Another analog filter question for DAC

[Sheldon]

I've got a Sony CDP C222ES that I like to try a discreet output stage on. I'd take the output from the DAC, which is a single ended voltage output (CDX9658n). I'll build a complete separate board and power supply and RCA's. I'd leave the current outputs in place for comparison. Attached is the schematic from the DAC to the outputs (minus some series resistors and mute transistors in the output). The op-amp for the sony filter is an njm2114. Any suggestions on filter options, prefereably passive?

Sheldon

Note: the input cap is 47uf. I couldn't edit the file and get it in the right format.

 

[Sheldon]

Someone?

Alright, one of you sharp folks must have a decent guess here?

Sheldon

 

[rfbrw]

Take a look at the schematic for jwb's CS43122 dac linked to below

http://www.diyaudio.com/forums/showthread.php?s=&threadid=6430&highlight=43122

The CS43122 is a voltage out dac like the CXD9658N. In his design, jwb uses a RC filter at the output and follows it with a FET source follower. Not completely passive but it would seem unlikely that CXD9658N was intended to directly drive downstream components.

 

[Sheldon]

Thanks,

I had certainly intended on using a buffer/line amp, if nothing else to get the output to the usual 2v. This version seems straightforward. My DAC is single-ended. See any reason I can't just connect in/out neg. on the follower to com (gnd)?

Sheldon

 

[rfbrw]

Each +/- output stage is made up of two separate followers so you can use one stage for both channels. In the absence of any data on the CXD9658N, the filter values in jwb's dac is as good a place as any to start. Just ditch the lower resistor and ground that end of the capacitor and double the remaining 300R to 600R.

 

[Sheldon]

Thanks,

I thought it might be a differential stage, but now that I look at it you're right, it's two independent followers. If I calculate correctly, it looks like the filter is around 100k. I guess that might be fine for starters. I have a lp in my amp input buffer at about 50k. That should give me enought attenuation above 100k. Easy to change anyway.

Sheldon