As a secondary regulator, I don't think PSRR is important because it is fed a DC voltage with very little ripple. Possible high frequency noise is already filtered by the CLRC filter in the LCDPS
I believe the choice was made because it has low inherent noise. 20uV is probably the lowest noise you can find for a fixed voltage regulator.
The real question is: Does 20 uV matters when powering electronics and designs with high PSRR? Maybe the noise that can pass through is below the noise floor of the DAC/output stage...
I believe the choice was made because it has low inherent noise. 20uV is probably the lowest noise you can find for a fixed voltage regulator.
The real question is: Does 20 uV matters when powering electronics and designs with high PSRR? Maybe the noise that can pass through is below the noise floor of the DAC/output stage...
jkeny said:
I think those figures are a bit high. I would say from different sources in the literature that the noise level of LM317 is in the 50-100 uV range. Lets say 100 uV. the secondary regulators have 50 dB PSRR, Which is an attenuation of 25X, so 100uV/25 yields 4 uV. Lets say the noise is 250 uV. Only 10 uV passes through.
In almost all cases, tweaking the prmary power supply makes no difference (in theory) because the predominant noise comes from the local or secondary regulator.
Bur removing the secondary regulator and replacing it with a lower noise design, may or may not yield any benefit because:
1- The noise that goes through (opamps have PSRR in the 100 dB) may be below the noise floor of the design
2- Bulkier design may pick up noise
3- Output impedance may be an issue
4- Other stuff...
glt said:
50 dB attenuation is ~300X (dB=20 log(X) ?), so the noise that can go through is just 1 uV...
So, in theory, lets say the local regulator feeds some opamp and it has a PSRR of 60 dB (that's 1000X attenuation). The resultant noise level is in the .05 uV level. Now suppose the signal output is 2V. The noise level is 40,000,000X below. That is -152 dB...
That I think is the theory....
Russ, show me where I'm being unreasonable or unfriendly? I've simply stated questions I've asked & your response. I stated that you didn't want to answer a question above that I asked about the LT regs. Am I wrong? What's with the warning - are you a moderator now?
All I'm doing is trying to get the best out of this DAC in the spirit of DIY. If it seems to encroach on your "business hobby", apologies, but this is a DIY forum after all, not a shopping outlet ! This is a tweakers thread you know, not your vendor's thread.
Again, I'm sorry if I ask questions which seem to criticise your design decisions but by asking questions one learns & I'm no expert which I freely admit.
Edit: I know it must be annoying to have somebody question your design, after all, you were the first to bring this DAC to the DIY community but that was the first phase of a products life-cycle. Now it's perhaps moving into the next phase where a more critical view is taken? Perhaps it would be more sensible to fully engage with this?
All I'm doing is trying to get the best out of this DAC in the spirit of DIY. If it seems to encroach on your "business hobby", apologies, but this is a DIY forum after all, not a shopping outlet ! This is a tweakers thread you know, not your vendor's thread.
Again, I'm sorry if I ask questions which seem to criticise your design decisions but by asking questions one learns & I'm no expert which I freely admit.
Edit: I know it must be annoying to have somebody question your design, after all, you were the first to bring this DAC to the DIY community but that was the first phase of a products life-cycle. Now it's perhaps moving into the next phase where a more critical view is taken? Perhaps it would be more sensible to fully engage with this?
jkeny said:
JKeny,
Listen fella. I think you are probably an alright guy. But I read a post of yours on the tired old washed up engineer-ish forum (DIY-whine-fi) where you referred to my good friend Brian and I as "Twisted Sister". I am sure you meant that affectionately. Now to be honest, I just laughed. But just don't expect much help from a bloke after something like that because I can't take you very seriously.
I think you can probably see by your recent treatment over there that that place is only for wasting time. They have no more interest in really helping you understand things then they have tact and grace.
Now for best results when talking to folks who work hard to design something just keep in mind that courtesy demands not cross posting to try to fish for an answer you might want to hear. A respectful tone and an open mind will take you a lot farther.
Now I can get past something like what you said pretty easily. But lets keep it civil and we can make a lot more progress.
You can get a lot more information out of me than you will ever get from those cry babies at the dung heap.
Cheers!
Russ
OK Russ, the name was a joke! I can't be ars**d with all these factions in audio - you all know you read one another's forums so why the pretence of separate forums?
I couldn't care less about the egos involved or all the muck slinging & bad feeling between both - I'm simply trying to advance my understanding in this hobby & if that entails cross posting, I don't see the problem.
As I said elsewhere (another cross post) I recognise that there are lots of smarter people with deeper knowledge & experience than me and if I had to do all the tests myself to verify everything I said I wouldn't be going anywhere fast!
I have the ability to learn from others just as much as from my own experiments, in fact this is how I thought most fields of investigation progressed but maybe in audio there is some other tenet in operation that I'm not aware of? If you know what this tenet is please let me know?
I couldn't care less about the egos involved or all the muck slinging & bad feeling between both - I'm simply trying to advance my understanding in this hobby & if that entails cross posting, I don't see the problem.
As I said elsewhere (another cross post) I recognise that there are lots of smarter people with deeper knowledge & experience than me and if I had to do all the tests myself to verify everything I said I wouldn't be going anywhere fast!
I have the ability to learn from others just as much as from my own experiments, in fact this is how I thought most fields of investigation progressed but maybe in audio there is some other tenet in operation that I'm not aware of? If you know what this tenet is please let me know?
Jkeny,
I fully understand your point of view.
