Jump to content

DAC - šta kupiti...

x1gor

Od x1gor
in Digital

 Share

Preporučeni Komentari

  • Odgovora 193
  • Kreirano pre
  • Zadnji odgovor pre

Aktivni članovi u ovoj temi

Popularni dani

Aktivni članovi u ovoj temi

Popularni dani

Popularni postovi

zbotic
zbotic

Nisam znao koji DAC kupiti pa sam kupio sve   AK4396, AK4495, CM6631A, 2 x CS4398 na izlazu HDAM sa NPD5564, odabir DAC-a u analognoj domeni da se izbjegne jitter. 3 USB i 3 SPDIF ulaza. 24bit/192k

Zflazar
Zflazar

...samo vas posmatram! Toliko se izmenio svetonazor na digitaliju u poslednjih par nedelja, da sad imam zabavno štivo za čitanje barem još izvesno vreme...

Eci
Eci

Posle pogona TCD2000, uspeo sam da nadjem i DAC iz iste serije TDA2000.

Postovane slike

aaa zmaju! Dugo je sa nama

aaa zmaju!

Napisano
Napisano

Pa ovi drugi linak sa KP je ustvari 4audio, ustvari cerka firma 4audija odnosno sestra ili mama firma ili tako nesto.

Link to comment
Podeli na ovim sajtovima
Mikorist Inventar foruma

Mikorist

Napisano
Napisano

Aj sad pogledajte ovu sliku i razmislite ko je blesav ovde.

Untitled.png

 

Koliko isti kosta ovde i u Nemackoj u FIIO-shop-u  sa i bez VAT-a  . Ako mama firma kupuje na veliko tamo je jos i jeftinije. A i oni svercuju do ovamo  preko JugoSvaba....DAC kantica/kutijica  je samo primer . . .   :buehehe:

Link to comment
Podeli na ovim sajtovima
Mikorist Inventar foruma

Mikorist

Napisano
Napisano

slab mu gain @32ohm, 20mW

 

Fiio ima ~200mW na 32 oma sa THD: <0.006% ... to govance raznosi bubne opne i na slusalicama od 150ohm. :to_become_senile:

 

plus sto radi na mac osx plug in play

 

imaju bolje sprave koje se meni svidjaju . . . cena im je 200-250$ ... ili imaju dve sprave odvojene DAC +AMP posebno.

 

al to sve komplikuje stvar. nije za poneti , trebaju posebna napajanja kablovi izmedju , trte mrtrte... neki i ne rade sa OSX ili Linuxom ....ovde samo ubodem USB u laptop. i peva.

 

nema na trzistu kartica audio koja svira kao ovo na mojim sluskama. (cak i one skuplje) a probao sam bar 20 raznih u zadnjih 5 godina (M-audio[2 modela] , ESI[bar 3 razlicite], Sound Blaster[jedno 3 modela], ASUS Xonar , Auzentech, AUDIOTRAK Prodigy HD2 gold .. ukratko a ima jos. ) :buehehe:

 

p.s.

 

Fiio X3 ima Wolfson DAC WM8740 takodje i E11K / E10 (bez K) nije lose to ....

 

al PCM5102 je ludilo mozga

 

u sustini najbolje bi bilo da mi Zen napravi Babela za slusalice - dzepni format - tipa 2x1W. a dac ovaj iz Fiio . . . :msn-rofl:

Link to comment
Podeli na ovim sajtovima
Mikorist Inventar foruma

Mikorist

Napisano
Napisano

ja nisam cuo klipovanje.

zavisi od implementacije chipa. naravno i sinergija sa op-ampom i njegova implementacija.

 LMH6643 op-amp ide do  130 MHz - opasnu brzinu ima (duplo vecu od AD8397 )  

odnosno toliko je jaka jacina zvuka da nikad ne mozes da odvrnes do klipovanja. jedino da ozbiljno ostetis sluh. 

Link to comment
Podeli na ovim sajtovima
Pchelac Dugo je sa nama

Pchelac

Napisano
Napisano

Ja nisam čuo ni 5102! :whistle2:  Ali pričalo se o tome na sveckom vorumu... Nešto o klipingu u digitalnom domenu i o tome kako to treba TI da reši, bla bla,... Da li je 5122 odgovor-ne znam.

 

http://www.diyaudio.com/forums/digital-source/185761-open-source-usb-interface-audio-widget-113.html#post2907899

Link to comment
Podeli na ovim sajtovima
Mikorist Inventar foruma

Mikorist

Napisano
Napisano

evo ti deo sa head fi. kako ovo radi ova kantica...

USB Receiver and S/PDIF Conversion

The E10K preserves the Galaxy Far-East Company TENOR TE7022L USB receiver --- one of the few parts carried over from the E10 of old (the E07K, E17, and E18 all also use this part, in addition to many other devices from other brands, such as the JDS Labs ODAC).

