The Subbu cooked with caps.
Hi,
I’ve got 2 Subbu V3. the second is to try to tailor, to cook the DAC in relation to my system. The Speakers have a 5” mid aluminium from 125 hz to 2600 hz then an aluminium tweeter said to climb to 35 khz but with a mechanical notch trap at 22 Khz. Those are the Boston Lynfield 400 L : very transparent (clear), very good sound stage, very fast, fabulous mid bass, tight with details (closed band pass with two 17” with double spiders : 35 hz to 125 hz.
The speakers are difficult : when I change a little cable, a capacitor, a resistor… everything I can hear it. And the treble is a little to brighty… So many things have to be cooked to match with a neutral sound with keeping all the qualities (superb transcient and subjective fastness, very analytic, none directive (they diseapper) and defaults : bad recording… They can be fatiguing after 1 or 2 hours in relation to a more steamed Kef 104/2 ref I own also.
Well what about the V3. first I think the BOM is very good… no doubt at all. Here it’s not to be better in the absolute but in regards to my system and personal tastes. At the beginning I with the Garry’s revision with increasing C35 with 3.3 Wima MKS2 (Garry: 4.7 uF) but come back to 1 uF. C4 is 1 uf MKS2, I put 8 tantal 4.7 uf (Garry: went to the 4x 10 uf X7R before the regs)
C8&C14 are Panasonic FM. & C21 Silmic2. The 100 uf caps are the very good UCC polymers of the BOM but C22 where 470 uF Panasonic SEPC was firstly choosed according the advises. Both Subbu have a 75 ohms BNC direct on the pcb (no wires). Sense of the wall plug are marked. Wires of the analog RCA are 0.6 mm² copper, silver plated with PFTE insulation. Tracks for the tests are always the same from many years and are classic, instrumental, rock, jazz, world music, lives with applauses also. Genuine PS is used : the same is shared with the twp PCB at each test !
I like cooking & I like cooking the devices with caps also because I think most of devices have to match with the rest of the whole system and the own flavour of the listener. There is no neutral sound… sound is always an interpretation of the tastes of the listener IMHO and what he can do with what he has at home.
I took also a JG buffer but those taste was made without it. First Subbu V3 is the reference.
The second was firstly tweaked in relation to the first for C8&C13 : after trying two polymers : the ASA serie from UCC and SEP and SEPF serie from Panasonic : 10 uf SEP won : better.
C4 : I swap MKS2 for 1 uf NPO in 603 case (Marce input : no // advised, low inductance). I stay with the same 3 x 100 uF polymers as my first taste showed they was best than SEPC or Nichicon R7 here (but after an error tested with 40 uf here not 100 uF) : the difference is huge according to me and the UCC of the BOM with SEPC at C22 is a very good match in my system. 2 SEPC is too lighty…too thin !
The tantal caps just before the regs were swaped for 1 uF acrylic : the same as the PS BOM. Caps of the XO at C17 : I put three : 1 nf, 100 nf, 1 uF, first NPO, others X7R, first close to the crystal.
All the 805 X7R were swaped for 603 case with the same values (more cracking or better inductance…time will say which won !). So the second board is very different from the first. How it sounds ? With the same 470 uf C22, C21 & 1 uF C35, with just 1 uF at C34 : the second win by a very good margin : better trade off, softer without lake of definition.
So all the tests was maid many times and always two times: with the first board as reference as populated above in relation to the second populated as above. And also with the first board with the same C22, C21, C35. Why ? Most of the time the second board was better, so it was needed to benchmark with its own ref alone the tests with the three caps (C22,C21,C35).
I have to propose also three confs in relation to your main type of sound to try at the end for the fellows who have two Subbu… something else. As it’s risky for the PCB who is fragile, it is not advised for the owner of a standalone Subbu. I repeat it : the official BOM is neutral which is hard to do. I never tried the SAL but surmise a very good match at C35 with the caps tested below :
Voila for the ”blabla” (“the bad literature”: in French) 😀
To be continued….pictures comes !
