Who said I want to equal the parameters? They sound bad for some applications. Understood its a can of worms to get into how regulators sound, but this is audio. Its to create an illusion for human entertainment. We don't have the technology to reproduce sound pressure at the ears of a listener perfectly like a live performance. What we can do is create a mental experience in the heads of many listeners that they like and enjoy. OTOH if we are designing instrumentation to meet measured performance specs, then that's another matter.
Difficult as I don't use ESS DACs at all. I only liked ES9023 for a combination of reasons. Recently I had a Didden/Jung regulator that did not perform OK. Took too much time to find out why so it ended in recycling. So that would mean I would need to buy an ESS DAC AND a Didden/Jung regulator. Even better I would need to design a new ESS DAC with Didden/Jung and LT regs integrated. Pfff that costs an awful lot of time 
As Jean-Paul said layout is very important. If this picture is a representation of how you evaluate regulators it is quite clear that your results cannot be used even as anecdotal evidence.
https://www.diyaudio.com/community/attachments/th_lt1763-jpg.710267/
* picture was part of this post https://www.diyaudio.com/community/threads/es9038q2m-board.314935/post-5577671
Other people in the forum have already done the experiment. LT304x always loses. Well implemented opamp buffers or Jungs sound better to most people. tps7a4700 sounds better than LT304x too.
EDIT: Some reading material on dead bugs: https://spectrum.ieee.org/with-the-...s-can-break-through-the-highfrequency-barrier
EDIT: Some reading material on dead bugs: https://spectrum.ieee.org/with-the-...s-can-break-through-the-highfrequency-barrier
Last edited:
The LT304x are sensitive to pcb layout for best performance as the AN points out and as Ive found. That could be one reason why they might be a sonic difference. I dont understand
why they dont apparently work well with the ESS if its not that but ok. It should be perfectly possible to see why if its affecting the AVCC. I was only pointing out that there other Linear regs out
there and I used the LT 304x as an example. There are others.
Just seen replies. Beat me to it.
why they dont apparently work well with the ESS if its not that but ok. It should be perfectly possible to see why if its affecting the AVCC. I was only pointing out that there other Linear regs out
there and I used the LT 304x as an example. There are others.
Just seen replies. Beat me to it.
LT304x can sound very good for certain audio applications. Not for AVCC.
One problem with regs ike LT304x for AVCC is that they typically need C-II ceramic output caps for stability. Such caps are highly nonlinear and piezoelectric. IIRC KSTR reported AVCC draws current pulses at half the clock frequency. ESS dacs are probably most often clocked at 45/49MHz. How does LT304x with a C-II output cap look to the load at 22/24MHz? Also, its further complicated that AVCC has zero PSRR. KSTR also reported that high impedance at half the clock frequency increases dac audio output distortion. What is exactly going on with LT304x and AVCC that sounds bad? So far I haven't seen any good, complete explanation of that but I and others have tried LT304x for the obvious reason that it looks good on its datasheet. Other regulators were tried too. LT304x did not get the design win. Sorry.
One problem with regs ike LT304x for AVCC is that they typically need C-II ceramic output caps for stability. Such caps are highly nonlinear and piezoelectric. IIRC KSTR reported AVCC draws current pulses at half the clock frequency. ESS dacs are probably most often clocked at 45/49MHz. How does LT304x with a C-II output cap look to the load at 22/24MHz? Also, its further complicated that AVCC has zero PSRR. KSTR also reported that high impedance at half the clock frequency increases dac audio output distortion. What is exactly going on with LT304x and AVCC that sounds bad? So far I haven't seen any good, complete explanation of that but I and others have tried LT304x for the obvious reason that it looks good on its datasheet. Other regulators were tried too. LT304x did not get the design win. Sorry.
Last edited:
Look, its true that people can make up stories. Its also true they could post fake measurements. That's why I suggest that people do their own experiments and form their own opinions. Don't assume a datasheet tells you everything you might need to know. Art of Electronics shows some measurements of op amps that the manufacturers apparently didn't want you to see. Manufacturers are in the business of selling chips. Datasheets are intended to influence engineer buying decisions, not simply as neutral and complete technical information. Don't expect manufacturers to tell you why you shouldn't buy their products, that's not their in their interests. What they will sometimes admit is that they fixed some old problem they didn't tell you about before, so now they are telling you so they can use the fix as a new selling point.
I have quite a lot of experience with LT30xx regulators and I am quite sure that the datasheet does not contain false information. And I have yet to see a properly implemented LT30xx regulator by the same people that make the claims regarding LT30xx's inferior sound with AVCC.
Well when you can't make up your mind, you can also use both types and go overboard. Because ..you will always eventually wonder what if I had .........Difficult as I don't use ESS DACs at all. I only liked ES9023 for a combination of reasons. Recently I had a Didden/Jung regulator that did not perform OK. Took too much time to find out why so it ended in recycling. So that would mean I would need to buy an ESS DAC AND a Didden/Jung regulator. Even better I would need to design a new ESS DAC with Didden/Jung and LT regs integrated. Pfff that costs an awful lot of time
At the end of the road, you now can say been there done that.
Attachments
This is where you've made a wrong turn. The correct term is audio reproduction.
