DAC + bridged gainclone idea...

[MWP]

Hello all,

Is there anything conceptually wrong this idea?



The DAC will be a BB item like the PCM1730, PCM1794, PCM1798, etc.
I realise there is no volume control, but that isnt a problem.

Thanks!

[MWP]

Originally posted by UrSv
Could it be that the DAC plays at a fixed level meaning that you will play at maximum all the time without possibility of adjusting the volume?

Yup.... but thats why i said to forget about the fact it has no volume control
Ill add a simple course resistive relay switched vol control between the I/V opamps and the LM3886, then use digital volume control before the DAC for the rest.
But that would have overcomplicated the schematic, so ive left it out for the moment.

I havnt built this yet... i would just like confirmation from some of the other brains here before i spend the time making the PCBs, building it, etc.

[borges]

I like your idea. Actually, I've been thinking about the same kind of thing, only with a digital crossover filter. But before I get there I have to get my design to work properly as a DAC.

I have my own oversampler up and running in a Xilinx FPGA, and modifying the filter coefficients for xover use shouldn't be that hard. If you trust the oversampler of the PCM1794, it should be possible to receive a digital 16bit/44.1 mono signal, filter it and present a "stereo" 16bit/44.1 signal to the DAC with one channel for each speaker element. Personally, I oversample on my own, so I'll need one converter for each element (no-df mode or whatever they call it.)

But you will still need some kind of volume control. I haven't yet started looking at digital potentiometers and programmable-gain amplifiers yet. But tell me if you find some promising ones!

Greetings,

Børge

[MWP]

Well this will actaully be used with digital xovers.
But from a PC (Linux + BruteFIR).
So, ill be making 3 of these PCBs (stereo out each) for Sub, Tweeter L&R, Low L&R.

Ive now decided ill be using PCM1798's in mono mode for the outputs.
The I/V opamps will be OPA627/637s (eek, ill need 12 of them!).

[banana]

TI suggest similar config with their new TPA6120 headphone amp IC. If it works on headphone amp, I don't see why it doesn't work on speakeramp.

TPA6120 datasheet

Even wilder idea in my brain. I was wondering using chip amp as the I/V directly!

[MWP]

Originally posted by banana
Even wilder idea in my brain. I was wondering using chip amp as the I/V directly!

Yeh, i thought about this too, but it would mean the LM3886 would need to have quite high gain and very limited low-pass filtering.

[edit]
BTW, thanks for the link to the datasheet...

[mdlover]

i'm thinking about the same thing and the design is almost done.

the implementation is like what Banana have mentioned.

[MWP]

Originally posted by mdlover
i'm thinking about the same thing and the design is almost done.
the implementation is like what Banana have mentioned.

Care to put a schematic up?

[MWP]

Oh.

One thing ive just thought of is that there is a Vcc/2 (2.5V) offset on the DACs current ouput pins.
With other DACs there is a VCOM or similar that is used to cancel this at the IV converter opamps.

With the PCM1798 there isnt such an output pin... they assume that the -Ve and +Ve offsets will be canceled at the ouput opamp.

To you guys have any ideas of how i could cancel this out without using series caps?

[banana]

Hi MWP,

Check out page 13 on AD1955 datasheet.

AD1955 datasheet

The concept is to inject DC nulling current into I/V's virtual ground summing node. Either current sourse or pull up resistor works. Or maybe servo works too, I suppose.

I've used the pull up resistor method in my AD1853, works great.