Hi,
I’ve got 2 Subbu V3. the second is to try to tailor, to cook the DAC in relation to my system. The Speakers have a 5” mid aluminium from 125 hz to 2600 hz then an aluminium tweeter said to climb to 35 khz but with a mechanical notch trap at 22 Khz. Those are the Boston Lynfield 400 L : very transparent (clear), very good sound stage, very fast, fabulous mid bass, tight with details (closed band pass with two 17” with double spiders : 35 hz to 125 hz.
The speakers are difficult : when I change a little cable, a capacitor, a resistor… everything I can hear it. And the treble is a little to brighty… So many things have to be cooked to match with a neutral sound with keeping all the qualities (superb transcient and subjective fastness, very analytic, none directive (they diseapper) and defaults : bad recording… They can be fatiguing after 1 or 2 hours in relation to a more steamed Kef 104/2 ref I own also.
Well what about the V3. first I think the BOM is very good… no doubt at all. Here it’s not to be better in the absolute but in regards to my system and personal tastes. At the beginning I with the Garry’s revision with increasing C35 with 3.3 Wima MKS2 (Garry: 4.7 uF) but come back to 1 uF. C4 is 1 uf MKS2, I put 8 tantal 4.7 uf (Garry: went to the 4x 10 uf X7R before the regs)
C8&C14 are Panasonic FM. & C21 Silmic2. The 100 uf caps are the very good UCC polymers of the BOM but C22 where 470 uF Panasonic SEPC was firstly choosed according the advises. Both Subbu have a 75 ohms BNC direct on the pcb (no wires). Sense of the wall plug are marked. Wires of the analog RCA are 0.6 mm² copper, silver plated with PFTE insulation. Tracks for the tests are always the same from many years and are classic, instrumental, rock, jazz, world music, lives with applauses also. Genuine PS is used : the same is shared with the twp PCB at each test !
I like cooking & I like cooking the devices with caps also because I think most of devices have to match with the rest of the whole system and the own flavour of the listener. There is no neutral sound… sound is always an interpretation of the tastes of the listener IMHO and what he can do with what he has at home.
I took also a JG buffer but those taste was made without it. First Subbu V3 is the reference.
The second was firstly tweaked in relation to the first for C8&C13 : after trying two polymers : the ASA serie from UCC and SEP and SEPF serie from Panasonic : 10 uf SEP won : better.
C4 : I swap MKS2 for 1 uf NPO in 603 case (Marce input : no // advised, low inductance). I stay with the same 3 x 100 uF polymers as my first taste showed they was best than SEPC or Nichicon R7 here (but after an error tested with 40 uf here not 100 uF) : the difference is huge according to me and the UCC of the BOM with SEPC at C22 is a very good match in my system. 2 SEPC is too lighty…too thin !
The tantal caps just before the regs were swaped for 1 uF acrylic : the same as the PS BOM. Caps of the XO at C17 : I put three : 1 nf, 100 nf, 1 uF, first NPO, others X7R, first close to the crystal.
All the 805 X7R were swaped for 603 case with the same values (more cracking or better inductance…time will say which won !). So the second board is very different from the first. How it sounds ? With the same 470 uf C22, C21 & 1 uF C35, with just 1 uF at C34 : the second win by a very good margin : better trade off, softer without lake of definition.
So all the tests was maid many times and always two times: with the first board as reference as populated above in relation to the second populated as above. And also with the first board with the same C22, C21, C35. Why ? Most of the time the second board was better, so it was needed to benchmark with its own ref alone the tests with the three caps (C22,C21,C35).
I have to propose also three confs in relation to your main type of sound to try at the end for the fellows who have two Subbu… something else. As it’s risky for the PCB who is fragile, it is not advised for the owner of a standalone Subbu. I repeat it : the official BOM is neutral which is hard to do. I never tried the SAL but surmise a very good match at C35 with the caps tested below :
Voila for the ”blabla” (“the bad literature”: in French) 😀
To be continued….pictures comes !
Eldam,
Thanks for sharing your results. You included a lot of different experimental results in your message. At first it was hard to determine your end point, so I will try to write a summary of what I think that you concluded. I'll use the same format that I used at the start of this thread - Tiers 1, 2, 3, and 4 and put your results into that framework. If you disagree with my priority, please comment. I'm using this framework because your results weren't always clear on what changes were most important.
