Katana runs on "Master " ( i think this is the first time that ess sabre runs on master) no Kali is necessary. No isolator is on board sorry but having a second ground (that cannot overlap with dac ground because you are creating a capacitor ) its impossible to do . We cannot swap ICs , power is different and IC footprint as well . Everything was optimized for ess sabre 9038q2mControl and voltage board!?!?
What's that for? Do we still need external power supplies?
The seperate output-stage board might be a nice approach. It'd perhaps allow to add another output stage in the future. We discussed this earlier. That's potentially quite nice and (hopefully) keeps the solution flexible.
I guess there won't be a need for a Kali or any other isolator board or similar!?!?
What I also like with this approach is its flexibilty. Perhaps there'd even be a chance to swap just the DAC chips (Q2M->Pro) and still use the other boards in the future.
Obviously all that will come at a cost. The Audiophonics ES9028Q2M
sells at 129€. And the Ebay ES9038Q2Ms at 40$.
Lets see how that will look for Katana.
Last edited:
I'm an Audio Fanatici, and I use Kali with high quality dacs like Buffalo, Soekris and Pcm1704 based. At today no Raspberry dedicated dac have a top audio quality only good quality. Your new product could be interesting only if it'll be top quality and the price must be in that League, no more. One or ten boards is not a problem, only top audio quality is the target. There are too much good raspberry audio dacs at 40 euro we don't need them.
Ciao
Guglielmo
Ciao
Guglielmo
Copying a post I just made on a different thread :
Top quality and HIFI sound was absolutely our priority . Price was secondary (but still it has to make sense)
""An output stage made of discreet opamps ( Class A... with a virtual ground for the ess sabre to feed current into...is basically ideal. I would suggest to peruse the website of Sparkos labs to understand why discreet opamps are better than monolithic opamps .
Of course you need very clean -+ 15v for everything to work. Right now we are in the middle of tweaking the DC/DC convertor (yeap we have RC snubbers across the output and on the input) to stop the RFI and EMI . We got 1.4mv ripple and noise thats unacceptable to us. Good news is that MCLK jitter is lowest we measured on a DAC.
In the end , allo is not (and never was ) a company that overcharged customers in the name of HIFI gods. Rest assured that Katana will reflect the true BOM ...but yes its ALLOs hi end dac.
One last thing. On the analog side, we have 2 super caps fallowed by 2x3 film capacitors LR (thx Greg for pushing us )..plenty of electrons indeed. But it was not enough....so we used the latest technology in supercaps to add one on the mains that can give 3A of power..again plenty of electrons on the PCB. I think that if you remove the PSU , it will still work for 30s at least.
So here you have it, plenty of power Master Ess Sabre (so 2 Xtals , thats a first) , with class A output opamps and great jitter.
Lets see how it sounds when we finished debugging.
Top quality and HIFI sound was absolutely our priority . Price was secondary (but still it has to make sense)
""An output stage made of discreet opamps ( Class A...
with a virtual ground for the ess sabre to feed current into...is basically ideal. I would suggest to peruse the website of Sparkos labs to understand why discreet opamps are better than monolithic opamps .Of course you need very clean -+ 15v for everything to work. Right now we are in the middle of tweaking the DC/DC convertor (yeap we have RC snubbers across the output and on the input) to stop the RFI and EMI . We got 1.4mv ripple and noise thats unacceptable to us. Good news is that MCLK jitter is lowest we measured on a DAC.
In the end , allo is not (and never was ) a company that overcharged customers in the name of HIFI gods. Rest assured that Katana will reflect the true BOM ...but yes its ALLOs hi end dac.
One last thing. On the analog side, we have 2 super caps fallowed by 2x3 film capacitors LR (thx Greg for pushing us )..plenty of electrons indeed. But it was not enough....so we used the latest technology in supercaps to add one on the mains that can give 3A of power..again plenty of electrons on the PCB. I think that if you remove the PSU , it will still work for 30s at least.
So here you have it, plenty of power Master Ess Sabre (so 2 Xtals , thats a first) , with class A output opamps and great jitter.
Lets see how it sounds when we finished debugging.
