Jonathan,
Good to hear from you. AND I'm glad the Oyaide cables were an upgrade over the Mogami... my experience of the Mogami 2534 is they make good basic cables, but nothing special.
I have not yet had a chance to try the balanced output on the Katana. As I mentioned a few posts back, I suspect I may need to adjust the distribution of the filtering in the Katana to make the balanced outputs match that of the single-ended ones and not cause me problems with ringing in my ears. As stock, the Katana balanced outputs MIGHT have a higher cutoff frequency than that of the single-ended outputs.
I'll take that on when I have a bit of time for extended experiments.
On you lifting the pads on the balanced output solder through-holes... how you fix it will depend on whether it is with the signal attachment through-holes or those for the ground.
If the ground ones, the entire area on the bottom of the board under all three balanced through-hole points is ground. For that one, I recommend:
1. Scraping off the solder mask between the hole and the edge of the board on the bottom.
2. Putting a 1" or so piece of thin solid copper wire (cut off leads from resistors work well) half-way through the hole.
3. Bending the ends together over the edge of the board.
4. Twisting the ends together tightly.
5. Then soldering it all together making sure you solder it to the scraped off area on the bottom.
You can then attach to the twisted wire ends extending past the edge of the board and worry less about pulling it loose again because it is physically attached. STILL you want to take care soldering and desoldering to it.
For the signal ones, the trace goes from the through-hole to an unused solder pad (in 3 instances, one where a resistor is soldered in the other instance) towards the center of the board. IF you pull that up, repair will be MUCH HARDER. To prevent that, I would:
1. First insulate the area on the top and bottom of the board between the through-hole and the edge of the board with a SMALL piece of Kapton or other high-temp tape.
2. Take a 1" or so piece of thin solid copper wire and bend an L into the end so that when the long end is inserted into the hole the end just lays on the pad OR touches the resistor.
3. Once you have the short end of the L adjusted EXACTLY RIGHT SO THAT IT DOES NOT SHORT TO ANYTHING ELSE, insert it and carefully solder it to the pad or resistor.
4. Bend it over the edge of the board (and the insulating tape) so that it lies against itself for about 1/2 of the way to the pad / resistor.
5. Solder it against itself.
This is MUCH LESS secure and you will have to be VERY careful to solder/desolder to it and not pull up the traces and pads. BUT MUCH better than soldering directly to them.
Once those pads / traces are gone, if you want to connect via the balanced outputs, you almost have to just buy another Opamp board.
Ideal Diode board... look in the Group Buy forum for these. UNLESS you can design and build circuit boards, you can't do this yourself.
The options you listed for linear 5V supplies all look good. But since ALL will require at a transformer (or 2 for the best results), I'd sure suggest waiting until reports get out on the new Allo Shanti which should cost not much more than the DIYINHK units and not much different than the MPAudio unit with one transformer AND may be better. Plus it'll be in a box and designed to work well with the Katana.
I hope this helps.
Greg in Mississippi
Good to hear from you. AND I'm glad the Oyaide cables were an upgrade over the Mogami... my experience of the Mogami 2534 is they make good basic cables, but nothing special.
I have not yet had a chance to try the balanced output on the Katana. As I mentioned a few posts back, I suspect I may need to adjust the distribution of the filtering in the Katana to make the balanced outputs match that of the single-ended ones and not cause me problems with ringing in my ears. As stock, the Katana balanced outputs MIGHT have a higher cutoff frequency than that of the single-ended outputs.
I'll take that on when I have a bit of time for extended experiments.
On you lifting the pads on the balanced output solder through-holes... how you fix it will depend on whether it is with the signal attachment through-holes or those for the ground.
If the ground ones, the entire area on the bottom of the board under all three balanced through-hole points is ground. For that one, I recommend:
1. Scraping off the solder mask between the hole and the edge of the board on the bottom.
2. Putting a 1" or so piece of thin solid copper wire (cut off leads from resistors work well) half-way through the hole.
