Just to clarify, I meant that I can see the voltage fluctuating from -0.02V to 0.01V while measuring between the wall ground and the Placid HD BP ground pin.
It is easy to mistake odd-order harmonic distortion for detail. At first it sounds good, but it is fatiguing and unnatural. Listen to human voices. Do you hear more chest resonance or details of sibilance? A well recorded voice doesn't sound overly detailed. Of course, every little detail may be there, but it should be more in the background, not in the foreground. The roundness of the fundamental vocal resonance should be primary. Again, only reliable for well recorded music.
@Markw4, Oups! I thought you were talking about my remark on the lack of 'richness' but if I understand you correctly, you meant that the "more dynamic, with much more "éclat" (or glow?), especially for instruments like piano" that I hear with my DIY XLR cable and/or Katana's balanced output might be unrealistically boosted? If so, do you think it's more likely due to the cable, amp input or Katana's balanced output lack of filtering that could be fixed by re-balancing the unbalanced output?
I'm hearing some interesting and promising music from my Katana tonight.. but more on that later.
First, sorry for being so scarce. A combination of:
1. A water pipe failure (not cold-related) a few weeks back and the aftermath of that water in our house plus
2. A set of projects with one of my other favorite DIY audio gear developers plus
3. This being the time of the year in Mississippi when I have a strong several-week allergic reaction to an early-blooming weed which (as it typically does) turned into a upper-respiratory infection needing antibiotics...
With al this, I haven't had a lot of time to be here.
To catch up...
@Jonathan P, interesting report on using the SE versus BAL outputs of the Katana. I suspect you could lay a lot of what you're hearing at the difference in cables. Are your 'MacKenzie' cables the line from Audioquest? If it is this line, I honestly can't comment on them as I don't have any experience with their cables. BUT I will say Audioquest has withstood the test of time better than all but a very few cable manufacturers, AND they have been generally well-received since the 1980s. OTOH, I have a lot of experience with Mogami 2534. I DO use the internal wires as hookup wire for less critical uses. AND I used a lot of their larger-gauge hookup wires in DIY projects back in the 2nd half of the 1980s. BUT when I tried the 2534 as an interconnect cable, I found it was light in the bass and mids and a slight bit strong in the trebles, though not obnoxiously so or nor in a tizzy or distorted manner. I don't attribute that to the copper (which is very good) or the insulation (which is not bad), but just that it does not appear to be as optimized of a cable for hifi interconnect purposes as it is as a microphone cable. So if you prefer what you are hearing from the MacKenzie SE cables, see if you can borrow a similar set of their BAL cables. I know a number of places will let you buy with a 15 or 30 or whatever day return period... maybe look for a retailer who does that.
Again, based on what Allo said, I expect the filtering to be roughly equivalent between the SE and BAL outputs, so I'd not blame that without more evidence.
On using a rebalancer, if you try it, I'll be curious what you find. I totally agree with Markw4 that a good opamp, good construction, and good power will be critical to it not just sounding poorer because of the build quality. AND I'd just solder it in at the connection pads of the existing RCAs with a short twisted-pair connection... the internal wire of the Mogami should work ok here.
I'm glad your Placid build went well. As for questions on the appropriate shunt current and extra/missing parts, the Twisted Pear support forums are the better place to ask than here. IF I had experience with the later Placid design, I'd chime in, but they've gone through several different versions and the only ones I have are from their 2nd configuration, not the latest.
AND I'll have to get a small quantity of the Furatech solder to try. I have heard good things about their other products and this may be one to adopt.
As for bypassing the USB-C connector AND the protection Diode, I connect the positive lead where the single ceramic cap next to the Supercap is soldered to the board. The positive side is the one on that cap closer to the edge of the board, BUT always use your meter to confirm both the positive and negative sides before you solder and power. I've attached pictures that should better show where and how I soldered the flying leads. BTW, the wire I tend to use nowadays for purposes like this is the Audience Auric 18g hookup wire from PartsConnexion, which is also the wire used in the leads of the Auricaps.
On where to go next on getting the best out of your Katana, I'd suggest looking at replacing the iFi's. See what other's have used and liked as possible alternatives. Also look at grounding... I commented on what I do in post #126. What you are observing on the ground of the Placid is a little concerning... also may be good to take to the TP support forums if it persists after you better attend to grounding.
