DAC DC blocking capacitor

[macgyver]

I'm almost done with my first dac, which is a balanced design based on AK4395. I need dc blocking capacitors in the output of the AK4395, so does anyone have some references for these capacitors? The capacitors should be about 47µF (maybe less, depending on resistors/input load). Cheap way would be to use bipolar or electrolytic capacitors, but maybe I'll go with the MKP/MKT ones (damn they are big).

I'm using OPA1632 opamp in the AK4395 output.
On page 24 is the capacitor which I mean

Datasheet

 

[lucpes]

Most AK4395 chips have under 2mV to no DC offset, so you might get away without the capacitors. If this is not an commercial design, among the best coupling caps I've encountered are old NOS motor run types - they're huge Or get some Black Gates NX type...

 

[macgyver]

Most AK4395 chips have under 2mV to no DC offset, so you might get away without the capacitors. If this is not an commercial design, among the best coupling caps I've encountered are old NOS motor run types - they're huge Or get some Black Gates NX type...

Datasheet says that during the reset there is Vcom voltage (2.5V DC) in the analog output pins.

 

[martin clark]

Can you reference the output stage to Vcom, rather than the DACs ground pin , and so do away with the caps entirely? Just a thought.

I've used, and like, EVOX MMK 47uF caps. OK, still big but they are damn fine sounding when used for coupling.

 

[Leolabs]

Can try to bypass with a smaller value cap.

 

[lucpes]

Datasheet says that during the reset there is Vcom voltage (2.5V DC) in the analog output pins.

Yes, make sure that you switch off the preamp when turning the DAC on, or power the DAC first. Reset time is very short, you will notice it as a medium loud plop in your speakers.

 

[macgyver]

Yes, make sure that you switch off the preamp when turning the DAC on, or power the DAC first. Reset time is very short, you will notice it as a medium loud plop in your speakers.

Yes that's clear. So the analog outputs doesn't have the bias voltage (Vcom=2.5V)?

Datasheet tells, that the analog outputs are centered around the Vcom and the Vcom is a signal ground of this chip. So I'm a little confused about bias voltage of the analog outputs. I think there is 2.5V DC at the outputs. But lucpes says there is hardly any offset...

 

[rfbrw]

Consider this. Given that the dac only has a positive voltage rail, is it possible for it to swing below 0V ?

 

[macgyver]

Consider this. Given that the dac only has a positive voltage rail, is it possible for it to swing below 0V ?

No it's not. That's why it should be biased to half of the positive rail. And that's why we need capacitor to block the DC at the output.

 

[rfbrw]

Problem solved. All you have to do now is decide the type of cap to use.

 

[macgyver]

Problem solved. All you have to do now is decide the type of cap to use.

Yeah. I was confused by lucpes posting. I got some plastic capacitor samples from wima. So the problem is solved.

 

[analog_sa]

Doesn't this DAC have a differential output? A differentially connected opamp will only leave a small offset. Many ways to deal with that.

 

[rfbrw]

Yeah. I was confused by lucpes posting. I got some plastic capacitor samples from wima. So the problem is solved.

You do understand that the capacitors are necessary for the circuit on page 24 to function correctly. The circuit in example 1 on page 23 is shown without capacitors because it does not need them. If you use the '1632 as shown on page 24 you will need the caps and if as shown on page 23 you will not.