I know that some books get translated into other languages and are then published in those languages, but I really don't have a clue how that works and what criteria a publisher uses for deciding if a book will be made available in a foreign language - or what language.
Cheers,
Bob
As you said, it's up to the publisher, but if you have any influence I like how pragmatic programming does it. For something like $10 they will give you the book in ebook (mobi/e-pub/pdf) after you answer some questions that only a holder of the physical book could answer (what's the first word of the 3rd paragraph on page 213). They protect the file somewhat by embedding your personal name into it to discourage theft.
I ordered it a few weeks back together with two of Self's books. Having now read Self's Power amp book, I am in the need for some "semiconductor balance" as I find Self too deep into the BJT world. Myself, I like JFETs and also MOSFETs. Self seems to dislike both JFETs and MOSFETs. Too bad.
Bob - I really look forward to your book and I hope it is pro MOSFET ;-)
Will be a nice Xmas present to myself
Best Regards,
Ingemar
Bob - I really look forward to your book and I hope it is pro MOSFET ;-)
Will be a nice Xmas present to myself
Best Regards,
Ingemar
MOSFETs
>Self seems to dislike both JFETs and MOSFETs.
His opinion on MOSFETs was only based on totally wrong simulations.
See: http://www.diyaudio.com/forums/soli...-interview-bjt-vs-mosfet-245.html#post1293783
>Self seems to dislike both JFETs and MOSFETs.
His opinion on MOSFETs was only based on totally wrong simulations.
See: http://www.diyaudio.com/forums/soli...-interview-bjt-vs-mosfet-245.html#post1293783
Hi Edmond,
My own experiences with Mosfets tend to bear this viewpoint out. With the same voltage amp stage and identical feedback, BJT stages tend to have lower distortion than Mosfet stages. This is not to say that Mosfet output stages can't sound good and have very good specifications, because they can (as Bob has proved). I think it's also true that many audio designers don't understand these devices at all and end up with a very poor design as a result. It seems that BJT design is better understood on average and those poor designs were created by people trying to cash in on the "Mosfet" fad.
Anyway, Doug's opinions are based on his experiences with Mosfets, not just some poor simulations. I think you are doing Doug Self a disservice by repeating this incorrect assumption you've made. Talk to the man first
-Chris
People are more comfortable with parts they use all the time. It's human nature. However, in conversation with Mr. Self, his opinions are based on his experience with the devices. Understand that until recently, commercially viable Mosfets were made by On Semi, IR and others. These are more suitable for inverters and such than music output stages.
My own experiences with Mosfets tend to bear this viewpoint out. With the same voltage amp stage and identical feedback, BJT stages tend to have lower distortion than Mosfet stages. This is not to say that Mosfet output stages can't sound good and have very good specifications, because they can (as Bob has proved). I think it's also true that many audio designers don't understand these devices at all and end up with a very poor design as a result. It seems that BJT design is better understood on average and those poor designs were created by people trying to cash in on the "Mosfet" fad.
Anyway, Doug's opinions are based on his experiences with Mosfets, not just some poor simulations. I think you are doing Doug Self a disservice by repeating this incorrect assumption you've made. Talk to the man first
-Chris
Mr Cordell stated his preference for MOSFETs long time ago, so it's not to be expected that he'll be squeamish with covering them in his book.
Long time overdue imo, and makes this publication rather special.
Downside is that, after receiving and reading it, none of the early buyers can slip up ever again, oops.
Long time overdue imo, and makes this publication rather special.
Downside is that, after receiving and reading it, none of the early buyers can slip up ever again, oops.
I ordered it a few weeks back together with two of Self's books. Having now read Self's Power amp book, I am in the need for some "semiconductor balance" as I find Self too deep into the BJT world. Myself, I like JFETs and also MOSFETs. Self seems to dislike both JFETs and MOSFETs. Too bad.
Bob - I really look forward to your book and I hope it is pro MOSFET ;-)
Will be a nice Xmas present to myself
Best Regards,
Ingemar
Hi Ingmar,
Thanks for your kind words. I've tried to be neutral in my book regarding BJT's, MOSFETs and JFETs. I have definitely not shy'd away from MOSFETs, but I'm not sure one would see it as pro MOSFET at the expense of BJTs. One can make a splendid amplifier out of BJTs, especially if they use the ThermalTrak devices wisely.
Each technology has its own set of advantages and disadvantages, and i've tried to cover those fairly in my book.
However, the answer to which technology is best is more complicated than a single observation, like MOSFETs tending to produce more distortion in traditional tests due to transconductance droop.
Having said all that, if I had only one chance to make the best possible amplifier that would be judged on listening tests under real-world conditions, that amplifier would have a JFET front end, and a vertical MOSFET output stage, with BJTs in the middle. This amplifier would certainly not cost less, and it would probably produce more heat (but I'm assuming that this best shot at an amplifier would not be allowed to be class A). I can also say that it would not necessarily have to depend on error correction, nor would it necessarily do better in conventional lab testing with a resistive load.