So what have I been doing that has me so excited? I'll go into full details over the next few weeks, I'm still listening to some of the alternative configurations and most are not are fully broken in yet. BUT what I've been playing with are:
- A couple of alternative capacitor types for the added filter caps across +15V rails.
- 2 Salas 5V Reflektor supplies for the DAC and Microprocessor 5V rails.
- A different version of the MP Audio paralleled LT3045 supplies, this one using 6 chips paralleled per rail versus 3 plus about 2x the pre-regulator capacitance.
- Variations on my modified Jameco Linear Regulated Adapters... adjusting the 12V ones up to 15V AND using the raw 19V out of the same units.
AND a couple of things I need to clear with Allo before I discuss them here.
More in the next few weeks. But some of the preliminary listening is VERY promising!
Greg in Mississippi
First, sorry for being so scarce. A combination of:
1. A water pipe failure (not cold-related) a few weeks back and the aftermath of that water in our house plus
2. A set of projects with one of my other favorite DIY audio gear developers plus
3. This being the time of the year in Mississippi when I have a strong several-week allergic reaction to an early-blooming weed which (as it typically does) turned into a upper-respiratory infection needing antibiotics...
With al this, I haven't had a lot of time to be here.
To catch up...
@Jonathan P, interesting report on using the SE versus BAL outputs of the Katana. I suspect you could lay a lot of what you're hearing at the difference in cables. Are your 'MacKenzie' cables the line from Audioquest? If it is this line, I honestly can't comment on them as I don't have any experience with their cables. BUT I will say Audioquest has withstood the test of time better than all but a very few cable manufacturers, AND they have been generally well-received since the 1980s. OTOH, I have a lot of experience with Mogami 2534. I DO use the internal wires as hookup wire for less critical uses. AND I used a lot of their larger-gauge hookup wires in DIY projects back in the 2nd half of the 1980s. BUT when I tried the 2534 as an interconnect cable, I found it was light in the bass and mids and a slight bit strong in the trebles, though not obnoxiously so or nor in a tizzy or distorted manner. I don't attribute that to the copper (which is very good) or the insulation (which is not bad), but just that it does not appear to be as optimized of a cable for hifi interconnect purposes as it is as a microphone cable. So if you prefer what you are hearing from the MacKenzie SE cables, see if you can borrow a similar set of their BAL cables. I know a number of places will let you buy with a 15 or 30 or whatever day return period... maybe look for a retailer who does that.
Again, based on what Allo said, I expect the filtering to be roughly equivalent between the SE and BAL outputs, so I'd not blame that without more evidence.
On using a rebalancer, if you try it, I'll be curious what you find. I totally agree with Markw4 that a good opamp, good construction, and good power will be critical to it not just sounding poorer because of the build quality. AND I'd just solder it in at the connection pads of the existing RCAs with a short twisted-pair connection... the internal wire of the Mogami should work ok here.
I'm glad your Placid build went well. As for questions on the appropriate shunt current and extra/missing parts, the Twisted Pear support forums are the better place to ask than here. IF I had experience with the later Placid design, I'd chime in, but they've gone through several different versions and the only ones I have are from their 2nd configuration, not the latest.
AND I'll have to get a small quantity of the Furatech solder to try. I have heard good things about their other products and this may be one to adopt.
As for bypassing the USB-C connector AND the protection Diode, I connect the positive lead where the single ceramic cap next to the Supercap is soldered to the board. The positive side is the one on that cap closer to the edge of the board, BUT always use your meter to confirm both the positive and negative sides before you solder and power. I've attached pictures that should better show where and how I soldered the flying leads. BTW, the wire I tend to use nowadays for purposes like this is the Audience Auric 18g hookup wire from PartsConnexion, which is also the wire used in the leads of the Auricaps.
On where to go next on getting the best out of your Katana, I'd suggest looking at replacing the iFi's. See what other's have used and liked as possible alternatives. Also look at grounding... I commented on what I do in post #126. What you are observing on the ground of the Placid is a little concerning... also may be good to take to the TP support forums if it persists after you better attend to grounding.
So what have I been doing that has me so excited? I'll go into full details over the next few weeks, I'm still listening to some of the alternative configurations and most are not are fully broken in yet. BUT what I've been playing with are:
- A couple of alternative capacitor types for the added filter caps across +15V rails.
