I absolutely agree that such a fancy external PLL is not required; that's why I said I'll show an example of 200 femtosecond jitter PLL only if I can get some reasoning if it is required.
However, I fail to understand the praise for the NDK clock on that particular JLM I2S/USB board, while other quartz for the same function are chastised. In particular if a (poorly as is "no specification") AK4137 SRC is used.
I looked up the LTC2387-20 (18 bit, right?) and it's a little out of my league (both the data sheet and the know how on how to use it with a FPGA for filtering). Also looking at the LTC2378-20 and ADC2500-20...
The big problem is the structure of the AD/LT demo board kit; it is another big investment, one needs to get not only the ADC daughter board ($300 canadian plus tax), but also the DC890B ($450 canadian plus tax) motherboard and probably more stuff. Perhaps when I'll get bored of the AK5572 demo board (that I got for under $100, second hand) and replaced the 5534 opamps with something more decent.
I happen to agree with his findings but I also judge that he gives an honest and open appraisal of what he hears when he makes PSU changes. In particular I concur with this assessment of sound quality changes as a result of LDO, Sulzer and Jung based regulators for the digital and analogue DAC supplies and I/V stages. To be perfectly honest the changes are not subtle and if you are dubious about them I would suggest trying them first if you haven't already. Much to gain and little to loose. I'd rather it wasn't so as multiple supplies from transformer to regulator and input pin are impossible to house and expensive.
FWIW I am also surprised by how much line regulation is needed to completely isolate these supplies from each other and the mains. I don't have the equipment to objectively measure such changes so I have no choice but to use lengthy listening after individual changes to confirm theory as I guess most people here do.
John
Yeah the part numbers are unfortunately similar, LTC2387-18 is the 15 MSPS part, LTC2378-20 is the 1 MSPS part. Easy to get them mixed up. The DC718C is the data acquisition board for the medium speed ADCs; I suspect it's different for the LTC2387-20 since it has a SPI interface instead of LVDS. Should be easier to work with.
Just an attempt, from someone who has not been beaten as child, to compensate for all the negativities.
Partially true: I'm only starting to look into these digital thingies. But then I did my homework on linear regulators, mostly discrete, and I listened to them, so I'm not surprised some IC sound better than others, regardless of their age, or what the datasheets say.
How lame of an insult. I was called elsewhere on this forum “abused by a simulator in my childhood” for my refusal to accept stupid simulation results, now that was funny.
Happy listening, bric-a-braque audio stories, and don’t let those sales numbers slump.
Did he tell you how his listening test was set up and is that the same method you used? If so, can you let me know so that I can try it? Thanks.
Sales numbers...
is he in audio business?Well, I also have extensive experience listening to op-amps powered by various regulators, including several LDOs, 78xx, LM317, etc. I happen to disagree, as do the measurements.
Chris,
I'm very interested in what you measure and in what way in concurs with your listening. I don't have the means to test. Do you measure line regulation and output impedance vs frequency and noise perhaps? Do you feel able to share your findings? I'd be very grateful- no axe to grind here.
John
I'm very interested in what you measure and in what way in concurs with your listening. I don't have the means to test. Do you measure line regulation and output impedance vs frequency and noise perhaps? Do you feel able to share your findings? I'd be very grateful- no axe to grind here.
John
I have to confess I have not done a direct listening comparison with the same circuit, but I have listened to very similar, and a properly used LDO sounds the same to me. I check for stability and expected load step response if using exotic parts, and residual switching noise if using SMPS. It’s not usually worth the time to continue hunting unless there are issues observed.
Chris, thank you for responding on that. May I ask what the rest of your reproduction system consists of (dac, amp, speakers, etc.), the volume at which you listen, and if using speakers how far away from them?
Don't intend say anything bad about what you use, just trying to understand a little more if there is a hardware reason that might help explain you don't hear small differences verses the particular way you focus attention when listening.
On the subject of focusing, some people listen for frequency response, noise, hum, switching buzz, etc., things we might also measure. Other people listen for reproduction accuracy of cymbals, bass instruments, percussion instruments, etc., things that are not normally directly measured. There are also people who listen for mix balance, quality of reverbs, compressors, and other things primarily under the purview of recording and mastering engineers. With all the things one might listen for, probably nobody can focus attention on the small details of all of it at once.
Last edited:
Which seems to be what happened in Pavel's preamp listening test which you questioned the value of?
Can you share the same info when you discovered the audible differences by changing DAC parts?
How do you listen for reproduction accuracy of those instruments you mentioned? You would have to do it in the same room and location where the microphone was placed to be able to compare side by side. I suspect that you haven't done such thing.
Sorry guys, have been busy with family, thanksgiving, work and my speakers/amplifiers. I've decided that the best way to go right now (for me) is to recreate the AK4499 Eval board in Kicad, then start making the changes I would like to implement, but it's too much to go from eval straight to my desired schematic without having a framework in place first. I'm going to be doing pretty much an exact copy, headers, I/V Stage and all. Likely not going to do any board work until the design looks a bit more like what I'm going for.
have not read whole thread but if i understand correctly you dont have time to do pcb.if eagle cad is option i could do pcb work for you in my spare time, even finish schematics if needed.
my issue is i have time to design pcb-s yet no money to test all of them, also i make mistakes as i am no professional.
attached is picture of 2 channel adc pcb that i never made with akm-s 5397 ADC with bit overkill regulation that i wanted to connect to "motherboard" that would contain power/mcu, my idea was to make it like this using ddr2 connector so i can prototype easy each part separately, dac/adc/mcu... if interested just say i would love to design pcb that people would actualy use.
Attachments
Yes, and some were even surprised to be told they had been listening in mono
I guess you missed the part about a perfect stereo setup would portray a good mono recording with exceptional focus and depth.......
Don’t hate on perfection Matt