Maybe elvee's denoizator for the lm317 could be interesting for you:
https://www.diyaudio.com/community/...-retrofit-upgrade-any-317-based-v-reg.331491/
https://www.diyaudio.com/community/...-retrofit-upgrade-any-317-based-v-reg.331491/
very interesting indeed I always had a feeling about this but without real basis
Anyway as to have dual voltage rails could be needed i have found this schematic attached that i like very much There is also filtering with bjts
( I dont understand well how to calculate the VDC out (here 28VDC) from VAC in (here 12VAC) But i will figure it out )
Any other usable arrangement to get dual rails from just one secondary ? thanks again
Attachments
Last edited:
Yes i guess so I have difficulty in understanding how things work However as i said i have always had this idea that using just one winding is good
Two secondaries will never be the same Different impedance for start ... also the phases i dont know I am just imaging
One secondary that is ... i much prefer this solution Then some very good and strong single diodes and that will be a good start
If i understand well to get a dual voltage suitable for +/-24VDC i should start from 12VAC
12VAC is a pretty common output for a mains transformer
I have found the case with a diode bridge even easier There will be a good filtering downstream
Two secondaries will never be the same Different impedance for start ... also the phases i dont know I am just imaging
One secondary that is ... i much prefer this solution Then some very good and strong single diodes and that will be a good start
If i understand well to get a dual voltage suitable for +/-24VDC i should start from 12VAC
12VAC is a pretty common output for a mains transformer
I have found the case with a diode bridge even easier There will be a good filtering downstream
Attachments
I would use a 2 x 9 volt or 18 volt transformer to avoid unnecessary reduction of the voltage via an active regulator.
In the case of a voltage for a preamplifier, passive reduction using a voltage divider can work. Or even with a fat Zener diode. Just try out what sounds best.
Or simply "open" the transformer and rewind the windings as required...-)
If the preamp is a PP concept: I would look for SE. There, too, two-voltage circuits are feasible, the usual "differential amplifiers". They sound better than PP concepts, and far better than complementary-transistor-pp concepts. Aside;-)
In the case of a voltage for a preamplifier, passive reduction using a voltage divider can work. Or even with a fat Zener diode. Just try out what sounds best.
Or simply "open" the transformer and rewind the windings as required...-)
If the preamp is a PP concept: I would look for SE. There, too, two-voltage circuits are feasible, the usual "differential amplifiers". They sound better than PP concepts, and far better than complementary-transistor-pp concepts. Aside;-)
i really like the idea of using just one secondary It makes sense to me I have to admit the most units use just a center tapped transformer and the two secondaries But i understand now that just one secondary could provide all that is needed
regulators are needed to keep the voltage constant with electrical grid voltage variations
maybe with varying currents they can be an issue but low level units are working in class A with constant current draw
I see this a little tricky for me
even more tricky
you mean single supply ? i dont understand
i have an old schematic that it could provide great performance wired as buffer or buffer with little gain ... are just two CFP made with complementary bjts
It was a TOTL preamp from Onkyo in the mid '70s .... i like its simplicity I will simulate it in the next days I think it could work also with reduced voltage
Attachments
Last edited:
hi thanks for the valuable info I will keep that in mind
I have noticed that some high end manufacturers offer for their best preamps the solution with a separated power supply A two boxes solution with the mains transformers separated from the audio circuitry
I really do not know if there is a good technical reason or it is just marketing strategy to charge more .... and they do charge a lot more
I have not still understood if in the case of using a power umbilical cord is better an AC cord solution or a DC cord solution
In the 1st case the separate box will contain only the mains transformer, some mains filtering and a fuse
Famous Tech Enzo always said (about people obsessing about minute amp details): "the amplifier is just that thingie between the power supply and the speaker"
With only one secondary winding, cleanliness and quietness finally come into the sound picture;-)i really like the idea of using just one secondary It makes sense to me I have to admit the most units use just a center tapped transformer and the two secondaries But i understand now that just one secondary could provide all that is needed
regulators are needed to keep the voltage constant with electrical grid voltage variations
maybe with varying currents they can be an issue but low level units are working in class A with constant current draw
I see this a little tricky for me
even more tricky
you mean single supply ? i dont understand
i have an old schematic that it could provide great performance wired as buffer or buffer with little gain ... are just two CFP made with complementary bjts
It was a TOTL preamp from Onkyo in the mid '70s .... i like its simplicity I will simulate it in the next days I think it could work also with reduced voltage
Regulators are not required. This is an unfounded fear that a voltage drop is also immediately audible. But I would also build regulators and listen to them. After all, it's about overall tuning;-)
Aside: If you test 10 different transistors, only one of them is suitable for audio. So the only option is to buy many different ones and test them. And: buy the right quantity, as another batch will already sound different;-)
The fact that high-end preamps, or even every conceivable stage, are equipped with their own power supply units is simply inexperience: they have never listened, or don't even know anything about speakers;-) Why? First and foremost, the ear perceives differences. And since different power supply units modulate the signal differently, we add up these audible, sound-degrading modulations. Anyone can try it out.