Cleaning the dual rails DC/DC convertor proves to be titanic work... our goal is to make something comparable to " The silent switcher".
The SilentSwitcher | Linear Audio NL
The SilentSwitcher | Linear Audio NL
Last edited:
Ok. It seems all you guys eat what's being put on the table.
Just to remind you - this is DIY-Audio!
What I did the day before yesterday, was adding two 33uF Blackgate NPs that I had around to the AVDD pins (both DACs).
Actually pretty much the same exercise as done with all my other HAT-MOD projects.
After two nights listening to it - all I can say is - this mod stays! The AVDD
pin is a key pin. Allo has taken much more care on that one on the BOSS
DACs. Treating this pin properly can make a helluva difference (in audiophile terms).
Perhaps there's more to gain, I'll see.
Enjoy.
Just to remind you - this is DIY-Audio!
What I did the day before yesterday, was adding two 33uF Blackgate NPs that I had around to the AVDD pins (both DACs).
Actually pretty much the same exercise as done with all my other HAT-MOD projects.
After two nights listening to it - all I can say is - this mod stays! The AVDD
pin is a key pin. Allo has taken much more care on that one on the BOSS
DACs. Treating this pin properly can make a helluva difference (in audiophile terms).
Perhaps there's more to gain, I'll see.
Enjoy.
Congrats! This is one proper board..
I'm sure that the other 2 "levels" will be just as well designed and built.
Thx you Dimitris
The biggest challenge was/is +- 15V. Lots of noise in a DC/DC convertor (thats feed by smps . We are down to 8mV at the output on both rails (of course we have further filtering before opamps)
BTW. We are the first in the world to run the ESS Sabre in Master mode
Last edited:
Just curious - will the Kali 2 product get released.
I recently built a Twisted Pear Buffalo 9038Pro dac, XMOS usb-I2S-Kali-Dac. I can't explain why the Kali makes a difference but it seems to help. Absolutely awesome sound.
So, of course, wondering how I can break it - er, improve it! Is it worth replacing the Kali clocks with CHD957 or should I wait on a new Kali?
I recently built a Twisted Pear Buffalo 9038Pro dac, XMOS usb-I2S-Kali-Dac. I can't explain why the Kali makes a difference but it seems to help. Absolutely awesome sound.
So, of course, wondering how I can break it - er, improve it! Is it worth replacing the Kali clocks with CHD957 or should I wait on a new Kali?
Thx you Dimitris
The biggest challenge was/is +- 15V. Lots of noise in a DC/DC convertor (thats feed by smps . We are down to 8mV at the output on both rails (of course we have further filtering before opamps)
BTW. We are the first in the world to run the ESS Sabre in Master mode
Cleaning up switching noise is not easy because at the same time you need to keep a decent power supply output impedance.. it helps if your op-amps have decent psrr..
Now that you have gotten the hang of writing drivers for Master mode DACs you guys are definitely ahead of the pack.. And you are definitely the first to get the 9028/9038 to work in Master mode.
About -110 db of psrr at audio frequency Even 5mv changes to something thats a few uV...Andrew from sparkolab has a nice whitepaper explaining all that.For the impedance...one of the biggest advantages of discreete opamps is that every fet can have its own decoupling capacitor next to it. From what I understood you dont even need decoupling capacitors (but still we included some extra capacitance) on opamp board.
However we were not worried about noise going to opamps only. The hi frequency sw can couple/radiate everywhere, DAC ic etc. We wanted to eliminate the sw noise at the DC/DC ic. Extra bonus is quiet voltage rails..
Yes we have tested master mode (working) , sw song frequency (working) 384Khz (working) , dop (working), volume control (working) , filter changes (working)
Still challenges exist. Some noise when sw from PCM to DOP. I am sure we will find a few extra bugs.
At last we want to further decrease the noise on each rail (and we have about 20 of them. About 10 have been measured and we are testing dampening networks , increasing inductance etc.
So still a lot of work...but we are much closer than we hoped to be at this time.
Looks very nice!
looking at the high quality connector terminals, pls. also use such for connection to the Raspberry Pi.
I think they are not compatible