This is what I got in simulator PMA used! It`s my version of complementary jfet/bjt preamp circuit with NO global negative feedback. The circuit needs buffer because the output is (10K) resistor loaded like in a JC2 preamp, so results of simulation are without output buffer.
That would be 0.00016% second harmonic for a 3V peak output for a single ende in/out and a much less for a balanced outout!
That would be 0.00016% second harmonic for a 3V peak output for a single ende in/out and a much less for a balanced outout!
Attachments
Last edited:
Well, it looks like everybody 'fell off the truck'. Could it be that schematics are intimidating? Or just too simple to bother with?
However, this is the question: How do you make a differential pair (of any active device) that is both lowest distortion and lowest noise (especially over a broad range of input impedance)?
Second question, how can you extend the S curve (always present in differential pairs) so that you are not hard limited by 2 Id ?
Three, who invented this first? and how many decades ago?
However, this is the question: How do you make a differential pair (of any active device) that is both lowest distortion and lowest noise (especially over a broad range of input impedance)?
Second question, how can you extend the S curve (always present in differential pairs) so that you are not hard limited by 2 Id ?
Three, who invented this first?
and how many decades ago?
John, I would like to ask a 'practical' question. Is there a big difference between my simulation results, for some 2.8Vpeak output amplitude, and your measurement on a real circuit? Also, is there a big difference between simulation and measurement in spectral content for SE and balanced input drive?
http://www.diyaudio.com/forums/showthread.php?p=1931538
http://www.diyaudio.com/forums/showthread.php?p=1931538
I don't think so, Pavel. I got .02% harmonic distortion at 2.5V out, with 3rd dominant and 2n'd about 20db below third, at 4KHz. This, of course, is beyond clipping level of my JC-1 or similar power amps or about 800+ watts output/channel into my loudspeakers with the JC-1, so it should be a high enough level for consideration.
Last edited:
Well, it looks like everybody 'fell off the truck'. Could it be that schematics are intimidating? Or just too simple to bother with?
However, this is the question: How do you make a differential pair (of any active device) that is both lowest distortion and lowest noise (especially over a broad range of input impedance)?
Second question, how can you extend the S curve (always present in differential pairs) so that you are not hard limited by 2 Id ?
Three, who invented this first?
I personally jumped out of the truck because I expected something exceptional, but found, like in a story, "a new patented way to screw the bulb", when the main device stands on the stool, puts the bulb into a socket on a ceiling, 4 devices helping it take each leg of the stool and start going in circles screwing the bulb; 4 (or more) additional devices go around in circles just to save the main screwing device from dizziness.
I personally would prefer just one device screwing the bulb by hands.
Well, it looks like everybody 'fell off the truck'. Could it be that schematics are intimidating? Or just too simple to bother with?
However, this is the question: How do you make a differential pair (of any active device) that is both lowest distortion and lowest noise (especially over a broad range of input impedance)?
Second question, how can you extend the S curve (always present in differential pairs) so that you are not hard limited by 2 Id ?
Three, who invented this first?
Hello John, I only know the third question
One question I'm wondering since some time....
Which of the shown variants of input differential would you prefer (I think you used both?)
Thanks, Tino
Attachments
To be fair, a differential pair of vacuum tubes, such as the 7788, with variable screen regulators and properly designed by a real engineer, can do amazing things in terms of distortion. If you what to know more about this configuration, please contact Bascom King, of Santa Barbara, CA USA, as he helped design this configuration. It is not as quiet as fets, but the dynamic range is at least, the same as the best 4 jfet configuration.
Even a single triode tube, operating open loop, can be very good, as shown by the Audible Illusions line stage, if you want to not use differential input.
The real advantage of the 4 quadrant configuration over other approaches is the self biasing aspect, so that the S curve can be extended, and linearized, especially at extreme current swings. The differential pairs are in parallel, so the noise floor can be 3dB lower.
The biggest weakness is that the p channel fets have more capacitance for the same amount of transconductance i.e. Gm, and this can get in the way of extremely high Z input operation.
Even a single triode tube, operating open loop, can be very good, as shown by the Audible Illusions line stage, if you want to not use differential input.
The real advantage of the 4 quadrant configuration over other approaches is the self biasing aspect, so that the S curve can be extended, and linearized, especially at extreme current swings. The differential pairs are in parallel, so the noise floor can be 3dB lower.
The biggest weakness is that the p channel fets have more capacitance for the same amount of transconductance i.e. Gm, and this can get in the way of extremely high Z input operation.
Last edited:
Hi John,
My guess for the first question. If you look at RF devices, they are almost always far more simple than audio parts if you look at # of active devices in the signal path.
This is all solidly before I was born in the late 50s.
-Chris
Edit: Hey John, we were posting at the same time.
