Balanced line driver design for output of AD1973 Codec

Hi Mark,

I'm an IC designer, not an apps guy, so I may be on thin ice here -- Most of our recent amplifier products that produce balanced outputs are tailored for various ADC driver applications, meaning 5v supplies. I don't know if those will work for you since they mostly have slightly less output swing than you need.

The concept of the pair of single-ended filters discussed with you in the audio forum is definitely an alternative. Some folks worry about balance errors due to filter mismatches, so it could depend on your sensitivity there.

Another alternative would be that we do have a balanced audio line driver in the SSM2142. As I recall, this is optimized for single-ended to differential conversion, but may have the output behavior you need. That could be possibly be used after a single-ended filter stage. In figure 33 of the AD1937 datasheet there's a differential to single-ended filter stage that, if followed by a single-ended to differential line driver with ssm2142, might give the best result. It's a bit of a longer signal chain, but seems like it could work.

There might be a way to make a discrete differential filter with balanced output but, if so, I don't think it would be a lot smaller in board area (if that's an issue).

Good Luck,

Dave

Hi Mark,

I'm an IC designer, not an apps guy, so I may be on thin ice here -- Most of our recent amplifier products that produce balanced outputs are tailored for various ADC driver applications, meaning 5v supplies. I don't know if those will work for you since they mostly have slightly less output swing than you need.

The concept of the pair of single-ended filters discussed with you in the audio forum is definitely an alternative. Some folks worry about balance errors due to filter mismatches, so it could depend on your sensitivity there.

Another alternative would be that we do have a balanced audio line driver in the SSM2142. As I recall, this is optimized for single-ended to differential conversion, but may have the output behavior you need. That could be possibly be used after a single-ended filter stage. In figure 33 of the AD1937 datasheet there's a differential to single-ended filter stage that, if followed by a single-ended to differential line driver with ssm2142, might give the best result. It's a bit of a longer signal chain, but seems like it could work.

There might be a way to make a discrete differential filter with balanced output but, if so, I don't think it would be a lot smaller in board area (if that's an issue).

Good Luck,

Dave

Hi Dave,

Thanks for your reply!  In hindsight, we would now prefer a unity gain balanced audio line driver.  We are happy with the +/- 1.27V voltage swings out of the AD1937 as this is close to the maximum input voltage the new amplifier chips we have can accept.

With this new information, do you have a product in mind (preferably a FDA) with differential inputs and differential outputs, to achieve unity gain to drive the audio output over balanced audio cables to an amplifier input with an input impedance of ~ 85 kOhm?  Similar functionality to the SSM2142 but a FDA.

I'd prefer to avoid going from differential to single-ended, and then back to differential again.  I have 24 balanced audio channels and the board will likely become too large if I do that!

Again, thanks for your support.

Kind regards,

Mark

Some of our older FDAs were optimized as balanced video drivers for cat5.. They're similar to FDA ADC drivers and might work? I'd check out the ad8131 & ad8132; they're basically the same amplifier but with fixed gain in the '31. Optimized for video so probably more bandwidth than you need.

Regards,

Dave

Thanks, Dave.  I think the AD8131 might work for me.  Are you able to offer some guidance as to a circuit design based on the information I provided in my original question and most recent reply to you?  

Hi Mark,

Sorry for the delay, it looks like I lost track of this. Are you still looking for guidance on this? I will review what you said in the morning, and discuss with an Apps engineer before responding. It shouldn't take us long.

Regards,

Dave

Thanks, Dave.  Look forward to hearing back from you.  Cheers!