AMB y2 balanced output module development
Nov 24, 2009 at 3:26 AM Thread Starter Post #1 of 16

mewrei

Head-Fier
Joined
Jul 23, 2009
Posts
87
Likes
0
I've grown to become a fan of Ti's products, and one thing I wished was that the gamma-2 DAC had balanced output. Well fortunately, while browsing the development threads, someone mentioned that the board did have balanced output, and turns out it does in a couple of through-holes (apparently development pins).

So my idea was to build a module for the y2 (yeah yeah yeah, module of a module) that allows for balanced output while not impeding on the normal function of unbalanced output.

Ti mentioned that just airwiring the balanced output bypasses the buffer and low-pass filter stage of the y2 which was really one of the big selling points of the DAC in my eyes at least.

So ultimately the module needs a good circuit. Now I've been mostly a digital up to this point, but I'm looking to break into analog with this (hopefully) not tremendously complex addon, but if possible I'd like to keep the experts close at hand while developing the board so I don't make some big mistakes.

First off (kind of a minor thing, but I just want to get it out of the way), is it worth using the AD797 op-amp? Reason being the entire circuit needs to be designed around the op-amp for it to work properly, and while it'd remove versatility of the unit, it'd certainly sound good and leave the coloring of the audio up to the amplifier.

Second off, designing the opamp circuit for balanced output. This may be the digital-electronics speaking in me but wouldn't you design the circuit to run standard polarity (L+ to + on the opamp, L- to the - terminal, etc) and then run a second circuit in inverted polarity (L+ to the - terminal, and L- to the + terminal), to get a balanced output? Or is there a better way of doing it?

I'd appreciate any help I can get in the endeavour (I'm pouring over opamp documentation right now to work out as many angles as I can in the theoretical).

EDIT: For anyone interested in doing this specifically on the y2, it looks like it can be done rather trivially by tying an output buffer like Twisted Pair Audio's IVY or Counterpoint I/V (the Counterpoint is a linestage for current-output DACs, which I don't believe the y2 is) to the L+, L-, R+, and R- (and ground) points on the y2 board, near U5 and C10, although I haven't verified this.
 
Nov 24, 2009 at 8:06 AM Post #3 of 16
Have you built the y2 already? If not, what about a the Twisted Pear Opus? It uses the same DAC chip and is designed with this kind of modularity/expandability in mind. I suspect you'll find that modifying the y2 will be a clumsy path to your objective, it simply isn't designed with this type of modularity in mind...
 
Nov 24, 2009 at 1:33 PM Post #4 of 16
I looked into the Opus, but the issue was the buffer/filter stage in that I didn't even see one (which was why I mentioned that this was one of the selling points of the y2; and also why I chose not to just airwire the balanced outputs). Its just coupled by a few capacitors, and doesn't have some of the other features which define the y2 (if I'm correct in assuming, I didn't see any kind of preamp for the Opus, and I'm assuming it could be added, but it also doesn't seem to have as much output flexibility as that of the y2)
 
Nov 24, 2009 at 2:28 PM Post #5 of 16
Well good luck with the work I think it would be a fun project for you. The y2 does indeed have easy points to tie into for balanced out, however you do bypass the buffer stage completely using these. So you would need to build that for the balanced out. I think it would be very doable and not a hack at all.

Looking forward to see how this progresses.
 
Nov 24, 2009 at 4:37 PM Post #6 of 16
Quote:

Originally Posted by mewrei /img/forum/go_quote.gif
I looked into the Opus, but the issue was the buffer/filter stage in that I didn't even see one (which was why I mentioned that this was one of the selling points of the y2; and also why I chose not to just airwire the balanced outputs). Its just coupled by a few capacitors, and doesn't have some of the other features which define the y2 (if I'm correct in assuming, I didn't see any kind of preamp for the Opus, and I'm assuming it could be added, but it also doesn't seem to have as much output flexibility as that of the y2)


The Opus in fact has much more flexibility than the Y2 since its modular. Twisted Pear makes the IVY output stage which filters and buffers both balanced and unbalanced outputs. The upside to the Opus though is that you can combine it with any other output stage you find. The main benefit of the Y2 is how small and integrated it is. Adding an balanced output stage probably means you're going to lose the small form factor and go for a bigger case anyways.
 
Nov 24, 2009 at 11:23 PM Post #7 of 16
Quote:

Originally Posted by mugdecoffee /img/forum/go_quote.gif
The Opus in fact has much more flexibility than the Y2 since its modular. Twisted Pear makes the IVY output stage which filters and buffers both balanced and unbalanced outputs. The upside to the Opus though is that you can combine it with any other output stage you find. The main benefit of the Y2 is how small and integrated it is. Adding an balanced output stage probably means you're going to lose the small form factor and go for a bigger case anyways.


Hmm that's interesting, I didn't see the output stage whenever I was looking into it. I'm still a little put off by what looks to be airwiring from module to module (is this the case?) Plus it looks like the IVY boards are out of stock. Wonder if the Counterpoint would work.

