Are you really fine with IC voltage regulators ?

Jump to Latest
M

MarcelvdG

Member
Joined 2003
jean-paul said:
Marcel, I recall that there is a difference between in which brand produced/produces LM723. Some are nothing special and some are very quiet.

Same goes for TL431 now often produced by strange brand names in the Far East.... Also for 78xx/LM317 that are made by various manufacturers. Something one will not see when using LT3045 for instance. So it is IMO important to mention brand and exact typename and maybe even a datecode.

Example: if I would use the list I would assume LM723 to have 86 µV of noise. In practice I have found versions that are always below 5 µV. I am not alone in this.

https://dg4rbf.lima-city.de/Rauschmessungen am LM723.pdf
Click to expand...

It could very well be that there are large differences between brands, but the data in that link are with low-pass filtering between the reference and the regulator. My list applies to the unfiltered references.

Partially updated list:

85A2 (glow discharge): 648.8 nV/sqrt(Hz) and 85 V, so 7.633E-9 Vref/sqrt(Hz)
National Semiconductor/Texas Instruments LM723 (Zener / avalanche diode), June 1999/April 2013 datasheet: 86 uV from 10 Hz to 10 kHz at 5 V output voltage, so 172.1E-9 Vref/sqrt(Hz)
LM317 (bandgap): 0.003 % of Vout from 10 Hz to 10 kHz, so 300.2E-9 Vref/sqrt(Hz)
uA7805 (bandgap): 40 uV from 10 Hz to 100 kHz at 5 V, so 25.3E-9 Vref/sqrt(Hz)
LT1236 (buried Zener reference): 2.2 uV from 10 Hz to 1 kHz at 5 V, so 13.98E-9 Vref/sqrt(Hz)
LT3081 (unknown, presumably bandgap): 5.7 nA from 10 Hz to 100 kHz at 50 uA, so 360.5E-9 Iref/sqrt(Hz)
LT3042 (unknown, presumably bandgap): 6 nA from 10 Hz to 100 kHz at 100 uA, so 189.7E-9 Iref/sqrt(Hz)
Diodes Incorporated TL431 (bandgap), June 2021 datasheet, not recommended for new designs: 230 nV/sqrt(Hz) at 2.495 V, so 92.18E-9 Vref/sqrt(Hz)
OnSemi TL431 (bandgap), August 2021 datasheet: 48 nV/sqrt(Hz) at 2.495 V, so 19.24E-9 Vref/sqrt(Hz)

1960's style bandgap reference (bandgap): 29.784 nV/sqrt(Hz) at 4.96 V, so 6.005E-9 Vref/sqrt(Hz)
 
Last edited:
C

chris719

Member
Joined 2004
gerhard said:
What a crock. Nobody forces you to use piezo-electric ceramic capacitors with LT304*. you can use
just as well WIMA 4u7 foil capacitors or/and NP0. And X7R do not go piezo unless you hit them
with a hammer. It's not that I haven't tested it in that alu-cast Hammond box. And I can measure
below -205 dBV. Yes, you can find worse for that straw man argument. China delivers.



First, try to verify better PSSR than LT304*.
I would be ashamed if I had to admit that I need more than 60 dB
because I have a cesspit on my board




At the half of the clock frequency, the regulator has absolutely no influence on the DC quality.
The only thing important are the output capacitors to enforce low Zsource. You can shove
your pseudo arguments. Sorry, that is the friendliest it can be formulated.

Gerhard
Click to expand...

Well said. I doubt most people have layouts that can even achieve or utilize the datasheet performance of LT304x. I also have never experienced a real-world problem with correctly sized and specified X7R caps.
 
B

bohrok2610

Member
Joined 2005
dreamth said:
Then why a precision refference and an op amp buffer with very high psrr and cmrr(if low impedance is needed) wouldn't be ideal for that?
Click to expand...
I'm using this approach in my ES9038Q2M DAC AVCC with LT3042 providing the reference followed by an op amp buffer. However the reason for this has nothing to do with how it sounds in subjective sighted listening tests. While LT30xx is the ultimate hammer it does not have ultra-low output impedance.
 