- 2 Salas 5V Reflektor supplies for the DAC and Microprocessor 5V rails.
- A different version of the MP Audio paralleled LT3045 supplies, this one using 6 chips paralleled per rail versus 3 plus about 2x the pre-regulator capacitance.
- Variations on my modified Jameco Linear Regulated Adapters... adjusting the 12V ones up to 15V AND using the raw 19V out of the same units.
AND a couple of things I need to clear with Allo before I discuss them here.
More in the next few weeks. But some of the preliminary listening is VERY promising!
Greg in Mississippi
Attachments
Mackenzie RCA cables also mention audio quest on the connectors.
The other crazy expensive RCA cables I tried were Audience AU24se. More fine grain details with these but still, more “éclat” (in french) or glow or spark, sorry I’m not sure what the proper term would be in english… with the mogami/neutrik XLR. Definitely doesn’t mean it’s better though… But with my current system and for the music type I listen to, I think I like the XLR sound better.
Thanks for the input on Mogami 2534! Maybe part of the “éclat” I hear in voice and piano is related to your comment on "slight bit strong in the trebles”…. I’ll try to get my hands on other balanced cables or maybe built another one with top notch furutech balanced bulk cable and connectors
Cool, will definitely let you know if/when I try.
By solder at the connection pads of the existing RCAs, you mean the 2 big solder spots under the board or the 3 small solder points on the top of the board?
Also, to feed the balancing circuit, do you think I could I split the power from the Placid HD BP between the output stage and the balancing circuit? If so, should I increase the shunt current a bit?
I’m definitely no soldering expert but as you said, furutech is well renowned for making very high quality products, including many that rely on high quality soldering, so…
I also hear that Cardas solder is very good but I opted for the furutech because the Cardas was hard to find from a trusted source and furutech mentions on their site that there has been several cases of counterfeit products shipping from china on eBay and others….
Ok thanks! In the meantime, based on @jonners suggestions, I have already soldered the way you did in an earlier post, under the MC board
https://www.diyaudio.com/forums/pc-based/329911-getting-allo-coms-katana-dac-20.html#post5666270
Just wondering the difference between these connection points (under the MC board) vs the the ones on top of the board. Do they both bypass the Common Mode Choke?
Also, about bypassing the usb connector on the Isolator, did someone post details on how to do it or did I dream? Can’t find it anymore
Thanks man, will definitely look into that! BUT maybe I’ll wait until you come back to us with your recent findings before I replace my 3 5V supplies (2 iPower + 1 allo)!
Finally, 2 more questions bothering me…
1- I noticed the top of the big cap on MC board is a bit swollen with some yellowish liquid on it. I cleaned it a few days ago and it doesn’t seem to leak anymore (maybe this came from somewhere else?). Do you guys think I should still replace it?
2- Did you guys experiment with [url="https://www.mouser.ca/ProductDetail/3m/ab5100shf/qs=gMfM3zrbS4r0qTugzC4tdw==&countrycode=CA¤cycode=CAD”]EMI absorbers[/url]?
It seems obvious this can make a difference but it also seems prone to make things worst by causing other problems if not used/positioned properly….
I was thinking to wrap some around to protect the Katana from external EMF but also internally, like say on the DAC and NDK oscillator chips and between the RPI board and the rest of the stack.
You should contact Allo warranty service about the cap. It is bad if leaking or changing its shape.
I believe that once one started to mod a purchased board (and documenting it on the Internet) the word 'warranty' can be forgotten... @Jonathan P mentions at some point:
A leaking capacitor after a documented work on the same board can be serious infrigement of the warranty conditions... On the other hand the capacitor replacement on the MC board is an easy and non-costly fix.
Mackenzie RCA cables also mention audio quest on the connectors.
The other crazy expensive RCA cables I tried were Audience AU24se. More fine grain details with these but still, more “éclat” (in french) or glow or spark, sorry I’m not sure what the proper term would be in english… with the mogami/neutrik XLR. Definitely doesn’t mean it’s better though… But with my current system and for the music type I listen to, I think I like the XLR sound better.
