AD1896 In slavemode how??

I have started a new dac project with:
CS8416->AD1896->PCM1798 -> Zapfilter 2.

CD player as transport

44Khz in to AD1896->192Khz sampling out.

External clock for AD1896 24.57Mhz clock.
AD1896 must therefore be in slavemode.

What does slavemode mean in relation to mastermode?

Do someone know how to connect a divider circuit so I can get 192Khz sampling out off AD1896 in slavemode?

Best regards
Kim
 
The AD1896 puts out data at 256Fs in master mode. So if you clock it's MCLK_I pin directly with a 24.576MHz clock, it will put out 96KHz.

In slave mode, the AD1896 can put out data at whatever rate you want, as long as the rate is slower than MCLK_I / 136... For 192KHz input/output, this means a MCLK_I frequency of at least 26.112MHz. You can run the AD1896 asynchronously; just tie a 27MHz crystal or oscillator to its MCLK pin(s).

The only issue here is that the PCM179x part will only run in I2S slave mode - this means you'll have to generate the I2S clock and frame sync signals yourself. But this isn't hard to do.

From the 24.576MHz clock, you'll have to generate a 12.288 (/2) I2S clock and a 192KHz (/128) I2S frame sync. You can do this by using a 74HC4040 counter, and negating the 12.288 clock with a 74HC04, 74HC86 or whatever so that its output is low when the FS changes state.

And of course, feed the PCM179x's MCLK with the 24.576 directly.
 
Hey

Thanks gmarsh for your time and great input.

The devider circuit can I use this, at page 7 with 2 74AC161?

Let´s take your schematic

I set AD1896 in slavemode (in and out)

About the generated signals:
Do I connect the 12.288Mhz i2S to pin 25 in AD1896 and leg 6 in pcm179x?

And the 192Khz frame sync to pin 24 in AD1896 and to pin4 at PCM179X?

The 24.57Mhz clock connecting to MCLK leg 7 at PCM179X.

Or is it only Sdout and Data that connects AD1896 and PCM179X,
So the generated signals go alone to PCm179X?

Thank

Kim
 
I2S clock and frame sync from the divider circuitry will go to both the AD1896 and PCM179x parts.

You can use two AC161's... but AC logic is evil (use VHC instead) and a 74VHC4040 will only require one chip to perform the division. You'll still have to invert the I2S clock to create a proper phase relationship between it and frame sync.
 
If i choose a 27.000Mhz clock for AD1896 /PCM179x in slavemode

AD1896 gives out 27.000Mhz / 136 = 198Khz sampling
Datasheet don´t say how to set AD1896 to /136??

And PCM1798 only has a 128 or 256FS? Forexample in PCM1798:
27Mhz clock as above will be 27Mhz /128fs = 210Khz sampling?
Pcm1798 can only do 200Khz?

With 24.576Mhz clock -> 180.7Khz out from AD1896
with PCM1798 = gives 192kHz (24.576Mhz / 128)

Can anyone clear this up for me?

I could also settle with hardware mode with 29.9Mhz osc and 116Khz samplefrekvens?

gmarsh where did you buy your X1 cb3 osc from?

Thanks
Kim
 
If you choose a 27.000MHz clock or crystal for the AD1986 and run its output in slave mode, its maximum output sampling frequency will be 198KHz. This 27MHz clock is only used internally to the AD1986 to perform calculations, it doesn't affect the input/output sampling rates in any way.

The actual output sampling frequency of the AD1986 in output slave mode is determined by the clock and frame sync you feed it - if you feed in a 12.288MHz clock and a 192KHz frame sync, it will put out audio at 192KHz.

The PCM1798, if it's anything like the 1794 or 1792, requires you to feed in data at a sampling rate of SCK/128, /256, /384, /512, and maybe other rates. It will automatically detect the rate. Check the datasheet on this one and confirm.

I've attached a 2 minute schematic of how you'll want to hook it up. I drew a 49.152MHz oscillator; you can use a 23.8MHz part instead but you'll have to use Q1,Q7 from the VHC4040 instead of Q2,Q8 to get a 192KHz sampling rate.

