USB DAC Design questions
Feb 6, 2006 at 10:12 PM Post #151 of 458
The AK4117 was used in AOS' DACs. It seems he didn't use a micro controller with them, perhaps do they run ok without one ? It could be good to send him an email about it.

See for the exemple the ally dac board schematic : http://www.aoselectronics.com/ally.html

Edit what worries me a bit is how to get the AK parts. They're not in digikey, mouser, farnell, allied electronics, farnell's stocks.
 
Feb 6, 2006 at 11:12 PM Post #152 of 458
Quote:

Originally Posted by 00940
The AK4117 was used in AOS' DACs. It seems he didn't use a micro controller with them, perhaps do they run ok without one ? It could be good to send him an email about it.

See for the exemple the ally dac board schematic : http://www.aoselectronics.com/ally.html

Edit what worries me a bit is how to get the AK parts. They're not in digikey, mouser, farnell, allied electronics, farnell's stocks.



I just contacted AOS by email. He lives in Vancouver, and it's still early here so perhaps we'll hear back from him soon. I'll post his comments when I get them.

If it's hard to get the Ak parts, then maybe we shoudln't use it. Maybe we just need to look for another SPDIF receiver?
 
Feb 6, 2006 at 11:52 PM Post #153 of 458
From aos:
Quote:

...I can tell you that AK4117 works well in hardware mode, without a
microcontroller because I used it in both PPA and Piccolo standalone. You might run into some problems as it seems to behave
sometimes in the ways not in the datasheets, but you shouldn't have trouble to get it working with at least 44/48 kHz but since you
likely want it for USB and possibly CD use, that is all you care anyway.


But since it's difficult to find this part, maybe we should consider DIR1703E? They have them on hand at digikey.. The only difficulty is finding a datasheet for it...

Edit: No, it's quite easy to find the data sheet for the DIR1703E. I'm an idiot. Anyway, I asked aos about acquiring the AK4117s. He said go to the AKM website. He thought it might be Allied Electronics or All Electronics though.

All American Electronics seems to carry it- but their website says that it is currently down for maintenance, sooo...? No - it's back up. So that's a potential source for the AK4117, which could be nice.

The DIR1703E from TI is a lower jitter part, 75ps versus 200 for the AK4117; so the DIR1703 is a better part, and it is still available from Digikey- though they only have 613 in stock, and it doesn't look like TI is producing them anymore, so...

Another possibility is that AK makes a part that is both a DIR and an ASRC is the ak4122. I haven't finished downloading the datasheet for it yet- but it might work then to replace both the CS8416, a mux, and the ASRC. But I',m talking out of my ass, so.. wait and lets look at the data sheet.
smily_headphones1.gif
 
Feb 9, 2006 at 4:57 PM Post #155 of 458
Looking a bit more closely at the AK4117 datasheet, it isn't technically necessary to digitally control it at all. However, that leaves you with using all default settings which are as follows:

-Low Power mode
-MCKO and SDTO source are recovered by PPL from RX input
-MCKO is 256fs where fs is 44.1kHz
-Clock/crystal used must be 11.2896MHz
-Input is RX0
-Output is 24bit, Left Justified

I'm assuming the SPDIF is only intended to be used for CD playback, so having fs at 44.1kHz is fine. The only thing that differs is the output mode. On the PCM1798, all that requires changing in order to have it accept left justified (as opposed to the I2S from the PCM2707) is to put FMT0 high. Since we are already planning on implementing some logic to cause the mux to switch, it would simply be an extension of this.
 
Feb 9, 2006 at 11:37 PM Post #156 of 458
Quote:

Originally Posted by ble0t
Looking a bit more closely at the AK4117 datasheet, it isn't technically necessary to digitally control it at all. However, that leaves you with using all default settings which are as follows:

-Low Power mode
-MCKO and SDTO source are recovered by PPL from RX input
-MCKO is 256fs where fs is 44.1kHz
-Clock/crystal used must be 11.2896MHz
-Input is RX0
-Output is 24bit, Left Justified

I'm assuming the SPDIF is only intended to be used for CD playback, so having fs at 44.1kHz is fine. The only thing that differs is the output mode. On the PCM1798, all that requires changing in order to have it accept left justified (as opposed to the I2S from the PCM2707) is to put FMT0 high. Since we are already planning on implementing some logic to cause the mux to switch, it would simply be an extension of this.



There is a good chance I'm just not getting something, but, doesn't that make this a less than ideal part to use then? Cause we wanted to use I2S to feed the DAC because it has less jitter, and the CS8416 can output I2S? Otherwise we have a PCM2707 feeding I2S into the mux, the mux feeding the PCM1798, and the AK4117 feeding left justified (left justified what?) into the mux to the PCM1798 and switch both the PCM1798 to accept a left justified input while switching the inputs on the mux?

What about the DIR1703? I can send the datasheet to everyone...?
 
Feb 9, 2006 at 11:55 PM Post #157 of 458
24 bit, Left Justified is just another form of transmitting serial data. It has a different format than I2S, but it isn't necessarily worse. Since we're using a mux, which is in essence a kind of switch, it doesn't really matter what we're sending through it whether it be I2S or something else. We would just need to make sure that the DAC chip is switched to accept either data format.

It's just a thought if we're looking at using the AK4117....
 
Feb 10, 2006 at 3:50 AM Post #158 of 458
Quote:

Originally Posted by ble0t
24 bit, Left Justified is just another form of transmitting serial data. It has a different format than I2S, but it isn't necessarily worse. Since we're using a mux, which is in essence a kind of switch, it doesn't really matter what we're sending through it whether it be I2S or something else. We would just need to make sure that the DAC chip is switched to accept either data format.

It's just a thought if we're looking at using the AK4117....



Ahh.
smily_headphones1.gif
Okay.
smily_headphones1.gif
I thought that it was something like SPDIF, but then I figured it wasn't..

I was just confused.. Thanks for clearing itup.
smily_headphones1.gif
 
Feb 10, 2006 at 3:50 AM Post #159 of 458
Quote:

Originally Posted by ble0t
24 bit, Left Justified is just another form of transmitting serial data. It has a different format than I2S, but it isn't necessarily worse. Since we're using a mux, which is in essence a kind of switch, it doesn't really matter what we're sending through it whether it be I2S or something else. We would just need to make sure that the DAC chip is switched to accept either data format.

It's just a thought if we're looking at using the AK4117....



Ahh.
smily_headphones1.gif
Okay.
smily_headphones1.gif
I thought that it was something like SPDIF, but then I figured it wasn't..

I was just confused.. Thanks for clearing itup.
smily_headphones1.gif
 
Feb 10, 2006 at 10:55 AM Post #160 of 458
I suggest we wait for John Westlake measurements on the DIR1703. It still is my favorite option for now, unless the measurements are very bad.

And don't forget we're not feeding the pcm1798 directly, there's the AD1896 in between the mux and the dac.

Edit : I suggest we start slowly to draw schematics to fill in the operating blocks. And there is the question of the power supply to consider then.
Edit2 : I'll upload a library with some useful parts later on today. It can be improved I guess and must be checked.
 
Feb 10, 2006 at 11:18 AM Post #161 of 458
Quote:

Originally Posted by 00940
I suggest we wait for John Westlake measurements on the DIR1703. It still is my favorite option for now, unless the measurements are very bad.

And don't forget we're not feeding the pcm1798 directly, there's the AD1896 in between the mux and the dac.

Edit : I suggest we start slowly to draw schematics to fill in the operating blocks. And there is the question of the power supply to consider then.
Edit2 : I'll upload a library with some useful parts later on today. It can be improved I guess and must be checked.



I think that it would be relevant to focus on the power supply and the output stage (obviously it implies the components of the DAC to be selected
biggrin.gif
).

From my point of view that's crucials part of the DAC,and they might be more limatating than the converter itself.

Fred
 
Feb 10, 2006 at 1:34 PM Post #162 of 458
Arg...I completely forgot about the little ASRC guy sitting in the middle there. Chalk that one up to a busy day at work...my mind must have been drifting
smily_headphones1.gif


I am also eager to see the results of the DIR1703 measurements. I was surprised when I finally cracked open the datasheet and saw that is uses a SPACT system as well
eek.gif
 
Feb 10, 2006 at 5:47 PM Post #164 of 458
BTW, 00940, nice find on that jitter article over at diyhifi. I guess I would definitely lean toward using the DIR1703 at this point, although I'm anxious to see what JohnW comes up with.
 
Feb 13, 2006 at 11:54 PM Post #165 of 458
Quote:

Originally Posted by 00940
here is the library; replace .txt by .lbr


Found a very small error in it -

In the library, the we have a device named TPS7933, but a symbol for TPS79333.

How did you add these parts to the library?

Cheers,
Clutz
 

Users who are viewing this thread

Back
Top