Different, separate psus have nothing to do with HighEnd, with AUDIO;-)
... the same story as with the single secondary winding;-)
Ergo: ONE psu for ALL - if possible;-)
Push-Pull: PP
Single Ended: SE
Your picture shows a complementary-transistors-pp-concept. (Important to distinguish from "classic" pp, which is a pp pairing with identical transistors (PNP/PNP or NPN/NPN). These sound less different than pairs of complementary transistors (PNP/NPN))
Your: One NPN and one PNP transistor per half-wave (+ and -). If you use two identical complementary pairs, both characters - these complementary transistors sound significantly different than identical transistors - can be found in each half-wave. This is more homogeneous.
well maybe being ossessive about details is evilish but the question is ... given a same power supply and two different circuits that have a same gain the resulting sound will be different ? the answer is of course Distortion comes from the circuit and even PSRR can be different for different design
If the PSRR is very high the difference between power supplies can be reduced ?
i would say that circuit is more important Power supplies have to be just decent
@Extreme_Boky I always felt that the psu should not contain any active devices in series with the supply rail to the amps output stages.
It should also be generously specified to be able to supply the loudest transients without drooping.
I have never used a capacitance multiplie which I regard as a cost saving 'cheat'.
An SMPS in an AB amp to save space and money, a regulated psu pretty much another cost saving technique, where the impedance and transient response of the regulator means it is not transparent.
Having said that, I have been listening to a pair of SMPS* driving a Class A amp and I'm in the process of replacing the Krell lcones with JLH ESL high power vesion. (* claimed audio grade versions)
It should also be generously specified to be able to supply the loudest transients without drooping.
I have never used a capacitance multiplie which I regard as a cost saving 'cheat'.
An SMPS in an AB amp to save space and money, a regulated psu pretty much another cost saving technique, where the impedance and transient response of the regulator means it is not transparent.
Having said that, I have been listening to a pair of SMPS* driving a Class A amp and I'm in the process of replacing the Krell lcones with JLH ESL high power vesion. (* claimed audio grade versions)
Like this? (from more than twenty years ago) https://sound-au.com/tcaas/jlhesl.htm
A tip about the JLH:
You can run it as a 3-stage, 2-stage and 1-stage SE amp, but also as an 1-stage SE follower. Simply connect the input to the respective transistor via a capacitor. Just try it out.
This way you can also hear the influence of stages/components: 3-stage: disco sound (a lot of detail is lost, it gets dark, fat, bumpy), 2-stage: party (cleaner, smoother, more detailed, more colorful), 1-stage (high resolution, very fine colors and fine contours) and 1-stage follower (usually even cleaner, milder, softer, even purer than the 1-stage amplifier).
I would remove the bootstrap capacitor. It brings a lot of uncleanliness and noise. If you want to get more flavor, simply connect a resistor (first 50 or 100 ohms) from the output to the base of the follower, i.e. the upper one, usually Q2.
You can run it as a 3-stage, 2-stage and 1-stage SE amp, but also as an 1-stage SE follower. Simply connect the input to the respective transistor via a capacitor. Just try it out.
This way you can also hear the influence of stages/components: 3-stage: disco sound (a lot of detail is lost, it gets dark, fat, bumpy), 2-stage: party (cleaner, smoother, more detailed, more colorful), 1-stage (high resolution, very fine colors and fine contours) and 1-stage follower (usually even cleaner, milder, softer, even purer than the 1-stage amplifier).
I would remove the bootstrap capacitor. It brings a lot of uncleanliness and noise. If you want to get more flavor, simply connect a resistor (first 50 or 100 ohms) from the output to the base of the follower, i.e. the upper one, usually Q2.
Hi thank you for the very interesting advice However i guess that the power supply requirements for preamps and power amps could be different
Most preamps draw a constant current so the impedance could not be a big issue An many preamps do use cap multipliers in their power supplies
Not that i do not like the idea of passive filtering before the regulation stages It could be very beneficial fwiu
That graph is fairly useless for predicting the ripple rejection at an Amp or so, with constant Vbe, not Ib
LTSpice is also inaccurate for this sort of modelling.
Add to that, a Darlington is more temperature sensitive than a discrete pair as both transistors are at the same die temperature, so more thermal modulation.