Go Back   Home > Forums > > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Vendor's Bazaar Commercial Vendors large & small hawking their wares

New FIFO buffer for RPI/SBCs
New FIFO buffer for RPI/SBCs
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 9th November 2018, 04:56 PM   #2621
pbarna is online now pbarna  Romania
diyAudio Member
 
Join Date: Aug 2017
Location: Marosvásárhely
In my opinion there would be buyers for a third option: one single board accomodating all stages, boards of Katana.
  Reply With Quote
Old 9th November 2018, 06:02 PM   #2622
ndriver is offline ndriver  Portugal
diyAudio Member
 
Join Date: Jul 2017
Quote:
Originally Posted by jrocker View Post
driver:

Which version of the Boss DAC and what power supply arrangement, please be specific, thx.
Sure, here are the details

Boss V1. I mod it according to the instructions manual to power it independently from the RPI.
RPI: 5V 3A SMPS (A generic medium quality one from an android box) soldered directly to the RPI.
BOSS: LT3042 chinese board + 15VA 6V transformer.

The Katana microcontroller is being powered by an IFI ipower with the negative side connected to mains ground and a 16000uF buffer cap close to the stack.
The Isolator is powered by a generic 5V 3A SMPS.

All this is fed directly to a MyRef rev C amp (another diy project) and then to my speakers: Tannoy Revolution XT8f.

Since most people think burn in helps, I refrained myself on comparison comments between Katana and Boss. But it is an upgrade, of course.
Worth mentioning that I am listening with the "Sound quality" output stage, I still have not tried the THD optimized version. I plan to do it this weekend.

Last edited by ndriver; 9th November 2018 at 06:10 PM.
  Reply With Quote
Old 11th November 2018, 04:48 AM   #2623
Greg Stewart is offline Greg Stewart  United States
diyAudio Member
 
Join Date: Jul 2003
Location: Columbus, Mississippi
New FIFO buffer for RPI/SBCs
Quote:
Originally Posted by ktham View Post
@cdsgames


Received my new replacement Katana today and will start listening to it tonight after the doorbell stops ringing when those little trick-or-treaters have gone home. Thank you.
Ktham,

I'm wondering how your Katana is sounding? I was looking for some comments I knew were posted earlier this summer for a post I'm working on and saw your power supplies waiting for a Katana. They looked good... can you give a description?

The Katana can sound VERY good with great supplies. I'm curious how others using good power are fairing.

Greg in Mississippi
__________________
Everything matters!
  Reply With Quote
Old 11th November 2018, 04:07 PM   #2624
Greg Stewart is offline Greg Stewart  United States
diyAudio Member
 
Join Date: Jul 2003
Location: Columbus, Mississippi
New FIFO buffer for RPI/SBCs
Catching up on some old posts last night and this morning, I see some I SHOULD have responded to and didn't.

First, on the 'Smurf' sound... As Allo said, that was fixed by a firmware update that is in the lastest boards. AND the great thing is that those boards will also support updating firmware by the user! YAY!

I never saw any connection between dirty or bad inter-board connecting pins (on the Katana or RPi) and the 'Smurf' sound issue. I get it pretty regularly with the early firmware boards. In my experience, it would occur when I'd power down and power up quickly, even using the recommended sequence of 'DAC Board -> Microprocessor Board -> RPi -> [Output stage board (when using a separate +-15V supply)]'.

I found I needed not only to un-power / re-power the boards, but also sometimes reboot my LMS server (I'm running PiCorePlayer, which uses SqueezeLite as player software and requires a music server process running either on the RPi or on a separate computer. IMHO, separate sounds better, so I've stuck with that).

AND it also helped to give the Katana stack a few minutes to fully power down. Remember, Allo has .33F supercaps on their low-voltage power rails on this and several other of their offerings. That amount of energy storage will keep some of the circuitry powered for several minutes before finally draining. NOT giving it time to drain would sometimes result in the 'Smurf' sound recurring on power-up OR I'd get a flashing green light on the Microprocessor board indicating an issue.

