Yes exactly, so I did it too on my chinese 9038q2m and the SQ improvement is really noticeable.
I just mean not using the usb plug and soldering directly PSU wires to katana microcontroller pcb like Greg did.
Hi terry22,
So far, I don't think cdsgames is convinced there just might be a better way. IMHO only, the difference in this case between what he and I think about dac design is that I have heard Katana and got it to sound as good as I good, and I have heard my modded dac which I have also got to sound as good as I can so far (although with a best-sound stability issue I think is very close to solved). Don't know if cdsgames has tried film caps with Katana yet, and I am sure he has not heard the modded dac. With the two of dacs sounding the best I could get them, the modded dac equaled or beat Katana in every aspect of sound quality I could hear. I attribute the difference mainly to maybe taking a little pretty inaudible hit on jitter (not sure about how much of one if any the way I am doing it) and pick up more benefits by upsampling to DSD. So far as I can tell, it works by both reducing jitter and by avoiding use of Sabre PCM interpolation filters, those 7 or so different filters one can choose, but that never seem to be just right. The other thing about that approach is the dac can work with RPi, USB, or SPDIF. And it is still possible to play the highest sample rates the dac chip can do, as well as native DSD. However, playing some sample rates higher than AK4137 can go might require bypassing its functions.
Eventually, if anyone besides me ever finishes all the mods, and there is lot of work involved, we will start getting some other reports of how it sounds. Of course, in commercial form, I think it would probably cost somewhat more to make than Katana, so any benefits would probably not be free.
The above having been said, I still recommend Katana as the best low-cost high-sound-quality dac I know of to anyone who does not want to mod. Katana with linear supplies and film caps is very, very good and probably the best dac many here will have ever heard.
Last edited:
Hi
I have no doubt about what Mark and Terry are reporting . Upsampling to DSD seems beneficial in SQ.
That being said , there are many paths to great SQ. I am a firm believer in low noise (rails) and low jitter. Katana was designed on that philosophy .
Still I think that ess9038q2m can reach even higher as per Greg , Mark reports . Its possible to improve another few percentages .
However you will need a properly designed PSU to go along with it.. (15v dual rail and 5V) LPS..
So efforts to bring the "best" 9039q2m must be divided between hardware implementation , analog stage , digital filters , jitter reduction and low noise power They are all important . Such DAC will cost easily 750$
I have no doubt about what Mark and Terry are reporting . Upsampling to DSD seems beneficial in SQ.
That being said , there are many paths to great SQ. I am a firm believer in low noise (rails) and low jitter. Katana was designed on that philosophy .
Still I think that ess9038q2m can reach even higher as per Greg , Mark reports . Its possible to improve another few percentages .
However you will need a properly designed PSU to go along with it.. (15v dual rail and 5V) LPS..
So efforts to bring the "best" 9039q2m must be divided between hardware implementation , analog stage , digital filters , jitter reduction and low noise power They are all important . Such DAC will cost easily 750$
LPS development is over . We are taking readings with thermal cameras after using a box. We started sourcing components .
Small changes to PCB , we will manufacture 500 units
Price will be less than 200$ for US and less than 210 EUR (quite a bit less but we still need to get some components pricing)
Small changes to PCB , we will manufacture 500 units
Price will be less than 200$ for US and less than 210 EUR (quite a bit less but we still need to get some components pricing)
True. Likely more if sold through distributors and retailers.
Unfortunately, making really good dacs is still hard. Towards the end stages of development, smaller and smaller improvements in sound quality tend to come with increasingly greater costs, complexity, and difficulty. However, if a dac is good enough, it can serve well for a decade or more. Cheap dacs tend to get replaced much more often.
Last edited:
Have you tested the unit on Digione Sig. or Katana, and what is the improvement over smps/battery supplies? Pre-orders and timeframe for sale?
This should provide for anyone with a pi based system and 5V dac - multiple configurations, all of a sudden 500 units could go very quickly.
We have tested the unit on Katana and got improvement of 1.5db on THD+N . Testing on digital will not be possible, however its a given that ultralow noise its a positive in any system .
