Gamma-2 (γ2) DAC Thread - Page 2

Considering I recently picked up the base components and pcb for another gamma1 to fill out an order from you I like this alot!

How far along is the development for the gamma 2? Are the boards ready to be ordered and beta tested or are you still working on it?


About the filtering... the dac chip has a built in digital filter? And the op amp is also being used as a filter? If that is correct then are the op amps strictly necessary? Perhaps their purpose is to further enhance the sound quality?

I also remember someone made a comment like "I think the twisted pear opus with zap filter sounds better than the stock twisted pear buffalo dac". I've heard lots of positive things about the zap filter, for your convenience, I've attached a schematic of it. Incase you could borrow some design ideas to improve the gamma 2 or to just compare the design.

I like!!

Nice work guys

The external filter is low pass only, which is required to filter out the high frequency artifacts of the D to A process. The internal filters, which operate in the digital domain, are much more sophisticated, and operate on the waveform as it is being "created" in the DAC(as I can understand it). Check out the Wolfson web pages for a more complete explanation. The internal and external filters do two different things. These digital filters, have been developed and implemented in very high end DACs, and the WM8741/2 bring their capabilities down to a much lower price range.

Very Exciting stuff Ti, looking forward to hearing more about your new baby as it comes closer to fruition..

*Raises hand to offer assistance if builders are needed for the proto process*

I should also point out that γ2 is designed more as an "integrated" solution with a specific case in mind (even though it could be used in a different case), rather than just a collection of boards. The default γ2 setup would involve no wires.

The WM874x have on-chip digital filters. The WM8741 and WM8742 have multiple filter responses that could be selected. There have been a lot of research into the behavior of different filters and their sonic characteristics. Like many things, each filter has its strengths and weaknesses, so we thought it would be great to allow the user to switch amongst them to hear the differences. It should be great fun.

The analog filter before the output has its corner frequency set further out-of-band (~100KHz), to clean up any residual noise and artifacts without adversely affecting the audio band. As I said, the opamp that implements this filter also serves as balanced-to-unbalanced converter and buffer. We do not recommend substituting other opamps than the two mentioned, they were very carefully selected for their rail-to-rail swing capability. Not only is rail-to-rail required, how close the opamp could swing to the rails is crucial, due to the low analog supply voltage (regulated to 4.5V or 4.75V depending on the voltage regulator used). Using another opamp might cause clipping at or near 0dBFS. These opamps were also chosen for their excellent sound quality.

We have not yet done any beta testing on a real board. Hopefully that would happen soon.

For those who want to read the details about the various digital filters onboard the WM8741 and WM8742, refer to the datasheets, and also to a Wolfson whitepaper. You could also google for more info on the 'net, there are plenty of discussions on the subject.

Looks like I've found my new DAC

Does that mean I will get the Gamma 1 Prototype back some time soon?

Never!!!


I'll try to get it to you soon.

Very nice!. I suppose there are no hard-wired limitations for s/w control with a microcontroller. The Apodizing filters and digital volume control are available through the s/w interface. I think the 3 filters available in h/w mode are: linear phase slow rolloff, minimum phase slow rolloff, and brickwall (filter 1,2 and 3). Filters 4 and 5 are minimum phase apodizing and linera phase apodizing. I did a little summary of the wm8741 filters here and here.

Also have you considered a dual frequency clock? (Silicon Labs makes them). With dual frequency you can upsample to the even and odd frequencies (44.1, 48, 88.2, 96, etc).

glt, thanks for your comments. There is no room on the γ2 board for a microcontroller, nor is there the power budget for it (the whole thing, γ1 and γ2 combined, must run off USB power). We're trying to keep things simple (i.e., no programming, no firmware, etc), so we run the chip in hardware mode. Also, I don't see a need to upsample to any frequency below 96KHz.

Fair enough. I was thinking external, with an option to set the chip in s/w mode and having the pins of the asrc available to switch the sampling frequency...

TPS60403 mounted on the "bottom" of the PCB to create dual rail?

You can consider also the LME49721 (dual, rail to rail CMOS opamp with low distortion and the best sound I've heard from low voltage CMOS opamps), and the OPA2727 (best sound I've heard from a 4V to 12V rail to rail CMOS opamp).

The OPA2365 sounds very good too, anyway. It's a pity, though, that it always makes that big sharp "pop" when powered on (and off). FIY, the OPA2376 sounds good and doesn't make it.