly written textbook for lecture.
But, we are getting off-topic. And, it seems I started that.
Well dear friends... come on
These gentlemen, Self and Cordell, were designing amplifiers when your mother and father began dating, and you were not even a future project, we now discuss these issues they've reflected on it at that time, decades ago ... do not think that you guys are so smart and that our teachers are mentally retarded ....... perhaps the opposite may be possible .... not dare to teach our masters, sit in the chair, answer the questions of your teachers and pay attention boys.
You mum maybe were using orthodontic appliances that time when they discussed some subjects you now trow to them.
Old folks does not become old...they become wise...they do not loose memory, the memory is selective...they remember the date
Powder without smoke was discovered long, long time ago, and now you discuss with Chinese (the inventors) why not to put the smoke back again.
regards,
Carlos
These gentlemen, Self and Cordell, were designing amplifiers when your mother and father began dating, and you were not even a future project, we now discuss these issues they've reflected on it at that time, decades ago ... do not think that you guys are so smart and that our teachers are mentally retarded ....... perhaps the opposite may be possible .... not dare to teach our masters, sit in the chair, answer the questions of your teachers and pay attention boys.
You mum maybe were using orthodontic appliances that time when they discussed some subjects you now trow to them.
Old folks does not become old...they become wise...they do not loose memory, the memory is selective...they remember the date
Powder without smoke was discovered long, long time ago, and now you discuss with Chinese (the inventors) why not to put the smoke back again.
regards,
Carlos
Sy,
Are those Ebers-Moll's the same ones who dig in my backyard, I thought they were just ground squirrels.........
That's not Ebers, I think it is Albert, See Albert Mol - Wikipedia, the free encyclopedia
Agreed, well said Carlos! Mr Self published his seminal preamp design in 1976, which just happens to be the year I was born
And while I was at uni in my EE classes, he published the now famous series on Distortion in Power Amplifiers which forms part of the book in question and which taught me far more about amplifier operation than any of the theory taught at my uni at the time.I think it's perfectly reasonable to disagree or question their work, as they have no doubt questioned other's before them - this is part of learning and potentially advancing the science/art. Nobody has all the answers. But it needs to be done with humility and respect (qualities which they both clearly display I might add), and not driven by inflated egos... that way the potential for learning is preserved. We are fortunate indeed to be able to discuss our favourite subject with these masters so let's make it worth their while to be a part of this and not potentially drive them away with negativity and disrespectful comments.
It's called debate. What started out by someone as a simple request to include a very popular topology was dismissed out of hand as being inferior, despite plenty of anecdotal and commercial evidence that suggests something quite to the contrary.
I will continue to buy their books - and Jan Didden's now that I have a Paypal account, if he will let me - but they are not infallible.
I will continue to buy their books - and Jan Didden's now that I have a Paypal account, if he will let me
- but they are not infallible.
TMC
Hi Douglas,
First, apologies for the late response (due trouble with my PC*).
As for TOIMC, admittedly, that would be a more exact description, though it's still ambiguous. As David pointed out, there are also other transition schemes possible: Transitional Input Miller Compensation (TIMC), for example. But that's not the only one. Also a combination of the two is possible: input AND output inclusion. How we call that? Also TOIMC or TIOMC? And what about the combination two pole compensation (TPC) and TMC: TTMC?
Anyhow, most of us know what is meant by TMC, so let's keep it that way.
The idea was born somewhere in the mid-nineties, when I was thinking how to make Cherry's output inclusive compensation more stable.
I sent you the emails around February 2002. Perhaps this helps to find them again.
We have never discussed a letter to the editor. Maybe you confuses it with another letter (see picture below, a shortened version). That was about input inclusive compensation. But also in this case we didn't have any discussion about this subject, though you know about it, as you have listed it on your website.
Anyhow, it was you who told me that TMC was already invented by Baxandall. By then, you were the only one who know this. So no one else could have told me about Baxandall. This leaves no other possibility than our emails were about TMC.
Much depends on minor circuit details and how it was built. Too long traces/wires to the output devices for example, will be devastating. Also notice that Cherry applied output inclusive compensation in conjunction with shunt compensation by putting (small) capacitors between the base of the drivers and the supply rails. I guess without the latter has amp will also be unstable.
When? As I said, somewhere it the mid-nineties.
Hope you will find back the emails.
Cheers,
E.
