Dual I/V stage implementation for Buffalo II & III

[squizcat]

Do, I'm very interested to hear more on your efforts. I am eagerly awaiting delivery of a Buffalo III board and Tube-I-Zator boards. Have been mulling the benefits of getting a second Buffalo III board for the dual mono configuration. Would love to see some shots of your build?

[qusp]

Originally Posted by pinnocchio

Hi Qusp,

I do appreciate the feedback but if you could also provide some positive feedback, maybe tell us how you would do it instead of just pointing out the bad.

If using relays is not good then there must be tons of people wasting their times building / implementing R2R volume control using relays...

After all, this way of thinking would have never brought us to the moon!

Please don't be a show stopper and see the positive side of anyone interested to use an I/V stage of their liking. I do love the Legato / IVY stages but there's plenty of people that won't give up their tube I/V for SS output...

Lets just be innovative and find ways around this issue.

Thanks for your understanding

you are missing the point, relay volume controls are dealing with voltage, not current and volume controls are designed to lower dnr by their very nature, that is not what you want from an IV stage.

i'm seriously happy you love the sound of your dac, thats what its all about, i really am, but the way i see it building an entire second dac to gain a few db performance and then throwing it directly away before the signal even gets anywhere seems a total waste of time and effort, i dont like to see people make such decisions, so i speak up when called for comment, but hey its your money. i think you could probably get better performance with one dac done properly than making compromises with 2. doing the dual mono thing, the only reason to do it is to squeeze the last bit out of the sabre dac, whether this has anything much to do with great sound is subjective

myself (and btw the way the designer of the sabre dac sees it) short line of sight current paths that keep the phases as close to each other as possible and direct signal returns are to be maintained for good performance from an iv stage with sabre. that simply is not possible with the setup you mention and its not possible with ANY setup that uses relays in the current path, the resistance of the relays will become the IV stage, sabre starts falling out of current mode with <1ohm

[pinnocchio]

Originally Posted by qusp

you are missing the point, relay volume controls are dealing with voltage, not current and volume controls are designed to lower dnr by their very nature, that is not what you want from an IV stage.

i'm seriously happy you love the sound of your dac, thats what its all about, i really am, but the way i see it building an entire second dac to gain a few db performance and then throwing it directly away before the signal even gets anywhere seems a total waste of time and effort, i dont like to see people make such decisions, so i speak up when called for comment, but hey its your money. i think you could probably get better performance with one dac done properly than making compromises with 2. doing the dual mono thing, the only reason to do it is to squeeze the last bit out of the sabre dac, whether this has anything much to do with great sound is subjective

myself (and btw the way the designer of the sabre dac sees it) short line of sight current paths that keep the phases as close to each other as possible and direct signal returns are to be maintained for good performance from an iv stage with sabre. that simply is not possible with the setup you mention and its not possible with ANY setup that uses relays in the current path, the resistance of the relays will become the IV stage, sabre starts falling out of current mode with <1ohm

Hi Qusp,

Would switches create the same scenario?

thx
pinnocchio

[pinnocchio]

Originally Posted by squizcat

Do, I'm very interested to hear more on your efforts. I am eagerly awaiting delivery of a Buffalo III board and Tube-I-Zator boards. Have been mulling the benefits of getting a second Buffalo III board for the dual mono configuration. Would love to see some shots of your build?

I will post pictures of the prototype shortly. Everything will be installed in a nice box. But I'm waiting on Brian and Russ to release the AC2 (Audio controller 2) before I do this, since I want to create a nice front panel using Front Panel Express!

Thanks

[pinnocchio]

Ok...

I'm disapointed but it looks like this thread is dead...

If you need help with a dual mono BII and external I/V just PM me.

Thanks

[jcarr]

Hi Qusp:

You mentioned the importance of low parasitics in the link from the 9018 chip to the I-V stage, which I fully agree with.

To give that point more attention, I'd like to ask your opinion (or experience) as to how much resistance, capacitance or inductance the link can have before the 9018 performance starts showing signs of degradation.

I've read elsewhere that ESS themselves strongly suggest that the link or current injection point of the I-V stage not exceed 10 ohms impedance, wide-band. But do you think that there will be a steady increase in performance with reduced impedance at the current injection node, or do you think that beyond some point there will be a plateau? As an extreme example, would 0.2 ohm input impedance allow superior performance to 0.6 ohm impedance, or is it that below 1 ohm (or 10 ohms, for that matter), performance differences are more or less inconsequential?