3. Bending the ends together over the edge of the board.
4. Twisting the ends together tightly.
5. Then soldering it all together making sure you solder it to the scraped off area on the bottom.
You can then attach to the twisted wire ends extending past the edge of the board and worry less about pulling it loose again because it is physically attached. STILL you want to take care soldering and desoldering to it.
For the signal ones, the trace goes from the through-hole to an unused solder pad (in 3 instances, one where a resistor is soldered in the other instance) towards the center of the board. IF you pull that up, repair will be MUCH HARDER. To prevent that, I would:
1. First insulate the area on the top and bottom of the board between the through-hole and the edge of the board with a SMALL piece of Kapton or other high-temp tape.
2. Take a 1" or so piece of thin solid copper wire and bend an L into the end so that when the long end is inserted into the hole the end just lays on the pad OR touches the resistor.
3. Once you have the short end of the L adjusted EXACTLY RIGHT SO THAT IT DOES NOT SHORT TO ANYTHING ELSE, insert it and carefully solder it to the pad or resistor.
4. Bend it over the edge of the board (and the insulating tape) so that it lies against itself for about 1/2 of the way to the pad / resistor.
5. Solder it against itself.
This is MUCH LESS secure and you will have to be VERY careful to solder/desolder to it and not pull up the traces and pads. BUT MUCH better than soldering directly to them.
Once those pads / traces are gone, if you want to connect via the balanced outputs, you almost have to just buy another Opamp board.
Ideal Diode board... look in the Group Buy forum for these. UNLESS you can design and build circuit boards, you can't do this yourself.
The options you listed for linear 5V supplies all look good. But since ALL will require at a transformer (or 2 for the best results), I'd sure suggest waiting until reports get out on the new Allo Shanti which should cost not much more than the DIYINHK units and not much different than the MPAudio unit with one transformer AND may be better. Plus it'll be in a box and designed to work well with the Katana.
I hope this helps.
Greg in Mississippi
Greg,
Good to hear from you too and thanks for your answers
Good to hear from you too and thanks for your answers
- Balanced output: Might consider your technique but I’m now wondering if migrating the small extra PCB from the “Sound Quality” board to an unused “THD” board could be an easier option?
- ‘Ideal diode’ board: Think I found something, thanks for the hint
- 5VDC PSU: I was late on the news! I’ll wait for Allo Shanti and ‘Audiophile’ RPI/USB Bridge
hello,
I'm nood and french people so please be tolérent
I read all the post and a thank everyone because a learn a lot.
for nood people like me who are looking for power supply with good value price/quality a purpose this one
for isolator: Module d'Alimentation lineaire DC regule LT1083 1.2V / 19.5V 2.5A - Audiophonics
for DAC and microcontrol: Module d'alimentation lineaire LT3042 pour XMOS
for +/- 15v Module Alimentation Lineaire Regule Double LM317/337 + TL431 +/- 5V a 37V 1.5A - Audiophonics
for cable: Cable XH 2.54mm Femelle Blinde 1 Connecteur 3 Poles vers Fils Nus Gris 30cm (Unite) - Audiophonics
I test it and it's work perfectly.
Have a good ... song
Hi Elba,
How many and what transformers did you use?
I updated post #1 with the status of the Katana unit I'm currently using in one of my setups.
I had also posted info on some of the things I do to minimize any chance of electrical and/or radio interference from the RPi player in post #699 in this thread:
IanCanada's Latest RPi GB Goodies Impressions... and your tweaks, mods and hints...
I do have some further mods planned, but am VERY happy with the results I'm currently getting from the modified Katana and 4 different power supplies (including a slightly modified Shanti). I likely will spend time on one audio projects before I get back to trying any additional improvements on my Katana.
Anyone else got anything to report?
Greg in Mississippi
I had also posted info on some of the things I do to minimize any chance of electrical and/or radio interference from the RPi player in post #699 in this thread:
IanCanada's Latest RPi GB Goodies Impressions... and your tweaks, mods and hints...
I do have some further mods planned, but am VERY happy with the results I'm currently getting from the modified Katana and 4 different power supplies (including a slightly modified Shanti). I likely will spend time on one audio projects before I get back to trying any additional improvements on my Katana.
Anyone else got anything to report?