I suspect it all had to do with initialization of the registers in the DAC, the Microprocessor, and the RPi (and the Katana driver) getting out of sync due to some portion of the the stack not fully powering down. Waiting, then repowering it all lets it reset them correctly.

Glad Allo addressed this in the firmware!



Then DimDim had a GREAT post on deemphasis:

Quote:
Originally Posted by Dimdim View Post
The deemphasis filter is meant to automatically or manually apply a certain attenuation to high frequencies, depending on the recording's characteristics. In auto mode the dac chip will try to automatically determine whether deemphasis should be applied or not. In my experiments with the 90x8 dacs this feature is not reliable, i.e. a lot of the time it fails at determining whether deemphasis should be applied or not. I choose to keep it at "disabled" since I don't have any recordings that would require deemphasis anyway.

The thing is, there is a right setting for this filter and a wrong one. It only has to do with the recording. <SNIP>
I had said similar a couple of times to Allo in email coversations. I DO have some recordings with deemphasis... early Sheffield Labs as I remember. Back in the early 90s, I did a monster CD player rebuild based on the Audio Amateur Pooge articles. One thing I did was tap the deemphasis control line and use that via a driver circuit to lite an LED. Just as DimDim said, almost NONE of my CDs at the time and for as long as I used the unit (through the early 2000s) showed as needing deemphasis. WHILE one can use it to cut the highs for sonic preference purposes, that is not what it is for. AND if it helps all the time, that indicates some other issues that should be addressed.



Finally, I see I missed a post from wtnh correcting one of my earlier statements:

I found it worthwhile and would not use a Katana without it, just as I won't use the RPi as a source for any other DAC (RPi mounted or separate) with an isolator board in place.

THANKS for the catch wtnh... corrected it would read:

I found it worthwhile and would not use a Katana without it, just as I won't use the RPi as a source for any other DAC (RPi mounted or separate) without an isolator board in place.

Greg in Mississippi
__________________
Everything matters!
  Reply With Quote
Old 11th November 2018, 04:22 PM   #2625
marudo is offline marudo  Philippines
diyAudio Member
 
Join Date: May 2018
Location: Mars
Quote:
Originally Posted by Greg Stewart View Post
First, on the 'Smurf' sound... As Allo said, that was fixed by a firmware update that is in the lastest boards. AND the great thing is that those boards will also support updating firmware by the user! YAY!

I never saw any connection between dirty or bad inter-board connecting pins (on the Katana or RPi) and the 'Smurf' sound issue. I get it pretty regularly with the early firmware boards. In my experience, it would occur when I'd power down and power up quickly, even using the recommended sequence of 'DAC Board -> Microprocessor Board -> RPi -> [Output stage board (when using a separate +-15V supply)]'.
Greg,

You nail it. I tried to clean the terminal with contact cleaning solution and file it slowly with fine paper sand, yet smurf or garble was still there. I think its more on the power sequence and firmware issues.
  Reply With Quote
Old 11th November 2018, 06:05 PM   #2626
Greg Stewart is offline Greg Stewart  United States
diyAudio Member
 
Join Date: Jul 2003
Location: Columbus, Mississippi
New FIFO buffer for RPI/SBCs
Default My recent Katana WOW!!!!

Some thoughts about the Katana and power supplies...

I believe Allo has done a great job developing a high-quality DIY DAC designed to fit on a Raspberry Pi. Of course, as a DIY DAC, the end-user needs to mate it with an RPi, appropriate power supplies for the RPi and the Katana, and put it in a case (if desired).

Most of those buying a Katana will want to use a limited number of fairly low cost supplies to power the RPi and Katana/Isolator stack... likely the minimum 2. Allo has also done a good job of making the Katana sound as good as possible in that configuration. AND that is pretty good.

BUT a lot of what distinguishes a $450 RPi / Katana / 2 low-cost power supplies DAC from a $1000 DAC from a $5,000 DAC from a $10,000 DAC are better and more sophisticated power supplies and a fancier case. That's not all... better and more input interfaces, better output stages, alternative filtering, or non-traditional non-DAC-chip setups also drive the price and quality up. We can't deal with these latter items without re-engineering the Katana. AND regarding the case, if you don't like the ones available from Allo and others that will fit, you're on your own. But of course, it IS a DIY DAC, right?

