According to page 29 of the datasheet, DSD is filtered by a digital filter. Figure 1 suggests it is in the path with or without volume bypass.
I guess it is a short filter with short integer coefficients, chosen such that the numbers that can come out are always valid input codes for the AK4499's multibit DAC. If the filter suppresses the idle tones around half the sample rate, the sensitivity to clock subharmonics is much reduced.
If my guess is correct, then the DEM cannot be bypassed.
Going back to the OP, the my statement in bold is not entirely correct.
The CXD2565 is one of several single-IC (all-in-one chips with both DF + D/A) "PULSE" devices Sony was using in the early 1990s.
If one searches ...
https://www.dutchaudioclassics.nl/the_complete_d_a_dac_converter_list/
... for "CXD2562", one gets 27 hits.
... for "CXD2561", one gets 12 hits.
... for "CXD2565", one gets 12 hits.
Often, in the high-end ES series, Sony will use CXD2561's digital filter only, and feed that into CXD2562
What's a bit weird is that is that while '61 and '62 were used extensively, one cannot find any datasheet PDF record of them (unlike, say, CXD2565 or
CXD2552 ).
Not even in Sony's annual data books:
http://www.bitsavers.org/components/sony/
https://www.manualslib.com/manual/1635119/Sony-Cxd2701q.html
https://ia903202.us.archive.org/0/i...94202/1993_Sony_AD_DA_Converter_Data_Book.pdf
So, the '61 and '62 might have been deliberately obfuscated by Sony because the competition was intense.
From my buildings, measuring and listening tests experiences
TDA1540 (All versions plactic and ceramic, build and test in NOS mode), (also TDA1543 made just in 16x parallel dacs)
is one of the best sounding DACs. Maybe slight better than TDA1541A in sam mode of operation Time simpultaneous format...
.
Zoran (or anyone). Did you investigate the "Zoran" processor? Vaddis, I think ... used in Oppo, Samsung and Toshiba DVD players, roughly in the mid aughts.
There were various popular mod pages by Vinnie and Swenson and others.
Some Zoran chips outputted current for I/V.
In certain players from Samsung, the Zoran chip fed I2S to a TI sd1608:
https://rocelec.widen.net/view/pdf/gjbkpfnfg4/dsd1608.pdf?t.download=true&u=5oefqw
Is outdated superseded technology really the holy grail in audio? When the multibit chips are finally unobtainable then the first 1 bit, first delta sigma DACs are the next old thing to cherish? Then the first DSD DAC chips?
The challenge is to compare such devices to modern devices and have an honest comparison.
The challenge is to compare such devices to modern devices and have an honest comparison.
Last edited:
And this is kind of what the low-cost Chinese PCB kits have been offering on Ali and eBay for over a decade: Some old dac models (new old stock) and some modern ones (ESS, AKM, TI, Cirrus, etc). I personally have not tried them all -- not by a long shot. But I prefer, say the TDA1305, kit to a comparable ESS9038 kit -- all else held equal, meaning unmodded.
After modding, anything goes.
So what "encapsulates" the DAC chip -- the magic dream spirit - -seems to matter a lot. And many still treasure the TDA1540. I wonder why?
Haven’t heard any ES9038 based DAC that I liked. Heard ES9028 based DACs that are really good. AKM based DACs can be excellent.
If you buy cheap you have to mod. If you buy old you pay more and still have to mod and replace caps.
If you have still working TDA1540 you are old and so is the chip
That magic dream spirit is between the ears and imaginary. It can very well be that a promising young engineer is designing a new AKM DAC chip as we speak. Full of passion, driven. Will we hear that? It is silicium, not peace love and understanding baked together.
If you buy cheap you have to mod. If you buy old you pay more and still have to mod and replace caps.
If you have still working TDA1540 you are old and so is the chip
That magic dream spirit is between the ears and imaginary. It can very well be that a promising young engineer is designing a new AKM DAC chip as we speak. Full of passion, driven. Will we hear that? It is silicium, not peace love and understanding baked together.
Last edited:
I have a 2020 Fiio M11 Pro DAP -- manuf just before the AKM fire -- that has dual AK4497 DACs and " two independent low phase noise crystal oscillators from Japan-based Rive Electec" and THX AAA amps, yada, yada. It sounds decent. But I can use the same headphone and plug it into a $100 DAP from 2010 (Wolfson DAC) and use a $90 dedicated headphone amp. And -- hands down -- the $100 DAP + $90 amp beats the Fiio. Now, using the Fiio's Line Out to feed the headphone amp changes the game. But not by much at all.
So I'm still searching for the source of the source dream magic. Maybe it's as you noted ... "peace love and understanding baked together." They are not emotional concepts in the hearts of the Fiio company. And, regardless of the fancy technology. Yeah, that 14-bit 1540 has some magic. If you're curious, DIYA has a good UFO thread somewhere
So I'm still searching for the source of the source dream magic. Maybe it's as you noted ... "peace love and understanding baked together." They are not emotional concepts in the hearts of the Fiio company. And, regardless of the fancy technology. Yeah, that 14-bit 1540 has some magic. If you're curious, DIYA has a good UFO thread somewhere
What technology is in that unknown DAP with typeless Wolfson DAC? Is it fed by an SMPS or linear PSU? What is the interface and interface chip? Is it used via spdif coax, optical or USB?
Is it coincidence that when it is about old stuff all minuscule details are mentioned and when it is about recent stuff no names and type numbers are given?
Is it coincidence that when it is about old stuff all minuscule details are mentioned and when it is about recent stuff no names and type numbers are given?
My impression is just the opposite - ES9038 is much better than ES9028 or even more than ES9018, and AK4497 is better than ES9038.
Romm BD34301 looks is between ES9038 and AK4497, but closer to AK. Especially in DSD.
14 bit is just a data register bit depth jf 1540, in reality (by it's THD, ENOB) - it has 13 bit.
1543 has 16 bit data register, but it's ENOB is just 11 bit.
We can not debate taste but I somewhat doubt the “much” (both ways). I probably had to use the word “yet”
Never heard the Rohm DAC chips but I am curious. Since I fish in the pond of cheap I maybe will never hear one.
Last edited:
The $12,000USD Jadis D/A (1995) used a pair of SAA7323 dual-diff with a Sovtek tube output. The Jadis designers -- according to the Stereophile review in March 1995 ?? -- supposedly tested out multi-bits for this expensive product. Finally decided on 1-bit. I'll try to dig up the schematic, but I doubt I'll find one.
Some inside images...
http://www.audiotecnico.it/riparazioni/riparazioni_J/jadis/Jadis_JS1/Jadis_JS1.htm
They are up to a Mk4 of this same D/A, now using ak4497. I think Jadis only changed the dac chip ... all else is vintage 1995.
Some inside images...
http://www.audiotecnico.it/riparazioni/riparazioni_J/jadis/Jadis_JS1/Jadis_JS1.htm
They are up to a Mk4 of this same D/A, now using ak4497. I think Jadis only changed the dac chip ... all else is vintage 1995.
Last edited: