Ad1852

[johnny1]

I have two of these chips and i'm thinking of how the will sound.
I will build a second DAC inside my cd player with these chips as all the signals are available from the decoder chip. Has anyone ever heard them and if not, does anyone knows a commercial player that use it?????

[Petter]

I am using 1853 which is a current output version of this with slightly better specifications.

Analog devices DAC's are generally of high quality (but everybody makes some bad apples as well). The problem with Analog devices is that they pretty much only design devices which require a serial port interface for control of internal registers (to have a low pin-count. This is probably OK and good for volume manufacturers but not for DIY'ers who want to concentrate on sound-enhancement rather than interface programming.

Hooking up an actual DAC chip is easy if you have L/RCLK, BCLK, SDATA and MCLK (or equivalent). If you don't have that and want do roll your own, consider purchasing an evaluation board instead.

Why did you choose that particular chip (it is not the best one around, but not bad either)? In general you will find current output chips are better. Also, if you are looking for something relatively simple to design with consider TI PCM1730 or some of the Cirrus models (43122 is pretty good). You should consider looking for units that can be set up without programming internal registers unless you know how to program PIC's etc.

I have used the 1853 eval board and was quite pleased with it. It made me the first person in my country to own a 24 bit audio DAC (I am pretty sure of that). It was easy to use -- sounds a bit like what you are trying to do.

Don't choose a chip based on what you have -- the cost of the chip itself is minimal compared with the cost of your time and assosicated parts you will purchase.

Petter

[johnny1]

The fact is that the signals L/RCLK, BCLK, SDATA and MCLK are allready available in the cd player. My first consideration when readind the datasheet, is that it does not need programming or external components (in the digital part of course). I will read it more carefully before starting pcb's.

Anyway, the chips are free samples and they are significally better than the stock PCM61 that the cd player has now (Pioneer PD6300 bought in 1990.....) . The purpose of the project is just to "play" with DAC chips and digital data, not to make an old player a hi-end unit.
For this purpose there are two PCM1704 in my drawer waiting a CS8414....

[Petter]

Well that is very cool and I wish you luck with it. It really is quite easy to go from the 4 signals into a stereo converer with internal digital filter. There is an excellent thread on decoupling in the digital section.

I am currently modding a Creative Extigy to take 3 ea. PCM1730. I am also designing a stereo DAC to take multiples of 1730's

Again, good luck! I think you will find that most of the hard work has been taken care of by the chip manufacturer.

Petter

[johnny1]

Thanks Peter!

Now, about the decoupling thing, i found similar solutions using the famous but effective "try and error method"!

Well, besides the decoupling of the DAC's, the digital filter and the output op-amps, there are many other tricks to improve a player's performance as a transport.

The purpose of the tweaks above, is to minimize the noise in the supply rails of those components-which definitely affect the sound.

I don't know how many of you people have ever try to impove the decoupling of the decoder chip, the RF receiver or the microcontroller. These units are also sensitive to noise. The microcontroller is ussally a noisy component.
Just put a 1uF MKP in parallel with the (existing) electrolytic in the RF receiver supply pins. The result - especially in the highs- will surpise you....

P.S : Sorry for my English

[tiroth]

try a 1uF ceramic...

[Peter Daniel]

....or check this link Bypass Caps for digital IC decoupling That guy is really serious about decoupling.

[capslock]

Johnny,

if it's an old Japanese CD player, I'd almost bet the signal (sdata, lrclock, bitclock) travels in Sony (i.e. right justified) format. With right justfied, you should know the number of bits. The CD player is going to give you 16 bits, the AD1852 expects 24 bits unless you have a micro to programm the word length.

For this reason, I use the AD1854 which has a pin-programmable 16 bit right justified mode. The K version has almost the same distortion performance as the AD1852. My impression is that the 1854 shares the analog section with the 1852 but, judging from the inferior stop band suppression, has a shorter and hence simpler digital filter which might even be the better choice sonically.

Still beats me why there are two selection grades. Seems to be the only sigma delta DAC with selection grades. A multi-bit DAC depends heavily on element matching. In a true 1-bit DAC, there are no elements to match. In newer multi-level sigma-delta DACs, element mismatch is taken care of by randomly selecting nominally identical elements, so it shouldn't matter either. So what is the reason for the two grades here??? Different process?


If you still want to use the 1852, here are some workarounds:
- use a micro
- use shift registers to shift sdata 8 bits to the left
- use a DIR chip and set it to I2S output but remember to run it in slave clock mode


Eric