Greg in Mississippi
The basic Katana dac is really a pretty good dac by any measure, good enough perhaps to use more to its fullest potential. I have thought of trying an alternate AVCC supply, as one possibility. Also, might be fun to control it from an Arduino rather than the built-in MCU, and feed it I2S from a USB board (although probably not so much need now that Revolution dac is nearing completion).
Hello people,
I received Ultrabib (3) and L-adapter (1) (Sala's Design) and assembled the boards.
My goal is to feed the katana in Quadruple Feed Mode (Pi, Mc, Dac, OpAmp), similar to the beautiful set of Alex Kosha. But I want to test in Dual Power Mode first.
Initially I was feeding Katana with only two regular switched PSU (Dual Power mode - Pi and Mc).
So I continued with Dual Power mode, and tested it with an UltraBib for the MC board (With a 30va toroidal - 9v output - C type)) and the L-adapter for the Pi (With a 50va toroidal - 9v output)
The L-Adapter worked fine, but UltraBib worked, but noisily. (Loud noise when I turned up the volume)
I'm wondering if Ultrabib is not enough to power the Katana In dual power mode (Mc) or if the noise was due to another factor.
Now I'm using the L-adapter for Pi and iFi iPower for Mc, it's working fine, and I'm relatively satisfied!
Any help is welcome !!
Antonio
I received Ultrabib (3) and L-adapter (1) (Sala's Design) and assembled the boards.
My goal is to feed the katana in Quadruple Feed Mode (Pi, Mc, Dac, OpAmp), similar to the beautiful set of Alex Kosha. But I want to test in Dual Power Mode first.
Initially I was feeding Katana with only two regular switched PSU (Dual Power mode - Pi and Mc).
So I continued with Dual Power mode, and tested it with an UltraBib for the MC board (With a 30va toroidal - 9v output - C type)) and the L-adapter for the Pi (With a 50va toroidal - 9v output)
The L-Adapter worked fine, but UltraBib worked, but noisily. (Loud noise when I turned up the volume)
I'm wondering if Ultrabib is not enough to power the Katana In dual power mode (Mc) or if the noise was due to another factor.
Now I'm using the L-adapter for Pi and iFi iPower for Mc, it's working fine, and I'm relatively satisfied!
Any help is welcome !!
Antonio
I use 3 Ultra BiB power supplies to power the Allo boards, and a medical grade SMPS for the RPi. Mine is (was) dead silent and free of noise.
If you wish to test the system out piece meal, you will have to change the BiB's for each set up. It is very important to use the right part values for your BiB's. Testing with different numbers of power supplies, means that you have to change the current rating each time.
Also, the manual that comes with the DAC is very poorly written. It is very easy to use the wrong current rating. Some of the posts in this thread about current rating are incorrect, and you will not get good results if you set up the BiB for the wrong current rating. Because of the lack of clarity, I contacted Allo directly and confirmed what the proper values should be and posted them in this thread a while back.
^^ Those numbers come directly from Allo tech support, if you are using a different value, it is likely not correct.
I would recommend using more than 1A for the RPi, they are famous for reacting badly to power sag.
I use this
TR36M050-12G02-Level-VI Cincon | Mouser
and I find it to be better than the iFi unit (that actually is pretty crappy quality wise).
Hope this helps.
-Josh
P.S. My "was" comment is because the Allo died. I have a new MCU board on the way, hopefully it didn't damage any other parts when it shorted out.
Hi Joshua,
Thanks
I believe that with respect to Pi, the L-adapter is working very well.
But actually using katana's Quad Feed mode is beyond my means.
I do not know how to regulate the current requirements of each Ultrabib.
I'm also afraid of damaging Katana , it's really a fragile device.
In any case I was wondering if using another L-adapter for Mc would be a good idea, since it has enough power (I think) to power Mc in Dual mode.
Any suggestion
Thanks again
Antonio
Thanks
I believe that with respect to Pi, the L-adapter is working very well.
But actually using katana's Quad Feed mode is beyond my means.
I do not know how to regulate the current requirements of each Ultrabib.
