Well I finally received my book today, excellent job Bob, worth its weight in gold 
congratulations and great success.
I went through Chapter 4, found a few minor corrections.
1) p97, 10th line up from bottom, 'discharges C5"
2) p100, 2nd line down from top, "Hysteresis is provided by R15(470K)
3) p116, 8th line up from bottom, "If JFETs are used, their drains can be connected to a +15-V supply to keep Vds ..." assuming n-ch jfets, I would think you need at + vs a - supply for the drains?
Lots to read, I'll be at it for a while.
Cheers
Rick
congratulations and great success.I went through Chapter 4, found a few minor corrections.
1) p97, 10th line up from bottom, 'discharges C5"
2) p100, 2nd line down from top, "Hysteresis is provided by R15(470K)
3) p116, 8th line up from bottom, "If JFETs are used, their drains can be connected to a +15-V supply to keep Vds ..." assuming n-ch jfets, I would think you need at + vs a - supply for the drains?
Lots to read, I'll be at it for a while.
Cheers
Rick
I don't think so.
The voltage drop across R3 for 5mA will be around 50mV and Q3 transistor Vbe3 voltage will be larger than Q1 Vbe1 voltage by 60mV. Therefore the voltage drop across R1 should be around 50mV+60mV = 110mV hence R1 = 110mV/0.5mA = 220R. But because of the fact that we have ignored the Q3 base current and Q2 current. We should slightly increase the resistor value.
But we'll see what Bob will say about it. I'm definitely not a master as Bob is. I'm just a hobbyist and never taken any electronics classes.
Your analysis is correct for ideal transistors with exactly the same Vbe; but this never happens in real life. Even matched transistors on the same die can have Vbe differences of up to 10mV. The chances of a random 2N5401 and a random 2N5551 transistor having the same Vbe is close to zero. In this particular instance I'd be surprised if 0.5mA through Q1 produced anything like 5mA through Q3 with either a 100R or 240R resistor unless Q1 and Q3 were very carefully chosen. The low current through Q1 and the large ratio of R1 to R3 make the current through Q3 very dependent on Vbe values. In the real world the rule of thumb that the inverse ratio of resistor values is roughly equal to the ratio of currents holds up pretty well. I took some measurements on one of my headphone amplifiers that has a diamond buffer output stage ( no parts selection ) and obtained the following results; resistors 120R and 10R, input current 3.0mA, predicted output current 36mA, actual output current 39mA.
The bias current in the second stage of a diamond buffer does not to first order depend on matching of Vbe between the PNP and NPN devices, since offsets of the two halves of the first and second stages tend to cancel out in creating the emitter-to-emitter difference at the second stage. Differences in Vbe between the NPN and PNP devices will, however, result in a DC offset between the input and output of the diamond buffer.
Cheers,
Bob
Hi Bob
With all due respect I think that your CFA design in figure. 9.19 on page 214 is killing the whole idea about a so called CFA.
Personally I would never design a CFA that way. (and never conect the collectors of the IPS BJT's to the Emitter of the second stage BJT's.)
To be honest I'm not a big fan of the CFA, but when I'm doing an analysis of VFA vs CFA I'm using good designs of both topologies.
What do others think? Bonsai, Scott, Syn08, Glen, Edmond and so on.
Stein
With all due respect I think that your CFA design in figure. 9.19 on page 214 is killing the whole idea about a so called CFA.
Personally I would never design a CFA that way. (and never conect the collectors of the IPS BJT's to the Emitter of the second stage BJT's.)
To be honest I'm not a big fan of the CFA, but when I'm doing an analysis of VFA vs CFA I'm using good designs of both topologies.
What do others think? Bonsai, Scott, Syn08, Glen, Edmond and so on.
Stein
Hi Stein,
Thanks for your feedback; I do appreciate it. That feedback may serve as a good starting point for a discussion. Can you be more specific in your criticisms, and describe in what way the simplified circuit of Figure 9.19 kills the whole idea of a CFA?