BUT we can buy and/or DIY power supplies. AND that's where a significant increase in sound quality is available.

In an earlier post, I'd suggested a recent set of related reviews posted on Computer Audiophile. Part of it was to refute the oft-repeated comment "Doesn't the DAC chip chosen determine a unit's SQ?". Quick tip... it doesn't! The review articles are:

Intro: My Quest for a New DAC - Introduction to the Series

Mytek Brooklyn DAC: My Quest for a New DAC Part 1 - The Mytek Brooklyn DAC+ with Uptone JS-2 Power Supply

Ayre QX-8 DAC: My Quest for a new DAC, Part 2 - Ayre QX-8

Denafrips Terminator DAC: My Quest for a New DAC, Part 3 - Denafrips Terminator

iFi DSD Pro DAC: My Quest for a New DAC, Part 4 - iFi Pro iDSD (and friends)


One key point of interest in these reviews are the comparisons between the Ayre Codex ($1,800) and QX-8 ($5,500), and QX-5 ($10,000). The Codex uses the previous generation ES9018K2M DAC chip, the QX-8 uses the same ES9038Q2M as the Katana. AND the QX-5 the ES9028PRO. BUT the biggest differences between those units (as outlined in the review) are more sophisticated and expensive power supplies as you go up in price and quality.

Another key point of interest is that several of the DACs (Mytek Brooklyn, iFi Pro iDSD) use external power supplies and the reviewer also tests them with higher quality supplies, finding significant SQ increases with more expensive and sophisticated supplies.

AND another key point is that the Ayre QX-8, using the same DAC chip as the Katana, beats the Mytek, base iFi Pro iDSD, and Schit Yggdrasil, roughly equals the Denafrips, and is not far behind the nearly $10,000 USD Ayre QX-5. THAT says a lot about how much EVERYTHING beyond the DAC chip contributes to the SQ of a DAC. AND about how good a DAC using the ES9038Q2M can be.

Can you make a Katana equal the SQ of the QX-8? I doubt it. Ayre has included a number of additional features in the QX-8 that aren't or can't easily be included on the Katana. BUT I suspect the best power supplies will take it a good bit of the way there!

I suggest you think of the Katana as a component in a DIY DAC system that can be anywhere from good to VERY GOOD depending on how you power it.

I say this after an experience this week, where I was testing one of the levels of power supplies I will be comparing for a final review on the Katana. The SQ with these supplies, after break-in, produced the best sound I'd gotten from the Katana so far AND beat what I had expected to be the best sounding set of supplies.

These supplies were as follows:

- RPi <- K&K 12 watt Low Voltage power supply modified with somewhat larger United Chemi-Con KYB series caps (I don't like the sound of the Panasonic caps they use) and a DIY C-L-C AC filter

- Katana DAC board <- 1/2 Stammheim 5V 3||LT3045 regulator board <- 13V raw DC from 47,000/16V Jensen 4-pole cap, 4 DSEP30-6A diodes as a full-wave bridge, 1 secondary of a Hammond 229D16 transformer and a DIY C-L-C AC filter

- Katana Microprocessor board <- both halves of an OPC 5V 3||LT3042 regulator board paralleled on the output and 2 separate raw DC supplies, one into each half, same supply as above, but each secondary of the transformer producing a raw DC supply... note that while the OPC board is no longer available, the equivalent Stammheim board will work as well or better

- Katana Output stage <- +-15V Sparkos Labs regulators <- +-23V raw DC from 2 47,000/25V 4-pole caps, 4 DSEP30-6A as a full-wave bridge, 1 65VA Chinese R-Core transformer with 2 15V/2.1A secondaries

The sound I got from this setup was VERY dynamic, expressive, with great pacing and PRAT. The overall directness, clarity and excitement of vocals, and great dynamic jump I'd attribute to the Sparkos Lab based Class-A output stages, which shined in this setup! I thought I had already heard the Katana at its best, which was VERY GOOD. I was seriously wrong!

