Darry I think that your schematic N°3 (Post #565) is by far the best one, it’s a low noise design (filtered reference) and moreover a non feedback design and the mosfet gate is fed by a very low impedance source.
If the goal is to achieve a close design to John Curl’s PSU, I think it could be something like this. But maybe now John has solved his computer problems, if not maybe SY can enlighten this.
Talking about PSU’s, John do you use (or SY does he use) 1 or 2 regs in series (317/337) per polarity for the first "low frequency" power supply regulation part, 2 for a better LF rejection?
Another question is about the phono preamp, I read what follows on the Stereophile review of the Vendetta by J. Gordon Holt in June 1988:
“The HF part of the RIAA playback equalization is passive, and there is no feedback around the first stage, but some can be added if desired to reduce the input impedance to below 10 ohms minimum. The second stage consists of complementary symmetrical FETs, around which negative feedback is used for the LF portion of the RIAA EQ.”
My question is: nowadays will you still design a phono preamp with negative feedback on the second stage around the LF time constant of the RIAA equalisation, or will you go for a non feedback design, in other words negative feedback around this equalisation is it still the best compromise for gain, noise, low capacitor values (best quality) for the LF time constant, or even for others reasons ? (As it’s well known that generally you prefer non feedback designs.)
If the goal is to achieve a close design to John Curl’s PSU, I think it could be something like this. But maybe now John has solved his computer problems, if not maybe SY can enlighten this.
Talking about PSU’s, John do you use (or SY does he use) 1 or 2 regs in series (317/337) per polarity for the first "low frequency" power supply regulation part, 2 for a better LF rejection?
Another question is about the phono preamp, I read what follows on the Stereophile review of the Vendetta by J. Gordon Holt in June 1988:
“The HF part of the RIAA playback equalization is passive, and there is no feedback around the first stage, but some can be added if desired to reduce the input impedance to below 10 ohms minimum. The second stage consists of complementary symmetrical FETs, around which negative feedback is used for the LF portion of the RIAA EQ.”
My question is: nowadays will you still design a phono preamp with negative feedback on the second stage around the LF time constant of the RIAA equalisation, or will you go for a non feedback design, in other words negative feedback around this equalisation is it still the best compromise for gain, noise, low capacitor values (best quality) for the LF time constant, or even for others reasons ? (As it’s well known that generally you prefer non feedback designs.)
Richard, I would like to make an all 'open loop phono preamp' but it is not easy to do right. Someday I hope to make an improved replacement for the Vendetta phono stage. Now I am just updating units.
I find the constant criticism of this preamp, virtually a joke. We are now building 5 of these units, and I can say that the case is a little larger than necessary for most versions, but my personal unit, with the Vendetta phono stage is fairly full up, actually, and it would be even more so, if I had selected to have balanced inputs as well as balanced outputs. These units do not generate any signficant hum pickup in actual operation.
I find the constant criticism of this preamp, virtually a joke. We are now building 5 of these units, and I can say that the case is a little larger than necessary for most versions, but my personal unit, with the Vendetta phono stage is fairly full up, actually, and it would be even more so, if I had selected to have balanced inputs as well as balanced outputs. These units do not generate any signficant hum pickup in actual operation.
john curl said:
John at that time I suppose you tested a nonfeedback solution for the second stage, which are the drawbacks and difficulties?
As I would like to improve my preamp power supply, wich in fact is an “active/passive” crossover with input selector and vol control, actually the PSU is just a Mosfet cap multiplier. I would like to improve it with something like the schematic I proposed here a few weeks ago (LM317/337 then shunt reg), I am really interested to know if a series prereg as on the LM317/337 datasheet example could improve the sound and if it worth the sorrow, “on paper” it provides about 60 dB more line rejection in the LF region carrying the total LF rejection to 120 dB, but maybe this is not necessary. Just want to know if you have play with or even done it, and in this case which are your conclusions.
As I said before, lineup, balanced inputs are overrated. Actually the electronics circuit is fully balanced on both input and output, but virtually all of my sources are single ended, and so it is better to have single ended inputs on my preamp. However, the same CTC team has designed a less expensive preamp for Parasound that has fully balanced input and output. This requires XLR connectors on input and output, as well as quad input pots and selector switches. Think of the extra cost and space needed for a preamp of the Blowtorch's quality. It just wasn't worth it in my case. Now what about hum? Well, there isn't any signficant hum, but then I don't put my power supply next to the preamp either. I put it on a separate shelf perhaps a meter or so away from the preamp. This is necessary, because it is the hum sources that must be avoided.
It is true that with an all balanced system, the power supply could be moved closer. This is a small improvement for me, but your needs may be different.
It is true that with an all balanced system, the power supply could be moved closer. This is a small improvement for me, but your needs may be different.
Yes, I believe you.john curl said:
There are few situations, where trouble of building a balanced amplifier is justified.
Just as good sound quality, if not better, is to get from a normal single ended signal.
If doing it in a proper way.
There is just too much fuzz of building 'fully symmetrical' circuits and amplifiers.
With non symmetrical not fully complementary devices .....