What would be interesting are some papers on this subject, rather than a look at how one man performed this design. Again, Bell Labs may have something here. Did Bascom King ever work for Bell? I'm not familiar with him or his work that I know of.
Fewer active devices in the signal path running at higher currents. Reduce supply voltages (so things run cooler and a selected optimal point) and fix D-S (or C-E) voltages to prevent modulation effects from the supply voltages varying.
My guess for the first question. If you look at RF devices, they are almost always far more simple than audio parts if you look at # of active devices in the signal path.
That's a good question, by I'd ask why there is a need to remove this limit? Just work within it, and if your bias currents are high, it's very unlikely you would even approach 1 ID. After all, limiting the signal current to a smaller fraction of the operating range buys you a more linear transfer curve anyway. Once you begin to approach the complete operating range of current swing, you are creating your own problems to solve. I think that may force you to a more complicated way of maintaining linearity, and that seems to be opposite to the direction you normally would design in John.
No idea. I would suspect that there are examples (in a Bell lab somewhere of) almost every way to deal with an audio signal. Remember that before feedback amplifiers were designed, everything they did was open loop and tubes.
This is all solidly before I was born in the late 50s.
-Chris
Edit: Hey John, we were posting at the same time.
What would be interesting are some papers on this subject, rather than a look at how one man performed this design. Again, Bell Labs may have something here. Did Bascom King ever work for Bell? I'm not familiar with him or his work that I know of.
Hi John,
I think this is an especially good point you make about the S-curve extension for the four-quadrant JFET design. It is a very nice feature. I am not aware of any BJT implementation of the complementary differential pair input stage that has this nice feature.
Cheers,
Bob
If one wants to make a design like the CTC Blowtorch, then operation near the current extremes merits attention. Handwaving this away serves no purpose, any more that critiquing each and every design approach on this website. I could certainly do it, but for what gain? To discourage a certain percentage of inputs?
Perhaps, Bell Labs has developed every permutation and combination in the known world when it comes to circuit topology. So, prove it.
Perhaps, Bell Labs has developed every permutation and combination in the known world when it comes to circuit topology. So, prove it.
Hi John,
Remember, you are teaching, I am learning. An answer to your question is simply my best guess as a student.
John, I am not your enemy, not do I wish you a hard time in any way. Please shift your view to something less threatening and take my questions simply as ... questions. There is no other motive than to learn.
BTW, thank you for the information you have shared so far. Any offers I have ever made to assist you have been earnest. One way I can maybe thank you for your efforts. Yes, I am that simple to understand. Ask Stuart if you are really unsure.
-Chris
Okay, but I was unaware that you pushed the operating point so far out. It just seemed to me that it would be counter-intuitive. Given that I haven't dissected your design, I simply applied what I know of a successful minimalist design.
Remember, you are teaching, I am learning. An answer to your question is simply my best guess as a student.
Not my intention, it was a valid question. I did post my reasoning for my answer so you would understand where I was coming from. Attacking your idea was furtherest from my mind. All you had to say that I was wrong, or that there was a better way John.
Wouldn't that depend on the current levels you set? You're the designer, so it's up to you to set your operating points and limits created from those decisions. This is another question John, I am not debating with you on this stuff.
John, I was not attacking you in any way. I was simply pointing out that Bell Labs has been responsible for most of the early knowledge. In other words, a good place to investigate if you want to know about "the earliest" of audio electronics. Wouldn't you agree? I have zero to prove, it was a suggestion on where a great deal of information was developed, that's all.
John, I am not your enemy, not do I wish you a hard time in any way. Please shift your view to something less threatening and take my questions simply as ... questions. There is no other motive than to learn.
BTW, thank you for the information you have shared so far. Any offers I have ever made to assist you have been earnest. One way I can maybe thank you for your efforts. Yes, I am that simple to understand. Ask Stuart if you are really unsure.
-Chris
Hi John,
Sigh ........
Look, I had not much to analyze up until I knew what was a legit circuit. I don't have time to dig through other sources on the internet, nor am I so interested in you that I wish to look up everything on the 'net about you. So yes, you are quite correct in that I did not analyze your circuit. I figured it would unfold in the thread.
As for my comments to you in the past, they were expressly focused on the way you reacted to other members around you. Not a position I enjoyed, but I felt strongly that your responses to most people were not professional and completely undeserved for the most part. I haven't really faulted your actual circuit as it was unknown to me - and still is if you ask me about details. So until I understand your circuit, and the reasons you made your decisions, I can't possibly have an opinion. Anything I've heard has been in isolation. Never in context to the rest of your design.
