Getting the best out of Allo.com's new Katana DAC...

Member
Joined 2003
Paid Member
@terry22,

LOL, I knew about the I2S being setup with bi-directional lines to support slave and master modes. Heck, I helped edit the User's Guide for the IsolatorPi! ;)

BUT what I don't know if the appropriate GPIO for I2C are bi-directional. That's what the Katana needs for communication between the driver and the microprocessor board and what I need to confirm.

I DO like that the IsolatorPi (like the original Allo Isolator, but NOT Allo's new Isolator) can be powered separately. I have found that to provide subtle positive benefits. But one can't have EVERYTHING!

In other Katana news I'm finalizing the modified Jameco 5V and 12V units for trials powering both the Katana 5V lines and the 5V inputs into the IsolatorPi and Kali under the Ian's ES9028Q2M stack. I also did the first power-ups on the 325F Ultracap supplies for 5V use, along with an alternative on-board regulator for Ian's DAC board. Should be listening in the next few days & have some fun with all that.

I also realized I can use a pair of the 12V Jameco for feeding either my DIY'd +-15V regulator HAT OR +-15V into the Katana directly. The raw DC from those units is about 19V, plenty of headroom for the Sparkos regulators I used. AND I believe (but haven't confirmed) that the regulators they use in those units are adjustable. Once I do that, I'll setup another modifed pair to output both the raw 19VDC and either the required +-15VDC from the on-board regulators or the same voltages via a separate regulator board (I have a Twisted Pear Low-Current Bipolar Power Supply here I can use). SO another option to try AFTER I get another couple of those from Jameco.

BUT while I was upstairs (the system I'm doing this listening in is my larger system in a dedicated room), I started listening to my benchmark Katana setup... AND started getting lost in the music again, listening to a couple of songs, realizing something else I wanted to hear, adding that to the playlist, rinse, repeat... until I was up there about an hour longer than I planned.

Really grooving on that setup!

Greg in Mississippi
 
Last edited:
BUT what I don't know if the appropriate GPIO for I2C are bi-directional. That's what the Katana needs for communication between the driver and the microprocessor board and what I need to confirm.

In order for I2C to go through an isolator at all, the isolator must be bi-directional. So, if I2C works with one HAT, it will work with any HAT, regardless of master or slave mode.
 
Member
Joined 2003
Paid Member
@terry22,

Early on I settled on the minimum phase slow roll-off. I really need to try them all again, but that's what I've been using for awhile. I did re-test the different filter options several times during the evolution of the Katana, but I did not re-test with the last couple of iterations.

Listening to the Katana with some alternative supplies tonight. Not sufficient break-in or listening time to report any findings, but some interesting and promising results so far. I suspect I will be able to report on a good lower-cost (but not universal) power option. AND 2 potentially better power options have not wowed me yet, though are not bad.

More after a few days of on-time and some time for dedicated listening.

Greg in Mississippi
 
Last edited:
Member
Joined 2003
Paid Member
@DimDim,

What you said makes perfect sense. I honestly don't know enough about I2C to comment intelligently. I was just repeating what Cdsgames said in post # 2012 in the main Allo thread:

Both isolators will be available. Old isolator will not work with Katana due to some changes in the i2c (before was unidirectional now its bi)


New isolator , that I am testing presently , will be available in 1 week. It was designed for 2 things...Katana and Boss only.The fact that we dont have to make it compatible with everything out there , made it for a very elegant layout . <SNIP>

Ioan, I'll contact you offline and Ian too to understand if the lines you mention are working the same way on Ian's isolator as yours.

Greg in Mississippi
 
That or maybe output impedance too high over much of the audio band. The regulators in the power supply I used are not all that fast. They rely on output filter caps to help transient response. But electrolytic output caps may have some dielectric absorption that affects step transient power supply performance. The file caps I used don't have that same type of characteristic.
 
Member
Joined 2003
Paid Member
@terry22 / @Markw4,

First, Terry, the Audiophonics unit you listed was also mentioned by @fabio1068 in post #51. I didn't have any experience with it directly, but looking at the comments in a thread he listed it sounded like a good bet. @fabio1068, do you have any updates?

SORRY, I don't have any experience with the other 2.

Second, I was going the mention the MPAudio units as possibles for the +-15V. I DO use his 3||LT3045 boards in a number of places with VERY good results. One of them is supplying the +5V to my reference Katana setup's Isolator output / DAC board feed. ALL of my other applications are 5V or 6V. BUT I have never used any of his for +-15V. So a good possible, but not a recommendation from experience. AND they are not inexpensive.

Third, I agree with Mark, the Super Regulator would also be a good choice. BUT I don't have a good feel for how expensive a build is including the board... $60 or so? Mark, can you chime in?

