Hi IWC hence its similar to what I've always say what sounds great in 1 location doesn't mean that it will sound good in others. So 1 just have to slowly try at 1 location with different combi & listen. At the end of the day its what our ears hear regardless of specs of components used. Btw though it's big in size, one of my very favourite bypass caps to use is 0.47uf Mundorf Supreme. When its in sync with a location its sounds wonderful very very open & natural sounding
Cheers
Cheers
Last edited:
Thanks for the response sumotan, I've seen some BG series auction for $$/$$$ but I can try to find genuine, I've experimented with a few more capacitors as paralleled filters and I think I still prefer KZ (with it's own trade-offs), the results I've had was a smoother sound but still retaining enough detail, however like you said too much of one thing may not sound optimal, as with it being smooth made the music sound dull and less expressive so I went back to keeping the KZ.
If I keep the KZs in my amp (a bunch of them are also being used for poweamp rail decoupling) do you think mundorf mlytic AGs would pair well for main filters? I know you said a balanced sound but If you can elaborate a little more, maybe describe the sound of KZ treble/resolution against mlytic or mlytic against super through, even though that's not an option for me as they can't fit, but the KG type II is an option so if you have any experience feel free to share as well.
Talking about it's microdynamics/details would cover smearing which is something I noticed too in certain capacitors that sounded smoother, smeared subjectively for me can also encompass what others describe as shut in, slow, not open or too smooth, etc. I would normally play some Kingo Hamada song like Dakare Ni Kita Onna for testing the treble, the ending part of that song mainly, I'm essentially in the same situation as gentlevoice but I don't think they've tried the MLytics before.
polypropylene capacitors which have the lowest dielectric absorption, among most types so the archetype should sound fast and the least smeared but I've never actually tested them against a top tier electrolytic like KZ so I will do that sometime and see if any are better.
Hi Skye,
If we're just focusing on your power amp, What ps caps are you using ? My most favourite cap to use is 0.47uf Mundorf Supreme
as bypass, when there's synergy the sound becomes very spacious open, natural & this helps compensate for the sizzling treble
of the Kz. Btw not all KZs are created equal ya, Id stay away from 25v types only 50v & strange there's a sweet spot in the range
as well, 22uf version very very dynamic, treble less sizzle & sounds more balance. Too many varieties & brands to test, besides
its system dependant. Only thing that I can say is when one mucks around enough, you'll get to know the sound of the components
which will give you an idea of how to combine them to try & achieve the sound that one is looking for.
Cheers
If we're just focusing on your power amp, What ps caps are you using ? My most favourite cap to use is 0.47uf Mundorf Supreme
as bypass, when there's synergy the sound becomes very spacious open, natural & this helps compensate for the sizzling treble
of the Kz. Btw not all KZs are created equal ya, Id stay away from 25v types only 50v & strange there's a sweet spot in the range
as well, 22uf version very very dynamic, treble less sizzle & sounds more balance. Too many varieties & brands to test, besides
its system dependant. Only thing that I can say is when one mucks around enough, you'll get to know the sound of the components
which will give you an idea of how to combine them to try & achieve the sound that one is looking for.
Cheers
The poweamp is part of an integrated amp and currently I have 2x 330uF 50V KZ and 4x 22uF 50V KZ for poweamp rails, I haven't replaced the 22uF capactors on the IPS complementary differential pair yet, those are used to short frequencies above their cutoff so emitter degeneration feedback isn't applied. For the preamps power supply they are 2x 100uF silmic II which I plan to replace with Kaisei if I like from testing. The preamp rail decoupling are a few silmic II and some Nichicon KA.
Aside from that it's a very obvious difference when bypassing the filters (after all they are nearly 50 years old). I found a decent enough smooth sound for now without loss or smearing of detail using the same 2x 330uF KZ and some 2x 2200uF capacitors I desoldered from the output filter of a TPA3255 amp. Only way really I guess is to try it in my system and see how it sounds myself while the general opinion aside from invoking an expectation bias gives a good idea.
