Looks like the audio design team in France is no more after COVID. Cdsgames, the chief audio designer has disappeared. His real name was Ioan, IIRC. Don't know what ever happened to him. We used to have some conversations about dac design and during one period of time he would send prototypes here for listening tests and our comments. Eventually that slowed his development time down too much. I suggested that he at least listen himself since if he could hear a difference then surely some of his customers could too. With some practice he found he could do it. Good for him 
Hi folks.
Has anybody put the Katana side by side to the PI using a ribbon cable or whatever cabling ?? What would that be ?? Any downsides - soundwise?
Why?
I am already using a 90° angle GPIO adapter.
1. I am getting the DAC slightly away from the Pi
2. It lets me run a much better Pi cooling.
3. And I can access the GPIO for attaching some buffer caps and filters
Has anybody put the Katana side by side to the PI using a ribbon cable or whatever cabling ?? What would that be ?? Any downsides - soundwise?
Why?
I am already using a 90° angle GPIO adapter.
1. I am getting the DAC slightly away from the Pi
2. It lets me run a much better Pi cooling.
3. And I can access the GPIO for attaching some buffer caps and filters
Using the stack as intended, I glued a small heatsink to the CPU and to the Katana controller chips. It seems sufficient.
I would also be interested to know whether separation would have more positive or negative consequences. I would like to build the whole thing into a flat housing at some point.
I would also be interested to know whether separation would have more positive or negative consequences. I would like to build the whole thing into a flat housing at some point.
Hey Mark, was going to mention your experience. A useful data point.
User42, a worthwhile experiment, let us know what you hear.
In my experience I2S signal integrity seems to be more important than about anything else. The GPIO connections are already not optimized to maintain integrity of that type of signal, so I expect lengthening them to reduce SQ.
In various DAC implementations using I2S over the last 15 years or so, some using u.fl, others less optimized leads, shorter leads always improved SQ. So my experience says no.
Related, with the Katana, a friend reported improved SQ when on a USBBridge Signature after removing Allo's Isolator board. I heard the same.
#1 - The isolation devices in an isolator board add some jitter to the i2S signals
#2 - The Katana does not reclock at the DAC board so no means to improve lost signal integrity
#3 - The USBBridge Sig is a lower noise source both due no switching DC-DC regulators shared across multiple chips but instead to all chips having dedicated linear regulators, & the Compute Module off to the side of the DAC instead of directly underneath, likely reducing radiated noise at the Katana DAC board
We both used the Katana with all linear supplies & regulators with the on-board DC-DC regulator providing +-15V to the output board turned off, so an already overall low-noise setup.
Of course this is not directly compatible, but perhaps another useful datapoint. In our lower-noise than typical Katana implementations, better I2S signal integrity was more important to SQ than the galvanic isolation of Allo's Isolator board.
If you are using the DC-DC converter for the +-15V to the output board or SMPS's for any of the supplies, I am not sure the differences will be anywhere near as audible.
Looking forward to your report.
Greg in Mississippi
User42, a worthwhile experiment, let us know what you hear.
In my experience I2S signal integrity seems to be more important than about anything else. The GPIO connections are already not optimized to maintain integrity of that type of signal, so I expect lengthening them to reduce SQ.
In various DAC implementations using I2S over the last 15 years or so, some using u.fl, others less optimized leads, shorter leads always improved SQ. So my experience says no.
Related, with the Katana, a friend reported improved SQ when on a USBBridge Signature after removing Allo's Isolator board. I heard the same.
#1 - The isolation devices in an isolator board add some jitter to the i2S signals
#2 - The Katana does not reclock at the DAC board so no means to improve lost signal integrity
#3 - The USBBridge Sig is a lower noise source both due no switching DC-DC regulators shared across multiple chips but instead to all chips having dedicated linear regulators, & the Compute Module off to the side of the DAC instead of directly underneath, likely reducing radiated noise at the Katana DAC board
We both used the Katana with all linear supplies & regulators with the on-board DC-DC regulator providing +-15V to the output board turned off, so an already overall low-noise setup.
