How to connect CS8414 to SRC4192? ( 5V->3.3V)

[dieringe]

Hi
I am trying to connect these chips. The CS8414 runs on 5V, the SRC4192 on 3.3V. The SRC is set to 24bit I2S, slave mode, the CS is at I2S out, it drives a DF1704 ok in this config.
I tried to pull down the 5V to 3 by an R-divider, but the SRC doesn't "get it".
Is there a proven way to make this connection, or is it impossible?
Can I run the SRC at 5V input without destroying it?

What I connect is (CS -> SRC):
FSYNC -> LRCKI
SCK -> BCKI
MCL -> RCKI
SDATA -> SDIN
...something wrong with that?

The !RDY pin stays on and there is no output signal.

thanks

[rfbrw]

The SRC4192 is a 3.3volt device and its i/o is not 5volt tolerant. You will have to buffer it. A 74AHC244 with 3.3v rails will do the job.

[gbyleveldt]

As a matter of interest, why don't you rather use the SRC4382? It is the same price as the SRC4192, but has an integrated AES receiver as well. You could even go for the bigger one, the SRC4392. This has, I think, a 4 or 8 way mux, so you can have multiple selectable inputs. This then feeds to the intgegrated transmitter, should you wish. Much more elegant and much easier. I received my samples a few weeks back, but have not had time to play yet.

Cheers

Gert

[ezkcdude]

Why bother with CS8414 and buffering? Just use CS8416 and SRC4192 (or AD1896). I'm not an expert, but that's what I did, and it works fine.

[gbyleveldt]

My point is, you can use a SRC4382 or SRC4392 and do away with the CS8414 or CS8416 altogether! They already support AES, so you don't have to deal with the CS part at all. Plus, the SRC4382 costs the same as the SRC4192!!! In total, it'll be cheaper going the SRC4382 route and use less chips/board space.

[ezkcdude]

Everyone has to make their own decisions about these two things, but for me, here's two reasons I would not use SRC4392:

1) Don't want to deal with the quad package.
2) Must be programmed by logic or microcontroller.

I think this approach will end up being much more complex, and probably take up as much, if not more, room than using CS8416, which can be programmed by pull-up,pull-down resistors.

[gbyleveldt]

Interesting point on 2, I do a lot of MPU work so I don't even consider that as a problem, especially if you consider the benefits gained from using an MPU.

To be honest, I believe that if you going to put the effort into something like this, you should go all the way. MPU's are really not that hard to use. People are prepared to battle for days with digital audio interfaces, but as soon as a MPU is needed, they shy away from it. WEIRD!! Anyway, if you want to control the digital interface you can do it with an 8pin micro for around $1 instead of a CS for $5...

[ezkcdude]

You should realize everyone comes into this hobby (and it really is a hobby for most of us) with different levels of skill/knowledge. For my first DAC project, a microcontroller was way over my head. Now that I've successfully completed that project, I am interested in learning more about controllers, FPGA's, etc. But, if one doesn't know much about digital logic, I would estimate a good 6 months to get up to a working level of knowledge on the subject.

[dieringe]

thanks for your answers.
I will not go MPU, because that is another part and I can't do the programming
(just what ezkcdude is saying).

I could use a DIR4096 or something as 3.3V receiver, but I would rather do it with the parts I have.
Don't know how/where to get the CS8416...

My question was: is it impossible to get the signals across with pulldown-resistor-logic?

Oh wait... the AD1896 can take 5V input? I will just wait for my sample!

[ezkcdude]

CS8416 is available at digikey and is actually much easier to get than CS8414, which is now obsolete. That's actually the main reason I would use it - aside from the voltage issue. If it helps take a look at my web page. I had to go through all these decisions in the last several months.