Hi Christer,
The "Rush" circuit cannot be called a cascode at all. I have seen it called "series differential pair" in comparison to the more conventional "parallel differential pair" before I became aware it was attributed to Rush.
I've seen used at the input of some NAD preamplifiers or amplifiers, in an intermediate stage in Cherry's NDFL amplifier, in a Hewlett-Packard signal generator and may be, with different types for the lower and upper device, in an Audio Resarch preamp.
http://www.4qdtec.com/opamp.html
Figure 4 :
The text has wrong comments : yes, it is an emitter coupled circuit and then a differential circuit. But where is the "long tail pair" ? I thought an LTP implies the use of a constant current source. One of the interests of the Rush's circuit is to avoid it, with less parasitic capacitances than with the LTP. However, cery simply, being not a push-pull circuit, the linearity of the series differential circuit is poorer than the LTP.
The "Rush" circuit cannot be called a cascode at all. I have seen it called "series differential pair" in comparison to the more conventional "parallel differential pair" before I became aware it was attributed to Rush.
I've seen used at the input of some NAD preamplifiers or amplifiers, in an intermediate stage in Cherry's NDFL amplifier, in a Hewlett-Packard signal generator and may be, with different types for the lower and upper device, in an Audio Resarch preamp.
http://www.4qdtec.com/opamp.html
Figure 4 :
The text has wrong comments : yes, it is an emitter coupled circuit and then a differential circuit. But where is the "long tail pair" ? I thought an LTP implies the use of a constant current source. One of the interests of the Rush's circuit is to avoid it, with less parasitic capacitances than with the LTP. However, cery simply, being not a push-pull circuit, the linearity of the series differential circuit is poorer than the LTP.
forr said:
I agree it is not a good name for it, but it is the only name I know for it. Your suggestion makes much more sense, I think.
I tried google on both. For "rush cascode" I found only two hits, both to this forum, but strangely not the old thread I linked to, but two occasions where Kanwar used the term. Searching for "series differential pair" gave more hits, however, none of them had anything to do with transistors.
I honestly didn't reread the text now, but posted the link just because it had an example of an all-BJT Rush cascode. It is of course not identical to an ordinary diff pair, but still very similar if looking at the equations. The lack of a CCS to set the current independently of the input voltages is of course a major difference.
An externally hosted image should be here but it was not working when we last tested it.
Are Q2, and Q7 held at 10V because it's the voltage of testing for the datsheets so have graphs to hand ?
Sorry I'm struggling to grasp the roles of Q2, and Q7!
Do they just stop the FETS from burning?
I see that the zener is 'capped' so I guess this combined with common mode rejection is enough not to introduce too much noise.
Hi,
even without the JLH credit the circuit is easily recognised as a JLH creation. Note, no place for Miller cap compensation here.
The cascode transistors sit with their bases tied to the fixed voltage of the Zener referenced to signal ground.
Both emitters will sit about 600mV below the Zener reference.
The input pair will have source voltages about 1V to 2V below signal ground. The net effect of the cascode connected LTP is that the FETs have a fixed and relatively constant voltage of about 11Vds. Some designers use a very low voltage here, only 2V to 3V (an LED or two) and others stretch up to near the voltage rating of the input pair.
The advantage of the second option is reduced Cob (or it's FET equivalent) due to high Vds and operating the input pair at below their safe working, but fixed, voltage.
The first option loses the low Cob part but relies heavily on the fixed Vds (Vce in BJTs) which improves performance.
even without the JLH credit the circuit is easily recognised as a JLH creation. Note, no place for Miller cap compensation here.
The cascode transistors sit with their bases tied to the fixed voltage of the Zener referenced to signal ground.
Both emitters will sit about 600mV below the Zener reference.
The input pair will have source voltages about 1V to 2V below signal ground. The net effect of the cascode connected LTP is that the FETs have a fixed and relatively constant voltage of about 11Vds. Some designers use a very low voltage here, only 2V to 3V (an LED or two) and others stretch up to near the voltage rating of the input pair.
