What works well for coupling doesn't always work well for decoupling. Film caps of any type often sound too bright (but not always) In coupling position. If you find the upper mids thin or the sound a bit bright Nichicon FW solves that problem. Nichicon Muse sound like film caps to my ears, so good caps technically but don't always sound good. I used to like the old Oscon SP for decoupling In Naim preamps but haven't experimented with them anywhere else. For feedback the best I've used In my Onix OA21 are the Vishay Sprague Tantalex 150D (as used by Avondale audio In their NCC200 power amp module)
Last edited by a moderator:
Sorry I meant KZ. I didn't mean they are bad capacitors (quite the opposite), just that there's a synergy thing with caps and even the best cap doesn't always sound good In certain equipment. For example, I use film coupling caps In my Quad 306 power amp and they sound excellent. The same film caps In my Onix oa21 sound bright and thin. I guess they are just too revealing of problems elsewhere.
Trust your ears, not the specs.
Trust your ears, not the specs.
Last edited by a moderator:
Yes, there's no better cap than no cap. But many will argue that on a DC servo you also have two film caps, usually between 220n to 1uF. I've been simulating several power amps that I plan to build, and originally they are "pure DC", with no decoupling cap on the NFB or anything. But it's close to impossible to build a completely stable DIY amp that will not drift in DC output. On the simulation, using a classic LF411 which Erno Borbely (a DC servo fan) considered to be a low distortion zero drift chip, THD increases slightly with the servo compared to using a cap.
What is the DC servo design you are suggesting?
Last edited by a moderator:
Totally different. Black Gate (Std/F) is great for supply decoupling. Tried/tested it once for coupling and never used them since for that purpose. OSCONs, are only good for feedback caps. The voltage loss of this cap is IME very high, not suitable for general supply bypass. But digital circuits (such as part of your DAC) have different requirements than analog circuits.
I think the Keltron is similar to Nippon GXE as they are both orange in color
. The GXE is specified for high temperature applications, but it is measured better than many suspiciously fake caps trying to act like audio caps.Attachments
When I think of the Muse parts for coupling, I would always use the bp and not the other, non polarized that are meant for power supplies. With the exception of a bit of treble response, those parts sound really good. Damping a regulator output (with digital circuits in particular) the audio specific parts can make for some odd results I’ve noticed. I use a dac which has a servo, and that uses a 1uf polystyrene, sounds great.
Haven’t heard of Keltron before.
Haven’t heard of Keltron before.
Last edited by a moderator:
Where did you get 1uF polystyrene? They must be gigantic! I found these:
1uF 250V 10% K71-4 POLYSTYRENE capacitors
For servos they should be great, even if hard to fit.
1uF 250V 10% K71-4 POLYSTYRENE capacitors
For servos they should be great, even if hard to fit.
Last edited by a moderator:
Is it good idea to use polymer solid electrolytic in the Vbe multiplier bypass cap, as these operate in high frequencies?
Here pictured, the Nichicon LV
Here pictured, the Nichicon LV
Last edited by a moderator:
I had been using the solid polymer type caps, not in that position, but for digital decoupling, however have started to use a .1uf cog + an xr7 10uf, preceded by whatever is on the output of the regulator. They tend to be grainy in comparison for circuits that require a low impedance at higher frequencies.
The solid parts seem to work better before the regulation it seems. They have too low esr to use directly on regulator outputs.
They are higher frequency parts but nearly as high as ceramics, a step up from tantalum, and without the risky characteristics of tantalum parts.
I haven’t tried those in the vbe position, have used a higher value Muse bp, with a .01uf film/foil bypass. It would be interesting to see how they work there, since the demand is low, might be surprising.
@Carlmart; those are the ones I’m using, have made custom cradle type brackets from plastic to secure them in the chassis, next to the servo, which is on the edge of the board.
The solid parts seem to work better before the regulation it seems. They have too low esr to use directly on regulator outputs.
They are higher frequency parts but nearly as high as ceramics, a step up from tantalum, and without the risky characteristics of tantalum parts.
I haven’t tried those in the vbe position, have used a higher value Muse bp, with a .01uf film/foil bypass. It would be interesting to see how they work there, since the demand is low, might be surprising.
@Carlmart; those are the ones I’m using, have made custom cradle type brackets from plastic to secure them in the chassis, next to the servo, which is on the edge of the board.
That's interesting. Nelson Pass used a pretty high value, 100uF I think, in his Class AB amp, but he's chosen a Tantalum type:
AB100 Class AB Power Amplifier
What BP value did you use?
Last edited by a moderator:
Hi all,
Inspired by Linear Tech. I recently did a series of measurements of various combinations of capacitors for digital circuitries. And - although the "best" solution likely also depends on the actually noise frequencies in a given circuitry - I found that typically 100 nF (e.g. C0G) combined with 10uF makes for a very clean noise spectrum. So apparently similar to what phase has started using ...
@Karl vd Berg:
Cheers,
Jesper
Inspired by Linear Tech. I recently did a series of measurements of various combinations of capacitors for digital circuitries. And - although the "best" solution likely also depends on the actually noise frequencies in a given circuitry - I found that typically 100 nF (e.g. C0G) combined with 10uF makes for a very clean noise spectrum. So apparently similar to what phase has started using ...
@Karl vd Berg:
Briefly studying the datasheet I would be a bit cautious about the LF performance - as mentioned previously I have found many polymers to have somewhat lacking LF performance, and without knowing, I at this point in time am assuming that it could be related to the high leakage currents.
Cheers,
Jesper
Last edited by a moderator:
Phase, Fatmarley and Johnego, thanks for the replies to my questions. Seems everyone agrees that an electrolytic sounds totally different in power decoupling duty versus signal coupling.I can't afford BG caps. I'm open to other suggestions for smoothing caps. Silmic II?
Last edited by a moderator:
The Black Gate NP do the job ! just replace Mundorf 10uF S/G/O on Pass B1
output with BG np 10 uF 50v , these BG need hundreds hours burning ,the one I use is at least 1khours minimum.
I gave all my Silmic 2 caps away. Sound too soft and cuddly for me (coupling and decoupling), but they were quite nice In the feedback position of my old Naim preamp when combined with Oscon SP for power rail decoupling (some kind of synergy thing going on).
