now that i have discovered (thank to you again) how good can be a basic combination cap multiplier + LM317 i am personally done
Moreover just recently i have even discovered a new generation darlington with some 5000 of hfe
the 2sd2390 
The king of cap multipliers
Actually, I am upgrading some older equipment with newer chips. Much easier and cheaper than trying to beat a 30 year old design into submission. I am also building a few new things and for sure am not going to use antique parts that don't meet the basic requirements. I am also looking to improve a couple new units I bought where they used an older cheap regulator and the performance is now what I want. So back to the original topic, YES, I am satisfied with IC regulators. Just select the one that is appropriate for your design.
the order of -60
Speaking of cap multipliers, you might like to have a look at this:
The K Multiplier
There is also a thread devoted to this mutiplier on this forum:
Keantoken's CFP cap multiplier
Noise rejection is of the order of 60dB.
Regards,
Braca
Speaking of cap multipliers, you might like to have a look at this:
The K Multiplier
There is also a thread devoted to this mutiplier on this forum:
Keantoken's CFP cap multiplier
Noise rejection is of the order of 60dB.
Regards,
Braca
As far as I remember, so far no-one has specified what the design goals are in the first place.
Of cause, there are pros and cons of both those options, but I'm not sure they are critical for audio. So there can exist both.
For example, if there is no need for low or ultra-low output impedance then the k-multiplier can be after some 'noisy' regulator. And vice-verca.
Good morning ! thank you very much again. I am completely sold on using "also" a IC regulation stage. After all these very kind and valuable advice i am also sold on using a cap multiplier stage ... i just do not know which one should go first after the rectification ... i mean cap multiplier > Ic regulator or viceversa ? IC are all very good at low Hz ... even the old but good half dollar LM317 that is what i am already using by the way.
Given that the idea is too start with a higher voltage smps i think that it would be better to put the cap multiplier first.
I have found a darlington 2sd2390 with a hfe of 5000 ... that would be a nice piece to start for building a cap multiplier.
Stubbie
I wish you enjoy what you are doing. This is the important part
When you upgrade a PSU, try to take some measurements before and after at the output of the PSU and at the output of the audio circuit. See what you get.
George
Hello
thank you so much again for let me know so many important aspects of power supplies design. And also that Spice sw ... it is marvelous. I will be trying some simulations for sure ... even if i am pretty sure that the circuit you simulated will work just fantastic. I think i will place the cap multiplier just after the rectification stage ... also because my intention is to start with a higher dc voltage smps
That cap multi will take care of any high Hz noise generated by the smps
After the CM a very basic IC regulator
Thanks a lot again
Have a nice day
Hi ! thanks a lot indeed. But i would get the most out of very basic circuits as a start ... i am just a beginner
The performance of a very basic IC based regulator are very ok for me
I understand that adding a cap multiplier will improve psrr at higher Hz inside the audio range.
A 70dB psrr wideband is the target.
Hi ! noisy regulator ?
my feeling is that IC regulators should not be that noisy. Maybe they have not excellent psrr wideband ... but they do not have a self-noise
However i have checked in the datasheet of the LM317 and found nothing about chip self-noise ... like EIN or so ... no spec
A very good test would be to check the noise spectrum of a battery and the spectrum after a regulator ... and compare the two spectra maybe ?
AC is noisy ... switching diodes are noisy ... maybe i am wrong
Last edited:
Hi ! thank you so much for the very important advice I did not see a figure looking for noise
I have found something about LM317 self-noise and how to reduce it ... i will study more
What is precisely the parameter i have to look at in the datasheet ?
another alternative is the LT108x family regulators ... are they better noise wise ? (not that 250uV of noise scare me ... i come from smps with hundreds of mV of ripple
)Another interesting thread here
What Linear Regulator Chips Have the Lowest Noise Figures?
