Someone could do us a great service
It would be fascinating and valuable if someone* could characterize the output impedances and code dependencies of contemporary and past DACs.
Particularly the ones that appear to be capable of the highest performance, and including ones that continue to be popular despite their not-quite-sweepstakes-winning specifications.
And include the capacitances. The potential downfalls of directly feeding a current-feedback opamp from a ponderable capacitance are manifold.
Brad
*As Mr. Natural says to Flaky Foont, when advising him to find a guru, "not me".
It would be fascinating and valuable if someone* could characterize the output impedances and code dependencies of contemporary and past DACs.
Particularly the ones that appear to be capable of the highest performance, and including ones that continue to be popular despite their not-quite-sweepstakes-winning specifications.
And include the capacitances. The potential downfalls of directly feeding a current-feedback opamp from a ponderable capacitance are manifold.
Brad
*As Mr. Natural says to Flaky Foont, when advising him to find a guru, "not me".
Didn't some guy write an AD application note about that?
And we haven't touched the use of an input transformer swiped from a ribbon microphone yet.
Thanks as always,
Chris
Well it appears that many of you have looked at the I-V conversion from the DAC's more than I have. My complaint is not just slew rate, or RFI, but the 'excessive' loading of a typical op amp that may be used, and its internal thermal feedback. Just one other problem, perhaps.
I am pleasantly surprised that many here, who criticize any further analog improvements, are sensitive to the D-A interface inside digital playback systems. Malcolm Hawksford's AES paper is a true revelation to me.
I consider Malcolm Hawksford a colleague. He and I have dined together, as well as his visiting my lab about 30 years ago. He pushes the limits of audio design, and I appreciate that.
I am pleasantly surprised that many here, who criticize any further analog improvements, are sensitive to the D-A interface inside digital playback systems. Malcolm Hawksford's AES paper is a true revelation to me.
I consider Malcolm Hawksford a colleague. He and I have dined together, as well as his visiting my lab about 30 years ago. He pushes the limits of audio design, and I appreciate that.
sigh ...
by loading, you mean forcing class A from an opamp? or loading by forcing them to do too much work at bandwidths not needed for output? not a big fan of trying to force class A from opamps either, well not by dropping a resistor to the rail. seems to lack finesse; especially the way its sometimes done by default as a 'tweak' to opamps that would spend most of their time there if used correctly anyway.
Malcolm is not a colleague of mine, I have great reverence for him, but I dont find not knowing him personally effects my efforts a great deal.
who ever criticized pushing for high analogue performance? seems you are lumping a whole lot of other objections together into that. some may not agree with the NEED for it, or the extreme nature of the claimed differences, but I doubt anyone here would object to it as an intellectual pursuit if the person had their eyes open.
by loading, you mean forcing class A from an opamp? or loading by forcing them to do too much work at bandwidths not needed for output? not a big fan of trying to force class A from opamps either, well not by dropping a resistor to the rail. seems to lack finesse; especially the way its sometimes done by default as a 'tweak' to opamps that would spend most of their time there if used correctly anyway.
Malcolm is not a colleague of mine, I have great reverence for him, but I dont find not knowing him personally effects my efforts a great deal.
who ever criticized pushing for high analogue performance? seems you are lumping a whole lot of other objections together into that. some may not agree with the NEED for it, or the extreme nature of the claimed differences, but I doubt anyone here would object to it as an intellectual pursuit if the person had their eyes open.
Last edited:
You know, qusp, I think that you are disappointed in me, and my engineering contributions here. However, I look at things in a somewhat different way.
You see, 39 years ago (when you were born), I hired my first PhD to design a special anti-aliasing filter for a digital time delay, and we NEVER went below 50KHz clock, and usually operated even faster.
I already had 8 years of professional engineering design under my belt, and had a fully equipped lab in Switzerland, including FFT analysis by mini computer.
Although I am much older, (more than twice as old, today) as I was then, I seem to get along with what I have learned over the ages. That does not mean that I am as 'up to date' as many of you, but I have my own brand of 'wisdom' that many others appreciate. That is why some people respond to my inputs here in a positive way, and it they, whom I am addressing, not you.
As you grow in experience, you too might gain similar 'wisdom'.
You see, 39 years ago (when you were born), I hired my first PhD to design a special anti-aliasing filter for a digital time delay, and we NEVER went below 50KHz clock, and usually operated even faster.
