At the least, any power filters should have competitive ripple current at lf, or risk sounding anemic and bass shy (if effect is audible). I like well-represented bass. Too high a hf ripple current figure here will have the effect of overly-energetic, fatiguing highs (if audible). Decisions such as these may be used to help take some off the top or add more to the bottom and vice versa as it relates to current @ frequency. This strategy can also be used to help tame shrillness by reducing available hf current.
Current is energy and energy is life. Amps don't lie.
Need a t-shirt with that quote on it🙂
Late to the party but this is a quality thread.
After 2 Years messing with preamps I can only agree that electrolytics affect sonics greatly in all audio applications. And it's down to specs and physical geometry.
Thanks for all comments here in this thread.
I am currently building list of capacitors for recap of my NAD c326bee . This topic is more then relevant to me.
Not sure if my question is related to the thread, you tell.
What bothering me is that if sound can be affected by way of capacitor is mounted on PCB.
For example in my amp all capacitors are sitting tight on PCB.
But when I try to see if I can fit into it KZ capacitors I see that they are much bigger then originals and they have bigger legs spacing (and legs width).
So, if I will install them, then they will sit in the air on their some time very long and curved legs.
Intuitively I understand that it is better to have really solid way of mounting, when cap fits exact holes on PCB and place it takes.
Do you think there can be a difference (immediate or when some time will pass) in sound between these two ways of mounting?
I am currently building list of capacitors for recap of my NAD c326bee . This topic is more then relevant to me.
Not sure if my question is related to the thread, you tell.
What bothering me is that if sound can be affected by way of capacitor is mounted on PCB.
For example in my amp all capacitors are sitting tight on PCB.
But when I try to see if I can fit into it KZ capacitors I see that they are much bigger then originals and they have bigger legs spacing (and legs width).
So, if I will install them, then they will sit in the air on their some time very long and curved legs.
Intuitively I understand that it is better to have really solid way of mounting, when cap fits exact holes on PCB and place it takes.
Do you think there can be a difference (immediate or when some time will pass) in sound between these two ways of mounting?
Well, maybe I was not clear enough.
Lets say the original capacitor is 5x11mm with 2mm legs spacing. And it sits with its bottom on PCB - I can not move it with hand.
I can freely fit there same value/voltage Nichicon KZ capacitor of size 10x16mm. There is enough space there for doing this. No close hot components, or any risk of shortening things.
But for such fit I will need to bend legs of KZ capacitor and when it is installed it will not sit as solid as original capacitor did.
It will "stand" on its legs above PCB.
Bottom of KZ capacitor will not sit on PCB and will be in the air.
So my questions - can it be a problem for sound? In any other aspect?
Lots of commercial gear, even high-end products, have small electrolytic caps soldered above the board with short legs. Larger ones need the support of sitting on the board usually.
The trend has been a reduction in package size with electrolytic caps, so it's less common to see someone trying to replace with physically larger caps unless increasing voltage and/or capacitance. I have a preamp where someone thought they were improving it by putting in caps that are too tall for the case, so they bent them over onto the differential pairs, forcing them to be bent. Not good practice in my view.
The trend has been a reduction in package size with electrolytic caps, so it's less common to see someone trying to replace with physically larger caps unless increasing voltage and/or capacitance. I have a preamp where someone thought they were improving it by putting in caps that are too tall for the case, so they bent them over onto the differential pairs, forcing them to be bent. Not good practice in my view.
My short experiments with lead lengths showed basically no difference. Other factors are far more important. You don't want them able to vibrate though.
5 figure tube preamps out there use point to point wiring with long capacitor leads sooo ..
I was just trying out KL in signal path and agree with some findings here. I prefer ES bipolar.
I think big blue epcos poly caps are the best at leaving no signature of their own, and like KZ for decoupling.
Biggest difference I've heard was swapping 100n 50v x7r to 100v versions. Still can't believe it.
5 figure tube preamps out there use point to point wiring with long capacitor leads sooo ..
I was just trying out KL in signal path and agree with some findings here. I prefer ES bipolar.
I think big blue epcos poly caps are the best at leaving no signature of their own, and like KZ for decoupling.
Biggest difference I've heard was swapping 100n 50v x7r to 100v versions. Still can't believe it.
.......................Wise Old Tech quietly sneaks out of this (yet another) capacitor thread before the crap hits the fan... 💩
Smaller parts, say less than 12mm in diameter, can usually stand off the board. Car ECUs even do this as standard practice.
The potential problem is when the leads are formed to fit the differently spaced holes in the pcb. The rubber seal in the bottoms of the caps are something that will fail eventually, so if the leads aren’t straight where they pass through this area, then you may have leaking problems prematurely.