You can get the CB3LV oscillator from Digi-Key.


kimschips said:
If i choose a 27.000Mhz clock for AD1896 /PCM179x in slavemode

AD1896 gives out 27.000Mhz / 136 = 198Khz sampling
Datasheet don´t say how to set AD1896 to /136??

And PCM1798 only has a 128 or 256FS? Forexample in PCM1798:
27Mhz clock as above will be 27Mhz /128fs = 210Khz sampling?
Pcm1798 can only do 200Khz?

With 24.576Mhz clock -> 180.7Khz out from AD1896
with PCM1798 = gives 192kHz (24.576Mhz / 128)

Can anyone clear this up for me?

I could also settle with hardware mode with 29.9Mhz osc and 116Khz samplefrekvens?

gmarsh where did you buy your X1 cb3 osc from?

Thanks
Kim
 

Attachments

  • slavemode.png
    slavemode.png
    67.8 KB · Views: 746
I Have decided to:

Use in AD1896 Mastermode with 27.0000Mhz-> Lcaudio homemade XO2 Clock or a Tentlabs XO (29Euro)

gmarsh is using one called CB3LV 29.9Mhz cheap and ok XO.
But I shall pay a least 30$ just to get it from USA (digi-key).

Because slavemode needs extra components->extra space
->supply.
And I don´t think the difference between 192Khz sampling and 105Khz will give that much difference (sound)

Pcm1798 can later be changed to PCM1794 if you want to reaach the top (also money)

Powersupply (3.3V digital section) can make it with LT1085 have heard a lot off it:)

Dac 5V PSU a kind off a TL431 in a tripple emittercoupling
Any suggestions here??

Iout will be with a kind off connector, so different I-V convertion circuits can be used

Thanks
Kim
 
kimschips said:
Powersupply (3.3V digital section) can make it with LT1085 have heard a lot off it:)

Dac 5V PSU a kind off a TL431 in a tripple emittercoupling
Any suggestions here??

[/B]

I think that would be nice if each chip has its own regulator. I think that there is a good article on guido tents site about that. So pre regulation for digital and analogue and than each component with regulator. All components that you are willing to use are "class A"
so a simple tl431 regulator can be used for 3.3 or 5V

I think if you are taking the effort to do this configuration, as I told you over the email, I think it is a good idea to go dual mono. So 1 chip (2 dacs) per channel.
What do you think. If you use 1978 it is just 8Eur for the other one is 20, but since the cost of enclosure switches transformers capacitors and other stuff will be much bigger than that I think it is worth it. Usualy when I thought to do something cheaper and than at later date go back to improve, it was to painfull and to costly than to start over again.

Do you have a schematics of a passive i/v stage without transformers? Does somebody knows if the "external" (not on the same board) i/v stage will limit the DAC performance?

Have a nice day,
Pred
 
Hello

An external board with preregs like LT1086 or super jung reg?

The smoothing capacitor could be a BG-F 2200uF.
Maybe panasonic FC?

About the TL431-> A local psu by each chip consist off a TL431 with a shunt transistor (BC550?), decoupled by a Blackgate N 10uF/50V in parallel with a PHE426 100nF capasitor, and a 47uH inductor??

I we go for a dual mono dac stage (2*pcm179X) is it possible to use that with only a AD1896, or do you also need a DF1706 to split the Left right data?

Any source for precision crystals in Germany or Netherlands?

Best regards
Kim
 
kimschips said:
Hello

An external board with preregs like LT1086 or super jung reg?

The smoothing capacitor could be a BG-F 2200uF.
Maybe panasonic FC?

I think that this should be left to others to think about, the preregulation and stuff, I was more thinking that the local regulation should be on the dac pcb.

About the TL431-> A local psu by each chip consist off a TL431 with a shunt transistor (BC550?), decoupled by a Blackgate N 10uF/50V in parallel with a PHE426 100nF capasitor, and a 47uH inductor??

I thought only tl431 with two resistors to set the voltage, one to sink the appropriate current, the inductor and capacitor. I think that none of the chips consumes more than 50mA so 100mA shunt should do.


I we go for a dual mono dac stage (2*pcm179X) is it possible to use that with only a AD1896, or do you also need a DF1706 to split the Left right data?