I'm also afraid of damaging Katana , it's really a fragile device.
In any case I was wondering if using another L-adapter for Mc would be a good idea, since it has enough power (I think) to power Mc in Dual mode.
Any suggestion
Thanks again
Antonio
Current regulation is done by changing out resistors.
From the build guide
So the strategy is to start with your current load (say 5V 20mA for the controller), add 100. divide 600 by the sum:
600/(20 + 100) = 5. You then search the parts supplier for a resistor with the proper characteristics, and then select the resistor with the closest resistance to 5 ohms.
I used the following resistors:
Katana Board:
3 ohm (100mA draw + 100mA overhead)
Vishay part # CPF23R0000JKB14
MCU:
5 ohm (20mA draw + 100mA overhead)
Vishay part # CPF25R0000FKB14
Opamp Board:
2.4 ohm (134mA + ~117mA overhead)
Vishay part # PR02000202408JR500
For dual rail, buy 10 and use the 2 that are closest to each other.
If you opt for different part numbers, I strongly suggest using mil spec parts for these (for all resistors in a PS really), and they need to be flame resistant (proof?)
I ordered the mini-kits from Salas because the diodes are different for the single rail and dual rail supplies in this usage.
I don't see any reason the L-adapter can't be used to power the the system other than it might not be as good a PS as the BiB.
Hello,
I'm really kind of a "Half Newbie" on the subject. I mount the boards, but with regard to type of each resistor etc ... I'm really "Newbie".
I handed it to a technical friend of mine to assemble the boards.
I don't know if the resistor R1 came in the kit or he installed it later. I really didn't research the kit.
I took a picture of the mounted boards (look next post). The first photo for Ultrabib 5v and the second for ultrabib 15v. I do not know if the resistors are of good quality and with the correct specification.
In any case thank you very much, and if you can help more looking at the photos I would appreciate it!
I'm really kind of a "Half Newbie" on the subject. I mount the boards, but with regard to type of each resistor etc ... I'm really "Newbie".
I handed it to a technical friend of mine to assemble the boards.
I don't know if the resistor R1 came in the kit or he installed it later. I really didn't research the kit.
I took a picture of the mounted boards (look next post). The first photo for Ultrabib 5v and the second for ultrabib 15v. I do not know if the resistors are of good quality and with the correct specification.
In any case thank you very much, and if you can help more looking at the photos I would appreciate it!
Here the photos:
1 resistor= ultrabib 5b
2nd resistor = ultrabib 15-+15 View attachment 772430 View attachment 772431
1 resistor= ultrabib 5b
2nd resistor = ultrabib 15-+15 View attachment 772430 View attachment 772431
It looks like the resistors in the dual rail are 3R3 (3.3 ohm), I can not make out the label on the single rail, but it appears to be the same resistor.
600/3.3 =~ 182mA
182 - 134 = 48 overhead. This is slightly lower than the spec of 50mA to 150mA overhead require for the board. Given some of the obvious part substitutions, there is no way to predict if this will be fine or not.
On the single rail, if the same resistor is used, you will have the same 182mA.
This will be fine to run just the MCU alone, but well short of what is needed to run the Katana alone, and will not run the pair together.
The BiB boards are not what I would call and advanced project, but they are not novice projects either. Most of the parts are very carefully selected. The diodes in particular are very important and you can not just sub in a random replacements. Even using the correct part with a different color has an effect. I know there was some issues sourcing diodes outside the US, and I seem to remember South Americans having particular issues.
Maybe not a bad idea to take a step back and power the Katana with something more "newbie" friendly while you learn more. I'm not in anyway saying toss it out, or that you are unable to master this. Almost every question you might have about the BiB is in the BiB thread. Read every post, you will understand as well as any layman when you are done (I speak as a layman myself). Once you understand how to make substitutions, ask in the thread if the proposed parts will work, get them on the way and do it
600/3.3 =~ 182mA
182 - 134 = 48 overhead. This is slightly lower than the spec of 50mA to 150mA overhead require for the board. Given some of the obvious part substitutions, there is no way to predict if this will be fine or not.