I'd appreciate hearing your thoughts.

cheers, jonathan

[pinnocchio]

Originally Posted by squizcat

Do, I'm very interested to hear more on your efforts. I am eagerly awaiting delivery of a Buffalo III board and Tube-I-Zator boards. Have been mulling the benefits of getting a second Buffalo III board for the dual mono configuration. Would love to see some shots of your build?

Hi squizcat,

For the Tube-I-Zator pics, please go to this thread Tube-I-zator Professional PCB

Thanks
Pinnocchio

[pinnocchio]

Originally Posted by jcarr

Hi Qusp:

You mentioned the importance of low parasitics in the link from the 9018 chip to the I-V stage, which I fully agree with.

To give that point more attention, I'd like to ask your opinion (or experience) as to how much resistance, capacitance or inductance the link can have before the 9018 performance starts showing signs of degradation.

I've read elsewhere that ESS themselves strongly suggest that the link or current injection point of the I-V stage not exceed 10 ohms impedance, wide-band. But do you think that there will be a steady increase in performance with reduced impedance at the current injection node, or do you think that beyond some point there will be a plateau? As an extreme example, would 0.2 ohm input impedance allow superior performance to 0.6 ohm impedance, or is it that below 1 ohm (or 10 ohms, for that matter), performance differences are more or less inconsequential?

I'd appreciate hearing your thoughts.

cheers, jonathan

Hi Jonathan,

Did you ever got an answer for that? I'd would really love to do this project but right now I'm kinda stuck not knowing...

Thanks
Do

[jcarr]

Hi Do: No, I haven't received a direct answer from anyone with hands-on 9018 design or build experience.

What I have been doing is working together with a buddy on a balanced-to-SE I-V converter design for the Buffalo II and III, that eschews monolithic devices, and has built-in DC cancellation and therefore requires no coupling capacitors. The overall topology assumes true differential outputs from the 9018, which means that the quantizer should be run in up to 8-bit mode rather than 9 (unless someone reading this knows how to configure a 9018 for true differential outputs when the quantizer is set to 9-bit operation).

Our present overall design is complex with multiple interlocking variables, and it isn't always easy to deduce that topological fragment A is directly responsible for distortion reduction B or phase improvement C.

Keeping those cautionary words in mind, at least down to about 0.2 ohms (achievable with fairly straightforward, although not necessarily simple topologies), it appears that there are direct and quantifiable improvements in distortion. Further reducing input impedances is requiring weirder topologies, and possibly as a result, doing so seems to be bringing diminishing degrees of distortion improvements.

At the end of the schematic design process we will then need to consider how to package the final circuit into a compact enough form so that the input impedance improvements aren't negated due to excess resistance, capacitance and inductance brought by a circuit that is physically too-large (or runs too hot).

kind regards, jonathan carr

[pinnocchio]

Originally Posted by jcarr

Hi Do: No, I haven't received a direct answer from anyone with hands-on 9018 design or build experience.

What I have been doing is working together with a buddy on a balanced-to-SE I-V converter design for the Buffalo II and III, that eschews monolithic devices, and has built-in DC cancellation and therefore requires no coupling capacitors. The overall topology assumes true differential outputs from the 9018, which means that the quantizer should be run in up to 8-bit mode rather than 9 (unless someone reading this knows how to configure a 9018 for true differential outputs when the quantizer is set to 9-bit operation).

Our present overall design is complex with multiple interlocking variables, and it isn't always easy to deduce that topological fragment A is directly responsible for distortion reduction B or phase improvement C.

Keeping those cautionary words in mind, at least down to about 0.2 ohms (achievable with fairly straightforward, although not necessarily simple topologies), it appears that there are direct and quantifiable improvements in distortion. Further reducing input impedances is requiring weirder topologies, and possibly as a result, doing so seems to be bringing diminishing degrees of distortion improvements.

At the end of the schematic design process we will then need to consider how to package the final circuit into a compact enough form so that the input impedance improvements aren't negated due to excess resistance, capacitance and inductance brought by a circuit that is physically too-large (or runs too hot).

kind regards, jonathan carr

Please let me know when you have this I/V completed

If you need some testers, I have some good gear to test it on.

BTW, someone asked me if instead of using relays to switch the BII/BIII current output to Legato or Tube I/V to use them for switching the power of the desired I/V stage and connect both I/V in parallel at all time. This way we could have really short traces.

The only thing I'm not too sure is if one of the I/V stage is powered off, current will still flow in it and go through passive components. I don't know if this could have an issue with the BII current output...

So basically, have the BII and Legato 3.1 sandwiched together like they're supposed to then run an extra set of cables to the tube?

Thanks
Do