Though you are right, it would kind of defeat the purpose of the y2, and would require a new case. Might be worth designing the circuit just for fun, though I may leave it up to someone else to prototype.
 
Nov 25, 2009 at 12:16 AM Post #8 of 16
Take a look at this type of circuit:

400px-Sallen-Key_Lowpass_General.svg.png


It is called a Sallen–Key filter. You could put one of these at each output of the DAC chip. One for + and one for - of each channel. So 4 in total for stereo. The active bit of the circuit can be an opamp, or even more interesting would be a buffer. A simple class-A SE buffer? A diamond buffer?

Also, there is a neat little app that is free called filterlab. You might want to take a look... FilterLab Filter Design Software
 
Nov 25, 2009 at 12:33 AM Post #9 of 16
Quote:

Originally Posted by Mr.Duck /img/forum/go_quote.gif
Take a look at this type of circuit:

400px-Sallen-Key_Lowpass_General.svg.png


It is called a Sallen–Key filter. You could put one of these at each output of the DAC chip. One for + and one for - of each channel. So 4 in total for stereo. The active bit of the circuit can be an opamp, or even more interesting would be a buffer. A simple class-A SE buffer? A diamond buffer?

Also, there is a neat little app that is free called filterlab. You might want to take a look... FilterLab Filter Design Software



I like the idea of the diamond buffer now that you mention it. It adds cost but I don't think anyone who'd be adding on balanced outputs to a standalone DAC cares that tremendously much.

And thanks for the circuit idea. Due to the exponentially increasing size (considering the unit would need 4 discrete buffers) and cost, it might be worth it to just design this as a generic output stage for any unit (y2, Opus, etc).

What about the inductance of airwiring these circuits? Is there a way to correct these issues in the circuit?

EDIT: Looks like the inductance of a 22 awg cable 25cm long would be about 330 nH, which at 20kHz, looks to be about 0.415Ohms (someone PLEASE check my math). This seems reasonable enough to not detract too much from audio quality (but please correct me if I'm wrong).

Is the purpose of the low-pass filter on the circuits to remove these high frequencies to lower inductance?
 
Nov 26, 2009 at 3:08 PM Post #10 of 16
Quote:

Originally Posted by mewrei /img/forum/go_quote.gif
What about the inductance of airwiring these circuits? Is there a way to correct these issues in the circuit?

...

Is the purpose of the low-pass filter on the circuits to remove these high frequencies to lower inductance?



You could use as few as 2 opamps instead. You use differential output opamps to get balanced output. I believe this is what the twisted pear output module does.

The low pass filter is needed as it reconstructs the original waveform from the digital like stepped output. It basically turns the output into a smooth waveform.

I've never heard any problems with inductance of hookup wires so I wouldn't worry about it.
 
Nov 26, 2009 at 6:25 PM Post #12 of 16
Quote:

Originally Posted by Mr.Duck /img/forum/go_quote.gif
You could use as few as 2 opamps instead. You use differential output opamps to get balanced output. I believe this is what the twisted pear output module does.

The low pass filter is needed as it reconstructs the original waveform from the digital like stepped output. It basically turns the output into a smooth waveform.

I've never heard any problems with inductance of hookup wires so I wouldn't worry about it.



I noticed that the IVY uses differential opamps. Part of me wonders about the project being moot though because the output stage would effectively be identical to other projects on the market, and I thought the diamond buffer would be a nice deviation from the norm (and possibly a nice upgrade). That's kind of why I liked your suggestion about it.
 
Nov 27, 2009 at 12:02 AM Post #13 of 16
Quote:

Originally Posted by cobaltmute /img/forum/go_quote.gif
Would a buffer work with no negative input for the feedback loop?


Yes I'm sure it will.

I agree opamps are pretty unappealing. That's Why I posted the sallen key filter. Ideal for a DAC output stage and can be build with a buffer so ideal for hi-fi as well
wink.gif
 
Nov 27, 2009 at 3:22 AM Post #14 of 16
Quote:

Originally Posted by Mr.Duck /img/forum/go_quote.gif
I agree opamps are pretty unappealing. That's Why I posted the sallen key filter. Ideal for a DAC output stage and can be build with a buffer so ideal for hi-fi as well
wink.gif



Well I guess that settles that argument
smily_headphones1.gif


What frequencies would I need to filter with the Sallen-Key Filter?

Also, out of curiosity, how would I use the diamond buffer in a balanced fashion? I'm a bit confused because it only has 1 input and I'm trying to figure out where I need to tie in the filter stage.
 
Dec 5, 2009 at 11:09 PM Post #15 of 16
no one else responded?

You set the corner frequency at around 100 KHz. You need a low pass filter. Use that program I linked to earlier to calculate the values for C and R.

The buffer is just part of the filter. You need to filter each output of the DAC. Both + and - of each channel. So you just build 4 identical filters. Each has as a buffer as part of the design. Attach each of the 4 filters to one of the 4 outputs of the DAC chip.
 

Users who are viewing this thread

Back
Top