Thanks for the input on Mogami 2534! Maybe part of the “éclat” I hear in voice and piano is related to your comment on "slight bit strong in the trebles”…. I’ll try to get my hands on other balanced cables or maybe built another one with top notch furutech balanced bulk cable and connectors
Cool, will definitely let you know if/when I try.
By solder at the connection pads of the existing RCAs, you mean the 2 big solder spots under the board or the 3 small solder points on the top of the board?
Also, to feed the balancing circuit, do you think I could I split the power from the Placid HD BP between the output stage and the balancing circuit? If so, should I increase the shunt current a bit?
I’m definitely no soldering expert but as you said, furutech is well renowned for making very high quality products, including many that rely on high quality soldering, so…
I also hear that Cardas solder is very good but I opted for the furutech because the Cardas was hard to find from a trusted source and furutech mentions on their site that there has been several cases of counterfeit products shipping from china on eBay and others….
Ok thanks! In the meantime, based on @jonners suggestions, I have already soldered the way you did in an earlier post, under the MC board
https://www.diyaudio.com/forums/pc-based/329911-getting-allo-coms-katana-dac-20.html#post5666270
Just wondering the difference between these connection points (under the MC board) vs the the ones on top of the board. Do they both bypass the Common Mode Choke?
Also, about bypassing the usb connector on the Isolator, did someone post details on how to do it or did I dream? Can’t find it anymore
Thanks man, will definitely look into that! BUT maybe I’ll wait until you come back to us with your recent findings before I replace my 3 5V supplies (2 iPower + 1 allo)!
Finally, 2 more questions bothering me…
1- I noticed the top of the big cap on MC board is a bit swollen with some yellowish liquid on it. I cleaned it a few days ago and it doesn’t seem to leak anymore (maybe this came from somewhere else?). Do you guys think I should still replace it?
View attachment 738818
View attachment 738819
2- Did you guys experiment with EMI absorbers?
It seems obvious this can make a difference but it also seems prone to make things worst by causing other problems if not used/positioned properly….
I was thinking to wrap some around to protect the Katana from external EMF but also internally, like say on the DAC and NDK oscillator chips and between the RPI board and the rest of the stack.
Contact us we will replace the board or provide you with a cap if you can install yourself.
@Jonathan P,
Ok then, your SE RCA cables are actually Audioquest cables from their Mackenzie line. Very curious what you find if you get a chance to build and try a set of BAL cables made with the Furatech components. Again I've heard some good reports on their raw cables and connectors.
On connecting in the rebalancer, you should check the spots you mentioned where the Katana Opamp board RCA connectors with your meter to confirm which of them are ground and which are hot. Then you'll know where to connect your rebalancer. I'm not trying to harp on this, but it is an easy thing to check and one of those things you ALWAYS want to be sure of before you start modifying or making external connections to a board. So just a friendly reminder. Let us know what you find.
In a similar vein, you wrote the following about where you connected your 5V external power feed into the Katana Microprocessor board:
I believe @jonners and you are referring to my post #217 (BTW, I use post numbers to indicate different posts because a link will not work correctly for everyone). If you are, what I wrote there was:
So I soldered to the BOTTOM of the DAC board and to the TOP of the Microprocessor board. You CAN connect the external 5V to the bottom of the Microprocessor board at the + & - of the supercap. There are not any other obvious 5V connection points on the bottom of the Microprocessor board.
All of this has me question where you actually connected your 5V to the Microprocessor board since I did not show a bottom connection for that board? Can you clarify, perhaps with a picture? Likely it is just a mis-communication.
Then 2 final things... you wrote
Go back and check the posts discussing the 'Common Mode Choke' again. They were referring to common mode chokes on the 5V power supply board used, not on the any of the Katana boards. I don't know that anyone has reported on bypassing any of the chokes on the Katana boards here. AND I suspect there are other higher-benefit mods than bypassing any of the Katana boards' chokes... Allo has spent a lot of effort developing effective power supply noise filtering methods the work well with little negative impact. With REALLY good power supplies, bypassing them entirely might provide some benefit. BUT one has to have the knowledge to identify where to connect power inputs and bypass their filtering networks.
AND AFAIK, there is NO easy way to bypass the USB connector on the Isolator 1.2. The purpose of that connector on the Isolator 1.2 is to provide an alternative 5V feed for the RPi and input side of the Isolator 1.2 INSTEAD of powering the combination via the RPi Micro USB. AND the output side of the Isolator 1.2 is designed to ALWAYS be fed from the DAC board above it, either the Katana DAC board or a Boss 1.2. So no.
