Mr White's "Opus", designing a simple balanced DAC

[fierce_freak]

I'd love to try the Ballsie out...in fact, once you guys get the pre-order or ordering up for the new Opus modules and the Ballsie, I'll probably spring on a set. I chose to go with the zapfilter for now because I've used one before and liked what it did for the DAC I used it in and because I really like its design (I actually haven't really looked at Ballsie yet).

A thought that just occured to me...I forgot to bypass the caps on the output of the DAC boards. I can't imagine that would have anything to do with the hum issue, but at least having to deal with the hum issue allowed me to remember to short the caps.

-edit- btw, the opus with the zapfilter (aside from the hum) sounds flippin' fantastic. some stuff sounds spookily real coming from it.

[wigginjs]

I'd be really interested to hear impressions of this DAC w/ and w/o the Zapfilter (and w/ and w/o the Ballsie too infact).

[wigginjs]

"The I2S bus separates clock and data signals, resulting in a very low jitter connection."

http://en.wikipedia.org/wiki/I2S

Hmm......

[neb001]

Originally posted by m0tion
I'd be really interested to hear impressions of this DAC w/ and w/o the Zapfilter (and w/ and w/o the Ballsie too infact).

Same here, I'm curious to hear how the Ballsie compares vs the Zapfilter. With my limited understanding, it seems the two are basically doing the same thing.

Btw, what exactly does the filter in the Ballsie do? I've been trying to access the wolfson datasheet to find out, but for some reason can't connect to their site right now.

[error401]

Originally posted by neb001


Same here, I'm curious to hear how the Ballsie compares vs the Zapfilter. With my limited understanding, it seems the two are basically doing the same thing.

Btw, what exactly does the filter in the Ballsie do? I've been trying to access the wolfson datasheet to find out, but for some reason can't connect to their site right now.

The datasheet doesn't explain the circuit at all, nor does it provide much information on the internal lowpass.

Without analyzing the circuit in question, I would guess that it's a lowpass tuned to a lower frequency than the ultra conservative (to keep responses flat) builtin analogue lowpass filter. This would serve to filter out additional digital noise and improve SNR and THD+N. How much, I'm not sure. I'd guess a few dB SNR at least.

My own question: how does using the builtin filter scaling affect sound quality? Are these digital filters going to lose bits when attenuating the signal, or are they past the stage where that matters (I'm really not totally clear on delta sigma operation...but is this filtering process not what converts the 24bit input to 6bits for the DAC itself anyway)? If so, with the impressive specs, I may add a passthrough input to my preamp project and seamlessly control the DAC IC instead of the PGA2310...

[Russ White]

Hi Folks,

Just a couple notes:

1) The main design goal for ballsie was to provide balanced and SE outputs at the same level from the DAC, and to allow for filtering all while presenting identical load to both parallel DAC outputs when used in dual differential mode.... phew... One other design goal was to be able to drive SE and BAL simultaneously from just one DAC while still providing optimal loading to the DAC and not compromising either signal. If you drive a Balanced amp from an Opus and tap one of the outputs(say out+ from each side) for something like a subwoofer amp, then you are loading the outputs of that channel differently. This sorta defeats the purpose of a balanced output.

2) Ballsie is not exactly the same filter as either of the filters (there are two) found in the datasheet (but very close). It is actually a hybrid of the two. I will graph the response and post in a bit.

3) I do not find the ballsie filter necessary at all, but some may like it better, so I chose to include it. In fact the DAC sound great just directly coupled to an amp using the DAC's internal volume control. I look forward to some long tests with both the filtered and unfiltered ballsie though.

4) The digital volume control works extremely well and does not appear, at least to me, to have any serious drawbacks. In fact I would say I am finding the presentation possibly a bit more dynamic then it was going through an attenuator. Now that said, you should know this... ANY analog device has a certain degree of noise present on the output. A DAC is no exception. One benefit of using an attenuator and/or preamp is that the output from the DAC is always full scale. So the audio signal is present at a much larger level than the natural device noise. To comfortably listen to the DAC you have to attenuate the signal, this attenuates both the noise and the audio signal and since the noise is already tiny and it will get even smaller you get lower noise floor with an attenuator than without. When you use the digital volume control you loose this "noise attenuation", the signal is attenuated, but not the noise. Now, this is not actually very worrisome as my testing so far proves that the noise output of the DAC even direct couple to an AMP is very very low. I can't hear it. These facts are also true of other analog devices (say a phono stage). What it means to you, potentially, is if you goal is the absolute lowest possible noise floor at very low listening level you will likely want an external attenuator like the Joshua Tree. Wow thats a long paragraph.

5) I spent yesterday doing some A/B tests with both the slow filter roll-off and the normal roll-off. So far I can't say I like one more than the other, but I am going to leave it in slow roll-off mode for a week or two, then switch back to normal and see if I detect a difference then.

Cheers!
Russ

[Russ White]

When configured with the suggested filter caps this is what the response of the ballsie should look like:

The signal is 3db down at 550Khz

Cheers!
Russ

[fierce_freak]

An update on my hum problem: I've spent hours trying to troubleshoot this. It seems the hum only appears when there's a connection between - phase signals from the DAC to the zapfilter. In every other case, there's no hum. This even happens when I pull out one of the DAC boards and run the remaining one regularly. Any other ideas?

[Russ White]

Originally posted by fierce_freak
An update on my hum problem: I've spent hours trying to troubleshoot this. It seems the hum only appears when there's a connection between - phase signals from the DAC to the zapfilter. In every other case, there's no hum. This even happens when I pull out one of the DAC boards and run the remaining one regularly. Any other ideas?

Sounds very frustrating... I know the feeling. It is still very likely a case/PS/device loop of some sort. Unfortunately its pretty much impossible to diagnose remotely.

One thing to do would to try grounding -in on the ZF. See if you still get any humm.

[wigginjs]

Hmm... Another very interesting thing. The WM8804 S/PDIF receiver has a PLL and re-generates the clock of the S/PDIF signal it receives based on a local crystal. Maybe S/PDIF signal transmitted in this manner is not quite so bad?

Below is interesting reading on a similar, but more complicated method of attempting to correct S/PDIF jitter labeled "clock injection":

http://peufeu.free.fr/audio/extremist_dac/spdif.html