Of course this is not directly compatible, but perhaps another useful datapoint. In our lower-noise than typical Katana implementations, better I2S signal integrity was more important to SQ than the galvanic isolation of Allo's Isolator board.
If you are using the DC-DC converter for the +-15V to the output board or SMPS's for any of the supplies, I am not sure the differences will be anywhere near as audible.
Looking forward to your report.
Greg in Mississippi
Thx guys. Seems that not much experience of relevance (beyond my own) on the subject
is available.
The Katana without reclockers/isolators etc made it back in my system after so-and-so
experiences with commercial stuff from e.g. Gustard, DIY IanCanada stuff and a lot more this and that.
Not that these solutions were bad. I consider most of them overkill and much too expensive for
what they offer. And for me overcomplexity of a system is also a NoGo.
Beside that all that HW tinkering has it's limitations. A lot more can IMO be gained
on the streaming (-network) or OS side.
A pity that Allo got of course. For a while - during the early days - they IMO were
on the right track. HifiBerry did a much better job - they remained successfully
in a niche market. They even seem to have survived that massive RPi shortage mess.
Perhaps one day IanCanada feels tempted to go for a rather minimalistic HQ
HAT-DAC solution (filling the Allo gap). He'd know how to get there.
Enjoy.
is available.
The Katana without reclockers/isolators etc made it back in my system after so-and-so
experiences with commercial stuff from e.g. Gustard, DIY IanCanada stuff and a lot more this and that.
Not that these solutions were bad. I consider most of them overkill and much too expensive for
what they offer. And for me overcomplexity of a system is also a NoGo.
Beside that all that HW tinkering has it's limitations. A lot more can IMO be gained
on the streaming (-network) or OS side.
A pity that Allo got of course. For a while - during the early days - they IMO were
on the right track. HifiBerry did a much better job - they remained successfully
in a niche market. They even seem to have survived that massive RPi shortage mess.
Perhaps one day IanCanada feels tempted to go for a rather minimalistic HQ
HAT-DAC solution (filling the Allo gap). He'd know how to get there.
Enjoy.
I noticed Archimago's blog had a review of the Innomaker DAC Pro HAT which I think is a clone of the Katana (Innomaker did the same with Allo's Boss DAC HAT as well). Just wondering if anyone has any thoughts on it? I was a little late to the Pi DAC HAT scene and the Katana seemed interesting so maybe this could be a way to get something close to it?
The blog post is here https://archimago.blogspot.com/2023/04/as-we-hear-it-innomaker-hifi-dac-pro.html
The blog post is here https://archimago.blogspot.com/2023/04/as-we-hear-it-innomaker-hifi-dac-pro.html
Sorry if the question is too trivial, but what is the name / brand of the white plastic power connector on the Katana circuit board ?
There is a corner broken at my connector, I would like to replace it and need to know the name of it.
For me it looks different from a Molex connector, or am I wrong?
Cheers,
Christoph
There is a corner broken at my connector, I would like to replace it and need to know the name of it.
For me it looks different from a Molex connector, or am I wrong?
Cheers,
Christoph
There are things like this: https://www.amazon.com/gp/product/B0731MZCGF/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1
However, they are a type of JST connector. There are a number of different sizes of the pictured series of connectors, depending in pin spacing.
That said, looks like Allo connector may be a little different. If so, don't know if they are compatible with the JST parts.
Some of the different pin spacing sizes look yet a little different from other pin spacings:
https://www.amazon.com/gp/product/B0B2R966ZY/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1
However, they are a type of JST connector. There are a number of different sizes of the pictured series of connectors, depending in pin spacing.
That said, looks like Allo connector may be a little different. If so, don't know if they are compatible with the JST parts.
Some of the different pin spacing sizes look yet a little different from other pin spacings:
https://www.amazon.com/gp/product/B0B2R966ZY/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&th=1