If I didn't think you were ok I wouldn't bother responding to you.
Here is my one word answer.
Courtesy.
Anything you do that looks or smells like you are joining muck slinging will make you perceived as part of the axis of evil.
I doubt you would have felt comfortable posting something like that here. Why should you do it there?
Cheers!
Russ
I fully understand your point of view.
If I didn't think you were ok I wouldn't bother responding to you.
Here is my one word answer.
Courtesy.
Anything you do that looks or smells like you are joining muck slinging will make you perceived as part of the axis of evil.
I doubt you would have felt comfortable posting something like that here. Why should you do it there?
Cheers!
Russ
jkeny said:
Russ, courtesy cuts both ways, you know & I found your quoted post above neither courteous nor helpful so maybe I was just reacting in kind!
Well it was actually meant to be kind.
I only did not give you all the details because it is a complex subject.
I will break it down for you as simply as I can right now.
Where those regulators are located (all digital ccts) the two most important factors are low noise and good load regulation out to the point where bypass caps and PCB parasitics will take over.
Also since it is intended to be pre-regulated it is even less important to have PSRR > 30db at 100khz.
It would not matter in the least.Now you may say back "but so and so tried such and such and heard a wild and dramatic difference". To that I say, good for them. I hope they are happy. But, I am not much into subjective anecdotes.
On the analog side I would say this. The LM317/LM337 are good enough to achieve a DNR not available in any other DIY DAC I know of. Now could shunt regs improve on this? They very well might. I have designed shunt regs and used them and find them to sound great. But I am going to be perfectly honest here. I nor any of the talented folks I know can hear the difference between 128db DNR and 130db DNR.
Now the perfectionist in me strives for 130db but I am perfectly pleased with 128db. 
What I don't like is hyperbole some "tweakers" tend to engage in. There are too many "such and such made an incredible difference" posts and not enough well reasoned thought.
You previous post to my short answer inferred that those LT regs were a poor design choice. Well if they are then its a poor design choice some of the best high end DAC designers on the market have made.
It is a pretty widely used regulator specifically because of its low noise. It was suggested to me by a name you would instantly recognize, but I will not drop it to protect their identity. Just do a little googling and you can find all you need.Do a little research and you will find it is a highly regarded IC regulator. I could not be happier with the way it is working in the role.
That better?
Cheers!
Russ
Reading your little discussion, I am wondering if you shouldn't rather talk about (frequency-dependend) impedance and slew-rates of the regulation, instead of talking about noise only.
Saving a couple of milliohms here and there do make a difference. Not to forget that it gets worse on higher frequencies.
Cheers
Saving a couple of milliohms here and there do make a difference. Not to forget that it gets worse on higher frequencies.
Cheers
Russ, I can handle the details, you know - no need to be "kind" - it feels patronising to me.
As soundchekk says, it may well be more to do with the impedance characteristics and transient characteristics of the regulator rather than just the noise! Without these measures then all that's left is subjective measurements!
In the final analysis that's what counts - which sounds better so I wouldn't dismiss these "field reports" - it may well improve your product
As soundchekk says, it may well be more to do with the impedance characteristics and transient characteristics of the regulator rather than just the noise! Without these measures then all that's left is subjective measurements!
In the final analysis that's what counts - which sounds better so I wouldn't dismiss these "field reports" - it may well improve your product
Jkeny, there was not any attempt to patronize. Its a matter of how much time I am willing to take to explain that which you could learn on your own. You not entitled to any explanations. I don't mind spending some time helping DIY folks, but lets face it, there is only so much to go around. Also keep in mind there are some details I may not want to reveal.
I always listen to feedback and then I apply what I learn like any sane person would do. I get field reports all the time. Many of them live and in person. I am lucky to live in Nashville and some of my friends (and relatives) are studio engineers and artists. This is wonderful for me because I have access to a lot of master recordings and skilled listeners.
Cheers!
Russ
I always listen to feedback and then I apply what I learn like any sane person would do.
I get field reports all the time. Many of them live and in person. I am lucky to live in Nashville and some of my friends (and relatives) are studio engineers and artists. This is wonderful for me because I have access to a lot of master recordings and skilled listeners.Cheers!
Russ
Buffalo option
Hello, I'm thinking to buy a buffalo DAC board to play with (maybe blow the chip as I'm used to... just kidding) but, for now, I have to choose between TP boards and NeoY2k's card. I've allready designed the DAC4392 board which I want to use it to feed my future ESS DAC. My question is: before buying the DAC card is there anyone who wants to test my '4392 with buffalo-32? (or NeoY2k's) and share impresions with me? I'll lend one/few board(s) to serve this purpose. I read somewhere on the net that ES9008 is pin-to-pin compatible with ES9018 so can be considered a drop-in replacement for ES9008?... Sorry for this newbies questions and for my english!
Thanx!
Hello, I'm thinking to buy a buffalo DAC board to play with (maybe blow the chip
as I'm used to... just kidding) but, for now, I have to choose between TP boards and NeoY2k's card. I've allready designed the DAC4392 board which I want to use it to feed my future ESS DAC. My question is: before buying the DAC card is there anyone who wants to test my '4392 with buffalo-32? (or NeoY2k's) and share impresions with me? I'll lend one/few board(s) to serve this purpose. I read somewhere on the net that ES9008 is pin-to-pin compatible with ES9018 so can be considered a drop-in replacement for ES9008?... Sorry for this newbies questions and for my english!Thanx!
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.