For larger capacitors, just get the ones that fit.
The potential problem is when the leads are formed to fit the differently spaced holes in the pcb. The rubber seal in the bottoms of the caps are something that will fail eventually, so if the leads aren’t straight where they pass through this area, then you may have leaking problems prematurely.
For larger capacitors, just get the ones that fit.
I've done a lot of testing on caps since I posted this original thread. The KL caps are good if you need a good stable very low leakage cap. However, they are not great sounding caps for an analog section as they are somewhat sterile and lackluster.
For signal path, the ES bipolar are good, but actually the FG fine-gold caps are quite good for this as well. I think the FG sound a bit sweeter and richer if you don't like the more transparent ES bipolar.
KZ make just about the best post-regulator filter cap for analog power supply lines (i.e. +/-15V DC line). They do not work well in post-regulator digital power supply lines. They are just too strong/firm. Think of it like holding a pure cast iron pan and slamming it down on concrete. It is very strong, but it's just way too strong/firm for digital power supply lines. The sound becomes stilted/firm and doesn't have any depth - it lacks emotional involvement as well,
I just recently learned that KW are really not great caps. They are actually colored and have this lush overtone and bass that isn't tight. Not transparent or neutral at all. They are not as colored as Elna Silmic, but still too colored for me. FG are neutral caps, but are slightly messy/soft in the high frequencies when used as a DC power line filter. They make excellent compensation caps in other areas of the circuit.
If you want a good transparent cap for power supply filter and the KZ is too large to fit, the VR caps are excellent. If you need a 105 degree, the VZ are great as well.
After many tests with the KA cap (Nichicon's 105 degree "audio cap") I have come to the conclusion that these are just crap. Just putting one KA as a post regulator for digital destroyed the sound and just made everything sound crappy and unappealing.
Not all Nichicon "audio" caps are great. KG is nice for those who like that sound, but it pushes the midrange too hard and has loose bass and no high frequency resolution. The LKS caps are tilted towards the upper mids and don't have good bass - they tend to sound shrill.
If you want a good snap-in, I found the Cornell Dublier 380LX to be excellent. Very neutral cap with excellent resolution and great bass. The 380LX has an incredibly high ripple current capability (which is likely why bass control is so excellent). The Nichicon LLS is probably okay (and looks to be Nichicon's version of 380LX if you look at the specs), but I have not tested it enough to give any good conclusions.
Avoid caps with more than a 1.15 ripple multiplier between 120hz and 10kz. The 105 degree caps you find tend to have a higher multiplier between 1.25 and 1.5. This is when the cap has a very high ripple current capability at 10khz when compared to low 120hz (which is what A/C current runs at). These tend to tilt the sound towards the upper mids/highs and lose bass.
The Nichicon VZ caps have an excellent ripple multiplier of 1.15 for a 105 degree cap (but they are not high hour caps like HE or KY - not that it would matter in an analog scenario).
For signal path, the ES bipolar are good, but actually the FG fine-gold caps are quite good for this as well. I think the FG sound a bit sweeter and richer if you don't like the more transparent ES bipolar.
KZ make just about the best post-regulator filter cap for analog power supply lines (i.e. +/-15V DC line). They do not work well in post-regulator digital power supply lines. They are just too strong/firm. Think of it like holding a pure cast iron pan and slamming it down on concrete. It is very strong, but it's just way too strong/firm for digital power supply lines. The sound becomes stilted/firm and doesn't have any depth - it lacks emotional involvement as well,
I just recently learned that KW are really not great caps. They are actually colored and have this lush overtone and bass that isn't tight. Not transparent or neutral at all. They are not as colored as Elna Silmic, but still too colored for me. FG are neutral caps, but are slightly messy/soft in the high frequencies when used as a DC power line filter. They make excellent compensation caps in other areas of the circuit.
If you want a good transparent cap for power supply filter and the KZ is too large to fit, the VR caps are excellent. If you need a 105 degree, the VZ are great as well.
After many tests with the KA cap (Nichicon's 105 degree "audio cap") I have come to the conclusion that these are just crap. Just putting one KA as a post regulator for digital destroyed the sound and just made everything sound crappy and unappealing.
Not all Nichicon "audio" caps are great. KG is nice for those who like that sound, but it pushes the midrange too hard and has loose bass and no high frequency resolution. The LKS caps are tilted towards the upper mids and don't have good bass - they tend to sound shrill.
If you want a good snap-in, I found the Cornell Dublier 380LX to be excellent. Very neutral cap with excellent resolution and great bass. The 380LX has an incredibly high ripple current capability (which is likely why bass control is so excellent). The Nichicon LLS is probably okay (and looks to be Nichicon's version of 380LX if you look at the specs), but I have not tested it enough to give any good conclusions.