One thing I try to do is keep my regulation as close to the driven circuits as possible. That's why I did the+-15V board using Sparkos regulators. BUT they are $47 each and another good alternative, the Belleson are $54 each. BOTH are good, I've used both to very good results in the past, but they are not cheap. BUT there's a recent alternative that is less expensive... the new Dexa UMB-2 which are now $26 each from PartsConnexion. I've only used them one time so far, but that build is one of my favorite setups to-date (TP BuffaloIII Pro 9028 with a Legato I/V stage). With smaller sophisticated discrete regulators like the Sparkos, Belleson, and Dexa UMB-2, you can do a small build that fits right on top of the Katana and directly feeds the +-15V input pins. OF course, you still need to supply the raw +-18V to +-23V DC supply, but there are a lot of options for those.

Let us know what you decide.

Greg in Mississippi
 
The biggest expense for populating theJung Didden Super Regulator board is the board itself. Most of the other parts are pretty generic and Jan Didden actually recommends using generic electrolytic caps. I will suggest trying some LME49710 op amps as they perform well there and are more stable than the AD797s. However if you can get a stable AD797, it is superior.
The Super Reg can also be converted into a shunt type for the shunt reg fans. This was one of the final suggestions by Walter Jung in one of his later articles on Super Regs. Unfortunately this article is no longer on the web without authorization.
The impedance levels that these regs can reach requires one to learn scientific notation to appreciate the zeros.

The remote sensing might actually be of significant benefit to the Katana based on what I have read about the routing of the IV supply section requiring an indirect route. This is one of the key parts of realizing the performance benefits of this reg.
 
Lots of builders since the late 90s have documented this. The LME49710 today is a safe bet but the AD797 even in the late 90s showed startup issues with some builders and was no longer recommended for startup issues and being finicky for some.

A series of tests was done by a diyaudio member Jack Walton documenting this and is available for download from linear audio publications. Super Regulator – diyAudio Store
 
Looks like a lot of LDOs. They are low noise, but not necessarily low output impedance. I don't think I would bother with it, but that's just me.

You might try this: Super Regulator – diyAudio Store
and then populate it with parts.

Thank you mark.
I read your post about impedance on the other thread.
Maybe I'll try to do the same fix when you 'll publish a definitive version of your film bank.
I had a look to the super regulator and maybe try to do it.

@terry22 / @Markw4,

First, Terry, the Audiophonics unit you listed was also mentioned by @fabio1068 in post #51. <snip>

Thank you Greg for all the info.

As Fabio already ordered the audiophonics, I ordered the LT1963AEQ module to give it a try. And I will have a look to the sparkos and Dexa regulators.

If I understand it well, shorter the connections from psu or discrete reg to the board, better the SQ will be?

Sorry I meant the one that uses Ti's NE5534. It is +-12v and not +-15v, but on some other thread some user said that it wasn't really critical.
 
Member
Joined 2003
Paid Member
@fabio1068... We look forward to your report!

@terry22... Can you point to the comment on +-12V being ok for the Katana? Or was it a comment that the NE5534 version can be adjusted up to +-15V? The reason I ask is that I don't believe +-12V will work well for the Katana output stage. Between the +-15V power input and the actual opamps, Allo has a pair of regulators providing +14V and -14V, which is what is fed into the opamp board. Feeding them with +-12V AFAIK won't produce good results. SORRY! I hope this doesn't throw a wrench into your plans.

Allo, can you confirm?

@Mikett & @Mikew4... thanks for all the info on the Super Regulators. I have many of the original articles in Audio Amateur and Audio Express. I'm very intrigued about the idea of a shunt Super Regulator. Was that article by Walt Jung in a later Audio Express, Linear Audio, or just published online? Also I remember some of the challenges presented by use of the AD797 in that circuit and the search for good alternatives, mostly from discussion here in DIYAudio.

While I've been aware of the Super Regulator for many years, with the advent of nearly-monolythic-sized regulators such as the Belleson and Sparkos units that SOME say have comparible performance, I've been less motivated to personally pursue the Super Regulators. I DO prefer to place my regulators as close as I can to the powered circuit, which is easier with these smaller units. AND I am willing to spend a few more dollars to buy a ready-made unit in trade against my time to build up and test Super Regulator boards.

Of course, this is only my personal opinion. Super Regulators are very good units, as witnessed by the results reported in the 'ES9038Q2M Board...' thread AND many other places. AND they will be among the best options available AND work VERY WELL powering any of the rails of the Katana.

I also look forward to reports from those using Super Regulators to power the rails of the Katana!

Greg in Mississippi
 
Last edited:
It is +-12v and not +-15v, but on some other thread some user said that it wasn't really critical.

Most IC opamps that work on 15v can also work well on 12v, mostly just with some reduced maximum output swing. Discrete opamps such as are used in Katana may have different power requirements. In any case, the manufacturer's literature and or advice should be followed.