It seems his observations are more about some interactions going on, with smaller combined values affecting performance of the dynamics.
So the total capacitance value is another aspect.
Little Dot MK8SE / MK6 Super Mods (All verified mods are on first page)
- Thread starter Redge78
- Start date
So the total capacitance value is another aspect.
I did try the impedance mod by installing four 680ohm resistors on top of the component side 330ohm power resistors.
The sound did change, but only the bass improved slightly and everything else sounded overly smooth and veiled. The bass was getting distorted at clearly lower volumes than before too. I did not have patience to keep them there for more than a couple of days before removing them though. Don't know why that mod didn't work for me. One thing that I had been wondering was that whether or not the decoupling caps would somehow affect the parameters in the Cavalli formula, or are those on an entirely different part of the circuit?
And I also did order the Raphaelite B10 impedance matcher from the latest AE sale. This is because I've had some good results with using 3.5mm impedance adapters on some amplifiers so I've been wanting to play with an XLR impedance adapter even before that was released.
My Tung-Sols are from this box. Was a good buy as I didn't have any of those particular brands' tubes before (had some Chathams though which surprisingly sound exactly the same as the Tung-Sols).
It probably should be mentioned there that things like high/low gain and also other components such as decoupling caps will have an effect on how the cathode bypass caps will function, possibly headphone type as well?
I didn't test the 550uf combination in high gain yet but might do that later too.
The smaller than 470uf capacitors are available in the 100V specification as well, which has lower esr (or "Tangent of loss angle") on the Nichicon KZ series data sheet. It seems to me that the faster capacitors means more detailed sound there.
Perhaps it would be interesting to test something like the Wima MKP line in ~500uf there as well but I have no plans to do that in the near future
. That would certainly require a bigger case...
baronbeehive
Headphoneus Supremus
OMG that is somwhere very near me, I could have gone over there, might try contacting him!
I can't remember but it would be easy to find out if the Chathams are just rebranded Tungsols.
I doubt that the decoupling caps would have any effect, their function is to provide a reservoir of fast power close by to help the PSU which isn't regulated. That's their only function, not to influence SQ as such.
It may be that the Cavalli formula isn't that precise, I saw in the paper that another factor was added in the equation last minute as a correction but I don't know how effective it is for WCF circuits.
Last edited:
baronbeehive
Headphoneus Supremus
Yes, not sure if WIMA would be worth trying because of their SQ. They seem fairly bog standard in that department.The smaller than 470uf capacitors are available in the 100V specification as well, which has lower esr (or "Tangent of loss angle") on the Nichicon KZ series data sheet. It seems to me that the faster capacitors means more detailed sound there.
Perhaps it would be interesting to test something like the Wima MKP line in ~500uf there as well but I have no plans to do that in the near future
I'm thinking of trying for a combined value of near to 550uF after your observations.
One thing that should be mentioned though that I don't think you use smaller bypasses? If so that would influence our choices because we do. I was wondering if that was because you don't like bright(er) sound? We have found bypasses to be necessary.
Last edited:
Okay so I discovered the reason why stock amp uses bipolar capacitor in the cathode position.
It is for better frequency response on the trebles end, so therefore a second "bypass cap" is not needed for the upper end, like what we have on first page.
Although is not verified here yet, it has been discussed elsewhere, on how a non-polarized can notice better tonality, when used in places like as a "coupling cap" (which of course we're not going to do.)
So in general, it is stated as being a bit better for tonality.
here is a statement by a manufacturer.
So anyways, we know that all ectrolytics have a weakness in the upper end trebles and details area, which is why we "bypass cap"(using the paralleled paper refa cap), but also why member copperfox noted tonality differences in details area, as he chose not to use a "bypass cap" on his lyctics.
So if not, then you have more tonality differences.
So I believe there were (three) multiple issues going on: tonality, dynamics, and detailing.
So since member copperfox is not using a any small "bypass cap" on his lyctic cathode caps (for even tonality), it was good choice for him instead select the nichion audio grade caps for his test to be more successful on tonality/detailing, and to use the smaller values in parallel for more dyamics.
