scott wurcer said:
Scott,
Thanks for the correction; my first example should have read "high negative feedback and low open loop bandwidth".
Cheers,
Bob
We've seen that GNFB is not needed at all to generate PIM.
Now, let's go to the version 3, with the 1'st order (first, literally) R-C, where C is variable It's a Ccb of one of amp's transistors. For simplicity of calculations you may assume that capacitance of the varactor linearly depend on a voltage.
Assume that opamps are ideal in terms of frequency response, but they have a limited gain.
Now, play with resistors R2 and R3 for different amplification factors, an open loop gain of the opamp, capacitance of the varactor, and value of the R1 resistor and see what conditions are needed for an ideal amp as shown to perform as a PIM generator.
Now, replace R1 by a current source then play again...
Now, let's go to the version 3, with the 1'st order (first, literally) R-C, where C is variable It's a Ccb of one of amp's transistors. For simplicity of calculations you may assume that capacitance of the varactor linearly depend on a voltage.
Assume that opamps are ideal in terms of frequency response, but they have a limited gain.
Now, play with resistors R2 and R3 for different amplification factors, an open loop gain of the opamp, capacitance of the varactor, and value of the R1 resistor and see what conditions are needed for an ideal amp as shown to perform as a PIM generator.
Now, replace R1 by a current source then play again...
Attachments
Now, play with this one: stable input capacitance of the output buffer, but non-linear high output resistance of what drives it... You may assume it is the 2'nd order for simplicity of calculations, i.e. it's output resistance linearly depends on the voltage.
What conditions do we need for such an idealized amp to act as a PIM generator?
Now, it's your time to play with real non-linearities and draw your conclusions...
What conditions do we need for such an idealized amp to act as a PIM generator?
Now, it's your time to play with real non-linearities and draw your conclusions...
Attachments
john curl said:
We would prefer you to explain it's meaning once instead of standing by many times...
The most recent interview with Otala (which I'm aware about) was published in 2003: http://mikropc.net/nettilehti/pdf/2711200320.pdf
Dimitri these things are fun to follow, this one was referenced by someone patenting an all npn output in 1990. Strange that folks were still non-complimentary then.
scott wurcer said:
There was a rule of thumb then: "PNP and NPN transistors can never be fully complimentary".
I have NOT read this whole thread in detail so far... but this is my share for the simplest PIM generator ever (I showed this one about a year ago in a discussion, IIRC, with Pavel... or Bob...).
Whenever a nonlinear static transfer function hits a net with non-constant group delay vs frequency we get into PIM troubles... or is this definition silly?
EDIT: found the original discussion... and my example isn't PIM, rather it showed that zero crossing shifts (as an indicator for PIM) can also result from someting else.
http://www.diyaudio.com/forums/showthread.php?postid=1537097#post1537097
- Klaus
Whenever a nonlinear static transfer function hits a net with non-constant group delay vs frequency we get into PIM troubles... or is this definition silly?
EDIT: found the original discussion... and my example isn't PIM, rather it showed that zero crossing shifts (as an indicator for PIM) can also result from someting else.
http://www.diyaudio.com/forums/showthread.php?postid=1537097#post1537097
- Klaus
Attachments
Bob Cordell said:
Hello Bob
I wonder if there actually is anything in those claims or is it just more marketing than substance ??
I have heard those sorts of claims since the days of the Rappaport amplifiers but it hasn't really moved the industry much.
Also I still can't get the anti-negative feedback protagonists to explain why the Halcro amps seem to receive unanimous praise in the press. According to their philosophies on zero or low global negative feedback the Halcro should be a bad amplifier from a subjective point of view but it doesn't read like that though 😉
regards
Trev
Hi Trevor;
there are many ways to skin the same cat.
3 main whales that support sound reproduction are: low power losses so less part of transfer function is used (less of stages, better impedance matching); assistance to expand the transfer function so it is more linear on wide power amplification ratio (bootstrapping, CCS loading, other error correction techniques), and feedbacks. Less known whales are more attention to more audible errors, and approximation. Blend and mix for optimal results, it sounds simple.
there are many ways to skin the same cat.
3 main whales that support sound reproduction are: low power losses so less part of transfer function is used (less of stages, better impedance matching); assistance to expand the transfer function so it is more linear on wide power amplification ratio (bootstrapping, CCS loading, other error correction techniques), and feedbacks. Less known whales are more attention to more audible errors, and approximation. Blend and mix for optimal results, it sounds simple.