Aside from that it's a very obvious difference when bypassing the filters (after all they are nearly 50 years old). I found a decent enough smooth sound for now without loss or smearing of detail using the same 2x 330uF KZ and some 2x 2200uF capacitors I desoldered from the output filter of a TPA3255 amp. Only way really I guess is to try it in my system and see how it sounds myself while the general opinion aside from invoking an expectation bias gives a good idea.
Last edited:
Aaah so its a class D amp that your playing with Skye ? Well this means you'll have to add some warmth unless youl like its paticular SQ. Me played with 3 sets of Hypex amps before think they pretty much sound in way similar
Sorry if I didn't clarify properly, the amp I'm refering to and the one that I use as my main is class AB, I just got the capacitors from the output low pass filter of a TPA3255 amp I had lying around. I do think MOSFETs (which is what all these class D chips use) sound more warm than BJTs but it is a generalisation.
Hi Skye,
Apologies I misread as well so is it a 50 watt power amp or less ? Cause re-reading, your main caps seems of low uf value.
May I re direct you to take a slightly different route, re bridge rectifier change them to Shindegen D6SBN20 & take it from there
Apologies I misread as well so is it a 50 watt power amp or less ? Cause re-reading, your main caps seems of low uf value.
May I re direct you to take a slightly different route, re bridge rectifier change them to Shindegen D6SBN20 & take it from there
Hi sumotan,
yes it's 55 watts into 8 ohms here's the service manual, model is Realistic SA-2000.
This amp uses a 200V 6A Shindengen S5VB rectifier with standard silicon diodes I believe and it's the integrated type with all the diodes in one case so the Shindengen D6SBN20 which is the same rating but with Schottky diodes would work but with it's package like that I would leave it as the last thing to change out as it would be the most difficult to solder and try to fit, I am also not sure if it would still need the snubber capacitors and if I'm happy with the sound after replacing the stock 8000uF filters with new 10,000uF.
It said the 15 MB service manual file is too large so I will send the schematic instead
I think this might be in the wrong thread
yes it's 55 watts into 8 ohms here's the service manual, model is Realistic SA-2000.
This amp uses a 200V 6A Shindengen S5VB rectifier with standard silicon diodes I believe and it's the integrated type with all the diodes in one case so the Shindengen D6SBN20 which is the same rating but with Schottky diodes would work but with it's package like that I would leave it as the last thing to change out as it would be the most difficult to solder and try to fit, I am also not sure if it would still need the snubber capacitors and if I'm happy with the sound after replacing the stock 8000uF filters with new 10,000uF.
It said the 15 MB service manual file is too large so I will send the schematic instead
I think this might be in the wrong thread
Attachments
Last edited:
The reason that Im suggesting change the bridge cause it changes the SQ. My way is always to make the original PS side
as clean as possible before moving on with other tweaks of mods, hence even trafos on my projects are custom wound.
Who knows after upgrading the bridge you won't want to tweak anymore cause the D6Sbdn20 is very clean & grainless
So Sa2000 a vintage intergrated amp, are the sliders the volume pot, if so too bad can't upgrade that part. Lol
Cheers
as clean as possible before moving on with other tweaks of mods, hence even trafos on my projects are custom wound.
Who knows after upgrading the bridge you won't want to tweak anymore cause the D6Sbdn20 is very clean & grainless
So Sa2000 a vintage intergrated amp, are the sliders the volume pot, if so too bad can't upgrade that part. Lol
Cheers
Yea they tried to integrate the balance into the volume pot, looks cool and unique but not pratical, unfortunately no stepped attenuator with this one lol. I'll see what I can do about the D6SBN20, I have a variable linear power supply I can probably test it in, I was originally planning to use soft recovery.
Hi sumotan, Austrian Printer,
What you are suggesting is valid as far as supply noise being important. There is zero wrong with the factory rectifiers, or normal rectifiers that are used as replacements. They are actually very highly engineered and only inexpensive due to production volumes. This is a concept where most "modifiers" are completely off the rails. Fast recovery diodes generate more HF hash. They were designed for high frequency switching supplies. Regular rectifiers outperform them at normal AC grid frequencies. Soft recovery types are only mitigating the error of fast recovery to begin with, in other words you are pretty much back where you started except you spent money and effort to install parts that don't fit. Fantastic idea!
I really wish people would actually understand what they were doing with various parts. Similarly, increasing capacitance unnecessarily will not improve sound quality, and greatly increasing capacitance generates even more HF noise.
I get equipment in regularly that has changes like these. To restore good performance and solve some issues, I am sometimes forced to remove these "better" parts and install the correctly specified parts used in the original manufacture. The pulled parts are returned to the customer along with equipment that now meets performance specs.
A part is designed for specific applications. Putting a part designed for a different job in the wrong application does not improve performance. A more expensive part does not mean it is better. It means that production quantities are lower, and maybe the part is more costly to produce. That does not mean the different performance is a benefit in your application. I find the crowd that decides it knows better than the original design engineers (without engineering training of their own) makes extremely poor choices when working on equipment. They typically decide if a part is described similarly to other parts, the more expensive one is obviously better and the manufacturer "cheaped out". This is actually not completely true and varies between manufacturers.
I very seldom see correct information on the internet about improving sound equipment. Apparently untrained folks know better.
What you are suggesting is valid as far as supply noise being important. There is zero wrong with the factory rectifiers, or normal rectifiers that are used as replacements. They are actually very highly engineered and only inexpensive due to production volumes. This is a concept where most "modifiers" are completely off the rails. Fast recovery diodes generate more HF hash. They were designed for high frequency switching supplies. Regular rectifiers outperform them at normal AC grid frequencies. Soft recovery types are only mitigating the error of fast recovery to begin with, in other words you are pretty much back where you started except you spent money and effort to install parts that don't fit. Fantastic idea!
I really wish people would actually understand what they were doing with various parts. Similarly, increasing capacitance unnecessarily will not improve sound quality, and greatly increasing capacitance generates even more HF noise.
I get equipment in regularly that has changes like these. To restore good performance and solve some issues, I am sometimes forced to remove these "better" parts and install the correctly specified parts used in the original manufacture. The pulled parts are returned to the customer along with equipment that now meets performance specs.
A part is designed for specific applications. Putting a part designed for a different job in the wrong application does not improve performance. A more expensive part does not mean it is better. It means that production quantities are lower, and maybe the part is more costly to produce. That does not mean the different performance is a benefit in your application. I find the crowd that decides it knows better than the original design engineers (without engineering training of their own) makes extremely poor choices when working on equipment. They typically decide if a part is described similarly to other parts, the more expensive one is obviously better and the manufacturer "cheaped out". This is actually not completely true and varies between manufacturers.
I very seldom see correct information on the internet about improving sound equipment. Apparently untrained folks know better.
Very much depends on the pricing of the device, I guess. Accuphase probably didn't cheap out, most small manufacturers of pricy items didn't cheap out. Mass produced items to my experience are a mixed bag - good parts where it matters, junk where it does not matter a lot. Leaving hardly a lot of room for real improvements, IMHO.
This sort of derating (e.g., using caps with voltage rating > 2x the actual applied voltage in-circuit, and only parts rated for 105 ° C) was standard practice when I was designing avionics & aerospace circuits. This was to meet reliability and life span requirements of less than one failure in 10^9 hours. So, yeah, it's an excellent idea.
@Austrian Printer : With reference to your post #3027 I have only just read the post right now (been away for some time).
As to your question I do not really have a replacement suggestion for the KZ series. IMHO this is a challenge currently as, at least to my ears, some of them are quite unique in terms of overall "realism" (although with many of the voltage/capacitance types having a sort of screamy upper midrange/treble, as has been mentioned here).
Regarding the Kaiseis I can say that I have listened to their 25V/470uF types both in the bipolar and polarized versions. And to my ears they are somewhat comparable to the nichicon FG tonally, however, their dynamics are more insisting and reasonably realistic - albeit without the realism that I find with some of the KZs. Their room rendering was quite up front giving a closer perspective than some other capacitors. The conclusion on my listening tests of these capacitors was that "yes" they are good but also pricey and I decided to stay with the KZs. It should be noted though that my current conclusion is that capacitors in general sound better/best in their 50-100 VDC model range - which I didn't know when I bought the Kaiseis - so likely higher voltage models could be better.
Cheers,
Jesper
As to your question I do not really have a replacement suggestion for the KZ series. IMHO this is a challenge currently as, at least to my ears, some of them are quite unique in terms of overall "realism" (although with many of the voltage/capacitance types having a sort of screamy upper midrange/treble, as has been mentioned here).
Regarding the Kaiseis I can say that I have listened to their 25V/470uF types both in the bipolar and polarized versions. And to my ears they are somewhat comparable to the nichicon FG tonally, however, their dynamics are more insisting and reasonably realistic - albeit without the realism that I find with some of the KZs. Their room rendering was quite up front giving a closer perspective than some other capacitors. The conclusion on my listening tests of these capacitors was that "yes" they are good but also pricey and I decided to stay with the KZs. It should be noted though that my current conclusion is that capacitors in general sound better/best in their 50-100 VDC model range - which I didn't know when I bought the Kaiseis - so likely higher voltage models could be better.
Cheers,
Jesper
Last edited:
Yes Jesper cant explain why perhaps lower esr or greater internal surface area etc me too only 50v above even if it low volatage application. For me the sweet spot is Kz 22 or 33uf very good sounding & dynamic
@sumotan : Hi 😉 ... I also don't have a conclusive explanation but I reckon that it may have to do with the tan delta specification which tends to drop (improve) at higher voltages. As I understand it this would lead to lower distortion levels at higher frequencies - which - everything else being equal - might explain "something" ... ?
Cheers, Jesper
Cheers, Jesper
Hi Jesper,
Yes perhaps your right, me don't want to get into tech nitty gritties, I've got very senistive ears so what ever things I change I can hear it even if its minute.
Cheers
Yes perhaps your right, me don't want to get into tech nitty gritties, I've got very senistive ears so what ever things I change I can hear it even if its minute.
Cheers
Hi AnahTech,
I went with 10,000uF 63V from 8000uF 50V for modern size convenience and the extra ripple current handling would help. Inrush shouldn't be much of an issue since the service manual specifies a tolerance of up to +50% Trying not to go too hard on the transformer.
Of course substituting a component in a circuit, if you know what you're doing it may or may not require adjustment of the circuitry for proper optimization and without issues, I'm not much knowledgeable in that area so I won't do it (in my main amp) and test it on a cheap supply I had instead. Just for fun of course, that wouldn't have any meaningful conclusion because variants like the PSRR would be different and it's feedback would not correct any TIM/IM/harmonic distortion caused by the leakage current or switching noise of non properly compensated diodes if it's extrinsic to the loop so my conclusion might be based on spurious factors, and like you said for a frequency of 120Hz the difference might be negligible but normally the advanges the other types offer (assuming if implemented well), can have better performance than standard silicon even for this application? Others have said they heard a difference even with well implemented blind testing and matched compensated circuits, meaning although the outcome is the same, there is a perceived difference which is pretty contradictory to me unless I consider it to be psychological/placebo. I don't know whether what all we can hear (assuming it's not psychological/placebo influenced) can be measured but surely it can only be measured to some extent, so if that's true I think you can only be pragmatic to the same extent as well.
Skye