Let us know on your 5V connections on the Microprocessor board. AND I hope this all helps.
Greg in Mississippi
Ok then, your SE RCA cables are actually Audioquest cables from their Mackenzie line. Very curious what you find if you get a chance to build and try a set of BAL cables made with the Furatech components. Again I've heard some good reports on their raw cables and connectors.
On connecting in the rebalancer, you should check the spots you mentioned where the Katana Opamp board RCA connectors with your meter to confirm which of them are ground and which are hot. Then you'll know where to connect your rebalancer. I'm not trying to harp on this, but it is an easy thing to check and one of those things you ALWAYS want to be sure of before you start modifying or making external connections to a board. So just a friendly reminder. Let us know what you find.
In a similar vein, you wrote the following about where you connected your 5V external power feed into the Katana Microprocessor board:
I believe @jonners and you are referring to my post #217 (BTW, I use post numbers to indicate different posts because a link will not work correctly for everyone). If you are, what I wrote there was:
So I soldered to the BOTTOM of the DAC board and to the TOP of the Microprocessor board. You CAN connect the external 5V to the bottom of the Microprocessor board at the + & - of the supercap. There are not any other obvious 5V connection points on the bottom of the Microprocessor board.
All of this has me question where you actually connected your 5V to the Microprocessor board since I did not show a bottom connection for that board? Can you clarify, perhaps with a picture? Likely it is just a mis-communication.
Then 2 final things... you wrote
... Just wondering the difference between these connection points (under the MC board) vs the the ones on top of the board. Do they both bypass the Common Mode Choke?
Also, about bypassing the usb connector on the Isolator, did someone post details on how to do it or did I dream? Can’t find it anymore
Go back and check the posts discussing the 'Common Mode Choke' again. They were referring to common mode chokes on the 5V power supply board used, not on the any of the Katana boards. I don't know that anyone has reported on bypassing any of the chokes on the Katana boards here. AND I suspect there are other higher-benefit mods than bypassing any of the Katana boards' chokes... Allo has spent a lot of effort developing effective power supply noise filtering methods the work well with little negative impact. With REALLY good power supplies, bypassing them entirely might provide some benefit. BUT one has to have the knowledge to identify where to connect power inputs and bypass their filtering networks.
AND AFAIK, there is NO easy way to bypass the USB connector on the Isolator 1.2. The purpose of that connector on the Isolator 1.2 is to provide an alternative 5V feed for the RPi and input side of the Isolator 1.2 INSTEAD of powering the combination via the RPi Micro USB. AND the output side of the Isolator 1.2 is designed to ALWAYS be fed from the DAC board above it, either the Katana DAC board or a Boss 1.2. So no.
Let us know on your 5V connections on the Microprocessor board. AND I hope this all helps.
Greg in Mississippi
@cdsgames, I contacted support but I'm still waiting for details as per what I dshould do. Here is the response I got so far:
Code:
Hi Jonathan,
Noted your feedbacks and we will take corrective actions.
This Capacitor we added as additional filter in addition to the on board filter circuit. This cap connected to the op-amp power input stage ( output of +/-14.1 LDOs). The yellowish liquid on capacitor shows high voltage (spike/noise) presence across the limbs for short duration. This spikes may be from External +/15V source. We will analyze the possibilities.
Now the external +/-15v in correct range and working fine.
Regarding MCBoard Replacement : Discussed this with Production team , the sticker pasted as part of internal marking. Normally will ship as a set , the replacement board taken from production line and shipped directly.
Regards
Sudeep
I've been listening to the differences between having my +-15V 'Studer 900' PSU connected direct to the large cap on the microprocessor board, as compared to going via the LDO regs. In both cases I have left my 110uF bank of film caps connected to the large cap.
Greg said that he thought some had tried taking the +-15V direct, but I haven't found any reports so far. I have a definite preference for bypassing the LDO's and I'll report on the differences I hear when some others have tried it.