Tier 1 - most important changes
Upgrade C22 to 470uf Panasonic SEPC - you agree with others on this change
Tier 2 - lower priority but still important
C35 (or C32) 1uf Wima MKS2 - you agree with JP's recommendation here
C4 - 1uf Wima MKS2 or 1uf NPO ceramic. I think you like the 1uf NPO better but it wasn't clear if you tested this change without making any other changes
50Mhz oscillator Vdd cap (C17)
Original BOM 4.7uf tanatalum, later updated to 1uf X7R
You followed Marce's recommendation of a parallel combination of 1nf (NPO) 100nf (X7R) 1uf (X7R) and liked this better. Again did you have a chance to listen to this change alone?
Tier 3 - even lower priority but still noticeable
WM8804 Vdd capacitors and ES9023 Vreg capacitor - C8, C13, C21
You prefer the Panasonic 10uf SEP polymer caps here
Tier 4 - in my experience slight changes but overall not that important
bypass capacitors for the input electrolytics (C5, C7, C19, C20)
you prefer the BOM 4.7uf tantalum caps vs. using 1uf or 10uf X7R ceramic caps
Does this summary capture your results accurately including the priority of the changes? Overall, it looks like your results are very similar to what others have found with a few changes of flavor, as you like to say.
---Gary
thanks Garry,
I prefer to write here because I receive too much often bad critics here and some infantilisant.
As you know my master of english is a limit and is a source of misanterstand, like in the method for example. It's subjective, and I more serious than you seem to believe often you and JP. My springfull personnality I surmise !😉 But as ancient IT manager and at 45 now you can have the faith of I can know what a serious job and method can be, even if here it stays an hobby for me and I'n not putting my life in it (humm it takes a lot of my free time in this moement). But If I write it's because I know difference are interresting enough for the ones who more than one both to acmodate the spices & flavor.
I think I will write the following tommorow, family first this week end. But the result will not be universal, I writed this already. But i will try to give three "enough" change to match with 3 different sound personnality for the transparent system as I think mine is good enough. It's without JG buffer with the genuine PS and just few caps : thinked to be non expensive.
get it as it is, try it if you want or have the courage. It's an hobby. maybe this tweak will be a good one in the flacor way, maybe not. But try before comment, if good say it. If not good say it also.
It's just for sharing my little work around cooking without pretention from me to be better than others.
Maybe JP or someone else maid this work, they will say if they want, here I share with the prcaution to say it's not universal but work with my system to have a better match after the excellent initial work of JB/Subbu.
So no podium rewarded for me, it's just sharing
I prefer to write here because I receive too much often bad critics here and some infantilisant.
As you know my master of english is a limit and is a source of misanterstand, like in the method for example. It's subjective, and I more serious than you seem to believe often you and JP. My springfull personnality I surmise !😉 But as ancient IT manager and at 45 now you can have the faith of I can know what a serious job and method can be, even if here it stays an hobby for me and I'n not putting my life in it (humm it takes a lot of my free time in this moement). But If I write it's because I know difference are interresting enough for the ones who more than one both to acmodate the spices & flavor.
I think I will write the following tommorow, family first this week end. But the result will not be universal, I writed this already. But i will try to give three "enough" change to match with 3 different sound personnality for the transparent system as I think mine is good enough. It's without JG buffer with the genuine PS and just few caps : thinked to be non expensive.
get it as it is, try it if you want or have the courage. It's an hobby. maybe this tweak will be a good one in the flacor way, maybe not. But try before comment, if good say it. If not good say it also.
It's just for sharing my little work around cooking without pretention from me to be better than others.
Maybe JP or someone else maid this work, they will say if they want, here I share with the prcaution to say it's not universal but work with my system to have a better match after the excellent initial work of JB/Subbu.
So no podium rewarded for me, it's just sharing
Well i will write as before the procedure more than a podium of changes. But I will explain in the procedure where are the change and where I believe people can start with the less modification in regards to the flavor.
Flavor is a biased word but here I write because I think my seconde version adapted with each flavor is twice better than the neutral stock of the BOM because fit perfectyly with the personality of the whole system and the flavor of the owner. It has the meritt to be said and explained. ! So can't comment really the sense of flavor but as I said in curry term: its' more here than hot or not, a good curry make the main good lamb better but don't improve a bad lamb.