Also, we read from member copperfox, that although 2x330uf did give the detailing, it did not give increased dynamics to the level of the other solutions.
This leads me to assume that going higher than multipled 220uf caps (paralleled) does not give more dynamics.
That was the only curious aspect that didn't work out as expected.
We can expect that the much higher single cap values (1000-2000uf) to not perform as well as the smaller multiple caps, and that a "single cap" solution should not be more than the 470uf observed by copperFox.
So with all this in mind, I decided to keep my 330uf and add the 220uf to it, as member copperfox optimally suggested.
The only difference is that I also decided to use non-polar since I am already using non-polar.
This can vary my results somewhat, but I expect it to be better regardless.
In order to keep the electrolytic caps as quality or audio grade as possible, I chose some kaisie 220uf caps.
Any differences in the tube harmonics from adding two different brand caps, would only add to the richness of the sound.
It's still a more optimal choice for richness, than cheaper solutions the other brands I discovered (using computer grade Nichicon cap).
It can swing both ways with good or bad quality choices.
I have also observed another manufacturer make an upgraded amp using higher grade caps, only to have a more sterile harder, darker, grittier sound!!
They were not nichicon audio caps, but instead a change from low-grade to mid-grade types. You will rarely see top grade caps even in $5000 amps.
So the point is, real world observation matters.
AudioNote kaisie caps are way more expensive for arguably little difference, but it is general concensus that they have good depth and tone.
The nichion caps have a general consensus of being more upfront clarity, while the silmic have the consensus of having detailed smoothness.
Being that these are all general consensus observations, they are to be taken lightly and not critically.
Bottom line, lots of things going on. Nothing is as simple as it seems when it comes to the variability of design choices.
It is for better frequency response on the trebles end, so therefore a second "bypass cap" is not needed for the upper end, like what we have on first page.
Although is not verified here yet, it has been discussed elsewhere, on how a non-polarized can notice better tonality, when used in places like as a "coupling cap" (which of course we're not going to do.)
So in general, it is stated as being a bit better for tonality.
here is a statement by a manufacturer.
So anyways, we know that all ectrolytics have a weakness in the upper end trebles and details area, which is why we "bypass cap"(using the paralleled paper refa cap), but also why member copperfox noted tonality differences in details area, as he chose not to use a "bypass cap" on his lyctics.
So if not, then you have more tonality differences.
So I believe there were (three) multiple issues going on: tonality, dynamics, and detailing.
So since member copperfox is not using a any small "bypass cap" on his lyctic cathode caps (for even tonality), it was good choice for him instead select the nichion audio grade caps for his test to be more successful on tonality/detailing, and to use the smaller values in parallel for more dyamics.
Also, we read from member copperfox, that although 2x330uf did give the detailing, it did not give increased dynamics to the level of the other solutions.
This leads me to assume that going higher than multipled 220uf caps (paralleled) does not give more dynamics.
That was the only curious aspect that didn't work out as expected.
We can expect that the much higher single cap values (1000-2000uf) to not perform as well as the smaller multiple caps, and that a "single cap" solution should not be more than the 470uf observed by copperFox.
So with all this in mind, I decided to keep my 330uf and add the 220uf to it, as member copperfox optimally suggested.
The only difference is that I also decided to use non-polar since I am already using non-polar.
This can vary my results somewhat, but I expect it to be better regardless.
In order to keep the electrolytic caps as quality or audio grade as possible, I chose some kaisie 220uf caps.
Any differences in the tube harmonics from adding two different brand caps, would only add to the richness of the sound.
It's still a more optimal choice for richness, than cheaper solutions the other brands I discovered (using computer grade Nichicon cap).
It can swing both ways with good or bad quality choices.
I have also observed another manufacturer make an upgraded amp using higher grade caps, only to have a more sterile harder, darker, grittier sound!!
They were not nichicon audio caps, but instead a change from low-grade to mid-grade types. You will rarely see top grade caps even in $5000 amps.
