john curl said:
Yes the lead designers are passed on, this fellow did some changes later on even later they eliminated some tantalum electrolytics from the original. He has sent me the original drawings, there is no money involved and there is nothing to be gained by subtrifuge. I will correct that to someone responsible for making design changes at some point.
BTW I've built two of these with some 50pF capsules and one of our new fast op-amps 2nV on 1.5mA. The whole thing runs on 2.5mA so it's great for phantom power and it is 100mV/PA to boot. Sounds great too!
Attachments
G.Kleinschmidt said:Hi Edmond
To beat jfet linearity with BJT's, we do not need to resort to a CFB input stage, we just need emitter degeneration
A good on-line reference:
http://www.dself.dsl.pipex.com/ampins/dipa/dipa.htm
5.1 DISTORTION 1. Input Pair Non-linearity
Cheers,
Glen
Hi Glen,
Maybe, but an ordinary BJT LTP with the same emitter degeneration (22 Ohm) and the same standing current per device (582uA) has only ~1/3 of the gain (gm) and produces at least ten times more distortion. If the CFB stage is well balanced, i.e. RE is exactly equal to 0.5 * VT / Ie, than the distortion is even 300 times lower WRT a LTP (at Vi = 1mV).
In this regard, nothing else can outperform a CFB input stage. The only minus are slightly more noise (from the bias circuitry) and a far larger offset voltage.
Cheers,
Edmond.
Scott, you are getting closer, but you are still compromising. In this case, you have to try a real discrete, first class, class A preamp and compare it to the IC version. I did that back when the 5534 was first introduced. The discrete preamp won. A pity really, because it was so much better than its predecessors, so much cheaper, and smaller in form factor.
For the record, I have made discrete electronics for at least 4 record companies, since the introduction of the 5534, with great success, I might add. I am even given a credit on several Album covers of Crystal Clear and Wilson Audio, and I designed an 8 ch discrete mixer for Elliot Mazer and 2 master recorders for Mobile Fidelity, etc, etc.
I also design microphones, on occasion at least their electronics
For the record, I have made discrete electronics for at least 4 record companies, since the introduction of the 5534, with great success, I might add. I am even given a credit on several Album covers of Crystal Clear and Wilson Audio, and I designed an 8 ch discrete mixer for Elliot Mazer and 2 master recorders for Mobile Fidelity, etc, etc.
I also design microphones, on occasion at least their electronics
GRollins said:
That is one of my points also. If you want to generalize the meaning of
negative feedback to include degeneration, then you must logically
decide that there is no such thing as "no feedback". In addition to the
"internal degeneration" in gain devices, there is an "internal loop"
between the Plate/Drain/Collector and the Grid/Gate/Base which
applies feedback to the device.
PMA, it is best, at this point, that I do not discuss the merits of, or problems with negative feedback with you (or anyone else) on this thread. That is OK, as you are doing your simulations, wonderfully well. Let us all just keep an open mind about feedback. Then we can proceed forward.
Scott, quality (hi end) capsules usually have 19pf such as the 1/2 inch B&K. Calrec is upper mid fi. Also, effective static capacitance is EVERYTHING in distortion generation from the capsule. Static vs dynamic capacitance. Check it out. Also, I first published my measurements of tantalum caps as coupling caps, 30 years ago, next month. It is about time that they removed a few. DUH!
john curl said:Scott, quality (hi end) capsules usually have 19pf such as the 1/2 inch B&K. Calrec is upper mid fi. Also, effective static capacitance is EVERYTHING in distortion generation from the capsule. Static vs dynamic capacitance. Check it out. Also, I first published my measurements of tantalum caps as coupling caps, 30 years ago, next month. It is about time that they removed a few. DUH!
High end bottle mikes are more like 50-100pF and B&K has several 50pF capsules. Neuman experimented with feedback to the capsule years ago but the FET op-amps they used didn't cut it. I don't know why they just didn't try the FET source follower in front of a low noise bi-polar. That circuit gets > 8.5 volts p-p of undistorted output on a 9V battery.
scott wurcer said:
Hi Scott,
You and Grey are right. I went back and read my letter and think I found the places you are referring to. I might have described it as indignation rather than emotion, but I see what you mean. These are the places I think you are referring to:
"That, as opposed to the joint condemnation here. Your four-on-one concept of fairness is, however, consistent with the nature and tone of the letter."
"First, the letter is very presumptuous. The unmistakable message which comes across is that together you consider yourselves the High Priests of TIM, and are appointed to sit in judgment of all TIM work not your own; that because you completely agree with each other all of your positions must be correct; that you must have the last word, even if it is a pompous 'we don't know'. There is very little technical disagreement to fuel your criticism, so two other forms of argument have unfortunately surfaced. The first is a form of one-up-man-ship; you criticise the few minor things you can, restate much of what I said, add a few wrinkles of your own, then hope that the volume of your commentary will convince the reader that there must be much wrong with what I've said. A common ploy of children and politicians, but not professionals."
I admit it, my statements there were born of some anger. I was being pretty haughty. I felt at the time that their letter was a thinly disguised expression of professional jealousy. It's unfortunate that my equally unprofessional response distracted from the substance of my reply. I plead guilty.
The important thing is that the letters reveal that there was not all that much technical difference between us at that point in time. I still think their letter was much ado about nothing, and that it certainly did not rebut the technical substance of my article. Some will disagree, and I'll be happy to discuss areas where anyone thinks I got it wrong in the article; that's how we all learn.
Cheers,
Bob
Nelson Pass said:
At this point I must ask: what exactly do you understand by negative feedback? Just state your definition and I may show how the emitter degeration or emitter follower matches your definition. And yes, some intrinsic device models as well.
As somebody else mentioned, devices don't know about NFB and so are simulators. From this perspective, there are only circuits and their properties. Each and every circuit can be analyzed with or without any feedback theory prior knowledge or understanding and the results are going to be precisely the same.
Then what about looking at feedback theory as a method of analyzing certain circuit topologies? That's the main reason why I prefer looking at emitter degeration or emitter follower in terms of feedback: it allows me to explain easily why emitter degeneration reduces THD and why an emitter follower can be unstable. From this perspective, feedback is just an abstraction that allows using a set of analysis methods (loop gain, stability criteria, etc...) and easily predict the circuit behaviour.
Yes, some internal device may appear as local feedback loops. Always remember that device equivalent circuits are only attempts to look at the device physics. Multiple models may exist and coexist and feedback elements may or may not appear in the device model.
Anyways, the fine line that isolates the possible internal device feedback loops and the external world is called pole/zero separation. If the internal poles/zeros are far away from the external poles/zeros than the internal and external feedback loops do not significantly interact. Therefore, from the circuit designer's perspective, the device is a black box with known parameters, independent from the external circuitry. This is almost always true for the audio modern devices. Hiowever, e.g. for very high frequency circuitry, there are better methods of design and analysis than feedback theory.
Edmond Stuart said:
OK, I agree, no BJT LTP will compete in linearity at that standing current. 5mA I(tail) is a different story however.
But anyway, all this nonsense argumentation over JFET or BJT input stages
Everybody knows that any proper amp uses sweet vacuum tubes here http://users.picknowl.com.au/~glenk/K10A.HTM
BTW, lots of feedback too, and clean 20kHz clipping also. Amazing, huh?
Attachments
- Status
- Not open for further replies.