From coming to the engeneer world I always found sterile tje way of process with logistics hints 1. 2. 3. because in a mixed world, words and story are important to let the people understand. In my professional life, 50% o my jmanagement job was to clarify and have pedagogic inputs for both workers and top managemen. Sometimes shortcuts are not understood and here for me it's difficult as my english is like you know more celtish than americaglish !
Well I hope than the procedure will be longer to wrote but give also for the last tweaker something to understand; But I'm the first to write again : don't go if one pcb only. When the lamb is too much cooked, it become wasted at the risk to have not meal anymore ! The only result with a broken pcb is your ears will starve. I make it because I like to cook and here had a good result so I share; take it like it is... enthusiasm!
Important : the story has to come... so please don't write it before and for me like above as I'm not totaly enterily agree with your last input above. People will pick in what they want even if the french cook book are always longer than in others language... sometimes for the best ! (cliché !?)
Flavor is a biased word but here I write because I think my seconde version adapted with each flavor is twice better than the neutral stock of the BOM because fit perfectyly with the personality of the whole system and the flavor of the owner. It has the meritt to be said and explained. ! So can't comment really the sense of flavor but as I said in curry term: its' more here than hot or not, a good curry make the main good lamb better but don't improve a bad lamb.
From coming to the engeneer world I always found sterile tje way of process with logistics hints 1. 2. 3. because in a mixed world, words and story are important to let the people understand. In my professional life, 50% o my jmanagement job was to clarify and have pedagogic inputs for both workers and top managemen. Sometimes shortcuts are not understood and here for me it's difficult as my english is like you know more celtish than americaglish !
Well I hope than the procedure will be longer to wrote but give also for the last tweaker something to understand; But I'm the first to write again : don't go if one pcb only. When the lamb is too much cooked, it become wasted at the risk to have not meal anymore ! The only result with a broken pcb is your ears will starve. I make it because I like to cook and here had a good result so I share; take it like it is... enthusiasm!
Important : the story has to come... so please don't write it before and for me like above as I'm not totaly enterily agree with your last input above. People will pick in what they want even if the french cook book are always longer than in others language... sometimes for the best ! (cliché !?)
Last edited:
You're doing a good job Eldam and your enthousiasm is unstoppable. For some reason it is hard to keep up with you (for me, could be the celtish/ameriglish which words are enough for a smile) but you convinced me to continue testing different parts. I particularly like the way you used a 75 Ohm BNC soldered straight to the board. I will have 2 more V3 kits sent to me by Subbu so I possibly will try out your most recent recommendations. A simple list or spread sheet with all 3 cap recipes would make things handy for others to try out.....I see Gary tries to give you a small push in the back to that direction too.
BTW our new PSU design is being reviewed by Subbu so we should have something new to try out for you soon as the first test-PCBs will be manufactured. As promised you will receive a board to try out if only to make up for the sometimes too harsh comments by me, I hope you will accept my offer and do an awful lot of beta testing after you made a BOM version first (of course) 😉
Quote of the day:
BTW our new PSU design is being reviewed by Subbu so we should have something new to try out for you soon as the first test-PCBs will be manufactured. As promised you will receive a board to try out if only to make up for the sometimes too harsh comments by me, I hope you will accept my offer and do an awful lot of beta testing after you made a BOM version first (of course) 😉
Quote of the day:
Last edited:
I know who Garry is, a kind man and a gentleman.
You people from netherland, America come from the "rigid" and little puritism protestant world. Hahahaha ! But I have to say than I had already said to you with "clin d'oeil" that some are rock and roll in yur countie like the one I like Erasmus... But the greater one from your countrie Spinoza show an excelent example of what I try to say with its : one, to , three arrengement ! Efficient, but hard to read ! But I like him as I say in french (I don't think you will undrrstand the word game) : "a little ethic is an etiquette !"