  • Like
Reactions: 1 user
jan.didden

jan.didden

AX tech editor
Joined 2002
linearaudio.net
Paid Member
ginetto61 said:
Hi ! looking for some information about dual voltage power supplies for preamps in an interview with the Conrad Johnson designer he states that they prefer to realize voltage regulators with discrete parts and not using single chip voltage regulators. In their opinion they can have an impact on the sound ... and not for the good.
Looking then at some preamps schematic i see that in the late '70s and early '80s many of the TOTL solid state preamps actually used voltage regulators made out of discrete parts.
I guess the reason was that this IC regulators were not available ? :rolleyes:
They are clearly handy ... but are they also completely fine to use in preamps ?
i would like to get some opinions ... i have a feeling that i was looking at the wrong part of the story ... the preamp circuit ... instead the secret of good sound could lie in the power supply design and construction :confused:
I am attaching one schematic just to explain what i have in mind ... more or less
Click to expand...
This is about the worst implementation of a pair of regulators I've ever seen, what with the common return resistors. If they say it 'sounds better' than an IC regulator, I would totally disregard everything they say.
My € 0.02.

Jan
 
M

Markw4

Member
Joined 2016
Measuring to -200dB is not necessarily measuring what is affecting the sound. If you have instruments that measure certain things, then every problem is a nail to be hit by those measurement hammers. When people are too stubborn to do the experiments and listen, then they will continue with their dogma and never learn they haven't' been measuring something or other very well, whatever that something may be. Probably its some dynamic in the time domain that doesn't show up well on a scope in real time, nor on an FFT expecting only steady state signals.

ESS has a term for such things, "Non-PSS." Starting around page 30 of the slides at: https://www.yumpu.com/en/document/read/23182504/noise-shaping-sigma-delta-dacs-ess-technology-inc My personal opinion is they are at least trying to think outside the box a little to understand problems, rather than sticking to the standard measurement dogma box.
 
M

MarcelvdG

Member
Joined 2003
bohrok2610 said:
I'm using this approach in my ES9038Q2M DAC AVCC with LT3042 providing the reference followed by an op amp buffer. However the reason for this has nothing to do with how it sounds in subjective sighted listening tests. While LT30xx is the ultimate hammer it does not have ultra-low output impedance.
Click to expand...

Why did you choose an LT3042 then? The big advantage of an LT3042 is that you can easily put a low-pass filter between its reference and its output buffer, but with an op-amp buffer, you could insert a low-pass between any other type of voltage reference and the op-amp buffer. On itself, the reference of the LT3042 is not particularly low in noise.
 
jean-paul

jean-paul

Disabled Account
Joined 2002
Well I have a bunch of LM723.... :D Just teasing, I heard that one too much from people using only known old parts they bought in bulk 20 years ago. Did you also try out TPS7A4700 and what is your opinion?

And do you have any other LDO regulator to recommend?
 
Last edited:
B

bohrok2610

Member
Joined 2005
I have only one LM723, NOS from the early 80s :)
I have not tried TPS7A47xx for AVCC. But I have a tip for TPS7A4701 with feedback resistors: for lowest noise add 10nF Cff in the same fashion as for TPS7A33. The TPS7A47 datasheet has no mention of this but it is needed to achieve datasheet noise performance. I hate the package though.
 
EUVL

EUVL

Member
Joined 2003
xen-audio.com
MarcelvdG said:
The big advantage of an LT3042 is that you can easily put a low-pass filter between its reference and its output buffer,
but with an op-amp buffer, you could insert a low-pass between any other type of voltage reference and the op-amp buffer.
Click to expand...

So can you suggest a suitable, commercially available opamp buffer with better performance (noise, bandwidth, stability with capacitive load, etc.) ?
I presume you have tested such a solution ?


Cheers,
Patrick
 
  • Like
Reactions: 1 user
You must log in or register to reply here.
Top Bottom