On the single rail, if the same resistor is used, you will have the same 182mA.
This will be fine to run just the MCU alone, but well short of what is needed to run the Katana alone, and will not run the pair together.
The BiB boards are not what I would call and advanced project, but they are not novice projects either. Most of the parts are very carefully selected. The diodes in particular are very important and you can not just sub in a random replacements. Even using the correct part with a different color has an effect. I know there was some issues sourcing diodes outside the US, and I seem to remember South Americans having particular issues.
Maybe not a bad idea to take a step back and power the Katana with something more "newbie" friendly while you learn more. I'm not in anyway saying toss it out, or that you are unable to master this. Almost every question you might have about the BiB is in the BiB thread. Read every post, you will understand as well as any layman when you are done (I speak as a layman myself). Once you understand how to make substitutions, ask in the thread if the proposed parts will work, get them on the way and do it
Hi Joshua, thanks again.
Actually I bought the mini kit from Gb.
Only the black capacitor and the resistors were sourced local.
I will read the ubib thread and see what I can learn from it.
For now I am testing Pi (l-adapter) and mc+day+opamp (from one lt1083 board)
But I will also test pi (ladapter) and iPower for the rest
And also test regular Smpsu for Pi and one Ldapter for the rest.
I would appreciate also comments on these options too.
Thanks
Actually I bought the mini kit from Gb.
Only the black capacitor and the resistors were sourced local.
I will read the ubib thread and see what I can learn from it.
For now I am testing Pi (l-adapter) and mc+day+opamp (from one lt1083 board)
But I will also test pi (ladapter) and iPower for the rest
And also test regular Smpsu for Pi and one Ldapter for the rest.
I would appreciate also comments on these options too.
Thanks
The official RPi site has this ominous statement
Note the 5.1V
They recommend the 2.5A PS because the RPi goes buggy when the voltage drops below 5V, and it needs that extra 0.1V for a 1.5A power supply to power it properly.
Keep this in mind when powering the stack.
I have found the iFi iPower to quite underwhelming in terms of build quality - in short, it is build like junk. After a very short time, I started having issues with the barrel jack not making good contact, add in Allo's (moronic) decision to use USB for power, and it just is a bad combo. I never felt like I was getting good power to the RPi via the isolator board with the iPower. removing the isolator seemed to make a small improvement, switching to the SMPS I listed above seemed to make more improvement, but I can not say for sure since the MCU board died shortly after (more bad build quality on Allo's part).
I suspect the L-Adapter will be better than anything from iFi, and I know the BiB is for sure way better.
Sorry if I sound a bit sour. I am really unhappy with Allo right now. So unhappy I am considering getting another DAC. The audio quality is really extremely good if you put enough effort into the power supplies.
There is also a very noticeable improvement in sound quality when the Katana is coupled with the B1 buffer pre-amp.
@altsouza,
IMHO, the UltraBiB will be a poor choice for a Katana Dual PS (1 PS for the RPi and Isolator input if used, 1 PS for the Katana boards and isolator output if used) setup. While Allo recommends a 3A supply for both sides in a 2-PS setup, I have gotten away with a 1.2A supply on the RPi (it boots most of the time) and 1.2A on the Katana stack (though that was close to the edge). 1.5A is safer.
The UltraBiB is a shunt regulator. Shunt regulators are typically designed with 2 stages, a current-limiting stage and a shunt element stage. You have to set the current-limiting stage to a value above your expected current draw. For the Katana in 2-PS mode that should be about 1.8A. THAT means that the regulator is drawing 1.8A ALL THE TIME AND both the raw DC and heatsinking must accomodate that. That is a VERY HIGH value for most shunt regulator, think large transformer and heatsinks!
Note that the current numbers @Joshua43214 posted won't add up to the current draw of the 2-PS configuration. In the 2-PS configuration, the microprocessor board will draw a lot more than the 20mA (uP board) + 134mA (+15V) + 134mA (-15V) because you are feeding it 5V which has to be converted to + and - 15V, which requires a much higher current at 5V than the +-15V target current.