Frenches are a melting pot from the antics times, a crossed country: barbarians, romains, galls, celtics, saxon, englishs (the Angles, the Pictes...). So we have humour like englishs. Most of the time it is for the sport, sometimes to send messages bu in a good way, because we are galls and bad way is my hand through your face like the Spartan and like an indonesian froma holland teach me during 15 years ! But here we are cool. In my job we joke if 50 servers on 100 are falling down, because we can master and jike in same time ! Frenchs are like that, sometimes sad, sometimes not friendly with turists more by complex than own personality. But we like the jokes and can be efficient also (a good match is th english humor : more in the reserved for english more rock and roll for the frenches but second degree also!).
Entusiasm in french have two siginification because two words :
First is "amateur" in the sens of non professional : can be pejorativ.
Second is"enthousiaste" : personn who are very imply, involve with volounteer attitud.
So all and its opposite ! french ilanguage is very efficient for the wordgames, the most art is the "contrepêt" but you have to be french to understand it and not all the french have the ear to do so ! But it's an art like to have a second degree word for an english gentleman !
Sheet, it's comming philosophical, where are my pills and botle of wine ?
Argh it's time to diner;;;;; sorry !
You people from netherland, America come from the "rigid" and little puritism protestant world. Hahahaha ! But I have to say than I had already said to you with "clin d'oeil" that some are rock and roll in yur countie like the one I like Erasmus... But the greater one from your countrie Spinoza show an excelent example of what I try to say with its : one, to , three arrengement ! Efficient, but hard to read ! But I like him as I say in french (I don't think you will undrrstand the word game) : "a little ethic is an etiquette !"
Frenches are a melting pot from the antics times, a crossed country: barbarians, romains, galls, celtics, saxon, englishs (the Angles, the Pictes...). So we have humour like englishs. Most of the time it is for the sport, sometimes to send messages bu in a good way, because we are galls and bad way is my hand through your face like the Spartan and like an indonesian froma holland teach me during 15 years ! But here we are cool. In my job we joke if 50 servers on 100 are falling down, because we can master and jike in same time ! Frenchs are like that, sometimes sad, sometimes not friendly with turists more by complex than own personality. But we like the jokes and can be efficient also (a good match is th english humor : more in the reserved for english more rock and roll for the frenches but second degree also!).
Entusiasm in french have two siginification because two words :
First is "amateur" in the sens of non professional : can be pejorativ.
Second is"enthousiaste" : personn who are very imply, involve with volounteer attitud.
So all and its opposite ! french ilanguage is very efficient for the wordgames, the most art is the "contrepêt" but you have to be french to understand it and not all the french have the ear to do so ! But it's an art like to have a second degree word for an english gentleman !
Sheet, it's comming philosophical, where are my pills and botle of wine ?
Argh it's time to diner;;;;; sorry !
That might all be true but please analyse our reactions to eachothers posts....you will see that neither of us answers to questions or specific remarks. We could conclude that we have a kind of communication misalignment 😉
Ach so !Communicubication is already hard in our mother language so forums and writting like me in an another language so badly do not help for sure 
.
I’m very suscipicious when a standalone use ‘ours’. It reminds me Cantor theory or the worst days of any dictaturs ! Are you talking for me now ?😛
No understood all what you wrote but more honored than everything else !
All those futil pursuit reminds me old chineese stories about loosing the face ! : “ I tell you my dog is older than yours!” ; “ Oh no my dog is older than yours I’m sure your dog is misaligned !”; “What ! my Miss is aliened ?” ….pif, paf, poum, boum ! 😉 “Banzaï !” (in Spartan greek translation, it is : “Whaou, Whaou!”).😀
Dringgggg time to drink !
.I’m very suscipicious when a standalone use ‘ours’. It reminds me Cantor theory or the worst days of any dictaturs ! Are you talking for me now ?😛
No understood all what you wrote but more honored than everything else !
All those futil pursuit reminds me old chineese stories about loosing the face ! : “ I tell you my dog is older than yours!” ; “ Oh no my dog is older than yours I’m sure your dog is misaligned !”; “What ! my Miss is aliened ?” ….pif, paf, poum, boum ! 😉 “Banzaï !” (in Spartan greek translation, it is : “Whaou, Whaou!”).😀
Dringgggg time to drink !