@jonners,
Thanks for the report. Based on what you found, I'm looking at how I can adapt my setup to bypass the on-board +-14V LDOs and run your test in my setup. As I said before, my gut-feel is that with better quality +-15V supplies (and I believe your supply well-qualifies as that), bypassing the LDOs should be a good choice. Glad to hear it confirmed.
On the items I teased in post #285, I've listened bit more to the alternative caps on the +-14V rails. First, I need to explain that none of the Katana I have here are exactly the same as the current shipping versions. One is very close, but has a different implementation of the SQ mod than the current shipping version, which is much easier for Allo to install than their original modification configuration, which is what I did on the ones I had here. Second, when they were first installing the 3 electrolytic caps on the +-14V rails they provided recommended connection points and caps and I used those. Their current shipping configuration uses one different connection point and different caps. Part of this was allow the the Katana to still fit in their case... the added caps that I've done definitely would not. OTOH, I tried very hard to minimize the lead length on the added caps, which should be a good thing.
SO my results on this MIGHT be different than what others would hear. Also, the original caps Allo specified were all Nichicon units, a 1000uF/50V UKW series and a pair of 100uF/35V UKA series. That latter is one of my current go-to caps in analog-only uses, such as a DAC output stage. My other go-to caps are Jensen 4-pole, several other lines of Nichicons (ES, FG, KZ, KG), digging into my dwindling stock of NOS Black Gates, and the United Chemicon KYB line, which I first encountered as the specified parts on OPC's dual 4||LT3042 regulator boards (I've also recently seen them on the well-regarded IcePower AS1200-2 modules). I had such good success with the KYBs there that I have used them in a number of other builds with similar good results. In addition to the OPC and my own builds of the MP Audio paralleled LT3045 regulator boards (I bought some bare boards from him back from when he still sold them that way), another of those applications is the +-15V Sparkos Regulator board I use to supply the Katana, so all of my results so far with the Katana have included use of these caps. I thought another place to try them was as the 3 added Cs to the +-14V rails.
I've been listening to an Opamp board with those caps installed for about 4 weeks now, so they should be well-broken-in. BUT my results are mixed AND I'd call it a draw. Dynamics and bass impact are definitely better with the KYB. Highs are a mixed bag, about the same resolution, but a tiny bit of hash added. BUT the main negative change is a reduction in the sense of depth over the caps originally specified by Allo. Honestly, I wouldn't chose one over the other... the things each do well are important to me. So IF you want to void your warranty and take a chance on replacing those caps either with those originally specified by Allo or the KYB, your own personal preferences should be your guide. I WILL spend some time soon comparing both to the Katana with the 'as-shipped' caps and report on them too.
I've also listened a bit more to a pair of Salas Reflektor-D regulators replacing the MP Audio 3||LT3045 boards (roughly equivalent to his DC-input only SD-HPULN PS from his current lineup). Here, I'll declare the MP Audio boards as a winner, but just by a nose. They won due to more delicate, detailed, and natural highs and a VERY slightly quieter background. BUT the Reflektor-Ds did provide a more solid, muscular, meaty, and palpable sense to the sound with slightly better bass impact and dynamics. So while I'll generally stick with the MP Audio units based on my listening preferences, others might as likely prefer the Reflektor-D's. AND if you build the Reflektors from the DIYAudio group-buy boards and semi-kits, the Reflektor-D's will cost you less out of pocket.
Both the MP Audio and Reflektor boards were powered from the same rails of my raw DC supplies as detailed in post #4 of this thread. Also, my Reflektor builts were slightly different than stock... you can see details in posts #1481 and #1499 of this thread:
Reflektor-D builds
On the 6||LT3045 MP Audio boards, while they were very promising when I first listened to them, they were less convincing to me than the 3||LT3045 versions during last night's listening session. I'm giving them more break-in-time before I do another comparison.
On the other items I teased, I've not taken the time yet to try the newly-modified Jameco linear-regulated adapters. Soon.
Greg in Mississippi
Thanks for the report. Based on what you found, I'm looking at how I can adapt my setup to bypass the on-board +-14V LDOs and run your test in my setup. As I said before, my gut-feel is that with better quality +-15V supplies (and I believe your supply well-qualifies as that), bypassing the LDOs should be a good choice. Glad to hear it confirmed.