As I saw in datasheet the using MONO, CHSL, FMT1 and FMT2 you can choose mono mode left or right channel and format I2s, left justify... I am not sure but I think that they put the output in the same time. The only difference is that you can not choose than the filter and it stays SHARP. I think I have read somwhere that the SLOW gives better sound...

So but if we plan to do the I/V conversion externaly we can just simply put the two DACs in paralell. You will than have 4 current out contact that you can use as you wish and if you do not want to use only one DAC. What do you think?

Predrag

PS
I wanted to wait for schematics and then we could make a new thread with better title to attract attention. I propose that we stay for the moment until we have a schematics and have a starting point with wich we should be able to minimize the noise of others.

PPS
I plan to work in paralel on PCB with you since I use Protel and you are using Eeagle if Iam not mistaking
 
I am thinking off making the pre regs at the dac board to avoid long vires

Tl431 I think will work best with a shunt transistor-> what technical function since you use that by a chip-> constant current-> dynamic sound??

I think it sounds great with dual mono dac. I don´t know if Sharp roll off will sound more bad than slow-> in how way should it sound bad?

Decopuling by each chip Preregs->tl31->inductor-> 22uF NX-HQ -> 0.1uF NX-HQ both blackgate

Look here for great prices for blackgates
http://www.acoustic-dimension.com/

Ps: Circuitmaker 2000 is an ealier version off protel 98-99 and so on.
What version are you using?

Kim
 
pred how do you define a passive I-V stage?
You could see Zapfilter2 it just sounds wonderfull.

New idea for PSU for the dac

All VD supplies (not dac chip(s) LT1086 or LM317

Dac PSU (3.3V and 5V) close to chip and onboard pcb The Sulzer reg look here

http://www.diyaudio.com/forums/attachment.php?s=&postid=15789&stamp=1011872634

And here for measurement
http://www.diyaudio.com/forums/attachment.php?s=&postid=15431&stamp=1011693000

Red trace is Sulzer reg

Light blue is lt1086 with lm329

Dark blue is Pooge circuit

So if we replace D1 in the sulzer with a Red led we could make a very vey low noise psu for the Dac chip

I have draw a circuit with a LM317 as a prereg, feeding a tripple emitter coupling with a 6.8V as reference
like a upscaled psu on TNT page

gmarsh and pred what do you think off a sulzer reg for digital circuits?

Kim
 
regs and i/v

kimschips said:
pred how do you define a passive I-V stage?
You could see Zapfilter2 it just sounds wonderfull.

Kim

I will look in to the zapfilter, I am not familiar with it. I thought maybe to have the same schematics as in datasheet, but instead of puting the resitors in the first stage (i/v conversion) "arround" the opamps, to connect them to ground and make the opamps just buffers for the next stage. Since we decided to put the I/V stage into separate PCB I think that my question is not critical for this stage of preparations.

Again I looked at the circuit and for me is too much components... I think that simple tl431 is enough. I mean it has around -85db of noise and if we look at the pssr of the dacs and components used I do not think that this small circuit will be the limiting factor for the -120dB performance. And one needs 3 resistors and 2 capacitors and an inductor with one TL431 so you can put pre regulated voltage to some value and then use 2TL for the CS chip 2 for the AD and 2 per DAC! Cost is minimal and I think that the decoupling between stages will be improved. Concerning noise I may be wrong but I think that BB is supplying the testboards for these AD converters with LM317 or 7805 and they are worse in noise than TL431 (look at the tnt site for measurements). There is also another shunt regulator with additional transitor, but this just complicates the thinghs IMHO.

Pred
 
Hello

Thanks pred.

I think we are on the way:)
Look here for an example off a psu with tl431

http://peufeu.free.fr/audio/extremist_dac/implement_2.html

Do you think is nice to transfer a current from a constant current generator on a seperate board to the TL431 on the dac board?

I mean instead off the usual low impedance direct from a LM317?

But I am still thinking off a very simpel sulzer reg for the dac chip it´s very important here for a great psu.

Else soon I will start to draw a schematic

Best regards
Thanks for all pred and your others:)

Kim