If you are looking for the best lab-measured performance for the dollar, a BJT amplifier will, I think, usually win. This does NOT mean that a BJT amplifier cannot be truly wonderful. Either technology can produce superlative results in the hands of a designer who knows the chosen technology well and who pays great attention to detail.
Picking a technology amounts to picking which dragons you choose to slay. Some folks are better at slaying one breed of dragon, while others are better at slaying a different breed of dragon. Pick your poison.
I might add that I have spent a lot of words on ThermalTrak transistors in my book, and that those devices really make it a horse race between BJT and MOSFET output stages. One of the key things they improve, that is an Achilles heel of BJT output stages, is dynamic thermal stability.
Cheers,
Bob
The Hitachi MOSFETs have been around a very long time. I built my very first amplifier with them (not for hifi use) around 1986 and they were by no means new devices then. No-one would consider using those devices for inverters (Rds =1.7ohms), they were explicity for audio. What Edmond says is correct - Doug Self criticized MOSFETs (in his E&WW articles) based on static Spice simulations which showed their poor transconductance compared to bipolars.
That may indeed be the case now, but it doesn't negate what Edmond said, which was not an assumption, it was based on evidence from his (Doug's) articles.
Congratulations, Bob, a life achievement realised and a credit to you and the community.
I agree that writing style is very important. Sentences must be short and pithy, and praise is due for this achievement too.
I often think about the adage that there is a book in everyone. Maybe there is, but few actually produce one. It's a huge task, lots of passion and hot coffee.
Hugh
I agree that writing style is very important. Sentences must be short and pithy, and praise is due for this achievement too.
I often think about the adage that there is a book in everyone. Maybe there is, but few actually produce one. It's a huge task, lots of passion and hot coffee.
Hugh
Hi abraxalito,
I'm not going to get into a long, protracted debate over what Doug's early articles contain. I will state that from my own discussions with Doug, he does not decide what to like or dislike based on simulations alone. This is especially true when everyone recognizes the shortcomings of SPICE models. What Edmond stated is not accurate today, I can't vouch for many years past. That is what I base my comments on. Keep in mind that I also have many, many years experience with these devices and that I have kept my comments on them fair and accurate. I do attempt to be impartial when talking about various technologies and devices. Keeping in mind what I know of Mr. Self's opinions gained from direct communication with him, Edmond's comments stated in the present tense are clearly not accurate. My comments on Mosfets come from direct experimental experience and experience repairing commercial amplifier products that were released to the consumer market. Any comment that suggests that any good designer shuns a particular device type due only to past simulator experiments (or even current ones) really does a deep disservice to the target of those comments. In other words, any comments directed towards any professional audio designer worth his salt must acknowledge that they have exercised "due diligence" and a somewhat high level of intelligence.
Keep this in mind as well. Lateral Mosfets have historically had no reliable second source suppliers. Knowing this, any manufacturer who relies solely on a variable supply of parts should expect unreliability in the parts chain. Having said that, I would not hesitate to use On Semi Thermal Trak devices in a product line, after confirming with the supplier that the device will not be cut off without a suitable substitute. In the case where On Semi is suddenly removed form this earth, those devices could be substituted with discrete devices, diodes mounted on the collector lead close to the body. Doesn't work as well, but equipment made using these could be kept in service. Lateral Mosfets do not have anything with the same characteristics in order to replace a discontinued parts supply. That is simply an assessment of risk. I've gone through that horror story in servicing Luxman "LV-105" (and similar) amplifiers where the device type vanished for a while.
-Chris
I'm not going to get into a long, protracted debate over what Doug's early articles contain. I will state that from my own discussions with Doug, he does not decide what to like or dislike based on simulations alone. This is especially true when everyone recognizes the shortcomings of SPICE models. What Edmond stated is not accurate today, I can't vouch for many years past. That is what I base my comments on. Keep in mind that I also have many, many years experience with these devices and that I have kept my comments on them fair and accurate. I do attempt to be impartial when talking about various technologies and devices. Keeping in mind what I know of Mr. Self's opinions gained from direct communication with him, Edmond's comments stated in the present tense are clearly not accurate. My comments on Mosfets come from direct experimental experience and experience repairing commercial amplifier products that were released to the consumer market. Any comment that suggests that any good designer shuns a particular device type due only to past simulator experiments (or even current ones) really does a deep disservice to the target of those comments. In other words, any comments directed towards any professional audio designer worth his salt must acknowledge that they have exercised "due diligence" and a somewhat high level of intelligence.
Yes they have - agreed. I've run into these devices quite often and for a very long time. This is not disputed here.
Well, of course not. They have a high rds(on) and inverters run the pass elements fully on and off (blocking). There are also other considerations for the pass devices, like the intrinsic reverse body diode for example. No arguments there either.
While that may be true, and I'd have to read the articles you are referring to (if I haven't already), it's up to Doug Self himself to clarify his position. No one can say for sure what the depth of his knowledge or experiences were at the time of those articles. I will bet that he used simulations to illustrate a point, rather than rely on anecdotal information only. This would certainly give the impression that this is where he got his information from. Any author of a technical subject will try to present their viewpoints in a dry and factual manner, rather than ignite a long running debate. However as humans, what they intend to say and what actually comes across to a cross section of readers may be two different things. As I said before - ask the man first before telling the world what he thinks.