So the point is, real world observation matters.
AudioNote kaisie caps are way more expensive for arguably little difference, but it is general concensus that they have good depth and tone.
The nichion caps have a general consensus of being more upfront clarity, while the silmic have the consensus of having detailed smoothness.
Being that these are all general consensus observations, they are to be taken lightly and not critically.
Bottom line, lots of things going on. Nothing is as simple as it seems when it comes to the variability of design choices.
Last edited:
baronbeehive
Headphoneus Supremus
Right! I have polar cathode caps for both amps, if we are to use bipolar we need to be looking at Nichicon ES in the Muse range instead of KZ for example. That means I can't parallel a 100uF to my existing for example, damn!
You could test this on the LD, just remove the bypasses from the Mundorfs to see before you parallel the 220uF's.
Perhaps we should define these terms to make sure we are talking about the same thing? For example I think of dynamics as large changes in volume in bass and mids frequencies handled well, I would think that for example fairly large caps around the 550uF range would bring dynamics out with the emphasis on lower down frequencies. But also I think of liveliness which is more of a treble thing. I would think of attack and so on which Copperfox noted on the smaller value caps up to 470uF.
From what we know about the high pass filter bypassing certain frequencies I would have thought that around the 2x330uF value dynamics in the sense of powerful bass and mids would be around the maximum. But liveliness in the sense of treble sparkle and upfront sound the smaller than the 2x330uF values would be more relevant. That may be why this wasn't expected. I know that on my other amp the 470uF Kaisei caps without a smaller bypass was much duller sounding and lacking in treble response and liveliness but the bass frequencies were good.
I don't know what you think, just trying to understand the findings. I know what I prefer.
Anyway I will have to try either 470uF + around 100uF polar Nichicons, or go bipolar around the same values to test this out.
Edit: Knowing what we know about the high pass filter in the cathode bypass I would think that the possibilities of tonality are pretty much endless. For example a single cap would get the tonality of the make of cap, if there were to be several makes of caps in parallel the tonality would take on characteristics of both caps, also the more the caps were paralleled together the more the tonality would be affected, possibly adversely due to the different harmonics coming out of the different caps, and possibly some THD would be involved as well. This could partly explain the smearing that can sometimes affect paralleled caps. And the cap values affecting the different bypassed frequencies would also be involved. Also as you say the linearity of negative feedback would iron out these effects so I can see that to use cathode caps or not to use them, or to use them in different combinations would produce interesting results which we may not have got to the bottom of yet!
So whether to use single or multiple caps looks to be a bit of a minefield, to be determined experimentally perhaps.
Last edited:
baronbeehive
Headphoneus Supremus
Should be interesting combining 2 different makes, Mundorf and Kaisei. I know that in my other amp combining Kaisei with small Audyn True Copper bypasses produced a remarkably good tone, I believe due to the mix of Kaisei and especially the Audyn bypass. I think there was a mix of the qualities of both types.
I agree with this, I think the Kaisei worked well with the other amp and complemented it's qualities with a fuller sound whereas the Nichicons worked well with the LD bringing out the clarity more which is what I was looking for.
I was thinking of the large film Wimas such as the 400v 400uf that I have on the psu output position, along the lines of @Redge78 's comments about the cathode bypass on the first page that "This 220µF value starts to be a significant value if you want to switch technology to some MKP or PIO caps. It's even worse if when you want to push the values two steps up (470µF). Those huge MKP/PIO caps would weight pounds (or even kilogrammes) each. It's feasable, it would be expensive, hard to manage ... but it could be better than even the best Lytics."
The Wima line seems to top at the 400uf value, but some other brands such as Vishay have 560uf 63V (part nr MKT1820756065 ). Those are quite large and expensive at $100 each and there is not much actual indication as to how they would actually sound so it would be an expensive blind experiment if it was possible to fit those in the case.