I already had 8 years of professional engineering design under my belt, and had a fully equipped lab in Switzerland, including FFT analysis by mini computer.
Although I am much older, (more than twice as old, today) as I was then, I seem to get along with what I have learned over the ages. That does not mean that I am as 'up to date' as many of you, but I have my own brand of 'wisdom' that many others appreciate. That is why some people respond to my inputs here in a positive way, and it they, whom I am addressing, not you.
As you grow in experience, you too might gain similar 'wisdom'.
Has anyone looked at the full spectrum coming from the output of a DAC chip? I see lots of discussion about handling very fast transients but no actual measurements of the spectrum and its energy levels? On the ESS chip with a master clock of 100 MHz you would expect 100 MHz and 200 MHz with some energy but even with fast CMOS there would be serious issues switching at those rates and getting accuracy. I have looked at a few and see strong energy to 5 MHz but I have not been comprehensive. . . With 100 MHz power bandwidth current feedback opamps is there still an issue?
And how, pray tell, does that impact audio design? Not to mention that he's retired, so he doesn't do any pushing anymore. Of audio, that is. Smart man.
jan
Built and tested, but I think I got to about John Curl stage three on that one. I forget, is that the one where we know about it but it does not matter?
Not really disappointed, this is the only John Curl I 'know' ... the problem is John, that instead of just sharing your knowledge and garnering respect, or being humble when the situation calls for it, you feel the need to stand atop of it pointing out your lineage ... looking down, firing off little snide comments (often shooting 'from the hip') about other people in the room under your breath. by extension, because of the crowd here, thats some of those people's life work.
Many you dont even seem to rate, I find quite generous with considerable knowledge and expertise; yet I get none of that from them.
Last edited:
Demian, re bandwidth, thats why I would rather band limit the signal or use parts that are more resistant to it; than try and actually handle it. on the other hand, I have listened directly to the ES9018 with headphones with nothing in the way, there isnt anything really objectionable in the audible spectrum, just the expected lack of good tight bass authority, which if the ESS has a character, the bass is certainly a high point IMO
I have looked but only to 5 MHz (the other spectrum analyzer is in storage). I was tuning a filter and it worked exactly as desired so I did not continue. Also it was not an ESS chip.
Before the filter there were several images of the fundamental band up to something like 4 MHz, I did not look higher, the filter worked fine with about 70 dB band rejection.
I would think that the source Z modulating would translate into a significant error signal, and could also produce DC shifts that are signal dependent (as well as the noise modulation with DC).
Before the filter there were several images of the fundamental band up to something like 4 MHz, I did not look higher, the filter worked fine with about 70 dB band rejection.
I would think that the source Z modulating would translate into a significant error signal, and could also produce DC shifts that are signal dependent (as well as the noise modulation with DC).
I agree with having passive CL filtering before the I/V stage. Lets keep the RF as low as possible without affecting the audio signal. Where I work, in the RF world, we always put a filter before the first active stage of a receiver. Too many signals coming in from the antenna to overload the front end.
Also, the 811 has an output current of over 100ma.
Also, the 811 has an output current of over 100ma.
No. Its the legitimate normal/differential mode HF that the antialiasing filter removes. Oversampling DAC (like all modern ones) move the aliases up to a much higher frequency but the amplitude is still high. The first alias would most likely be at the same level as the fundamental unless there is an analog low pass filter process in the chip.
Just come across this mob: About : Empirical Audio. They're making the right 'noises' - understanding that being fanatical about all aspects of the replay mechanism is key, and are quite intelligent in their explanations of what they're doing ...
I did say 'quite intelligent' ...
The bizarre world of audio marketing means that there seems to be always some level of spin injected - you just can't get away from it, especially if you want to actually sell some product ...
I saw this post - Baffled about computer power - Page 14 - on another forum, from the company, and it sent me the message that the people there are looking in the right places ...
The bizarre world of audio marketing means that there seems to be always some level of spin injected - you just can't get away from it, especially if you want to actually sell some product ...
I saw this post - Baffled about computer power - Page 14 - on another forum, from the company, and it sent me the message that the people there are looking in the right places ...
I think this guy is quite good at making assertions, but less so at providing any evidence, or even explanation of the reason, to substantiate the assertion/s. I am interested to know why leading nulls would affect the replay of a track when there doesn't seem any reason for this to occur. I may well be wrong but I would like to know why.
I am also not keen on the dismissive last words.
- Status
- Not open for further replies.