Eldam,
If even JP can't resist your enthusiasm, who I am to try. Your latest efforts have convinced me to relisten to different parts for C22 (Avcc), C21 (Vreg), and C35/C32 (Vneg) - the different power supply caps for the ES9023. I took the V3 DAC that I just boxed up last week and removed C22, C21, and C35 and installed machined pins so that I could quickly swap parts back and forth and do these experiments without having to unsolder anything. Here are a few pictures of the DAC with the machined pins installed and then the "standard parts" that I used plugged into the sockets. I'll let the newly soldered connection settle in for a few hours and then see what I can hear when quickly moving parts in and out - of course, always with the power off when swapping parts.
First up - comparing caps for C35/C32. My current part is Wima MKS2 4.7uf. I will compare to Wima MKS2 1uf in the C32 location and a 1uf BG NX cap. A very quick listen shows subtle but noticeable improvement moving from 4.7uf to 1uf.
All the best.
---Gary
Attachments
Told you so 😀 This is the case when the Wima is in // with the ceramic 1 µF caps so effectiively we are around 3 µF capacity in total. As you know 1 µF in total is not enough and above 10 µF things seem to deteriorate. The ceramic caps can not be left out in any case as the DAC will perform worse without them. The pins of the chips need to be decoupled with very nearby low ESR caps. I know we are in the flavors/colors area but that may count for C32/35 only.
So performance seems best with a total value between 2 µF and 10 µF....there is room for experimentation with 0 (no film cap), 1, 2.2, 3.3, 4.7 and 6.8 µF film caps but I only have those in 5 mm pitch.... Will you also try out the 2.5 mm 1 µF Wima MKS2 ? I thought of giving Evox MMK 5 mm caps a second chance.
Last edited:
(*) it must be said that finding the optimal value was done with the old 100 µF value for C22.
Last edited:
I surmise at C35, if existing in 1 uF : SAL RPM Vishay to be good with a clear cap at C22 (Nichicon R7 than many have?) or like one of my 3 on the short list at C22 ?
At C22 : 100 uF SEPC was worst than 470 uF. After that I tried less with the 39uf Nichicon R7 or BG N 33 uf : bad result. I never went back to try 100 uF because the price of the caps and only tested near 470. I surmise the value no to be important after 100 uF at C22 nore the legs pitch or ESR but more the construction of the cap !
Thank you for the Evox ref.
At C22 : 100 uF SEPC was worst than 470 uF. After that I tried less with the 39uf Nichicon R7 or BG N 33 uf : bad result. I never went back to try 100 uF because the price of the caps and only tested near 470. I surmise the value no to be important after 100 uF at C22 nore the legs pitch or ESR but more the construction of the cap !
Thank you for the Evox ref.
Gary, nice idea to solder in those machined pins for part swapping. However they will significantly increase inductance. So when you find your subjectively preferred combination just keep the pins 😉
kind regards, Daniel
kind regards, Daniel
Hi Garry,
Seeing your photograph, I believe it should be not hard to desoldering your two C19/C5 to give a try at 1 uF/16V Cornell acrylic (Panasonic have an equivalent plastic also) with your next visit at Mouser/Digikey. New Ipdia glass cap is promising also, but hard to sold (soldering must not touch the body and the pads are below the cap not on the sides also like the more classic smt caps: not DIY friendly.
Here Garry & JP : are you talking about the same DAC ? Garry has the JG buffer & another PS, concistency of the try have not the same value between you... I surmise the caps of the JG Buffer to hide the final result maybe !
I worked for a cheap mod with the stock version and the 2V output of the DAC.
PS has a lot of influence : like I see e.g. with the little 0.1 uf at the output or its bigger brother. I swaped a green 0.1 vishay i used to like with wima BOM & the same fm 4.7 who excell here if no BG of course...
I give it a try if I have more time with the PS by swapping all the Pan FM caps with Lelon 125° cheap aluminium for the first big cap, Pan FC or many else for the second big after the comon chocke...
2 cents idea... it's an infinite wheel, we have certainly to stop it for finding the Nirvana... Heard this week in Family : Astonishing eyes of Rita by Anouar Brahem, Porgy& Bess, Armstrong/Fitzgerald version and some opera's songs... nice moments...