The sound was in many aspects at the same level as the best SQ I get from my systems with much more expensive DIY DACs, used with the same (or similar) supplies.

I'll say more about power supplies and some aspects of my setups that help produce this level of SQ in some later posts. BUT after listening several nights to this setup, I just had to share my findings and excitement!

Greg in Mississippi

P.S. This was not intended as a 'how-to-duplicate-this' article, though experienced DIY'ers could do so. BUT to share my experience that while Allo has done a great job making a DIY DAC that works well with budget supplies, better supplies can take it so much farther. This is NOT a patch on what Allo has done. Again, read the DAC reviews and see that at least in the Ayre line of DACs, a lot of the difference between a $2000, a $5000, and at $10,000 DAC is the power supplies.
__________________
Everything matters!

Last edited by Greg Stewart; 11th November 2018 at 06:13 PM.
  Reply With Quote
Old 11th November 2018, 07:49 PM   #2627
Greg Stewart is offline Greg Stewart  United States
diyAudio Member
 
Join Date: Jul 2003
Location: Columbus, Mississippi
New FIFO buffer for RPI/SBCs
Default SMPS versus Linear Supplies

I know some have had some good results with Switched Mode Power Supplies. AND I have too, though mostly with amplifiers powered by SMPSs. BUT my experiences generally put average-to-good linear supplies above the same rough-quality SMPSs. Some SMPS experiences (mine and others):

- I compared the iFi 5V to the modified K&K Low Voltage supply (described in a previous post) powering both RPis and my SDTrans384 SD card player. Interesting that on short listen, I ccouldn't hear much of a difference between them... but on extended listening the SQ with the K&K just sounded more like music and was more engaging.

- I lent my iFi 5V to a friend along with some modified Jameco 5V/1A linear wallwarts. His impressions were about the same as mine above. (AND YES, someday I'll do a quick how-to on the Jameco mods).

- I used the supplied SMPS when I first started running Uptone Audio LPS-1 Ultracap supplies. In my upstairs system, while I could not hear a difference in the SQ between that supply and an appropriate linear one, I DID get some buzzing through my S&B TX102 transformer volume control using the SMPS.

- In the articles I linked in a previous post, the iFi iDSD Pro was first tested with the supplied iFi SMPS, then with an Uptone Audio JS-2 and a Paul Hynes SR-7. Both were judged to be better.

- I also have an iFi Phono 2, which has gotten good reviews and which I think sounds pretty good. There are comments on the Audiogon forums where one poster says he gets a significant improvement with a good linear supply.

OTOH, there are a number of highly regarded high-end manufacturers using SMPSs besides iFi... Linn and Emperical Audio quickly come to mind, but there are many others, along with David Berning who has been using SMPSs to good effect in his well-regarded tube gear since the late 1980s! AND a growing number of amplifiers use SMPSs, including ones based on the Class D B&O Icepower and the various Hypex power supplies and modules. AND I have used both to good effect, although my current main amplifiers (PS Audio BHK 250s) DO have linear supplies.

Why do some like SMPSs & others not? At one level, I am sure much can be attributed to system differences along with listening preferences and person-to-person hearing differences.

I think another factor may be a persons's AC environment. In my case, I have done a lot around our home to reduce AC line noise into my audio gear:

- Replacing flourescent lights with halogens, sticking with incandesent bulbs instead of going to CFLs or LEDs.

- Adding clamp-on AC filter chokes on all potentially noisy equipment AC lines and other parallel AC filters on those lines and on the same lines as the audio gear. AND unplugging known noisy SMPS-powered equipment when not in use.

- Reducing the number of SMPSs in the house. EVEN my phone chargers use Jameco 5V linear supplies!

- Using DIY'd linear supplies for all of my audio-associated networking gear and computer gear, with each including a DIY'd C-L-C AC filter to reduce equipment noise feeding into the AC lines.

