DAC output mods - suggestions please.

Hi all,

I have recently decided to do something about the sound in this unit I have.
I have sketched the basic circuit of the output from the DAC to the RCA on the back of the unit.

I plan to replace the following:
Electrolytics with Blackgate NX (size and space is a premium in this unit),
The mute transistor will become (magically?) a relay,
and the op amps I plan to use are OPA2134 SMD's.

I'm not too sure about the filter (first stage op amp) as it's a funny (to me) configuration. and I may be convinced to leave out a few of the decoupling caps. The first one I think is needed to remove the offset from the DAC output, while the others I think are unneccessary, but correct me if I'm wrong.

I really don't want to re-design the entire output stage, as it's a SMD board (except the volume pot and electros) and I only know the values from carefully measuring them in-circuit or by reading with a magnifying glass.

anyways, here are some pics:


  • alpine 701 circuit_sml.jpg
    alpine 701 circuit_sml.jpg
    22.3 KB · Views: 575
Thanks for the replies.

The offset is too great to omit the 22uf on the input - so that stays.
Is a lower value better, or a higher value? The other three will likely go.

I don't have the LM4562 in a SMD - do you know a cheap source of adapters? - i.e. through-hole IC onto SMD pads? This could be done in the available space, but then there's no chance of changing the caps to polyproylene/mica, etc..

The volume control is an integral part of the functionality, so it must stay for now. If I decide later, then the whole lot will go including the DAC. In this case, the DAC and filter circuit will be made outboard from the box it's in - I'll send the I2S direct to a daughter board with DAC's, etc.

Just to keep things interesting, I'll tell you it's a car audio product which includes active filters with time correction and EQ controlled from the CD transport/head unit. Currently being used in a 4-way active system (3-way + subwoofer) for stereo-only


Paid Member
I'd just get some LM4562 in the soic package, much more mechanically reliable than trying to bodge something together with the dip version.

How about a dc servo to null out the dc and get rid of all coupling caps? :devilr: Or is there way too much dc ahead of that cap? (Like half vcc?)

I second the notion to get rid of the volume control IC, and everything after it.
I was thinking about the DC servo - an excellent Idea, but at this stage I want to keep the circuit as much the same as possible.

really, I'm more interested in changes to component values, or removing/bypassing some parts.

paulebappie/kevinkr - thanks for the suggestion of the LM4562 in SOIC, I'll look into that.
yes, the DC bias of the DAC is 1/2 VCC (or 2.5V). I'll be looking at the 5v supply also, and adding caps to the supply at the op amps.

ohhh, so much to do... I'm not lloking forward to replacing the SMD parts with leggy parts, maybe I'll bypass some of the tracks to reduce the chance of lifted pads ;)


2007-07-30 5:38 pm
What about ripping ANYTHING out behind the DAC and substitute for I nice tube , like a 12B4. Give's about the same gain as the first op-amp ( 6 x ) and with R-loaded and decoupled kathode the Ruit will be approx. 1 K ! beat's every op-amp.
You only need a power supply of 200-250 VDC for the anode and 6,3 VAC for the heater ofcourse !

( I did it )

lol Bouquet ;)

I don't think a 200V supply and tubes will fit into the casing of the device, it's only 20mm thick above the board.

also, the volume control is required.

I have decided in the long term to completely redesign the output section, with different IC's for the volume, and separate balanced DACs - but that will be later.