[brj]

Given a need for X amount of capacitance, what practical differences result when using 5 caps that can store X/5 instead of 1 cap that can store the total X?  (My use of 5 as the divisor is an example.)

Speed of charge and discharge is the one argument that I suspect is coming, but how universal and measurable is this?  If this is the only significant difference, how completely does the use of a bypass cap resolve the difference?

What other differences are measurable, and, more importantly, typically observable?

Thanks for any insights!

[Christopher Witmer]

I have often heard it said that a bank of small capacitors is preferable to a single large one of the same capacitance, but I can't remember offhand the details why. I think it depends in part on the application too. I doubt it makes much difference in a Felicia or similar power transformer, but inside the amplifier maybe a different story.

[brj]

Thanks for the response, Chris!

I was actually thinking about a speaker crossover rather than an amp or power conditioner.

[peranders]

Paralleling small caps are better but eventually this combination gets worse again compared to a single caps.

The question is:Where is the limit?

The answer can not be found as a general answer. You must evaluate  it from case to case.

[BobM]

There's a good argument to be made for using a variety of caps of differing values. Different values effect different levels of the sound spectrum, and such. However you do have to worry about going too far - it can cause oscillation problems in some equipment.

One question: Are you talking about a power supply, or a crossover or what? That can make a difference in the answers you get.

Enjoy,
Bob

[peranders]

I was at least refering to PS applications.

[brj]

....and I was referring to crossovers!

[Gordy]

...and I'd like to learn more about caps in battery supplies  , I'm thinking of trying multiple 500uF's and/or 1.0 k's as going from 4.7k to 9.4k sure didn't do anything positive for me!  Any thoughts/info out there?

Thanks,
Gordy

[Christopher Witmer]

I am inclined to think that if you could hear differences among capacitors anywhere, it would be in the speaker crossover network.

I would think that as long as you have capacitors from the same line of the same manufacturer, there would be an advantage to getting your total capacitance from a variety of different sizes -- within "reason."

Trouble is, "reason" is hard to define, so you are probably back to using your ears to do comparisons.

As we say here in Tokyo, "Rotsa Ruck!"

[Occam]

I am inclined to think that if you could hear differences among capacitors anywhere, it would be in the speaker crossover network.....

Indeed, one would think this is obvious, that a capacitor in a signal coupling capacity would be the most discernable. I think it a tremendous example of cosmic humor that I've found an equally influential position for cap impact is the shunt position on a frigg'n power conditioner. Go figure...
JoshK, Tianguis, and I have evaluated about 5 different 'accross the line' big sucka caps, isolated from its fed component by 2 transformers that is only affecting the AC prior to it getting to a component. Each had a very different sonic signature. And smaller capacitors used to bypass that cap have a massive impact on subjective perceptions, IMO, as significant as in any other position, ps or signal bypassing. I'm so confused.

For those attending PhilNYC's June NY Rave, we'll demonstrate this via the miracle of DPDT switches, on the fly.

[ctviggen]

How do you know it's not the variation in capacitance causing this?  For instance, if one capacitor says that it's X when it's really X+Y and another capacitor says it's X when it's really X-Y, then you'd be dealing with different capacitances.  When I was in enginerding labs, I learned to test everything, because small errors can make a big difference in certain situations (and no difference in others).  In your case, aren't you making a filter?  If so, then difference in capacitance and not the difference in capacitors could be what you're measuring (or maybe not).  We had capacitor testers back in the lab, though.  I wonder if it would be worthwhile to buy something like this:

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&category=4662&item=7521725647&rd=1&ssPageName=WDVW&tc=photo

Also, I'm not sure I understand the theory behind bypass capacitors.  Is a bypass capacitor a smaller capacitor placed in parallel with a larger capacitor?  If so, what are the benefits of this?

[Occam]

How do you know it's not the variation in capacitance causing this?  ...

'Cause I already have the capability to measure those capacitances quite accurately, and do so. Motor caps are amazingly consistant and I've found them to be spot on for consistency within the accuracy of my instumentation, 2%.
Additionaly, the attenuation characteristics of a 14uf cap vs. 15uf cap isn't going to be materially different given their calculated poles at the relevant frequencies, 1khz+.
Prior to my career worshiping at the feet of Mamon, I spent 2 years working in a lab for a Nobel Laureate, and he taught me the same thing any good carpenter teaches an appentice - 'Measure twice, cut once.'

Re: explaining the theory behind bypassing - do a Google seach on 'capacitor bypassing'. They all say pretty much the same thing, far better and eloquently than I could.

[andyr]

Aha, here's my 2c worth on using more than 1 capacitor!!  I'd be interested on other people's input:

1.  Power supplies (for amps, source components etc. - ie. not power conditioners!):
As I understand it, the optimal filter cap arrangement will use a variety of sizes - for instance, in a recent phono stage PS, I used, for each power rail:
* 2 x 1,000uF
* 1 x 470uF
* 1 x 10uF
* 1 x 150nF.

The reason is: one large cap has a higher ESR than smaller caps (which can negatively affect high frequencies), so using a "spread" of several caps of different values - including small ones - minimises this negative affect.

2.  Other power rail (ie. electrolytic) caps:
It's always a good idea to bypass electrolytic caps with a 1/100th value (which doesn't have to be an electrolytic).

3.  Coupling caps:
Here the picture seems to be clouded!

Some say that 1/100th bypassing is still a good idea.

Others say that having the signal pass through 2 caps of different construction and value causes the signal to "smear" because it is delayed by one cap more than the other!!??  

I dunno!

Regards,

andy