- Using a PS Audio P15 power regenerator in each of my 2 systems to provide clean AC for the 'audio' gear.

- Using off-the-grid supplies (such as Uptone Audio Ultracap supplies) where they make sense and provide audible benefits.

- Using an AC noise meter to provide some level of confirmation that my DIY'd supplies don't feed much noise back into the AC lines. For example, I built some linear supplies to power my Uptone Audio Ultracap supplies. After checking them with the AC noise meter, I made some mods to them that took them from making the AC line noisier to actually reducing noise on the AC line.

I suspect in my likely quieter than normal AC power environment, any noise transmitted back into the AC line or into the powered audio or audio-related gear MAY be more audible than in a setup with less heroic measures.

Still, SMPSs are here to stay and I am hopeful that as they are further developed they will get better and better. For example, I have a lot of hope for the goodness of the in-development Allo Nirvana, based on what they've done with their other products in the past including the Katana. Once they are available, I will be comparing them against my DIY'd linears for powering the Katana, likely along with Allo's current low-cost SMPS supply.

So while I am hopeful that SMPSs can someday produce the same SQ I'm hearing with linear and off-the-grid supplies, IN MY EXPERIENCE testing the Katana in my systems AND to my ears, the best results I've gotten so far have been with fairly high-test linear supplies.

Greg in Mississippi

P.S. I wanted to post this today after my post above BECAUSE I have 1 theory as to why the setup I described sounded so good with the Katana AND why it beat my previous best off-the-grid setup. I think a lot of that has to do with the PS Audio P15s, which are definite upgrades over the P10s they replaced. ALL of the comparisions I'd done in the past between my DIY'd linear setups versus my off-the-grid setps were using the P10s. This was the first time I did that comparison after getting the P15s... and now realize I need to go back and re-do a number of my previous comparisons for a possible new benchmark.
__________________
Everything matters!
  Reply With Quote
Old 11th November 2018, 09:47 PM   #2628
jonners is offline jonners  United Kingdom
diyAudio Member
 
jonners's Avatar
 
Join Date: Jan 2007
Location: Gloucestershire
@Greg

Your 'WOW!!!!' applies to Katana with the 'Sound Quality' output stage, is that right?
  Reply With Quote
Old 12th November 2018, 03:13 AM   #2629
Greg Stewart is offline Greg Stewart  United States
diyAudio Member
 
Join Date: Jul 2003
Location: Columbus, Mississippi
New FIFO buffer for RPI/SBCs
@jonners,

Yes. I actually only have Katanas with that style output stage here, I'll update once I get a production unit with both. I DID compare earlier with / without the Rasmussen filter, but at the end I was comparing various versions of that configuration.

I should add that I also have not tried any Katana on a linear-regulator-modified RPi using LDOVR.com's RPi Mezzanine board. Those are currently my reference RPi sources and I expect a further SQ lift with them.

Greg in Mississippi
__________________
Everything matters!

Last edited by Greg Stewart; 12th November 2018 at 03:29 AM.
  Reply With Quote
Old 12th November 2018, 05:13 AM   #2630
drgnanam is online now drgnanam  India
diyAudio Member
 
Join Date: Mar 2014
Location: chennai
Hi greg,
Is it possible for you to measure the power supply for ripple ,noise and impedance ,so we will have idea or benchmark for for PS
  Reply With Quote

Reply


New FIFO buffer for RPI/SBCsHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
RPi 3 ifi ipower eis PC Based 10 12th October 2016 02:28 PM
RPI: Multichannel PCM over HDMI keylimesoda PC Based 0 28th January 2016 08:55 PM
RPi B+ and ES9018K2M DAC: can't get it to work! Feddo Digital Line Level 0 18th April 2015 02:41 PM
Volumio and Rpi DAC tjaekel PC Based 1 5th February 2014 03:34 PM
Using large buffer FIFO on SPDIF fed DAC wa2ise Digital Source 8 2nd February 2006 04:43 PM


New To Site? Need Help?

All times are GMT. The time now is 07:22 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki