Regarding powerbank, battery powered 5v USB sources use switching regulators to produce a steady 5v from whatever the battery voltage is. In my experience they can cause the same sound quality problems as other switching regulators.
Hi,
Do you think 2 of these low noise ps will be great to power katana and microcontroller?
I will use a separate +-15v for output stage.
TPS7A4700 Ultralow-noise (4mVrms) Power Supply 1.4-20.5V,1A for Audio DAC | eBay
Do you think 2 of these low noise ps will be great to power katana and microcontroller?
I will use a separate +-15v for output stage.
TPS7A4700 Ultralow-noise (4mVrms) Power Supply 1.4-20.5V,1A for Audio DAC | eBay
I'm more likely to trust this type, since the layout is pretty much taken directly from the LT3042 applications notes. Low Noise LT3042 Linear Regulator Power Supply Board DC Converter Overvoltage | eBay Such circuits can be sensitive to layout differences.
Hi.
I've been using TPSA4700 devices for RPI and Kali/Boss/Piano.
All I can say is it's waste of money. You don't need a top level regulated first stage. It IMO makes things worse. IMO the more regulation stages you put in the path the worse things get - for sure on the analog pins of the systems.
You don't need an LPS either. "LPS better then SMPS" I consider a myth.
What matters is the implementation. That discussion we had more than once and actually led to the "Nirvana" SMPS development ( I guess cdsgames will confirm that).
I do use Anker charging batteries for testing purposes. These are actually quite good - despite the noise they are causing.
On the other side there are others which are quite bad, really bad.
Getting the right one can get a bit tricky.
With these batteries, the whole ground loop issues and HF noise feedback and looping into the mains just don't happen. That alone helps a lot.
Right now I'm using iFi iPowers buffered with a 47k cap. I do have several quality LPS in the workshop - just to mention it. That setup sounded best so far. And gives me best performance for the buck.
Just my 2cts.
Looking forward to Nirvana!
Enjoy.
I've been using TPSA4700 devices for RPI and Kali/Boss/Piano.
All I can say is it's waste of money. You don't need a top level regulated first stage. It IMO makes things worse. IMO the more regulation stages you put in the path the worse things get - for sure on the analog pins of the systems.
You don't need an LPS either. "LPS better then SMPS" I consider a myth.
What matters is the implementation. That discussion we had more than once and actually led to the "Nirvana" SMPS development ( I guess cdsgames will confirm that).
I do use Anker charging batteries for testing purposes. These are actually quite good - despite the noise they are causing.
On the other side there are others which are quite bad, really bad.
Getting the right one can get a bit tricky.
With these batteries, the whole ground loop issues and HF noise feedback and looping into the mains just don't happen. That alone helps a lot.
Right now I'm using iFi iPowers buffered with a 47k cap. I do have several quality LPS in the workshop - just to mention it. That setup sounded best so far. And gives me best performance for the buck.
Just my 2cts.
Looking forward to Nirvana!
Enjoy.
Does it replace the existing Isolator?
When will it be on the Allo HP?
What about Nirvana?
Thanks
Matt
Last edited:
Hi ,
Thank you for the detailed answer.
Could you explain in simple words why it makes things wirse to put regulated stages before the board?
I’m just a newbie and was thiking it was a very good thing.
Maybe a little odd if a good regulator were to make a dac sound worse. Would seem to suggest some sort of pathology that might warrant investigation. My own dacs generally sound more accurate and more to my liking with better power quality.
EDIT: Also, there are some very good, very low noise switching power supplies that have been made, including those in Benchmark DAC-3 and AHB2. General purpose off-the-shelf switchers seem to be another matter. Have never seen one that didn't need a lot of work to suppress EMI/RFI. Nirvana may turn out to be a welcome exception, however.
In addition, I just ordered an iFi switcher to check out. Be interesting to see how it turns out.
EDIT: Also, there are some very good, very low noise switching power supplies that have been made, including those in Benchmark DAC-3 and AHB2. General purpose off-the-shelf switchers seem to be another matter. Have never seen one that didn't need a lot of work to suppress EMI/RFI. Nirvana may turn out to be a welcome exception, however.
In addition, I just ordered an iFi switcher to check out. Be interesting to see how it turns out.
Last edited:
That is my very little experience too.
Moving from non hifi smps to lps was what improved sq the most for now.
And having different linear psu(s) for rpi,kali reclocker and dac board opened the soundstage and clarity more.
Glad to see I am not the only nut with 4 separate linear supplies for SBC, Isolator, Reclocker, and DAC. My ears agree with all of these statements:
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 . There is one major problems with isolators that no one is talking about. Grounds.
You have 2 grounds , dirty and clean. Problem is that grounds should not overlap eachother because that makes a capacitor passing the noise from one side to another. RPI layout makes it that both grounds are on same side (bottom RPI connector is dirty and top RPI connector is clean) basically fully overlapping ...so imagine what tetris puzzle we had)
While we took great care in the first isolator to minimize this , new isolator has fewer lines and we decreased by 96% the overlapping (to something like 0.9sqmm)
I don't know if thats what I am hearing with the new isolator (or maybe better layout) but Katana sounds fantastic with it and right now its close to a solution that costs 100 times more.. (sorry I wont name it) but understand thats my subjective and biased opinion.
For all customers that want to use isolator with Katana (and bought old isolator) we offer a 50% reduction (from 47$ msrp ) Release time 1 week.
Nirvana....the way to enlightenment is full of dangers. Thousands of test ..capacitor explosion and main engineer hurt in the eye...still noise was present
We decided to go fully in.Design was changed to Half Bridge resonant with PFC stage . On the prototype noise is at 40uV with no Active noise Filter (that has about 40db of attenuation) from 1A to 3.5A (yes more power)
So proto is done , but project was delayed. Still its going on , target price is still under 100$ but project will use a more complex circuit with Zero Voltage Switching (very low EMI) . I would expect in less than 1 month to have news. Traget is no more than 500nV of noise at 0-20Khz.
It's the chaining of the regulators which causes the trouble.
There's a lot of feedback and buffering involved that also slows current delivery down.
Many of us experienced "the best reg is no reg" at that very first position.
That led to e.g. unregulated LiFePO4 supplies or supercap supplies as the first choice. I went for a compromise and hooked up 47kUF Rifa caps I had around to my iFis. I used to run LiFePO4 supplies too.
A supercap supply like the Uptone LPS-1 is not what I'm talking about though. That one is a great (IMO too expensive) supply but still has the regulation stage at the output. If that output would feed a VDD pin of a DAC directly that'd be fine. But feeding a two stage regulation doesn't make any sense. That'd IMO be the wrong choice of supply or just a waste of a lot of money and potential.
And again. Just talking boldly about LPS is better than SMPS for whatever purpose - and then comparing whatever high quality LPS (>100$) with a whatever $5 SMPS just doesn't make any sense.
Last edited:
To all reviewers.
While listening to Katana a few observation (most of you know that)
.
1 Day. Katana will be "in your face" . Class A amplification is different and most of us are used to listen in class AB. More rough edges and while details etc are still outstanding , seems to fall a bit short.
2 Day . Seems that things are settled a bit. Music seems to gain (slightly) in space and some rough edges are gone.
3 Day. Listen and tell us about it.
10 day. For me the DAC stopped improving around that time.
While listening to Katana a few observation (most of you know that)
.
1 Day. Katana will be "in your face" . Class A amplification is different and most of us are used to listen in class AB. More rough edges and while details etc are still outstanding , seems to fall a bit short.
2 Day . Seems that things are settled a bit. Music seems to gain (slightly) in space and some rough edges are gone.
3 Day. Listen and tell us about it.
10 day. For me the DAC stopped improving around that time.
What is this about an isolator? Does Katana need one to sound it's best?
Starting to look something like a certain brand of motorcycle that everyone can have their own custom version of. Chrome air cleaner, high compression pistons...
Starting to look something like a certain brand of motorcycle that everyone can have their own custom version of. Chrome air cleaner, high compression pistons...
Good question!
Is it still needed even if all katana elements are powered by different psu?
Katana with multiple PSUs sound very good.THD+N numbers were taken whiteout isolation board. However it has been my experience and that of others that adding isolation on grounds betwen RPI (thats notoriously noisy) and DAC can improve slightly the sound quality . We cannot see any improvement on THD+N but subjectively I hear one.
So we offer the isolation board .
So we offer the isolation board .