I haven't had the smaller bypasses installed after installing the decoupling and psu output wimas as after that they would make the sound way too trebly. So I would think that whether or not the smaller bypasses are useful depends on the other components like the psu output and decoupling caps as well.
baronbeehive
Headphoneus Supremus
Yes, you're right I forgot that one or two members did use those huge film caps in the cathode position, as well as for decoupling. From what I can gather the Wima sound is fairly soft I think, BUT, as film caps, in the cathode position they should be worth trying. I remember the favourable comments of the members that had used them.. in a MASSIVE custom chassis lol! The cathode position is one where speed is good and this is where film caps score above lytics.
Edit: Yes I would definitely think in your case these would be well worth trying. As I said the cathode position really needs fast caps and I think that it would benefit enormously if you did try them, bass would be snappier and transients also. This is the area that the amp would improve massively potentially as previous members have noticed.
Last edited:
baronbeehive
Headphoneus Supremus
Yes I thought so, I remember your description after using those fairly huge, fast decoupling caps which would obviously make a difference in your case. I believe your system set up is optimised to the max to get the best possible sound from the amp!
Hmm, maybe I could be able to fit those 560uf Vishays into the case. It would be quite a puzzle though. They are actually of similar 35mm thickness as the 100uf 500V Wimas but larger in height and width.
For comparing the speed of capacitors that have it specified in ESR and those that have it specified in tangent of loss angle I found an useful and informative post on the DigiKey support forum that has an equation to convert ESR into tangent of loss angle and vice versa.
By that equation the 0,07 tangent of loss angle at 120Hz of the Nichicon KZ capacitors would equal an ESR of ~422mOhm for the 220uf and ~281mOhm for the 330uf. That is probably not entirely accurate though as the Nichicon datasheet says "tan δ (MAX.) 0.07" for the <470uf caps, so it is a maximum value for all caps smaller than 470uf and so it is possible that the smaller caps could have lower than 0.07 tan δ at 120Hz.
Those Vishays' datasheet says that the 560uf version has 0,007 tan δ at 100Hz so by that equation their ESR would be ~20mOhm which is a lot less than the Nichicons but more than those film Wimas which have about 1 to 10 mOhm. So even there it might be beneficial to use two 275uf Wimas to reach ~550uf but no store seems to stock those and a minimum order to have them produced is 40 units.
Last edited:
baronbeehive
Headphoneus Supremus
Those Vishays' datasheet says that the 560uf version has 0,007 tan δ at 100Hz so by that equation their ESR would be ~20mOhm which is a lot less than the Nichicons but more than those film Wimas which have about 1 to 10 mOhm. So even there it might be beneficial to use two 275uf Wimas to reach ~550uf but no store seems to stock those and a minimum order to have them produced is 40 units.
No those Wima's can be tricky to get hold of, the only thing I could find would be this: https://www.conrad.com/o/film-capacitors-0245680?ATT_NUM_ELECTRIC_CAPACITY_any=100 µF,150 µF
which would be 3x150uF + 1x100uF, which is a lot of capacitors!
I can't remember what film caps the other member for cathode bypasses used, I don't know if anyone else could help.
So the Vishays might be OK.
Edit: I've also seen these 330uF :https://eu.mouser.com/WIMA/Passive-...s/_/N-9x371Z1yzvvqx?P=1z0zldhZ1z0x7p2&FS=True
to which you could add the 220uF's also at mouser.
I've just checked the member who used film caps and he mentioned AVX.
If I remember right you are using film caps for PSU as well as decoupling? What capacitance are you using for those caps?
Last edited:
baronbeehive
Headphoneus Supremus
@CopperFox sorry, copying the search results at mouser didn't work properly, try this: https://eu.mouser.com/ProductDetail/WIMA/DCP4G064009JD4KSSD?qs=RB4whv9F6rxb7mMvMPxktA==
or enter WIMA in the search box with the 400uF filter and then the 120uF, or 3x115uF.
or enter WIMA in the search box with the 400uF filter and then the 120uF, or 3x115uF.
Last edited:
Users who are viewing this thread