As such, the E10K remains an isochronous USB device. Upon inspection of the PCB board, I discovered a ROM part (see Fig. 2) that most likely serves as an elastic digital layer buffer for I/O. In that sense, the USB transfer protocol is likely more of an "adaptive" variant of isochronous transfer. The TE7022L uses a 12 MHz oscillator to generate its transport clock; the XO can be found in the top right-hand corner of the TE7022L in Fig. 3.

The E10K is an USB Audio Class 1.0 (UAC1) device, and thus only supports up to 96 kHz sample rates. To support sample rates beyond 96 kHz, FiiO would have to use an USB Audio Class 2.0 part (e.g. XMOS, C-Media CM6631A) and thus necessitate installation of device drivers in the Windows environment (OSX natively supports UAC2).

900x900px-LL-a32cf7a3_IMG_0086a_E10K_PCB

The largest drawback of the TE7022L is its lack of support for 88.2 kS/s sample rates. It's difficult to fathom why it's so, but if you intend on playing 24/88.2 music with the E10K, you'll have to perform a downsample with a decimation factor of two, from 88.2 to 44.1 kS/s, using a plugin such as SoX resampler.

I asked FiiO about the possibility of switching to an asynchronous USB receiver part, such as the affordable Bravo SA9027 from SaviTech (used in products like the JDS Labs C5D, Stoner Acoustics UD110, and HiFiMeDIY U2 Sabre DAC), but FiiO stated that they'd tested the SA9027 and decided the TE7022L was still a more stable, mature part. With their long-standing experience with the TENOR, and multiple product offerings that use the part (i.e. lots of ready stock, lower bulk prices), it makes sense that FiiO wanted to stick with the TENOR part. If well-implemented, in an ideal situation and board layout, I've heard that the TE7022L can get down to the order of a couple hundred picoseconds of jitter --- not world-class, but certainly not bad at all.

It also possesses built-in S/PDIF transceiver capabilities, making for extra functionality that carries over from the original E10, allowing for a penalty-free preservation of the coaxial output in the device.

There is the possibility that the E10K might work with Android devices over OTG, but I have neither confirmed this functionality with FiiO, nor have I tested it myself.

Digital-to-Analogue Converter

The DAC chip has been changed, from the often-used and venerable Wolfson WM8740 to the TI/Burr-Brown PCM5102A.

The PCM5102, like the WM8740, is a voltage output DAC chip. However, since it has a charge pump integrated into the IC, the PCM5102 is capable of delivering a true, Red Book friendly, ground-referenced 2 Vrms at the output. As such, the design methodology centered around this chip solution is less complex than with other DAC designs, simplifying both PCB layout and the number of components necessary. The built-in PLL mode of the PCM5102A also allows the E10K to function without a master clock oscillator or a clock generator (e.g. PLL1707), further simplifying digital design by accepting just a three-wire I2S input from the TE7022L. The simplified design also reduces possible EMI. Using a discrete clock design probably just complicates design a lot more and balloons the BOM cost. Thus, these parts choices inform us of the following: one of the design goals of the E10K was to improve performance without raising costs (if not outright cutting costs). Switching to this design helps get FiiO to that goal.

During the development process, I requested that FiiO look into implementation of the low-latency IIR minimum phase interpolation filter built into the PCM5102; both the Meridian Explorer and the JDS Labs C5D have the filter enabled. In the end, FiiO chose not to enable the filter, opting for the normal FIR linear phase filter. While pre-ringing is not eliminated, there are no phase distortions introduced.

Set at 0.55fs, the low-latency filter would have also introduced additional aliasing images, suppressed by about -40 dB, and thus aliasing would perhaps be audible at very high frequencies. Oversampling in PLL mode is conducted at a lower system rate, so the E10K wouldn't have the luxury of using shallower reconstruction filters.

Analogue Design

This section is all-new for the E10K. The current buffer, as mentioned in marketing materials, is the LMH6643, a voltage-feedback opamp capable of delivering 75 mA current, and is employed in both the E18 and the X5. Full scale output is referenced to 2.6 Vrms, which equates to roughly 200 mW into 32 ohms.

There's an LM2663 switching power regulator that serves to double/invert the power supply to meet different power needs across the board.

There is a gain stage to the E10K, which is responsible for switching between the +1/+9 dB gain selection, the duties of which are performed by the OPA1642, located right behind the ALPS potentiometer and surrounded by a bevy of resistors (different groups of resistors for feedback for different gain?). The volume pot is rated for <0.4 dB of tracking differential.

There is also an unknown opamp located right next to the LMH6643 in the middle of the board; at first I thought it was the OPA1642, but it was not the case. The printing is too faint to decipher. It is a TI part (FiiO uses TI heavily) in an SOIC package, but I cannot decipher anything else beyond that (see Fig. 2). It's possible it's used in the bass boost feedback loop, but I cannot say for sure. While I've improved my knowledge in audio design, I can only say I'm a mere dilettante and not a learned EE.