Keep this in mind as well. Lateral Mosfets have historically had no reliable second source suppliers. Knowing this, any manufacturer who relies solely on a variable supply of parts should expect unreliability in the parts chain. Having said that, I would not hesitate to use On Semi Thermal Trak devices in a product line, after confirming with the supplier that the device will not be cut off without a suitable substitute. In the case where On Semi is suddenly removed form this earth, those devices could be substituted with discrete devices, diodes mounted on the collector lead close to the body. Doesn't work as well, but equipment made using these could be kept in service. Lateral Mosfets do not have anything with the same characteristics in order to replace a discontinued parts supply. That is simply an assessment of risk. I've gone through that horror story in servicing Luxman "LV-105" (and similar) amplifiers where the device type vanished for a while.
-Chris
Well I don't know how this will come across in Bobs book, but what I find and found at the time fascinating in Dougs work and many articles was that no mention was ever made of actually "listening" to designs as part of the design process.
The goal was clear... to push the technical aspects ever higher toward the "blameless" and then be confident that the resulting design has to be beyond reproach sonically too.
You were left wondering whether the designs were ever listened to at all.
The late JLH on the other hand was perhaps the "opposite" of that approach.
Dougs technical ability is really beyond question, and he exploded many myths about amplifier design, and only recently have I detected in his writings a softening in that approach.
So it will be interesting to see where Bob stands on that... and whether listening tests form any part of the design procedure... guess we are going to have to wait till the book appears.
The goal was clear... to push the technical aspects ever higher toward the "blameless" and then be confident that the resulting design has to be beyond reproach sonically too.
You were left wondering whether the designs were ever listened to at all.
The late JLH on the other hand was perhaps the "opposite" of that approach.
Dougs technical ability is really beyond question, and he exploded many myths about amplifier design, and only recently have I detected in his writings a softening in that approach.
So it will be interesting to see where Bob stands on that... and whether listening tests form any part of the design procedure... guess we are going to have to wait till the book appears.
Hi Chris,
I simply was referring to what he has written. And that was his 'experience' with poor device models.
His sims were based on a 2SK135/2SJ50 pair. These MOSFETs were made especially for audio applications, thus NOT for switching.
That's right, but MOSFETs are a different kettle of fish. You can't simply replace BJTs by MOSFETs and expect the same results. They deserve a need and a different approach.
BTW, I wonder how BJTs would be judged if power MOSFETs were invented first. Maybe as totally useless, because of the low input impedance, second breakdown and hole storage effect.
I'm afraid the latter is very true.
Don't worry. I'm a big fan of Self. Please read one of my other comments:
http://www.diyaudio.com/forums/soli...-interview-bjt-vs-mosfet-246.html#post1294015
Why talk to him? Although wrong, he has written what he has written and that is crystal clear: "Figures 12 to 15 show spice gain plots, using 2SK135/2SJ50 devices".
Besides, I'm not the only one who disagrees with Self: See below, Marcel van de Gevel, EW, Feb. 1996, p.141
Cheers,
E.
Attachments
That's totally fine because there isn't really any room for debate over them, the articles can be read and studied and they're generally very clear with few if any 'variant readings' reasonably possible.
That's fine too because what he stated clearly uses the past tense : "his opinion was...".
Well it now looks like the apparent disagreement was due to your reading of Edmond's remarks as present tense.
Then I must have misinterpreted the sense of your orignal remarks. You said:
Understand that until recently, commercially viable Mosfets were made by On Semi, IR and others. These are more suitable for inverters and such than music output stages.
Hitachi MOSFETs clearly are an exception to your claim - they're both available prior to 'recently' and definitely NOT suitable for inverters. Or were you indirectly claiming that Hitachi MOSFETs were in the category 'not commercially viable' ?
... snipped out stuff agreed with, but found this point...
Not so the intrepid Mr Self in his chapter on 'subjectivism'.
<edit> just enough time to point out that its not 'everyone' who recognises the limitations of Spice models - Mr Self clearly did not in his 1996 book
Last edited:
How curious that you wrote this as I was chuckling at Mr Self's own words from his 1996 edition :
I must admit to a sneaking feeling that if practical power BJTs had come along after FETs, they would have been seized upon with glee as a major step forward in power amplification.
That was precisely the reason why I wrote that.
Excellent stuff Edmond.
While I was reviewing the book, I did come across these telling words from the same chapter (chapter 10: 'FET power amplifiers' in the first edition):
The PSpice simulation shown here was checked against manufacturers curves for the devices, and the agreement was very good - almost unnervingly so. It therefore seems reasonable to rely on simulator output for these kinds of studies - it is certainly infinitely quicker than doing the real measurements, and the comprehensive power FET component libraries that are part of PSpice allow the testing to be generalised over a huge number of component types without actually buying them.