the 1st rule in engineering is to try brute force and only resort to "finess" if the obvious doesn't get the job done
i think anywhere you might use the Finesse circuit, a cap multiplier would likely work
further it is very difficult to improve noise or psrr by putting a cap across a zener, the zener has low incremental impedance and pushes the RzC freq up, splitting the source R to filter the zener bias is more effective, likewise output RC filtering is better to reduce zener noise
simple RC attenuation in a ps is valuable due to the good RF rejection possible - when active junctions become rectifiers and gain to provide local negative feedback is history
a complete design would look a cfp stability, base stopper and series output lossy ferrite would be good things to evaluate when putting C directly across the output
i think anywhere you might use the Finesse circuit, a cap multiplier would likely work
further it is very difficult to improve noise or psrr by putting a cap across a zener, the zener has low incremental impedance and pushes the RzC freq up, splitting the source R to filter the zener bias is more effective, likewise output RC filtering is better to reduce zener noise
simple RC attenuation in a ps is valuable due to the good RF rejection possible - when active junctions become rectifiers and gain to provide local negative feedback is history
a complete design would look a cfp stability, base stopper and series output lossy ferrite would be good things to evaluate when putting C directly across the output
Attachments
Deja Vu or not for you?
Jeff Rowland used to use this topology for the regulators in the Coherance preamp and another forum member I know has worked quite extensively with that topology. They require great effort to get stable and are very twitchy to design. This is only a topology for the very experienced and I would not do it without at least a 100 MHz scope and preferably good Spice modeling. A bipolar and a mosfet may give better results and more flexibility in design for high frequency stability.
Jeff Rowland used to use this topology for the regulators in the Coherance preamp and another forum member I know has worked quite extensively with that topology. They require great effort to get stable and are very twitchy to design. This is only a topology for the very experienced and I would not do it without at least a 100 MHz scope and preferably good Spice modeling. A bipolar and a mosfet may give better results and more flexibility in design for high frequency stability.
Thanks for all the answers.
Jcx: Is there something wrong with the finesse? It gives higher output impedance but that is taken care of by the follower.
Yes i know that the active elements doesn't go well with the higher frequencies but they do good in the audio band which i found neat.
I am aware of the zener problems but the finesse circuit did a good job here. I would like to know if it is a workable solution, do not think i have the equipment to measure the zener noise with the circuit applied.
Yes i intend to use a lossy ferrite on the input of the regulator but i think i can avoid it on the output if i am to power a opamp.
Fred:
Not experienced in analog design at all and definitely not with other designs available.
Do you mean the Sziklai? I found i needed to have a very small cap across the 600ohm resistor, didn't include it in the schematic, to get a stable performance but other than that it seems to do a good job. I suppose the resistor is very sensitive to inductance.
The fineesse circuit with the sziklai simplifies the choice of the emitter resistor since the Rbe is more ideal.
Jcx: Is there something wrong with the finesse? It gives higher output impedance but that is taken care of by the follower.
Yes i know that the active elements doesn't go well with the higher frequencies but they do good in the audio band which i found neat.
I am aware of the zener problems but the finesse circuit did a good job here. I would like to know if it is a workable solution, do not think i have the equipment to measure the zener noise with the circuit applied.
Yes i intend to use a lossy ferrite on the input of the regulator but i think i can avoid it on the output if i am to power a opamp.
Fred:
Not experienced in analog design at all and definitely not with other designs available.
Do you mean the Sziklai? I found i needed to have a very small cap across the 600ohm resistor, didn't include it in the schematic, to get a stable performance but other than that it seems to do a good job. I suppose the resistor is very sensitive to inductance.
The fineesse circuit with the sziklai simplifies the choice of the emitter resistor since the Rbe is more ideal.
Your post is the first I’ve seen of the Finesse circuit so I haven’t thought long or hard about it but my 1st thought is that feeding noise into the input of a feedback amplifier can’t be the best approach for a robust noise reduction circuit
This doesn’t mean the circuit may not be useful in the right application but I like the Cap multiplier’s larger input noise amplitude range (it will soak up narrow positive spikes up to the transistor rating, or a few hundred mV of lower frequency noise) and provide good attenuation up to frequencies where Cce feedthru limits performance
modeling of simple active circuit psrr and talk of adding 100 dB attenuation to circuits with 40-60+ dB rejection already is somewhat pointless, if you’ve ever tried really getting even 16 bit (96 dB) performance out of a ADC system you quickly realize that the little triangle ground symbol is a misleading fiction and ground currents and ps inductive coupling effects are the real problems
Stability with Cap loads is an issue for single emitter followers, the added poles in a Sziklai follower make the situation worse, series ferrite can decouple and add damping to the output bypass C, the Sziklai follower is a feedback amplifier and much has been written about output C driving, stability and impedance peaking in regulators (= feedback amplifiers)
The super-regulator thread(s) illustrate both the ground noise and stability problems
With a low noise, high gain bandpass filter op amp circuit you should be able to use even a cheap pc soundcard to look at audio frequency noise – of course there are issues and limitations to measuring noise with a computer’s ground connected to your circuit but batteries, isolation xfmrs and/or audio coupling xfrmrs could get you to a pretty good level of resolution
This doesn’t mean the circuit may not be useful in the right application but I like the Cap multiplier’s larger input noise amplitude range (it will soak up narrow positive spikes up to the transistor rating, or a few hundred mV of lower frequency noise) and provide good attenuation up to frequencies where Cce feedthru limits performance
modeling of simple active circuit psrr and talk of adding 100 dB attenuation to circuits with 40-60+ dB rejection already is somewhat pointless, if you’ve ever tried really getting even 16 bit (96 dB) performance out of a ADC system you quickly realize that the little triangle ground symbol is a misleading fiction and ground currents and ps inductive coupling effects are the real problems
Stability with Cap loads is an issue for single emitter followers, the added poles in a Sziklai follower make the situation worse, series ferrite can decouple and add damping to the output bypass C, the Sziklai follower is a feedback amplifier and much has been written about output C driving, stability and impedance peaking in regulators (= feedback amplifiers)
The super-regulator thread(s) illustrate both the ground noise and stability problems
With a low noise, high gain bandpass filter op amp circuit you should be able to use even a cheap pc soundcard to look at audio frequency noise – of course there are issues and limitations to measuring noise with a computer’s ground connected to your circuit but batteries, isolation xfmrs and/or audio coupling xfrmrs could get you to a pretty good level of resolution
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.