*PS: I still have PC troubles. Just bought an Z77 Asrock motherboard. Regrettably, this MB doesn't work together with my PCI sound cards (a Lynx L22 and Waveterminal 192X). I get a lot of dropouts. It's due to the fact the chip set on this MB does not support (good ol') PCI. Instead, the PCI slots are connected via a flaky interface chip to the PCIe bus. Hence the dropouts. So anybody who's using PCI sound cards, be warned! When upgrading your PC, make sure PCI is supported sufficiently.
Hi Douglas,
First, apologies for the late response (due trouble with my PC*).
As for TOIMC, admittedly, that would be a more exact description, though it's still ambiguous. As David pointed out, there are also other transition schemes possible: Transitional Input Miller Compensation (TIMC), for example. But that's not the only one. Also a combination of the two is possible: input AND output inclusion. How we call that? Also TOIMC or TIOMC? And what about the combination two pole compensation (TPC) and TMC: TTMC?
Anyhow, most of us know what is meant by TMC, so let's keep it that way.
The idea was born somewhere in the mid-nineties, when I was thinking how to make Cherry's output inclusive compensation more stable.
I sent you the emails around February 2002. Perhaps this helps to find them again.
We have never discussed a letter to the editor. Maybe you confuses it with another letter (see picture below, a shortened version). That was about input inclusive compensation. But also in this case we didn't have any discussion about this subject, though you know about it, as you have listed it on your website.
Anyhow, it was you who told me that TMC was already invented by Baxandall. By then, you were the only one who know this. So no one else could have told me about Baxandall. This leaves no other possibility than our emails were about TMC.
Much depends on minor circuit details and how it was built. Too long traces/wires to the output devices for example, will be devastating. Also notice that Cherry applied output inclusive compensation in conjunction with shunt compensation by putting (small) capacitors between the base of the drivers and the supply rails. I guess without the latter has amp will also be unstable.
When? As I said, somewhere it the mid-nineties.
Hope you will find back the emails.
Cheers,
E.
*PS: I still have PC troubles. Just bought an Z77 Asrock motherboard. Regrettably, this MB doesn't work together with my PCI sound cards (a Lynx L22 and Waveterminal 192X). I get a lot of dropouts. It's due to the fact the chip set on this MB does not support (good ol') PCI. Instead, the PCI slots are connected via a flaky interface chip to the PCIe bus. Hence the dropouts. So anybody who's using PCI sound cards, be warned! When upgrading your PC, make sure PCI is supported sufficiently.
Attachments
Last edited:
My own suggestions (refering to the 5th edition),
I would like to read some more words on :
Page 119 Figure 5.4f
Bootstapping VAS load R using an emitter follower
I rarely met this configuration but I found its simplicity rather smart.
It seems a very simple enhancement to buffer the VAS and may provide the same advantage as a triple output stage without the drawback for the bias of having six Vbe to compensate.
Page 356, Figure 12.12 (circuit of the class G amplifier)
A resistor of 2.2 kOhm (R11) is insterted in the Constant Current Source (Q6) for the input differential pair.
Some amplifiers have a resistor (1 kOhm or so) in this place, most of them have not. I never saw many explanations on its role.
Here, wiht high voltage power supply rails, one aim is certainly to lower the heat dissipation in Q6 wich is about 6 mA * 50 V = 0.3 W without it.
Has it any other effect ? I think of a compensation of the not purely resistive behaviour of the CCS at high frequencies.
Its presence may also help to control the CCS output current when debugging or repairing.
Amplitude and frequency distributions
http://www.diyaudio.com/forums/solid-state/133168-amplitude-frequency-distributions-wanted.html
For me, it's worth of more investigations.
I would like to read some more words on :
Page 119 Figure 5.4f
Bootstapping VAS load R using an emitter follower
I rarely met this configuration but I found its simplicity rather smart.
It seems a very simple enhancement to buffer the VAS and may provide the same advantage as a triple output stage without the drawback for the bias of having six Vbe to compensate.
Page 356, Figure 12.12 (circuit of the class G amplifier)
A resistor of 2.2 kOhm (R11) is insterted in the Constant Current Source (Q6) for the input differential pair.
Some amplifiers have a resistor (1 kOhm or so) in this place, most of them have not. I never saw many explanations on its role.
Here, wiht high voltage power supply rails, one aim is certainly to lower the heat dissipation in Q6 wich is about 6 mA * 50 V = 0.3 W without it.
Has it any other effect ? I think of a compensation of the not purely resistive behaviour of the CCS at high frequencies.
Its presence may also help to control the CCS output current when debugging or repairing.