The IPS is essentially a diamond buffer. Connecting the collectors of Q1 and Q2 to the emitters of Q4 and Q3 respectively is one way to arrange a diamond buffer; it bootstraps the collectors of the input transistors with signal. Connecting those collectors to the rails instead would make little difference in the performance of the CFA.
In any case, recognize that this is a simplified CFA intended to capture the essence of a CFA design. It sounds like you do not think that it captures the essence. In what way does it not capture the essence?
If you could provide a schematic of how you would do a CFA of similar simplicity it would help.
Thanks again for your feedback.
Cheers,
Bob
Thanks for your feedback; I do appreciate it. That feedback may serve as a good starting point for a discussion. Can you be more specific in your criticisms, and describe in what way the simplified circuit of Figure 9.19 kills the whole idea of a CFA?
The IPS is essentially a diamond buffer. Connecting the collectors of Q1 and Q2 to the emitters of Q4 and Q3 respectively is one way to arrange a diamond buffer; it bootstraps the collectors of the input transistors with signal. Connecting those collectors to the rails instead would make little difference in the performance of the CFA.
In any case, recognize that this is a simplified CFA intended to capture the essence of a CFA design. It sounds like you do not think that it captures the essence. In what way does it not capture the essence?
If you could provide a schematic of how you would do a CFA of similar simplicity it would help.
Thanks again for your feedback.
Cheers,
Bob
Cross connecting the input stage collectors per fig. 9.19 is likely a good idea; set aside it could be interpreted as a local (current) feedback loop, it also keeps the Vbc of the input transistors more or less constant, lowering the Early effect distortions. Kind of bootstraping.
What I don't understand in that schematic is the huge R1/R3=24 ratio. This translates to a 12mA current through the current mirror and 24mA through the VAS (current gain is 2), which is in my view excessive. The schematic shows 5mA through the current mirror, so the issue could be a typo somewhere.
Since the two page discussion of CFA for audio is apparently intended more for completness rather than for going into the details, I don't think Bob would qualify as a CFA fan too
.
Last edited:
Hi Stein,
I only have the 1st edition of Bob's book, where page 214 is totally blank
, so I can't comment on that circuit.But generally, I have no problems with tying collectors to emitters of the next stage, see: MCP Front-End
Cheers, E.
Not to mention that (e.g.) Accuphase is manufacturing audio amplifiers (both preamps and power amps) using the standard CFA topology since the late 80's early 90's, so for about the famous 40 years now
.
Yes, you and I know it’s been that way since Elantec and Comlinear came out with parts 40 years ago but there remain troglodytes amongst us who will never accept CFA’s for audio or claim there is no such thing. As I have repeatedly stated, it’s no good fixating on these things - just design with the topology you are most comfortable with and if you do it right (buy Bob’s book for starters), I am quite sure it will sound excellent.
But, let’s stop this talk because it will soon degenerate into a CFA bashing session and this is Bob’s thread about his book.
Nice, requesting to stop the talk after calling people disagreeing with you about "CFAs for audio" as "troglodytes". And some are still wondering who's fueling this CFA war.
BTW, there are people around that were closer, 40 years ago, to the original Comlinear CFA development than yourself. As close as Scott, or even closer.
Last edited:
You’re twisting my words again. Typical. And polluting another thread with nonsense.
I’m sure you are Wally in disguise. Precisely the same tactics.
No - that indeed never comes from you
I got emails berating my use of them. I get folks accusing me of peddling them over VFA (our dear Waly, syn08, Mikeks, other folks as well).
Ridiculous.
Suggest we close this down on this thread.
I’m sure you are Wally in disguise. Precisely the same tactics.
No - that indeed never comes from you
I got emails berating my use of them. I get folks accusing me of peddling them over VFA (our dear Waly, syn08, Mikeks, other folks as well).
Ridiculous.
Suggest we close this down on this thread.