Because of the need to understand and balance target current, regulator current, raw DC current capacity, and output heatsinking, I personally consider the Salas shunt regulators advanced projects. They are less advanced when you are using them to power a low-current draw DAC such as the Boss 1.2 (>200mA) and using it in a way many others have done, so you can copy them. Also, the UltraBiB setup for the DAC side of at 2-Katana PS setup is a wildly different configuration than those needed for powering either the DAC or microprocessor board or the +-15V for the Opamp board for the Katana.
Salas' L-Adapter is a MUCH better choice for the DAC side of a 2-PS Katana.
As for Allo making a quad PSU, they are a small company with a very small design engineering staff. The Katana was a significant project for them with many new areas and challenges that consumed them for over a year. I don't believe they had the bandwidth to do the Katana along with a 4-output PS in that timeframe, especially one at the quality of their just-released Shanti. I know they learned a LOT during the Katana gestation period and you can see that with work on 3 significant projects simultaneously (Shanti, Revolution DAC, USBBridge Signature). So in addition to the Shanti, I'm hopeful we'll see a +-15V supply from them soon too. Plus there are a lot of non-advanced PSU options available with many discussed in this thread (and outlined in post 1).
Finally I am not clear on what you are trying to do when you ask "get the ground the 0 from the dual rail +15 -15 board". Specifically what 'board' are you asking about (I'm guessing the UltraBiB +-15V boards) and what are you trying to do with them?
I hope this all helps!
Greg in Mississippi
IMHO, the UltraBiB will be a poor choice for a Katana Dual PS (1 PS for the RPi and Isolator input if used, 1 PS for the Katana boards and isolator output if used) setup. While Allo recommends a 3A supply for both sides in a 2-PS setup, I have gotten away with a 1.2A supply on the RPi (it boots most of the time) and 1.2A on the Katana stack (though that was close to the edge). 1.5A is safer.
The UltraBiB is a shunt regulator. Shunt regulators are typically designed with 2 stages, a current-limiting stage and a shunt element stage. You have to set the current-limiting stage to a value above your expected current draw. For the Katana in 2-PS mode that should be about 1.8A. THAT means that the regulator is drawing 1.8A ALL THE TIME AND both the raw DC and heatsinking must accomodate that. That is a VERY HIGH value for most shunt regulator, think large transformer and heatsinks!
Note that the current numbers @Joshua43214 posted won't add up to the current draw of the 2-PS configuration. In the 2-PS configuration, the microprocessor board will draw a lot more than the 20mA (uP board) + 134mA (+15V) + 134mA (-15V) because you are feeding it 5V which has to be converted to + and - 15V, which requires a much higher current at 5V than the +-15V target current.
Because of the need to understand and balance target current, regulator current, raw DC current capacity, and output heatsinking, I personally consider the Salas shunt regulators advanced projects. They are less advanced when you are using them to power a low-current draw DAC such as the Boss 1.2 (>200mA) and using it in a way many others have done, so you can copy them. Also, the UltraBiB setup for the DAC side of at 2-Katana PS setup is a wildly different configuration than those needed for powering either the DAC or microprocessor board or the +-15V for the Opamp board for the Katana.
Salas' L-Adapter is a MUCH better choice for the DAC side of a 2-PS Katana.
As for Allo making a quad PSU, they are a small company with a very small design engineering staff. The Katana was a significant project for them with many new areas and challenges that consumed them for over a year. I don't believe they had the bandwidth to do the Katana along with a 4-output PS in that timeframe, especially one at the quality of their just-released Shanti. I know they learned a LOT during the Katana gestation period and you can see that with work on 3 significant projects simultaneously (Shanti, Revolution DAC, USBBridge Signature). So in addition to the Shanti, I'm hopeful we'll see a +-15V supply from them soon too. Plus there are a lot of non-advanced PSU options available with many discussed in this thread (and outlined in post 1).
Finally I am not clear on what you are trying to do when you ask "get the ground the 0 from the dual rail +15 -15 board". Specifically what 'board' are you asking about (I'm guessing the UltraBiB +-15V boards) and what are you trying to do with them?
I hope this all helps!
Greg in Mississippi