Avoid caps with more than a 1.15 ripple multiplier between 120hz and 10kz. The 105 degree caps you find tend to have a higher multiplier between 1.25 and 1.5. This is when the cap has a very high ripple current capability at 10khz when compared to low 120hz (which is what A/C current runs at). These tend to tilt the sound towards the upper mids/highs and lose bass.
The Nichicon VZ caps have an excellent ripple multiplier of 1.15 for a 105 degree cap (but they are not high hour caps like HE or KY - not that it would matter in an analog scenario).
I have done many projects where I have bent the legs of capacitors during installation so that the caps sits sideways down on the board. I have also mounted where there is about a 5-6mm length of lead between the cap and the board. Functionally, this all works fine. I'm sure there are going to be engineers that argue about the impedance impact of this extended lead length, but I have seen circuits where the path to the power supply capacitor is actually a 10-20mm tracing away from the main trace line (i.e. it is a 20mm length of trace between the cap and the actually power supply trace line on the board - like a "T").
But if you want to have all your caps mounted flat to the board, use KZ where you can (except for DC-blocking signal caps) and then use FG for compensation caps and VR/VZ for power supply filter caps. I should state that the 105 degree VZ caps should be used with care and you should not over-do the amount of these caps in a circuit. It can be a "modifier" cap to help dry out the sound (if it's too wet/lush/colored). I think it's also excellent for post-regulator power filter in digital circuits for transports and similar types of equipment.
That being said, caps that are standing on legs and not mounted firmly to the board can suffer more from electrical and mechanical resonance. This does affect sound quality.
But if you want to have all your caps mounted flat to the board, use KZ where you can (except for DC-blocking signal caps) and then use FG for compensation caps and VR/VZ for power supply filter caps. I should state that the 105 degree VZ caps should be used with care and you should not over-do the amount of these caps in a circuit. It can be a "modifier" cap to help dry out the sound (if it's too wet/lush/colored). I think it's also excellent for post-regulator power filter in digital circuits for transports and similar types of equipment.
That being said, caps that are standing on legs and not mounted firmly to the board can suffer more from electrical and mechanical resonance. This does affect sound quality.
Last edited:
Wow.. Thanks.. But..
I try to build a list of Nichicon capacitors for recap of c326bee.
With all info you provided it will be near to impossible to complete this list.
There they use in all places some General Type (GS) capacitors and some special types in few places.
So I thought to replace all GS with some Nichicon "audio" capacitors like KW/KA/FG and special capacitors with their analog from Nichicon and finally be happy.
But you made my life miserable 🙂
I would suggest to base your work on science and proven solutions (e.g. look at what is used in good commercial amplifiers), rather than subjective impressions from individual listening sessions.
Anecdotal evidence can be worse than no evidence at all 🙂
I am a conspiracy theorist myself in other fields (software, electromagnetic radiation).
I saw and felt things that other people just can not.
I used scientific method to make it real. To make sure that this is not my imagination and not glitch in the matrix.
I have my own experience with hearing huge differences between capacitors and resistors in crossovers in speakers.
Also I did blind test for people who are not related to audio in any way and they heard same difference as I did.
That's why I feel there is something behind opinions of people that listen to cables, caps, resistors, etc. in audio and hear differences.
In between I bought set of UKW/UKA capacitors to replace them in driver and power amplifier sections of c326bee.
They are really not many of them there.
Will try them after I solve idle currents problem.
UKZ are not fitting there exactly (will hang in the air on their legs and I don't like it).
I wish I could buy UVR Standard series, because NAD have some KSC Standard capacitors installed all other the place in c326bee.
But can not find a stock (mouser and other huge sites are not available for me).
Currently I can only buy any of UKW/UKA/UKZ.
I don't have any knowledge on what is used in commercial amplifiers.
Maybe I should open separate topic on recap of c326bee, with list of all capacitors and ask for suggestions for replacements of each one and later feedback on installed capacitors.
At least this is the case for CDE 380s (Pass, Audionet) and Nichicon UVR (Accuphase, 2000s Onkyos)/UKA. So he can't do much more wrong...
Can confirm: exact same impression! I use either post regulator, beside the load but never both. They are excellent.
Has anyone used KZ on a power amplifier?
Capacitor ripple current & ESR effect on the frequency range vs. the power supply output impedance, for me, is key.
True. And/Or those that focus on fine details and nuance. Not a bad area to focus obsessions on ... it makes a difference and hey, our brains are fundamentally difference engines🙂