Last edited:
The parameter is called "noise" and the rule is that when it is not mentioned in a regulators datasheet it is very high It should scare you as it is one of the more important parameters to obtain good audio. Using PSU's with hundreds of mV of ripple or noise means that it is not good audio. A device is as good as its PSU
In short: almost all old regulators that have been designed decades ago have high noise numbers except the rare few like LM723. Many newer LDO designs are way lower in noise and a few are truely exceptional like LT3045. Unfortunately the best among the few extremely good ones have limiting features like only 0.5A, max. 20V input etc. This is because modern design makes use of locally placed regulators in each section of a device. In older designs there was one PSU feeding the whole device. So: designing a 5V 1A PSU with ultra low noise is relatively easy, designing a 15V 3A ultra low noise LDO PSU is harder.
There are "tricks" to improve existing designs with older regulators but for new designs one better chooses better regs. Simple. The only reasons people still use LM317 for instance is that they "know them", "already have them", " they are cheap", "they are through hole" etc. All reasons that make it strange to combine these with ultra modern DAC chips .....
When designing a new device it is better to adapt to the newer way of placing local regulators at each section. Then the modern regulators have exactly the right features.
It should scare you as it is one of the more important parameters to obtain good audio. Using PSU's with hundreds of mV of ripple or noise means that it is not good audio. A device is as good as its PSU In short: almost all old regulators that have been designed decades ago have high noise numbers except the rare few like LM723. Many newer LDO designs are way lower in noise and a few are truely exceptional like LT3045. Unfortunately the best among the few extremely good ones have limiting features like only 0.5A, max. 20V input etc. This is because modern design makes use of locally placed regulators in each section of a device. In older designs there was one PSU feeding the whole device. So: designing a 5V 1A PSU with ultra low noise is relatively easy, designing a 15V 3A ultra low noise LDO PSU is harder.
There are "tricks" to improve existing designs with older regulators but for new designs one better chooses better regs. Simple. The only reasons people still use LM317 for instance is that they "know them", "already have them", " they are cheap", "they are through hole" etc. All reasons that make it strange to combine these with ultra modern DAC chips .....
When designing a new device it is better to adapt to the newer way of placing local regulators at each section. Then the modern regulators have exactly the right features.
Last edited:
Thank you so much again. Maybe they think that anyone is using them with opamps with 100dB psrr wide band ...The parameter is called "noise" and the rule is that when it is not mentioned in a regulators datasheet it is very high
not my case. But i cannot explain why ... Yes indeed. Actually i have become intrigued by psu after some listening tests ... the difference between smps and linear one is quite evident at the listening with the linear (LM317) being quite better.It should scare you as it is one of the more important parameters to obtain good audio. Using PSU's with hundreds of mV of ripple or noise means that it is not good audio. A device is as good as its PSU
But the biggest discover has been the cap multiplier circuit. The shunt de-noiser instead i do not like it much ... i do not know why.
thank you very much for these valuable info. The 20V max is more limiting than the current. I am only interested in psu to power line preamps.
As i am only interested in line preamps at all.
I think that a solution can arrive by more careful filtering ? again the cap multiplier build around a very low active device can be a god's gift and placed also after the regulator stage and get rid of its noise.
Changing my idea i could put only one low impedance high psrr cap multiplier after the IC regulation stage
Another very good thing is that i could use also a lot of capacitance at the cap multi output without problem ... like 2200uF bypassed with smaller caps
But i have to check
For now i will dedicate myself to understand better the cap multiplier circuits alaso through the simulation of some circuits
Another problem is to see the spectrum of the noise ... how it is distributed in frequency.
Thanks a lot again
Hi ! noisy regulator ?
my feeling is that IC regulators should not be that noisy. Maybe they have not excellent psrr wideband ... but they do not have a self-noise
However i have checked in the datasheet of the LM317 and found nothing about chip self-noise ... g
Every part has noise, except absolute cold ones ( about absolute zero temperature).
If it is too 'noisy' or not depends on exact circuit requirements (but not on absolute noise level).
LM317 and it's sisters have RMS output noise voltage of 0.003% Vo (output-dependent). It can be less than 0.1 mV RMS at 12 VDC (with 5 MHz range).
Last edited: