Those resistors are optional, Digital Out will work ok with most DACs without them.
Shigaclone has a DAC already built into it, by the way.
Some DACs may not like the high (TTL) level signal the shiga puts out. So unless you are sure, its better to keep the resistors or some other means of dropping the signal.
EUVL: no other A/Bs yet, I got distracted by an itch but will be back to it soon. The only DAC I have left to compare is the Pass D1 clone....... soon!
Fran
EUVL: no other A/Bs yet, I got distracted by an itch but will be back to it soon. The only DAC I have left to compare is the Pass D1 clone....... soon!
Fran
I decided to publish my conclusions from my "big capacitor tests" so far. I will still be buying and testing more caps, as I stumble upon them, and will post if any of them are worthwhile.
I don't think it makes sense for me to write all the long individual descriptions, as most people will want a simple answer - which is the best one.
So, here goes:
PSU caps
The best overall is still Black Gate FK 2200/35 -> BG STD 1000/25.
The best non-Black Gate combination is Elna Starget 2200/25 -> Sanyo Os-Con SA 1000/16. Elna Starget has a pleasant, natural sound with good detail. Os-con has a very matter-of-factly presentation - neutral and detailed.
Of other caps:
Panasonic FC: grainy and uninteresting sound. Silmic II: good tonality, but unbearably poor transient performance. Cerafine: not entirely correct in midrange and sound artificial somehow. Nichicon FG: Warm and pleasant, but detail loss due to sound "glow". Nichicon FW: rich bass, warm sound. Nichicon KZ: also warm-ish presentation. Nichicon BP: ok, but unremarkable. Rubycon ZLH: good detail, but dry presentation and grainy bass. Rubycon ZL: still dry, but better bass and overall presentation more agreeable; combined with Starget took the second place. Panasonic AM: average detail, but remerkably free of major flaws - it is the only capacitor in the test that sounded ok to me when combined with itself (ie put both before and after the reg).
C916
The winner overall is Black Gate N 47/50.
The non-Black Gate winner is a tie between Sanyo Os-Con SA 47/16, and... no cap at all. Os-Con is neutral, accurate and has very good detail, "no cap" loses some resolution, but gives a much more uncoloured tonality.
Close third place goes to Elna Starget 47/16 for its natural and pleasantly detailed sound.
Comments above, regarding other cap types, are broadly applicable to the small caps too, so I won't be repeating them. If anyone would like a more specific description of a certain cap, please ask.
So my currently recommended complete "Shigaclone kit" would consists of:
1x RC-EZ31 / RC-EZ51
4x MSR860 diodes
1x Elna Starget 2200uf/25V
1x LM7808
1x Sanyo Os-Con 1000uf/16
1x Sanyo 47uf/16
It sounds very good, not quite the Black Gate level of exquisite nuances, but very detailed and pleasant nonetheless. There is just one problem with this combination - there is a certain... distortion?.. inter-modulation?... in soprano range. It's not exactly a frequency peak, it's just certain soprano voices result in an unpleasant sensation that only borders on hearing. This sensation goes away when C916 is removed or replaced with BG N 47/50. So if you find that high voices have a certain annoying quality that is hard to pin down, try removing C916.
Finally, I have just sent one such "kit" to our forum colleague syklab - I hope he likes it, and will share his opinion with us when he receives it.
An externally hosted image should be here but it was not working when we last tested it.
I don't think it makes sense for me to write all the long individual descriptions, as most people will want a simple answer - which is the best one.
So, here goes:
PSU caps
The best overall is still Black Gate FK 2200/35 -> BG STD 1000/25.
The best non-Black Gate combination is Elna Starget 2200/25 -> Sanyo Os-Con SA 1000/16. Elna Starget has a pleasant, natural sound with good detail. Os-con has a very matter-of-factly presentation - neutral and detailed.
Of other caps:
Panasonic FC: grainy and uninteresting sound. Silmic II: good tonality, but unbearably poor transient performance. Cerafine: not entirely correct in midrange and sound artificial somehow. Nichicon FG: Warm and pleasant, but detail loss due to sound "glow". Nichicon FW: rich bass, warm sound. Nichicon KZ: also warm-ish presentation. Nichicon BP: ok, but unremarkable. Rubycon ZLH: good detail, but dry presentation and grainy bass. Rubycon ZL: still dry, but better bass and overall presentation more agreeable; combined with Starget took the second place. Panasonic AM: average detail, but remerkably free of major flaws - it is the only capacitor in the test that sounded ok to me when combined with itself (ie put both before and after the reg).
C916
The winner overall is Black Gate N 47/50.
The non-Black Gate winner is a tie between Sanyo Os-Con SA 47/16, and... no cap at all. Os-Con is neutral, accurate and has very good detail, "no cap" loses some resolution, but gives a much more uncoloured tonality.
Close third place goes to Elna Starget 47/16 for its natural and pleasantly detailed sound.
Comments above, regarding other cap types, are broadly applicable to the small caps too, so I won't be repeating them. If anyone would like a more specific description of a certain cap, please ask.
So my currently recommended complete "Shigaclone kit" would consists of:
1x RC-EZ31 / RC-EZ51
4x MSR860 diodes
1x Elna Starget 2200uf/25V
1x LM7808
1x Sanyo Os-Con 1000uf/16
1x Sanyo 47uf/16
It sounds very good, not quite the Black Gate level of exquisite nuances, but very detailed and pleasant nonetheless. There is just one problem with this combination - there is a certain... distortion?.. inter-modulation?... in soprano range. It's not exactly a frequency peak, it's just certain soprano voices result in an unpleasant sensation that only borders on hearing. This sensation goes away when C916 is removed or replaced with BG N 47/50. So if you find that high voices have a certain annoying quality that is hard to pin down, try removing C916.
Finally, I have just sent one such "kit" to our forum colleague syklab - I hope he likes it, and will share his opinion with us when he receives it.
If there are audiophile gods, they smiled upon me brightly last night 
It started with me thinking about C906 - it is clearly a critical part, and I have always felt that the stock electrolyte is a poor component. The commercial film capacitors I tried as a replacement proved disappointing, so there was only one option left - I had to make my own capacitor.
I went for the stacked foil geometry as it is the only "true" capacitor, with no side effects. For conductor plates I cut 70x90mm soft OFC copper 0.075mm foil rectangles. For dielectric I used a very fine natural silk fabric, soaked with cosmetic-grade linseed oil. Alternating layers are then sandwiched between two wooden planks and everything clamped together with aluminium screws. For leads I used 22awg solid core tin-plated copper wire in impregnated cotton insulation.
Last night I finished and tested it - capacitance measured at 11.3nF.
So I plugged it in and... it caught me completely off guard. Remember, I have been testing capacitors for weeks now, and every single one of them so far was a disappointment.
What I got this time was not an upgrade. It was a transformation. I don't even know where to begin describing it.
Everything suddenly sounds so real. Voices exhibit all these minute features and imperfections that you hear from a person singing without amplification; this brings incredible intimacy, sweetness and emotion. Instruments brim with beautiful overtones. Dynamic passages have staggering slam and dynamic range is much wider. Stereo and depth perception took a quantum leap.
I could go on, but I think you get the picture I can recall only a few instances in my HiFi building history when changing a single part made such a fundamental improvement. So I decided to call my capacitor immodestly "VNP Cap" - Very Nearly Perfect Capacitor
And best of all - what I tested was just a prototype. I already have plans for significant improvements to its construction
It started with me thinking about C906 - it is clearly a critical part, and I have always felt that the stock electrolyte is a poor component. The commercial film capacitors I tried as a replacement proved disappointing, so there was only one option left - I had to make my own capacitor.
I went for the stacked foil geometry as it is the only "true" capacitor, with no side effects. For conductor plates I cut 70x90mm soft OFC copper 0.075mm foil rectangles. For dielectric I used a very fine natural silk fabric, soaked with cosmetic-grade linseed oil. Alternating layers are then sandwiched between two wooden planks and everything clamped together with aluminium screws. For leads I used 22awg solid core tin-plated copper wire in impregnated cotton insulation.
Last night I finished and tested it - capacitance measured at 11.3nF.
So I plugged it in and... it caught me completely off guard. Remember, I have been testing capacitors for weeks now, and every single one of them so far was a disappointment.
What I got this time was not an upgrade. It was a transformation. I don't even know where to begin describing it.
Everything suddenly sounds so real. Voices exhibit all these minute features and imperfections that you hear from a person singing without amplification; this brings incredible intimacy, sweetness and emotion. Instruments brim with beautiful overtones. Dynamic passages have staggering slam and dynamic range is much wider. Stereo and depth perception took a quantum leap.
I could go on, but I think you get the picture
I can recall only a few instances in my HiFi building history when changing a single part made such a fundamental improvement. So I decided to call my capacitor immodestly "VNP Cap" - Very Nearly Perfect Capacitor And best of all - what I tested was just a prototype. I already have plans for significant improvements to its construction
An externally hosted image should be here but it was not working when we last tested it.
If there are audiophile gods, they smiled upon me brightly last night
It started with me thinking about C906 - it is clearly a critical part, and I have always felt that the stock electrolyte is a poor component. The commercial film capacitors I tried as a replacement proved disappointing, so there was only one option left - I had to make my own capacitor.
I went for the stacked foil geometry as it is the only "true" capacitor, with no side effects. For conductor plates I cut 70x90mm soft OFC copper 0.075mm foil rectangles. For dielectric I used a very fine natural silk fabric, soaked with cosmetic-grade linseed oil. Alternating layers are then sandwiched between two wooden planks and everything clamped together with aluminium screws. For leads I used 22awg solid core tin-plated copper wire in impregnated cotton insulation.
Last night I finished and tested it - capacitance measured at 11.3nF.
So I plugged it in and... it caught me completely off guard. Remember, I have been testing capacitors for weeks now, and every single one of them so far was a disappointment.
What I got this time was not an upgrade. It was a transformation. I don't even know where to begin describing it.
Everything suddenly sounds so real. Voices exhibit all these minute features and imperfections that you hear from a person singing without amplification; this brings incredible intimacy, sweetness and emotion. Instruments brim with beautiful overtones. Dynamic passages have staggering slam and dynamic range is much wider. Stereo and depth perception took a quantum leap.
I could go on, but I think you get the picture
And best of all - what I tested was just a prototype. I already have plans for significant improvements to its construction
An externally hosted image should be here but it was not working when we last tested it.
I made a cotton-copper capacitor some time ago after testing so many capacitors that all sounded wrong and finally the right sound came from diy cap.
I am happy that you discovered that kind of sound unobtainable with any other cap in my experience. Everything sounds so right.Try another cotton cap on digital out ,sound is amazing.Another transformation like with c906.Just try to eliminate input cap in the dac if there is some placed for security reason by dac designer ,check the circuit.
Last edited:
Hi nemo1968, thanks for your comments. There were input caps in my dac originally, but I removed them. I consciously chose silk over cotton for dielectric, as in my experiments silk had clearly superior sound (more firm, but more delicate at the same time).
Hi Erik, don't worry I will share my capacitors with the worldIncredible fun to read !!!!
And now the question you were expecting....what do I have to do to get me one of your homemade caps (ofcourse willing to pay all costs)..........or do I have to get myself a Duelund...
I just want to do some more experimenting first. At the minute this capacitor is in the "proof of concept" stage and has quite a few rough edges (literally).Also, I need to think of a way to build it a bit more efficiently, because in its current form it takes so much time to make, that I would have to charge nearly Duelund prices for it...
Although one has to keep in mind, Duelund don't use wood and silk, but paper and.. paper in their capacitors There are two potential problems with VNP Cap:
First, physical size. It is really quite large. The prototype is 9x10x1.5cm and I don't think I will be able to make it much smaller. In fact, I want to use thicker mahogany boards instead of current thin pine planks, which will pretty much double the thickness.
Second, linseed oil is a "curing oil", meaning it cures (becomes solid) over time. I have little idea how this may affect capacitor sound. In my experiments with cable insulation, curing further improved the sound, but I don't know how this relates to capacitors.
On my list of things to try is:
- using wax instead of (or mixed with) linseed oil. I already have some natural, unprocessed wax from a beehive keeper.
- the "electron transfer principle". I want to see if I can incorporate some of the famous Black Gate technology in my caps by making a multi-layered dielectric, containing graphite particles.
This is going to be fun
uncle_leon ,
I use cotton primarily because of it's low dielectric constant value of 1,3-1,4 and silk have 2,5-3,5 so I never tried it but as you reported different performance between two materials I must try silk too.The dimensions of my cap for 11nF value are 90x60x25mm and I like the low loss sound of dry cotton (the closest material to air characteristic) and the dimensions are not the problem to me.I am tightening the cap in four angles of the cap body and I use mechanically inert and strong nonmagnetic material for clamps so the cap can be made very rigid.
I am hopelessly bad photograph but I guess that you will be able to see how I managed the cap clamping.
I use cotton primarily because of it's low dielectric constant value of 1,3-1,4 and silk have 2,5-3,5 so I never tried it but as you reported different performance between two materials I must try silk too.The dimensions of my cap for 11nF value are 90x60x25mm and I like the low loss sound of dry cotton (the closest material to air characteristic) and the dimensions are not the problem to me.I am tightening the cap in four angles of the cap body and I use mechanically inert and strong nonmagnetic material for clamps so the cap can be made very rigid.
I am hopelessly bad photograph but I guess that you will be able to see how I managed the cap clamping.
Attachments
Although one has to keep in mind, Duelund don't use wood and silk, but paper and.. paper in their capacitors
Duelund VSF-Cu Copper Capacitor - 100VDC – 2% tolerance
Technical specifications (according to manufacturer): “The Virtual stack foil capacitor, a capacitor made specially for loudspeaker crossovers. This new capacitor is the realisation of a lifelong ambition. It consists of copper foil and high-density paper, soaked in pure mineral wax, coated with pure cocoon silk and finally, treated with special lacquers made from natural materials. It is without question a "green" product. The actual construction has been overseen by Mr. Steen Duelund Chief Engineer of Duelund Coherent Audio. Mr. Duelund has dictated the capacitor's special characteristics, i.e. its flatness, giving the opportunity to create its stack-foil working manner ensuring very low inductance, forming a solid block resulting in a form free of internal resonance. Furthermore, the flatness makes it easier to achieve greater values by connecting the capacitors in parallel”.
......
Hi Nemo, I have seen your capacitor pictures back in this thread, and I actually used them as an inspiration to a degree. One big difference is that I rejected the use of any plastics. I prefer natural materials, because they are more pleasant to work with and have more pleasant sound as well - and are friendly to the environment. Only problem is, they are more expensive, and actually hard to get sometimes.
I was wondering how come you had to use twenty-two layers in your cap to get to 11nF?... I only had to use between two and six, depending on how hard I clamp them.
Duelund VSF-Cu Copper Capacitor - 100VDC – 2% tolerance
Technical specifications (according to manufacturer): “The Virtual stack foil capacitor, a capacitor made specially for loudspeaker crossovers. This new capacitor is the realisation of a lifelong ambition. It consists of copper foil and high-density paper, soaked in pure mineral wax, coated with pure cocoon silk and finally, treated with special lacquers made from natural materials. It is without question a "green" product. The actual construction has been overseen by Mr. Steen Duelund Chief Engineer of Duelund Coherent Audio. Mr. Duelund has dictated the capacitor's special characteristics, i.e. its flatness, giving the opportunity to create its stack-foil working manner ensuring very low inductance, forming a solid block resulting in a form free of internal resonance. Furthermore, the flatness makes it easier to achieve greater values by connecting the capacitors in parallel”.