The supposed magics of balanced systems belongs to this imperfect thinking.
thanks, john
I tend to not complicate designs more than what I think is necessary. I don't know if the more complicated regulator would really help, but I would doubt it, in my case.
This is because I am making fixed regulators that are only optimized to filter input hum down to the noise floor. I know that these IC regulators 'fall apart' over approximately 1KHz and get worse and worse as regulators. They also have marginal transient response, both input and load. They will also ring as they generate a synthetic inductor at their output. A very high Q cap can make things actually worsen at the output of these regulators.
However, let's say that I wanted to make a 'lab' supply: Then the situation might be different. I might try to remove any hum or noise to the highest degree possible so that I could test a prototype preamp design without any extra regulators downstream.
Personally, I find the IC regulators marginal and I like to separate them from the actual working circuits. However, they DO reduce the hum and set a respectably constant reference voltage. To reduce transients, RFI, and audio noise, I use other circuits or passive parts either before or after the IC regulator.
This is because I am making fixed regulators that are only optimized to filter input hum down to the noise floor. I know that these IC regulators 'fall apart' over approximately 1KHz and get worse and worse as regulators. They also have marginal transient response, both input and load. They will also ring as they generate a synthetic inductor at their output. A very high Q cap can make things actually worsen at the output of these regulators.
However, let's say that I wanted to make a 'lab' supply: Then the situation might be different. I might try to remove any hum or noise to the highest degree possible so that I could test a prototype preamp design without any extra regulators downstream.
Personally, I find the IC regulators marginal and I like to separate them from the actual working circuits. However, they DO reduce the hum and set a respectably constant reference voltage. To reduce transients, RFI, and audio noise, I use other circuits or passive parts either before or after the IC regulator.
JOHN CURL
---I know that these IC regulators 'fall apart' over approximately 1KHz and get worse and worse as regulators. They also have marginal transient response, both input and load. They will also ring as they generate a synthetic inductor at their output. A very high Q cap can make things actually worsen at the output of these regulators.---
Very curiously, many amplifiers are studied for excellent transient response but this aspect is almost neglected in the design of power supplies. The output of an ideal power supply should behave like a low value resistor, not necessarily the lowest as possible, over a wide bandwidth. I think this could be an audiophile niche for some engineers.
However, as far as I remember from very old tests, the parasitic inductance of a few inches of wire ahead of a transient load having local decoupling can do some good for the response of a regulated power supply.
~~~~~ Forr
§§§
---I know that these IC regulators 'fall apart' over approximately 1KHz and get worse and worse as regulators. They also have marginal transient response, both input and load. They will also ring as they generate a synthetic inductor at their output. A very high Q cap can make things actually worsen at the output of these regulators.---
Very curiously, many amplifiers are studied for excellent transient response but this aspect is almost neglected in the design of power supplies. The output of an ideal power supply should behave like a low value resistor, not necessarily the lowest as possible, over a wide bandwidth. I think this could be an audiophile niche for some engineers.
However, as far as I remember from very old tests, the parasitic inductance of a few inches of wire ahead of a transient load having local decoupling can do some good for the response of a regulated power supply.
~~~~~ Forr
§§§
Originally posted by forr
Very curiously, many amplifiers are studied for excellent transient response but this aspect is almost neglected in the design of power supplies. ~~~~~ Forr
````````````````````````````````````````````
to me, most amps look like regulators, that a signal,
is then run down the middle of
Very curiously, many amplifiers are studied for excellent transient response but this aspect is almost neglected in the design of power supplies. ~~~~~ Forr
````````````````````````````````````````````
to me, most amps look like regulators, that a signal,
is then run down the middle of
TOMTT
--- most amps look like regulators---
We could also say that most regulators look like amps.
Some differences : most of the time, a power supply works in a single ended class A mode ; its load is capacitive due to an usual output capacitor which plays a major role regarding stability (it even could - it even should - be made the stabilising element) and due to local decoupling of the following circuits.
I would be happy to see studies on the transient behaviour of power supplies, a too often "neglected factor" in audio electronic devices.
~~~~~~ Forr
§§§
--- most amps look like regulators---
We could also say that most regulators look like amps.
Some differences : most of the time, a power supply works in a single ended class A mode ; its load is capacitive due to an usual output capacitor which plays a major role regarding stability (it even could - it even should - be made the stabilising element) and due to local decoupling of the following circuits.
I would be happy to see studies on the transient behaviour of power supplies, a too often "neglected factor" in audio electronic devices.
~~~~~~ Forr
§§§
In the case of a shunt reg why don’t “reverse” the BJT current source polarity, I mean, usually we use the fig1 design, in this case the BJT CCS present a low AC (Zin) load to the 3 term reg… (not really interesting), and where we need a low Z is at the output, in the second case fig2, in AC mode, the 2 devices BJT-Mosfet are in // referred to ground, so the Zout is lower than in the first case and the line regulation is also better.
Curiously it looks like a PP. audio amplifier output stage… (fig3).
Curiously it looks like a PP. audio amplifier output stage… (fig3).
An externally hosted image should be here but it was not working when we last tested it.
- Status
- Not open for further replies.