On the items I teased in post #285, I've listened bit more to the alternative caps on the +-14V rails. First, I need to explain that none of the Katana I have here are exactly the same as the current shipping versions. One is very close, but has a different implementation of the SQ mod than the current shipping version, which is much easier for Allo to install than their original modification configuration, which is what I did on the ones I had here. Second, when they were first installing the 3 electrolytic caps on the +-14V rails they provided recommended connection points and caps and I used those. Their current shipping configuration uses one different connection point and different caps. Part of this was allow the the Katana to still fit in their case... the added caps that I've done definitely would not. OTOH, I tried very hard to minimize the lead length on the added caps, which should be a good thing.
SO my results on this MIGHT be different than what others would hear. Also, the original caps Allo specified were all Nichicon units, a 1000uF/50V UKW series and a pair of 100uF/35V UKA series. That latter is one of my current go-to caps in analog-only uses, such as a DAC output stage. My other go-to caps are Jensen 4-pole, several other lines of Nichicons (ES, FG, KZ, KG), digging into my dwindling stock of NOS Black Gates, and the United Chemicon KYB line, which I first encountered as the specified parts on OPC's dual 4||LT3042 regulator boards (I've also recently seen them on the well-regarded IcePower AS1200-2 modules). I had such good success with the KYBs there that I have used them in a number of other builds with similar good results. In addition to the OPC and my own builds of the MP Audio paralleled LT3045 regulator boards (I bought some bare boards from him back from when he still sold them that way), another of those applications is the +-15V Sparkos Regulator board I use to supply the Katana, so all of my results so far with the Katana have included use of these caps. I thought another place to try them was as the 3 added Cs to the +-14V rails.
I've been listening to an Opamp board with those caps installed for about 4 weeks now, so they should be well-broken-in. BUT my results are mixed AND I'd call it a draw. Dynamics and bass impact are definitely better with the KYB. Highs are a mixed bag, about the same resolution, but a tiny bit of hash added. BUT the main negative change is a reduction in the sense of depth over the caps originally specified by Allo. Honestly, I wouldn't chose one over the other... the things each do well are important to me. So IF you want to void your warranty and take a chance on replacing those caps either with those originally specified by Allo or the KYB, your own personal preferences should be your guide. I WILL spend some time soon comparing both to the Katana with the 'as-shipped' caps and report on them too.
I've also listened a bit more to a pair of Salas Reflektor-D regulators replacing the MP Audio 3||LT3045 boards (roughly equivalent to his DC-input only SD-HPULN PS from his current lineup). Here, I'll declare the MP Audio boards as a winner, but just by a nose. They won due to more delicate, detailed, and natural highs and a VERY slightly quieter background. BUT the Reflektor-Ds did provide a more solid, muscular, meaty, and palpable sense to the sound with slightly better bass impact and dynamics. So while I'll generally stick with the MP Audio units based on my listening preferences, others might as likely prefer the Reflektor-D's. AND if you build the Reflektors from the DIYAudio group-buy boards and semi-kits, the Reflektor-D's will cost you less out of pocket.
Both the MP Audio and Reflektor boards were powered from the same rails of my raw DC supplies as detailed in post #4 of this thread. Also, my Reflektor builts were slightly different than stock... you can see details in posts #1481 and #1499 of this thread:
Reflektor-D builds
On the 6||LT3045 MP Audio boards, while they were very promising when I first listened to them, they were less convincing to me than the 3||LT3045 versions during last night's listening session. I'm giving them more break-in-time before I do another comparison.
On the other items I teased, I've not taken the time yet to try the newly-modified Jameco linear-regulated adapters. Soon.
Greg in Mississippi
@Greg Stewart
So far I haven't managed to follow your recommendation of very short psu leads. At present the leads from my +-15v supply to the 1000uF cap are about 10-12cm long. With some reorganisation I could probably get them down to half that length, but I don't know if it's worth the effort.
It might be interesting to try different supply voltages, but since the 1000uF across the rails is only rated at 35V I would be reluctant to go higher than 15V.
So far I haven't managed to follow your recommendation of very short psu leads. At present the leads from my +-15v supply to the 1000uF cap are about 10-12cm long. With some reorganisation I could probably get them down to half that length, but I don't know if it's worth the effort.
It might be interesting to try different supply voltages, but since the 1000uF across the rails is only rated at 35V I would be reluctant to go higher than 15V.