The advantage of the second option is reduced Cob (or it's FET equivalent) due to high Vds and operating the input pair at below their safe working, but fixed, voltage.
The first option loses the low Cob part but relies heavily on the fixed Vds (Vce in BJTs) which improves performance.
AndrewT said:
The second option also reduces Cob nonlinearity, which is, generally speaking (BUT not for all FET architectures!) higher at lower Vds. Low gm FETs work better with higer Vds for this case, as they require a larger Vgs change for equal net Id change, which also means Vgd change is larger. Cgd nonlinearity with Vgd would be more pronounced for lower Vds, so a higher Vds is beneficial, and of course the usual decrease in Cds also helps.
Note however that this is not an easy decision to make in practise as the behaviour of capacitances of a FET with respect to applied voltages and currents depends heavily on the channel architecture of the FET, and there are several in use.
The art of linear electronics / John Linsley Hood , 1998
The above is an outstanding book which analyzes the historical
development of gain modules, from just feedback via the common element
on a transistor, to two stage feedback, to three stage feedback.
Finally you get to a modern OP AMP circuit.
Another excellent book which analyzes distortion stage by stage in your basic three stage gain block is:
High-power audio amplifier construction manual : 50 to 500 watts for the audio perfectionist / G. Randy Slone ,1999
The most extensive treatment I have seen is:
The Audio Power Amplifier Design Handbook, by Douglas Self
http://www.dself.dsl.pipex.com/ampins/books/book.htm
The reason for going to JFET inputs would normally have to be an expectation of lower noise or distortion.
But this thread started with a discussion of slew rate, citing Walter Jung, another excellent audio electronics author. Some modern monolithic op amps have degraded emitters to give individual emitter resistances to the members of the input diff pair. This does give better slew rates.
All of these factors are complex and interrelated. But they can be analyzed. Simulations will help to verify the analysis, but they will not substitute for it.
Some types of open ended topics will do better if a parallel discussion is held here:
Audio Explorations:
http://groups.google.com/group/audioex_amps_atob
http://groups.google.com/group/audioex_power_supplies
http://groups.google.com/group/audioex_thermal_mechanical
I've just opened up a thread for this. Just make sure you are registered with Google, then join this group, and start posting.
http://groups.google.com/group/audioex_amps_atob/browse_thread/thread/63d70b698ba97db6
The above is an outstanding book which analyzes the historical
development of gain modules, from just feedback via the common element
on a transistor, to two stage feedback, to three stage feedback.
Finally you get to a modern OP AMP circuit.
Another excellent book which analyzes distortion stage by stage in your basic three stage gain block is:
High-power audio amplifier construction manual : 50 to 500 watts for the audio perfectionist / G. Randy Slone ,1999
The most extensive treatment I have seen is:
The Audio Power Amplifier Design Handbook, by Douglas Self
http://www.dself.dsl.pipex.com/ampins/books/book.htm
The reason for going to JFET inputs would normally have to be an expectation of lower noise or distortion.
But this thread started with a discussion of slew rate, citing Walter Jung, another excellent audio electronics author. Some modern monolithic op amps have degraded emitters to give individual emitter resistances to the members of the input diff pair. This does give better slew rates.
All of these factors are complex and interrelated. But they can be analyzed. Simulations will help to verify the analysis, but they will not substitute for it.
Some types of open ended topics will do better if a parallel discussion is held here:
Audio Explorations:
http://groups.google.com/group/audioex_amps_atob
http://groups.google.com/group/audioex_power_supplies
http://groups.google.com/group/audioex_thermal_mechanical
I've just opened up a thread for this. Just make sure you are registered with Google, then join this group, and start posting.