Ah, my bad then - I missed the "coated with silk" part
Well, at least it confirms my choice of materials I did some more testing on Sunday, and it seems that, unlike with the PSU caps, the actual capacitance matters quite a lot here. Increasing capacitance to 18nf resulted in a stark loss of resolution, pushing it back almost to the stock cap level.
One side effect I noticed while using VNP cap is that it decreases the ability of the drive to cope with scratched or warped CDs. I'm guessing it drives the optics harder, resulting in less tolerance.
Also, the graphite dielectric seems to be a dead end - I made another capacitor with silk layers soaked in mixture of linseed oil and micronised graphite. I used the following geometry (I included greaseproof paper layer to prevent excessive leakage current via graphite):
WOOD
=====
...
Copper foil
Silk / Oil /Graphite
Greaseproof paper
Silk / Oil /Graphite
Copper foil
...
=====
WOOD
The resulting sound had a strange midrange colouration, and a "heavy" sound (as opposed to "airy") - lacking the higher harmonics. The resolution also seemed a touch less than the original cap.
It's amazing how much difference all the different details make, and how right I got the first one, doing it by intuition.
Relisted for £30......
JVC RC-EZ51H Shigaclone donor Finally on eBay (end time 18-May-11 15:21:33 BST)
JVC RC-EZ51H Shigaclone donor Finally on eBay (end time 18-May-11 15:21:33 BST)
I learned over the years to still keep listening to my "gut-feeling" and more than often I get myself an "out-of-the-box" reward for that
This Shiga contraption is capable (on a fairly good system up to a cost-no-object system) of revealing ANY change you make on it........
I know that this changed my point of view(s) on DIGITAL in more than one way...
..
Hi Uncle
There are
Nichicon FX here 1000uf 63v @ US$ 2.5
Illinois caps 100uf 35 @ US$ 1
BGF 1000uf 35V @ US$ 25
wanna try ?
Last edited:
Hi Nemo, I have seen your capacitor pictures back in this thread, and I actually used them as an inspiration to a degree. One big difference is that I rejected the use of any plastics. I prefer natural materials, because they are more pleasant to work with and have more pleasant sound as well - and are friendly to the environment. Only problem is, they are more expensive, and actually hard to get sometimes.
I was wondering how come you had to use twenty-two layers in your cap to get to 11nF?... I only had to use between two and six, depending on how hard I clamp them.
Hi, uncle_leon
I have used 43x90mm Cu layers instead of 70x90.My dielectric is dry ,non soaked type and the copper thickness before clamping is 0,27mm resulting in much bigger distance between electrodes and less capacitance.Linseed oil dielectric constant is 3.2-3.5 so you easily get 11,3 value with less copper.For example only the aluminum oxide in an electrolytic cap have dielectric constant of 10 and ethylene gylcol which is part of electrolyte solution and is needed to slow down the electrolyte solution evaporating , have the constant of 37 and the cap loose its value with electrolyte evaporation during years.The larger is dielectric constant of insulation with less electrode distance between them the greater is the capacitance.
My idea was to get the closer sound to air capacitor without inter electrode resonances after been listening to some vintage variable cylindrical air caps on c906 place which I liked very much so I have not used wood because Cerrock is dead quiet in that application and I don't need another singing cap
The material is actually used in furniture workshops and is much more pleasant to work with because it leaves clean corners and holes after sawing and drilling and above everything I get scrap-material in local workshop which would otherwise finish in dustbin without being recycled
My alternative choice would be ebony but I leave it to string-piano instruments manufacturers to be able to listen to it recorded in the future seeing the national souvenirs industry in Kenya have nearly wiped out that wood
Have you already tried your cap on s/pdif out and what is the result?I have used 10nf as sufficient value.
