Jump to content

zenelectro

Members
  • Content count

    440
  • Joined

  • Last visited

  • Days Won

    1

About zenelectro

  • Rank
    Member

Profile Fields

  • Location
    NSW
  • Country
    Australia
  1. Haha... Well aware of this Rob, just trying to keep minds open around here.
  2. WRT 1704 - probably a bigger shame is Arda not following through with development of AT1401 R2R DAC which preliminarily out paced 1704 in every regard. Implementation would have been easier too given its high OP Z and current swing. WRT your fascination - or should that be obsession (LOL) with R2R - this DAC shootout's results pretty much sum up my opinions: It is implementation that will get results. http://www.superbestaudiofriends.org/index.php?threads/denafrips-r2r-dac.3845/page-4#post-142650 The DAC that won is, I believe, based on ES9018. How could that possibly be?? Implementation! So there you go! Terry
  3. TBH I haven't had a really close look at the board Clay. When the Singxr first came out I looked carefully at it due to the fact that it could do 352/384k with 22/24M clocks. They acheived this by doing reclocking inside FPGA. As such jitter is dependent on FPGA. After doing quite a bit of research it looks like there are better ways. T
  4. Unfortunately going to all these heroic lengths and then using SPDIF OP is......?????? Must use I2S / PCM (simultaneous) mode OP or all the money thrown at USB conversion is very much wasted. T
  5. After a 1 minute google search..... Quote from a Principal Engineer, major FPGA manufacturer: "even with the -90 degree clock trick, you will not get below 35 ps P-P. If that is a problem, then you should fix the system and layout, as nothing you will do will make it any better without NOT having the clocks go through the FPGA device" IOW most implementations will have jitter much worse than this. T
  6. I looked at Singxer but reclocking of I2S is done inside Xilinx FPGA. Not a good idea and I'm surprised they do something like this.
  7. What's wrong: a/ Mismatched resistors. b/ Variable OP resistance of logic driving those resistors c/ Glitch energy when bits change. d/ There also may be some AD limitations with measurement setup. Generally a loop through baseline is a good idea or at least a single tone 1 or 10kHz with VLD oscillator of which many are available nowadays. The Lavry DAC was as good as it gets WRT R2R technology. Many mastering engineers felt it was the gold standard but these days that is not so much the case. These discrete R2R DACs are becoming a fashion accessory these days for audiophiles. Once we fully exploit the latest DS DACs people will realize that this is just an interesting diversion. T
  8. What's the $ on this puppy?
  9. Yes and it's easy to see why. The Chinese do a some things really well. a/ copy great ideas. b/ low cost of manufacturing particularly labor intensive tasks. That pretty well wraps up the ladder DAC scenario. T
  10. George, I believe they are the same DAC. WRT price, you have to consider these are built in US, not China and are a lot more sophisticated than these Chines DAC's. I believe the whole DAC ladder is inside a temperature stabilized unit. Then there's the self calibration circuitry and custom implemented digital filter. You also should consider that these Chinese ladder DAC's are all knock offs of recent designs like Soekris. The Lavry is well over 10 years old. It was a landmark design. Whether this all translates to a subjective result that is worth the coin is obviously entirely up to the (potential) buyer. T
  11. I know this is an old post however just wanted to point out WRT heroic attempts at matching resistors: Resistor matching is only 1 part of equation. The biggest issue is the logic switch (driving resistor) OP resistance which is in series with the resistor. Even with theoretically perfect resistors there will still be this error and it can be significant. One solution is to switch the resistors with discrete high speed BJT transistors which are in turn driven from (cmos) logic. Much lower OP resistance then your fine tolerance resistor matching will make a real difference. After that you have glitch energy which is a consequence of the bits not being able to switch instantaneously. I think this is the typical high order harmonics you see on all these discrete R2R DACs. One solution to this is to use a sample and hold circuit at OP of ladder. The S+H waits for all bits to settle and then transfers that value to I-V opamp. It's worth noting that all these issues were addressed many years ago in the Lavry Engineering DAC DA924. Every bit had a small instrumentation DAC driving the resistor to adjust that bit on power up in a self calibration process. Pretty damn smart that man. Also the OP had a S+H circuit on it to take care of glitch energy. The DA924 got very high reviews (stereophile class A) and was used by many mastering guys. T
  12. Theoretically an asynchronous USB that is re clocked is as good as a FIFO.... theoretically. I guess it all comes down to the quality of isolation, the reclock chips, their power supplies, the clock power supplies, grounding, yada yada yada... This stuff is all so easy from a straight technical POV, ask any technician.... but all so very difficult in the real world when you start listening to results and what is the mechanism behind those subjective results. Digital - perfect sound forever.....hahahahaha! T
  13. Jlsounds already has reclocking on board so I'm not sure how much the FIFO / isolator / reclocker etc will bring to that party. Having said that the FIFO and Jlsounds will both be more dependent on what clocks are used. The Jlsounds needs 45/49 MHz clocks and the FIFO can work with 22/24MHz clocks which are generally a better option.
  14. Trivial labour? The Jlsounds is probably best option from a simplicity POV, it already has galvanic isolation, Amanero does not, you have to add it. Having said this I don't know how compatible Jlsounds is with existing hardware / pinout / layout. Good luck! T
×