http://groups.google.com/group/audioex_amps_atob/browse_thread/thread/63d70b698ba97db6
i see no difference in fet or bjt, however bipolar have a more
dynamic effect musically compared to fets
this observation has also occured to me when comparing
all fet stages like those of np the sound is warmer
in hybrid class-a's i have found to have both a kind of third
effect which i cant really define
john
dynamic effect musically compared to fets
this observation has also occured to me when comparing
all fet stages like those of np the sound is warmer
in hybrid class-a's i have found to have both a kind of third
effect which i cant really define
john
PS:
If anyone wants to talk about the use of JFETs, in a broader context, these are excellent places:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1016880#post1016880
And at audio explorations
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/844f6608b617ae28?hl=en
If anyone wants to talk about the use of JFETs, in a broader context, these are excellent places:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1016880#post1016880
And at audio explorations
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/844f6608b617ae28?hl=en
As long as people are making comparisons does anybody want to comment on the use of mosfets in input stages? They where very popular with Nelson Pass for a while, but much (not all) of his new amps have jfet input stages. I know they suffer from high gate capacitance so input impedances of 10k – 45k (depending on mosfet ) are needed to prevent high end roll off.
jerluwoo, I largely agree with what your book says.
Might you be able to give us a link to it, so we can read more of it?
Might you be able to give us author and title at least?
MOSFET inputs for op amps are popular when the very highest input impedance is desired. But this is totally unnecessary for any audio applications.
MOSFET noise is much higher than JFET of BJT.
JFETs will have lower noise than BJT when driven by high impedance. This is important for the built in amp of a condenser microphone, or it can be with a moving magnet phono cartridge, or even a guitar pickup.
JFETs will still have a higher mismatch than BJT.
There is some appeal in using JFET inputs, to get lower order, and even order distortion.
But this issue is far from resolved in my mind.
I have threads to deal with this specifically:
http://groups.google.com/group/audi...26d9c/9801c11d6fad8c4b?hl=en#9801c11d6fad8c4b
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/681ed762f9c75517?hl=en
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/681ed762f9c75517?hl=en
ZMB
Might you be able to give us a link to it, so we can read more of it?
Might you be able to give us author and title at least?
MOSFET inputs for op amps are popular when the very highest input impedance is desired. But this is totally unnecessary for any audio applications.
MOSFET noise is much higher than JFET of BJT.
JFETs will have lower noise than BJT when driven by high impedance. This is important for the built in amp of a condenser microphone, or it can be with a moving magnet phono cartridge, or even a guitar pickup.
JFETs will still have a higher mismatch than BJT.
There is some appeal in using JFET inputs, to get lower order, and even order distortion.
But this issue is far from resolved in my mind.
I have threads to deal with this specifically:
http://groups.google.com/group/audi...26d9c/9801c11d6fad8c4b?hl=en#9801c11d6fad8c4b
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/681ed762f9c75517?hl=en
http://groups.google.com/group/audioex_preamp_source/browse_thread/thread/681ed762f9c75517?hl=en
ZMB
"Might you be able to give us author and title at least?"
Hi zenmasterbrian
the book is AUDIO POWER AMPLIFIERS DESIGN HANDBOOK
BY DOUGLAS SELF
this book is really good, no other book comes close
of course it doesnt cover everything and the language
used by the author is a little bit mystique, it covers the
basics and very conventional topology some might tell
you is an advanced book not at all but it isnt as easy
some diy books on amplifiers either
grab a copy of this book if you want to learn some good
amplifier design practice, i dont agree with everything in
the book the author is against subjectivism which i find
rather amusing i myself am a subjectivist and i am not
against objectivists infact good audio is both that way
you get a wider understanding
lastly i have read many books this got to be the best
ive ever read it is truly close to a masterpiece
regards
john
Hi zenmasterbrian
the book is AUDIO POWER AMPLIFIERS DESIGN HANDBOOK
BY DOUGLAS SELF
this book is really good, no other book comes close
of course it doesnt cover everything and the language
used by the author is a little bit mystique, it covers the
basics and very conventional topology some might tell
you is an advanced book not at all but it isnt as easy
some diy books on amplifiers either
grab a copy of this book if you want to learn some good
amplifier design practice, i dont agree with everything in
the book the author is against subjectivism which i find
rather amusing i myself am a subjectivist and i am not
against objectivists infact good audio is both that way
you get a wider understanding
lastly i have read many books this got to be the best
ive ever read it is truly close to a masterpiece
regards
john
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.