 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
218.215.12.167
Hi fellow inmates
Soon I'll be swapping/upgrading caps in a SS MC phono stage and have resolved most technical issues already.
One aspect, though, has drawn mixed opinions from experienced circuit designers, viz, adjacent decouplers to high-bandwidth, high-gain op-amps. I'm aware of the need to minimize inductance in these caps and the logical, technically-astute option is stacked polyester (Epcos 0.1uF 250V at the ready).
However, I've read that some DIYers are pleasantly gob-smacked by the use of Black Gate's in "L cancelling - super e cap" decoupler format. I also have some BG NX 0.1uF at the ready.
One circuit designer is absolutely resolute in his opposition to BGs in this situation, as there "will always be some small level of residual inductance and one cannot consider electrolytic decouplers for high-bandwidth op-amps". Meaning oscillation or worse I wonder?
So who's compared stacked polyester and Black Gate as decouplers in this situation? Any clear sonic evidence to support one over the other?
spindrifter
P.S.: Please don't bother suggesting ceramic or other types. Thanks.
Follow Ups:
You make an important distinction in your post that holds the key to understanding. I'm reffering to what self proclaimed designers and or DIYers who post on internet electronics tweaker forums might present as fact.
Your job is to try and take the 10,000 foot view and seperate the wheat from the chaff.
The real answer is that each situation is unique and the real requirements to achive the desired level of perfection is different for everybody and depend alot on what the rest of the design is made up of and taste too. This is why it is hard to answer your question.
The highest priority when choosing DC power supply decoupling caps in a design which already has a decent power supply design and decent PCB layout design is choosing the decoupling cap type which displays the proper amount of capacitance across the frequncy range necessary to pull off expected level of performance. In most of todays audio circuitry this means best high frequency performance and thus highest self resonant frequency. By doing this you have ensured that the part is still acting capacitively during local high frequency transients which is exactly why a decoupling cap is there in the first place.
Problem is this can often be harder than it sounds.
This frequency response and SRF aspect all capacitors display becomes important in selecting decoupling caps. What you realize is there has to be a tradeoff between actual capacitance and voltage ripple at the load due to this non ideal decoupling cap frequency response.
Limitations of todays commercially available packaging technology and material types means there are trends which rules of thumb can be formed by. Capacitor performance is directly related to material. Capacitor performance is also directly related to it's packaging configuration.
Honestly man if you really do need decoupling solutions Black Gates are not the answer, way too expensive to waste there and not ideal anyway. I like to think about it in terms of what fits the job: Black Gates are AEL meaning wrong material choice for optimum frequency response and they also come in in through hole cans which means to me they come in the wrong package for optimum frequency response. Granted the black gates are obviously an exceptionally good AEL as far as frequency response and ESR goes but....no where near what a true decoupling cap contender could offfer. Aluminum electrolytics are great for bulk capacitance needs so obviously the Black Gates will shine whenever there is a need for bulk AND linearity. In this app we need neither bulk nor linearity so Black Gates are the wrong choice.
It is kind of the same case with polyesters. Unless you go out of your way to find a surface mount version designed for good high frequency response chances are most of what you do find is through hole. As with the black gates the through hole polyesters good for what it is intended for ie low distortion on time varying fairly low frequency signals but due to the relatively large bulky through hole packages you find parts in typically pretty much worthless for real high frequency work like decoupling caps.
In summary neither choice is axactly stellar but I bet you'd have better luck with the polyesters.
It is extremely clear that in the job of effective DC power supply decoupling is best accomplished by surface mount ceramics for most audio circuitry I've seen but I won't recomend them.
Thanks Ah-glee.
Unfortunately SMDs are not an option in this instance. Also, an issue with ceramic caps is capacitance varies noticeably according to heat dissipation, does it not?
Size-wise both proposed options make good sense. Epcos stacked polyesters are 2mm thick and about the size of your little finger nail. BG NX are about as small as cans could get before they'd become unwieldy to work with (micro-surgery anyone?).
Either way, another audio mod/tweaking journey will have been traversed.
spindrifter
Re the heat variance problem: All caps have capacitance that varies with heat, including ceramics. There are different grades of ceramic caps available which have varying levels of heat stability much like other capacitor types. NPO/COG is a type of ceramic capacitor dielectric material which claims to have nearly perfectly zero temperature drift, but you pay for it in part size and cost.
Re micro surgery and mini tweaks: it is the wave of the future man....smaller, faster, better. As you get smaller feature sizes the nonideal parasitics get less and less. I don't see this trend changing anytime soon unless of course Gordon Moore was actually wrong which doesn't appear to be the case.
Agreed, R-Gli, there is an undeniable trend in the mass-manufacturing of compact modular electronics. Appears as though you're a young, spritely fellow, most capable with nano-tweaking. Unfortunately for the rest of us tinkerers, advancing years inevitably take their toll on our physical acuity. For practical reasons I cannot consider SMDs and the like; stuck with through hole components, fumbling my way across my workbench.
spindrifter
Hello spindrifter
I have found the BG-NX Hi-Q excellent for decoupling the +/- supply rails of op-amps, they are no good a million miles away and should be as close to pins 4 & 8 as possible (dual).
For this application I have found the Hi-Q superior to any polyester or polypropylene I have tried although I can't claim to have tried them all. For example, I found the Hi-Q superior to a good quality dipped coated polypropylene but have not tried expensive 0.1 - 0.47 types because the increase in body size rules them out.
Psychologically, the small Hi-Q caps by comparison don't instil much confidence in a world where 'bigger is better', or so it seems. But once in the circuit and after 100 hours of burn-in you can clearly hear the improved detail, spatial quality and clarity of the music with the Hi-Q.
lovekylie
Thanks LK.
Obviously you have heard positive sonic changes with BG decouplers, as have several others in AA threads on the subject.
Two burning questions. Have you ever tried twin BGs in L cancelling Super E Cap format in exactly this application, compared with single BG only?
Secondly BG product data appears sketchy, and seems to infer BG caps measure/perform beyond the normal realm of even the best film caps. One of these claims is that BG capacitance remains effective up into MHz, even GHz regions. Has anyone been able to validate these claims?
(Just curious - I have no quibble with BG sonics. Enough end-users attest to this!)
spindrifter
Hello spindrifter
I have used the BG-N as an L cancelled pair for the same decoupling application but preferred a single Black Gate. I can find no blurb in the BG literature that suggests using the BG-N (NX) this way for localized decoupling; the Super-E cap seems to apply to main power supply circuits. For main power supply circuits I have used a pair of N type as an L cancelled pair with superb results, but never configured them as non L cancelled to make A/B comparisons.
As for the 10GHz decoupling bandwidth claims, I have only ever been able to find the technical notes published by Jelmax; http://www.acoustic-dimension.com/blackgate/techEcap.htm To be honest I find such claims hard to believe since at 10GHz the lead-wire itself would be inductive enough to cause problems.
lovekylie
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