Noise/Hum

Residual noise of the E10K's line output is at a healthy <10 uV RMS, while the headphone output is <15 uV RMS. That's admirable performance for an USB-powered DAC at the ~$100 price point. Take, for example, the HRT MicroStreamer, a well-received USB DAC that competes at a similar (but higher) price point to the E10K and has 18 uV RMS. Those little linear regulators that you see littered across the E10K's PCB board, while not boutique parts, are definitely doing their job.

The E10 of old had an issue of lots of noise rustling through when plugging/unplugging headphones (too much DC bleeding through?). That issue has been resolved in the E10K. While plugging in headphones isn't a completely silent affair, it is quiet enough not to be a bother for anyone.

Sound

Overall, I find the the FiiO E10K to be a very competent unit and generally a very good improvement over its predecessor. The E10 was a warm, intimate unit that sounded pleasant but rolled off on its ends. The successor E10K is a far more neutral-sounding device with little roll-off and more detailing, though it does still generally hint at warmth.

Presentation-wise, the soundstage and center image are cast somewhat intimately, but not quite forward. Overall, it feels laid-back. The E11K sounds more spacious a bit brighter (more later).

Bass boost this time around is more subtle than it was in the E10, aimed at bringing presence to music, rather than boosting the "boom" of added mid-bass response. Thus, the effect is somewhat subtle but still present.

A/B-ing between the E10K and the Resonessence Labs Concero HP, the E10K decidedly lost the competition (and with good reason --- the Concero HP costs more than eight times that of the E10K). The primary goal wasn't whether the E10K could acquit itself; it was to subjectively measure the E10K against a well-implemented jitter rejection paradigm of the Concero HP, a DAC that employs low-latency asynchronous USB transfer, discrete clocks for 44.1/48 kHz multiples, additional FPGA sub-routines for jitter timing, and ESSTech's built-in asynchronous sample rate converter. The E10K sounded a bit splashy compared to the Concero HP in the high frequencies. I tried testing some "jitter-sensitive" tracks, such as the evolution of the sound of the shakers in Rebecca Pidgeon's 'Spanish Harlem', and I did find the E10K to be a bit more indistinct with less dimensional movement across the soundstage, but if you weren't specifically looking for problems, I doubt anyone would even notice. I also found subjective jitter performance to be a little better on the AudioQuest Dragonfly 1.2, but the Dragonfly sounded a little boring to me overall.

The E10K performed well with low-impedance, full-sized headphones such as the NAD Viso HP50 and the Focal Spirit Classic. Neither headphone is exactly "hard-to-drive", however.

With IEMs, aside from the smaller center image cast, I found the relaxed, very slightly warm presentation of the E10K ideal for my collection of earphones, ranging from the UERM to the Noble 4C. Never did I ever feel the E10K was somehow "slow", "lacking weight", or anything like that.

 

a ni ovaj lik (admin sa foruma) ne zna sve (ili nece da kaze - poslovna tajna). odnosno ima i unknown opamp.  

 

pun naziv DAC chip-a  je TI/Burr-Brown PCM5102A  (postoje 3 verzije istog)  PCM5102-Q1 (NRND) , PCM5102 (NRND)  i  PCM5102A koji je isto sto i 5122 . Stim sto  5122 ima ROM based fixed processing (nema HW mode)

 

jedino što samo ima da se nadje implementacija za Raspberry...

 

PCM5102A provides a HW mode, all can be configured via pins and soldering on PCB and DAC pins. No I2C or SPI communication needed.

 

The PCM5122 needs I2C or SPI communication which is necessary in order to enable, configure etc. the DAC. It results in:

RPi has to talk via I2C to DAC, P1 allocated for I2C, driver in Linux kernel needed, not transparent (RPi-DAC-RCA can use the HifiBerry as well as the RPi-DAC driver).

 

 

p.s.

 

Postoje i lazni  (FEJK) FiiO E10K   zato sam i narucio od distributera u Nemackoj. Nije samo bitno koji DAC kupiti nego i gde.... :buehehe:

Link to comment
Podeli na ovim sajtovima

Kreiraj nalog ili se prijavi da daš komentar

Potrebno je da budeš član DiyAudio.rs-a da bi ostavio komentar

Kreiraj nalog

Prijavite se za novi nalog na DiyAudio.rs zajednici. Jednostavno je!

Registruj novi nalog

Prijavi se

Već imaš nalog? Prijavi se ovde

Prijavi se odmah
 Share
Idi na listu temâ

  • Članovi koji sada čitaju   0 članova

    • Nema registrovanih članova koji gledaju ovu stranicu
×
×
  • Kreiraj novo...