Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Getting Rid of DAC DC Offset
New Posts  All Forums:Forum Nav:

Getting Rid of DAC DC Offset

post #1 of 7
Thread Starter 
These past couple of days I've been working on putting my Opus DAC into a permanent case. Right now I am just using two caps in line with the left and right outputs to get rid of the DC offset in lieu of any more sophisticated buffer or filter. However, I have 4 JISBOS boards lying around that never became an amp and I want to use them as the output stage for the DAC.

The datasheet for the WM8741 says the output is filtered some already and the TP's IVY manuel says filtering isn't strictly necessary. Additionally I've been using it without filtering for a while now and it has been working well.

My problem then is how to deal with the offset. I've always wanted to take the midrail reference and use it for ground. Could I simply buffer this and the outputs to get DC free, capacitor free output? Could I get by without buffering it if the output were fed into a amp with a high input impedance? Is there a better way to get rid of the offset?
post #2 of 7
you would have to design a dc-coupled buffer if you don't want to use a cap.
post #3 of 7
So you've bypassed the on-board coupling caps?

Pg 62 of the WM8741 datasheet shows Wolfson's recommend circuit. It needs op-amps, not buffers.
post #4 of 7
Quote:
Originally Posted by mugdecoffee View Post
Could I simply buffer this and the outputs to get DC free, capacitor free output?
not that I know of.
post #5 of 7
Thread Starter 
So the output of the left and right channels of theWM8741 is the signal plus VCC/2 (2.5V with a 5V analogue supply). The midrail output is also at 2.5V. What I'd like to due is to buffer this 2.5V output and use that as the output signal ground. This would give me a DC coupled output buffer. However I've never seen any DAC designs that do this even though it seems like an obvious solution to me which makes me think there's a reason not to. I don't plan on recreating Wolfson's recommended circuit for now but if I can't use the JISBOS boards I may just go with the TPA IVY.
post #6 of 7
There is no conceptual reason why you can't use the mid-rail reference voltage and buffer it to create a "virtual ground". But since a DAC is typically used to drive preamps, headphone amps or other active devices rather than passive headphones, you could run into complications. You have to think about what happens in the whole chain of components that (possibly, and likely) share AC ground. Mixing the "ground" metaphors could mean a short circuit from your "virtual" ground to a "real" ground through the interconnect cable...
post #7 of 7
Thread Starter 
Quote:
Originally Posted by amb View Post
There is no conceptual reason why you can't use the mid-rail reference voltage and buffer it to create a "virtual ground". But since a DAC is typically used to drive preamps, headphone amps or other active devices rather than passive headphones, you could run into complications. You have to think about what happens in the whole chain of components that (possibly, and likely) share AC ground. Mixing the "ground" metaphors could mean a short circuit from your "virtual" ground to a "real" ground through the interconnect cable...
So you were absolutely right. I first tried the mid rail virtual ground and intended to solve the different grounds by not connecting the DAC to the ground lug on the IEC socket. However, the USB cable connected the grounds through my laptop. Even when the laptop charger's ground was floating (probably a terrible idea to have so many floating grounds) there was a bad hum.

Next I tried putting a nice film cap between the DAC output and the JIBOS input which worked but it didn't sound all that great and I had some problems with DC on the output of one channel. The caps also went against my original goal of a DC coupled output.

What I finally did was cut a trace on the JISBOS board to use the inverting input (there was a topic about this a little while back - "JISBOS - Future Gainstage" I think it was called) where I connected the inverting DAC output. From there, I added a couple 2K resistors to the feedback look then replaced R1 and R2 (the input voltage divider) with 2K resistors to make the JISBOS into a differential discrete op amp. The resistors/feedback loop were copied from the balanced to single ended converter section of the IVY schematic. So basically I have one JISBOS board per channel whose output is the difference between the positive and negative DAC outputs.

If I hook up headphones to the output directly, they sound very nice. Running the buffered output through my Bijou however gives a very harsh sound with tons of sibilance. Does the JISBOS need a low impedance load to sound best?

My trouble with this project highlighted how little I understand grounding. Does anyone have a guide or article on it? Specifically I don't understand why earth grounding is so important and also what happens when you connect a signal from one board to another if they are running off of separate transformers and don't share a ground.
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Getting Rid of DAC DC Offset