If a person X does an experiment one way and a person Y does it another way, there is a mismatch of experiences thus they will be talking in different languages. To form a common language, the experiment method needs to be brought to light. Please share the details of the experiment setup that you are referring to so that people can try the same and find out themselves.
Yes. Sure. A purpose built integrated regulator.
Designed to solve the problems of a 1000 engineers.
50 professionals see more than a redneck claims
to hear in his man cave, built to amplify his illusions.
Gerhard
Last edited:
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
I wonder why most talks on ic regulators lately revolve around lt304x no matter the application.
Now it seems that people try to solve every regulation problem with one chip that seem to be made for very specific applications no matter the power or voltages involved or at least that is what i read and you probably know that when you have a hammer in your hand every problem looks like a nail...
I found usual lm78xx regulators working fine in most audio applications and they aren't sensitive to how high the output capacitors are or if they are too low esr but you need to settle with their wider voltage drop . They also allow for a wider rage of powers and voltages being true chips for lower frequency domain where you want quick results, but i also know that modern dacs need more than just one 5v rail and there's a whole range of ADP ultra low noise 5 pin regulators made specifically for low voltage rails for digital apps.I saw them used with ess dacs and I've also seen the early ess9018 measured on AP top of the range equipment while suplied by analog devices ultra low noise chips so I know there's real competition outhere.If you're supplying a modern digital setup you should know that you need multiple regulators anyway...
An old friend of mine told me once that in the end most of the jobs of a regulator is actually done by the capacitors, inductances and resistors around it and i think he's right.
Now , because people talk about ESS dacs in particular i wonder what regulation ESS engineers used when they made their final measurements and how sensitive are those dac's to their power line regulation.Well actually i don't need to wonder about that once i signed an NDA for their way too misterious datasheet to be released to me at the time.It should be common wisdom by now as their datasheet made their way to the general public one way or another.
As i read here a lot about engineer's perspective, do ess engineers recommend some regulators over another? What do you think?
A while ago i launched a thread about NEVE's RNHP depicting it as a final provocation to the whole audio engineering world.
As it was expected my remarks quickly became a new source of fun when mentioning that the guy who tuned a now quite famous brand of audiophile headphones told me that he was impressed by RNHP and asked me to study it for a bit.
A year or so after that a technical review showed that its distortion measurements were simply sublime while the thing used cheap smps regulation and the simplest form of virtual ground BUT what Neve did first and foremost was to implement actually a lot of the tpa6120 datasheet in the real world.I wonder why no other manufaturer commercial or diy didn't consider implementing the recommended tpa6120 datasheet . We might have found that no fine audiophile ultrafast regulator was needed as well with a 1200v/us slew rate chip...
Honestly i filter a lot of audioholic misconception about supplying circuits lately removing tons of regulations all together where I can and lm78xx or 78lxx chips still serve me well in a lot of circuits so i'd say i'm fine with chips in general...
Now it seems that people try to solve every regulation problem with one chip that seem to be made for very specific applications no matter the power or voltages involved or at least that is what i read and you probably know that when you have a hammer in your hand every problem looks like a nail...
I found usual lm78xx regulators working fine in most audio applications and they aren't sensitive to how high the output capacitors are or if they are too low esr but you need to settle with their wider voltage drop . They also allow for a wider rage of powers and voltages being true chips for lower frequency domain where you want quick results, but i also know that modern dacs need more than just one 5v rail and there's a whole range of ADP ultra low noise 5 pin regulators made specifically for low voltage rails for digital apps.I saw them used with ess dacs and I've also seen the early ess9018 measured on AP top of the range equipment while suplied by analog devices ultra low noise chips so I know there's real competition outhere.If you're supplying a modern digital setup you should know that you need multiple regulators anyway...
An old friend of mine told me once that in the end most of the jobs of a regulator is actually done by the capacitors, inductances and resistors around it and i think he's right.
Now , because people talk about ESS dacs in particular i wonder what regulation ESS engineers used when they made their final measurements and how sensitive are those dac's to their power line regulation.Well actually i don't need to wonder about that once i signed an NDA for their way too misterious datasheet to be released to me at the time.It should be common wisdom by now as their datasheet made their way to the general public one way or another.
As i read here a lot about engineer's perspective, do ess engineers recommend some regulators over another? What do you think?
A while ago i launched a thread about NEVE's RNHP depicting it as a final provocation to the whole audio engineering world.
As it was expected my remarks quickly became a new source of fun when mentioning that the guy who tuned a now quite famous brand of audiophile headphones told me that he was impressed by RNHP and asked me to study it for a bit.
A year or so after that a technical review showed that its distortion measurements were simply sublime while the thing used cheap smps regulation and the simplest form of virtual ground BUT what Neve did first and foremost was to implement actually a lot of the tpa6120 datasheet in the real world.I wonder why no other manufaturer commercial or diy didn't consider implementing the recommended tpa6120 datasheet . We might have found that no fine audiophile ultrafast regulator was needed as well with a 1200v/us slew rate chip...
Honestly i filter a lot of audioholic misconception about supplying circuits lately removing tons of regulations all together where I can and lm78xx or 78lxx chips still serve me well in a lot of circuits so i'd say i'm fine with chips in general...