Seeing your photograph, I believe it should be not hard to desoldering your two C19/C5 to give a try at 1 uF/16V Cornell acrylic (Panasonic have an equivalent plastic also) with your next visit at Mouser/Digikey. New Ipdia glass cap is promising also, but hard to sold (soldering must not touch the body and the pads are below the cap not on the sides also like the more classic smt caps: not DIY friendly.
Here Garry & JP : are you talking about the same DAC ? Garry has the JG buffer & another PS, concistency of the try have not the same value between you... I surmise the caps of the JG Buffer to hide the final result maybe !
I worked for a cheap mod with the stock version and the 2V output of the DAC.
PS has a lot of influence : like I see e.g. with the little 0.1 uf at the output or its bigger brother. I swaped a green 0.1 vishay i used to like with wima BOM & the same fm 4.7 who excell here if no BG of course...
I give it a try if I have more time with the PS by swapping all the Pan FM caps with Lelon 125° cheap aluminium for the first big cap, Pan FC or many else for the second big after the comon chocke...
2 cents idea... it's an infinite wheel, we have certainly to stop it for finding the Nirvana... Heard this week in Family : Astonishing eyes of Rita by Anouar Brahem, Porgy& Bess, Armstrong/Fitzgerald version and some opera's songs... nice moments...
Last edited:
With ears check, i found the inductance between many pitch to have very few influences with the final result... many caps with 2 mm pitch were bad !
I assume also than it's less important because smd decoupling exist already and I choosed to go 0603 case everywhere (any courage for 0402 case !)
Daniel,
Thanks for raising that concern. I agree that this is a drawback that could affect the results. Fortunately all these caps are bypassed with at least a 1uf ceramic cap so that will partially ameliorate the bad effect at high frequencies.
I agree that these pins are only useful for quick experimentation and that the results will then need to be confirmed with parts that have been soldered directly to the board.
Regards,
---Gary
JP,
We seem to hear things the same way, i.e. your ears and my ears agree on what sounds good. I was impressed by the sound of the original BOM for the V3 DAC - it was a big step up from what I'd been able to achieve with the V2.6 DAC. The fact that I tried to improve it doesn't take away from what you achieved with the original BOM. I mention this to explain why I give your recommendations more priority for my experiments than other advice on this thread 😉. Other people report results that I am sure are good in their own systems but I often can't repeat the good results in my system.
Your recommendation of 1uf MKS2 for C32/C35 is good advice. So far I have swapped in 1uf Wima with 2.5mm leads, 1uf Wima with 5mm leads, 2.2uf Wima with 5mm leads, and 4.7uf Wima with 5mm leads. In the short time that I had to listen last night, I found that using parts with 2.5mm lead spacing in the C32 location is better. I like the 1uf Wima and also a 1uf BG NX cap that I have. These parts interact with the choice of C22 and C21, so the "best" choice depends on the entire suite of parts.
More to come.
---Gary
What is the function of the RC with C21 : a ref but for what ? Low noise ? filter ?
I tried without at a price of lowered tonal balance : more warm, more boomy also, low-medium&medium senf in front of the speakers..like some bad horns ! A little distord and resonant also !
This air load of few pF is funny to try, also to understand the role of C21. I tried with the first pcb 20 uf for laughing... it was... no comment ! But what could be the result with some cap between few nf to 10 uf ?
That's why I as for the exact role of C21 ? If a noise ref what happen with a very clean PS board like the one Gary use with the JG buffer after !
It coul be a way to tune this chip to the taste of the listener ? Or very bad idea !
Hou la la... I said I stoped !
I tried without at a price of lowered tonal balance : more warm, more boomy also, low-medium&medium senf in front of the speakers..like some bad horns ! A little distord and resonant also !
This air load of few pF is funny to try, also to understand the role of C21. I tried with the first pcb 20 uf for laughing... it was... no comment ! But what could be the result with some cap between few nf to 10 uf ?
That's why I as for the exact role of C21 ? If a noise ref what happen with a very clean PS board like the one Gary use with the JG buffer after !
It coul be a way to tune this chip to the taste of the listener ? Or very bad idea !
Hou la la... I said I stoped !