In my previous post that you are responding to (1534), I expressly told you that I was learning from you, and thanked you for the knowledge you have shared so far. In effect, I was attempting to smooth the waters as I have in the past. You know, if you wish to force an antagonistic relationship, I really do not have any choice but to respond as I have. I will of course step in where I feel you are abusing people around you. I'll also mention I have doubts when you expect everyone to suspend what they know of electronics or physics. I do not relish this, and I certainly do not enjoy it either. It is a waste of time and energy on my part, but I will not support anyone abusing other members around you.
It's as simple as that John, and you are the driver. I am very sorry that you can't accept my position. There is nothing else to say for me except that I offered you an olive branch.
-Chris
Sigh ........
Look, I had not much to analyze up until I knew what was a legit circuit. I don't have time to dig through other sources on the internet, nor am I so interested in you that I wish to look up everything on the 'net about you. So yes, you are quite correct in that I did not analyze your circuit. I figured it would unfold in the thread.
As for my comments to you in the past, they were expressly focused on the way you reacted to other members around you. Not a position I enjoyed, but I felt strongly that your responses to most people were not professional and completely undeserved for the most part. I haven't really faulted your actual circuit as it was unknown to me - and still is if you ask me about details. So until I understand your circuit, and the reasons you made your decisions, I can't possibly have an opinion. Anything I've heard has been in isolation. Never in context to the rest of your design.
I didn't, nor did I request this from you. That comment is both condescending and a bit insulting. A bit unfair really, especially if you read post #1427, http://www.diyaudio.com/forums/showthread.php?p=1930591
In my previous post that you are responding to (1534), I expressly told you that I was learning from you, and thanked you for the knowledge you have shared so far. In effect, I was attempting to smooth the waters as I have in the past. You know, if you wish to force an antagonistic relationship, I really do not have any choice but to respond as I have. I will of course step in where I feel you are abusing people around you. I'll also mention I have doubts when you expect everyone to suspend what they know of electronics or physics. I do not relish this, and I certainly do not enjoy it either. It is a waste of time and energy on my part, but I will not support anyone abusing other members around you.
It's as simple as that John, and you are the driver. I am very sorry that you can't accept my position. There is nothing else to say for me except that I offered you an olive branch.
-Chris
Chris;
let me rephrase you: "Johan Sebastian Bach used all combinations of musical notes long time before", but people still compose new and new examples of great music.
Especially today when modern solid state devices and many wonderful passive components are available, it is possible to build affordable thingies Bell Labs did not even dream of.
It is easy to say, "Ah-hah! That Great Dude in 1812 who I ride with in his carriage pulled by 3 pedigree horses used such a chord once when he played piano for Her Majesty!", but what's the point? To put down everyone on this forum who uses such a chord (tubes, diffpairs, does not matter) suggesting that actually he/she is stealing somebody's previous work?
"Even a single triode tube, operating open loop, can be very good, as shown by the Audible Illusions line stage, if you want to not use differential input. " -- is Audible Illusion the single company that has shown something good? Or what's the point? Why such a disrespect of fellow DIYers on the DIY forum, John?
I see John, you are trying to heat up emotions, to bring back attention to the thread such a way, that has nothing new and interesting by itself...
Chris, when you find out 'where the bear sits' when it comes to the 3 designs put up here in the past few days, you and I can have a better understanding. Since, you don't consider me a 'design engineer' it might surprise you that all 3 designs are developed by me, over the decades, and that the basic topology, the self biasing comp diff jfet input stage, was not in print in any 'normal' electronics textbook, magazine or journal, to the best of my knowledge, with is extensive, before I introduced it in the early '70's.
I might just say that I have, until recently, scanned virtually every well known engineering magazine since the 1930's, including 'Electronics', 'EDN', 'Electronic Engineering', and a dozen others, at minimum, the entire JAES, and hundreds of journals from the IEEE, which I am a life member, after being continuously a MEMBER for the last 43 years. And you insist that I am not an engineer.
I might just say that I have, until recently, scanned virtually every well known engineering magazine since the 1930's, including 'Electronics', 'EDN', 'Electronic Engineering', and a dozen others, at minimum, the entire JAES, and hundreds of journals from the IEEE, which I am a life member, after being continuously a MEMBER for the last 43 years. And you insist that I am not an engineer.
Last edited:
All depends on many factors. Like in any car, most significant is a gasket between a front seat and a steering wheel, linearity of any stage, no matter which active components it uses, depends on the designer's specifications and how well he / she had met that specifications. Each active device has own character, and linearity depends on how properly it is used, according to it's character.
Wavebourn, I don't have any understanding of what you are talking about most of the time, and I have more than 40 years experience in electronic design. Others must have similar or even more difficulty in understanding what you are getting at. Perhaps, you should present a schematic that actually does anything useful and is unusual in its concept.
- Status
- Not open for further replies.