Right now there is the Chinese Lunar year so PCB house and some supliers are closed. I would expect it to have it before 3 months for sale
Of course noise is not the only improvement , we are outputting V directly from suprecaps + about 15.000uF of capacitance that should give plenty of very low impedance in the rails
Frankly , I never expected the result we got , I am yet to see published test data that show noise lower than our LPS. What I am trying to say ... allo LPS is the lowest noise 5V power supply ..in the world.
As simple as that.
Right now there is the Chinese Lunar year so PCB house and some supliers are closed. I would expect it to have it before 3 months for sale
Of course noise is not the only improvement , we are outputting V directly from suprecaps + about 15.000uF of capacitance that should give plenty of very low impedance in the rails
Frankly , I never expected the result we got , I am yet to see published test data that show noise lower than our LPS. What I am trying to say ... allo LPS is the lowest noise 5V power supply ..in the world.
As simple as that.
CDSGAMES,
Is it better than Sbooster??
Also, will Allo obtain UL certification for this power supply?
Boss testing with LPS
So almost the same phenomenon was seen (as Katana) . THD stayed exactly at the same level , Noise however (from THD+N) decreased 0.6db (from 21.46Uvrms to 20.17uVrms)
Since Boss THD+N is dominated by the THD no change was detected
In case of Katana THD+N is dominated by noise so change in noise results in better THD+N
Seems to me that a better DAC will probably gain more from this LPS , but we can see reduction of noise on both Katana and cheaper Boss.
So almost the same phenomenon was seen (as Katana) . THD stayed exactly at the same level , Noise however (from THD+N) decreased 0.6db (from 21.46Uvrms to 20.17uVrms)
Since Boss THD+N is dominated by the THD no change was detected
In case of Katana THD+N is dominated by noise so change in noise results in better THD+N
Seems to me that a better DAC will probably gain more from this LPS , but we can see reduction of noise on both Katana and cheaper Boss.
Wow.
I had to read this twice!
"Issues will happen" - the talk was about overheating devices if I'm not mistaken!
Do you know what you're spreading here !?!? I could read it:
"Hey folks - buy a Katana - burning down your house is a random feature - we call it Allo's ""Russian Class-A roulette "" "
cdsgames - are you actually in a top position over @Allo!?!?
You better clarify what you've been saying!
Last edited:
yeah.. please control the quality.Wow.
I had to read this twice!
"Issues will happen" - the talk was about overheating devices if I'm not mistaken!
Do you know what you're spreading here !?!? I could read it:
"Hey folks - buy a Katana - burning down your house is a random feature - we call it Allo's ""Russian Class-A roulette "" "
cdsgames - are you actually in a top position over @Allo!?!?
You better clarify what you've been saying!
Wow.
I had to read this twice!
"Issues will happen" - the talk was about overheating devices if I'm not mistaken!
Do you know what you're spreading here !?!? I could read it:
"Hey folks - buy a Katana - burning down your house is a random feature - we call it Allo's ""Russian Class-A roulette "" "
cdsgames - are you actually in a top position over @Allo!?!?
You better clarify what you've been saying!
C'mon aren't you a bit overdramatizing
Attachments
Wow.
I had to read this twice!
"Issues will happen" - the talk was about overheating devices if I'm not mistaken!
Do you know what you're spreading here !?!? I could read it:
"Hey folks - buy a Katana - burning down your house is a random feature - we call it Allo's ""Russian Class-A roulette "" "
cdsgames - are you actually in a top position over @Allo!?!?
You better clarify what you've been saying!
Hi Soundcheck
What I am saying is that we have over 350 Katanas at this point on the field , one unit has reported a heating issue due to some short. Any electronics that you buy can have an issue , like shorting , then it will become hot but probably not burn the house. .Marudo reported a bit higher temp since the beginning , in our test data it shows heating is not an issue.
Please do not change my words or translate my message. I am just observing that all electronics can have shorts , Katana is not a Russian roulette and all units in the field , including yours , are working correctly without burning houses .