Amplitude and frequency distributions
http://www.diyaudio.com/forums/solid-state/133168-amplitude-frequency-distributions-wanted.html
For me, it's worth of more investigations.
Yes, that's why I wanted to use TIMC and TOMC - to keep I for Input and O for Output so that TIOMC was clear
I have started to analyse the entire family but don't have much on TIOMC yet.
I am happy to use TMC as the default when there is no risk of confusion.
Like JCX, I am unaware of any systematic name convention and some of the schemes (mostly in op-amp literature) are very wordy.
The obvious way is to specify the admittance matrix or similar network analysis matrix. Can't see that catch on!
Best wishes
David
or If TMC is short for TOMC then
Dave
Last edited:
Doug,
I really hope you will stick with TMC in your upcoming edition. No three letter acronym is a perfect, unambiguous description, including TPC. Edmond deserves credit for popularizing it and giving it a pretty descriptive name, and I followed suit by using the term Transitional Miller Compensation in my text (which I believe is the first text to describe it, and where due credit is given to Edmond). To have two competing names for TMC would just add to the confusion. I urge you to stick with TMC.
Cheers,
Bob
weak-inversion
Hi Douglas,
Forgive me if I say it a bit bluntly, but you for ask additions and/or improvement on the 5th edition: Chapter 14 on MOSFET output stages should be rewritten, completely.
The simulations of the (vertical) MOSFETs doesn't take into account the so called weak-inversion (or sub-threshold conduction). This has far reaching consequences, not just regarding the the graphs, but also with respect to your conclusions about MOSFETs. The point is that they perform much better then you and your simulator might think.
In this chapter, you stated:
"However, the most important difference may be that the bipolar gain variations are gentle wobbles, while all FET plots seem to have abrupt changes that are much harder to linearize with NFB that must decline with rising frequency. The basically exponential Ic/Vbe characteristics of two BJTs approach much more closely the ideal of conjugate (i.e. always adding up to 1) mathematical functions, and this is the root cause of the much lower crossover distortion.
A close-up examination of the way in which the two types of device begin conducting as their input voltages increase shows that FETs move abruptly into the square-law part of their characteristic, while the exponential behavior of bipolars actually gives a much slower and smoother start to conduction."
Well, the truth is that at low currents, that is, in the transition region (sorry for using the T-word again) MOSFETs show a similar exponential behavior. So the X-over between the two output devices isn't that abrupt. Admittedly, WRT to BJTs, the gain 'wobble' is higher, but equally important, it is also wider. That means that the spectrum of harmonics from this wobble is smaller (i.e. less HF). Consequently, it's easier to linearize it by means of feedback.
Actually, I'm surprised that you didn't revise this chapter already in the 5th edition, as Marcel van de Gevel pointed this already out in 1996, see below.
Cheers,
E.
PS: You really don't like MOSFEts, do you?
Hi Douglas,
Forgive me if I say it a bit bluntly, but you for ask additions and/or improvement on the 5th edition: Chapter 14 on MOSFET output stages should be rewritten, completely.
The simulations of the (vertical) MOSFETs doesn't take into account the so called weak-inversion (or sub-threshold conduction). This has far reaching consequences, not just regarding the the graphs, but also with respect to your conclusions about MOSFETs. The point is that they perform much better then you and your simulator might think.
In this chapter, you stated:
"However, the most important difference may be that the bipolar gain variations are gentle wobbles, while all FET plots seem to have abrupt changes that are much harder to linearize with NFB that must decline with rising frequency. The basically exponential Ic/Vbe characteristics of two BJTs approach much more closely the ideal of conjugate (i.e. always adding up to 1) mathematical functions, and this is the root cause of the much lower crossover distortion.
A close-up examination of the way in which the two types of device begin conducting as their input voltages increase shows that FETs move abruptly into the square-law part of their characteristic, while the exponential behavior of bipolars actually gives a much slower and smoother start to conduction."
Well, the truth is that at low currents, that is, in the transition region (sorry for using the T-word again
) MOSFETs show a similar exponential behavior. So the X-over between the two output devices isn't that abrupt. Admittedly, WRT to BJTs, the gain 'wobble' is higher, but equally important, it is also wider. That means that the spectrum of harmonics from this wobble is smaller (i.e. less HF). Consequently, it's easier to linearize it by means of feedback.Actually, I'm surprised that you didn't revise this chapter already in the 5th edition, as Marcel van de Gevel pointed this already out in 1996, see below.
Cheers,
E.
PS: You really don't like MOSFEts, do you?
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.