Nemo, where did you get such thick audio-grade copper foil? I wanted to try ribbon speaker cables at some point, but the Mundorf copper foil i have, while perfect for capacitors, is a bit too thin for speaker cable.
I intend to try dry silk at some point too, I will report back as soon as I do.
I would love to use ebony (and probably will at some point), but as you say, it's not just expensive, it's rare too. I picked mahogany as possibly the next best thing (although Baltic Birch would be worth considering too).
No, I haven't tried my cap at D-Out, simply because it works ok without any capacitors at all
Hi Pocoyo, I could try the Nichicon FX and the Illinois cap (have you got any values lower than 100uf?).
I don't think there is much point in me testing BG F. It wouldn't help anyone because they are practically extinct, and in any case they are very unlikely to be better than FK / STD. And besides, it's way too expensive to test it just for charity - unless you are willing to let me borrow two for a week.
I intend to try dry silk at some point too, I will report back as soon as I do.
I would love to use ebony (and probably will at some point), but as you say, it's not just expensive, it's rare too. I picked mahogany as possibly the next best thing (although Baltic Birch would be worth considering too).
No, I haven't tried my cap at D-Out, simply because it works ok without any capacitors at all
Hi Uncle
There are
Nichicon FX here 1000uf 63v @ US$ 2.5
Illinois caps 100uf 35 @ US$ 1
BGF 1000uf 35V @ US$ 25
wanna try ?
Hi Pocoyo, I could try the Nichicon FX and the Illinois cap (have you got any values lower than 100uf?).
I don't think there is much point in me testing BG F. It wouldn't help anyone because they are practically extinct, and in any case they are very unlikely to be better than FK / STD. And besides, it's way too expensive to test it just for charity - unless you are willing to let me borrow two for a week.
uncle_leon ,
Sorry for typo.I intended to write cotton but have came out copper.
Did you completely eliminated capacitors in s/pdif chain ,transport and dac side? I thought that one must be left for dac dc protection.
I use 12AWG copper foil for my PSU cables so why don't you try Mundorf on your speakers.12AWG foil is equal to 3,31mm2 and you can go bi-wire if you feel that you miss something with single run.You probably have 10AWG foil which have area of 5,27mm2.
Sorry for typo.I intended to write cotton but have came out copper.
Did you completely eliminated capacitors in s/pdif chain ,transport and dac side? I thought that one must be left for dac dc protection.
I use 12AWG copper foil for my PSU cables so why don't you try Mundorf on your speakers.12AWG foil is equal to 3,31mm2 and you can go bi-wire if you feel that you miss something with single run.You probably have 10AWG foil which have area of 5,27mm2.
Last edited:
I know, most audio designers think that a DAC has to have capacitors and/or transformers at the input, but this, apparently, is not always true. I removed mine after a suggestion from Peter (who designed my DAC) and it worked fine.
I think the potential danger here is much less than with amplifier input, because any DC at input will not get amplified further. So, as long as voltages stay within limits acceptable to the receiver chip, it should be ok to remove the input caps on any DAC. I know that theoretically all sorts of bad things can happen if the two devices are not galvanically separated, but hey - as long as it works and sounds good, I don't care about the potential problems it entails
Tried the 2.5V regulator at the C906 position.
Simpler to setup than I initially thought-the 5534 op-amp does not require a dual power supply or any power supply at all!
Shigaclone still works, so no issues there! (Simply removing C906 resulted in the clone not reading the table of contents)
The sound?
Giving it a day or so to burn in and will report back here.....
Simpler to setup than I initially thought-the 5534 op-amp does not require a dual power supply or any power supply at all!
Shigaclone still works, so no issues there! (Simply removing C906 resulted in the clone not reading the table of contents)
The sound?
Giving it a day or so to burn in and will report back here.....
