Power Supply filtering question

[schoolie]

  I read in Maerlin Blencowe's book excerpt on power supplies that adding a second capacitor in parallel which is smaller by a factor of ten (Like 40uF and 4uF) will improve the filtering of a reservoir capacitor.  This is cheap enough to add, but is there any down side to this and is it worth adding?  I've never seen this done in an amplifier circuit. Maybe this just decreases effective ESR for high frequency signals?  I had always assumed that it would just function as a 44uF cap

One more question, if you don't mind.  If I want to bump up the capacitance, can I add a 10uF capacitor in parallel with a 40uF capacitor, so long as both are rated higher than the voltage requirement?  I want to do this mostly because of size restrictions in the chassis. 

Thanks!

Rob

[PRR]

> read in Merlin Blencowe's book excerpt on power supplies that adding a second capacitor in parallel

I can not find that passage. Can you point at it?

There are related theories in fancy hi-fi.

It is certainly true that older electrolytics with very high-frequency amplifiers need additional bypass for high frequencies. (However the usual ratio is 100:1, the second cap is usually a different type such as film or ceramic, and this trick "can" cause great trouble when parasitics of two capacitors resonate in the passband.)

Since this idea first started, electrolytics have got a lot better. When I was young a tweeter crossover cap might turn inductive by 20KHz. Large modern caps have an extended range of low ESR and the inductive rise is deferred to much higher frequency.

All capacitors distort. When stressed, somewhere in the range of 0.5% to 0.000,1%. Oversize caps in typical circuits don't get as high as 0.5%. People who chase more-zeros on the distortion meter have to choose capacitors or capacitor-arrays carefully.

> I've never seen this done

Simple single electrolytics ARE "good enuff" for most guitar amplifier designers/builders. I suspect it would be unlikely for anybody to reliably know a difference in a _blind_ A/B test. I suspect it would be better to save your pennies toward better speaker or faceplate. Speaker has very obvious effect on tone, and faceplates have large effect on the player.

[Colas LeGrippa]

I have once bypassed a cap with a smaller one ( 50pf or so ) and the amp developped an oscillation to the point where I just could not play the amp ! Such a small stupid ceramic disc was causing such a big problem. If you are not very familiar with am building, better follow to the letter the proven-to-be-good schematics, if not, when oscillation occurs, good luck to chase the trouble. Lead dress ? Grid stoppers ? Capacitor, condenser ? You can spend weeks on that for finally de-building everything and start all over again !

Colas

[kagliostro]

I was writing this the same time PRR and Colas were writing his replay

adding a second capacitor in parallel which is smaller by a factor of ten (Like 40uF and 4uF) will improve the filtering of a reservoir capacitor

I have that Merlin's book, but I'm sure I've read that in some other place

If I remember well as to avoid better ripple must be used big and good quality capacitors

adding more capacitors (of lower quality) in parallel will help to reject ripple

can I add a 10uF capacitor in parallel with a 40uF capacitor

Yes

Maybe this just decreases effective ESR for high frequency signals?

Look to the little 500v 10nf capacitors that are in parallel with the 450v 22uf in the attached Vox AC30CC schematic (you can find the complete service manual in the library of schematics by Doug)

Kagliostro

[schoolie]

Thank you PRR, Colas, and Kagliostro.  I was thinking of adding a smaller Solen fastcap in parallel with an electrolytic.   The other application would be using two caps of same type, but different values to create the capacitance needed.

I could see how there could be strange modes of oscillation between the two components.  I should just follow the schematic and experiment later:)   I was just curious, so thanks for entertaining my question.

This is the link to the excerpt (I hope this is OK with MB to quote):

http://www.freewebs.com/valvewizard1/smoothing.html

This is the quote:

The reservoir capacitor needs to have a low reactance at frequencies up to 40kHz, because the HT at this point will contain high frequency harmonics caused by rectification. Its performance at these frequencies can be greatly improved by adding a capacitor in parallel with it, ten times smaller in value. (Another can also be added, ten times smaller than the second, but this would be usually be considered an extravogance.)

[stingray_65]

Merlin is correct,(imagine that!) the smaller cap has a higher reactance and will filter  high frequencies more efficiently, but he also states it is extravagant.

A lot of guitar amp builders fall prey to design features of HiFi amp builders. there are many similarities in designs, and many of the components are the same, in many ways the general purpose is the same too.

The BIG difference between HiFi amps and guitar amps is this.

HiFi is by nature designed to amplify a signal FAITHFULLY, with as little influence in coloring the signal from the components.

Guitar amps by nature is designed to amplify a signal and color the signal in a controlled way, the component choices we make decide how we color the signal.

so a HiFi amp builder MAY use the extra small cap because he cannot afford to have a 20Khz buzz lying under his output.

Most guitar ot's and most musical instrument speaker won't reproduce those frequencies, so we need not bother at all

[schoolie]

Those are good points, thanks.  I want a Bassman, not a Roland Jazz Chorus

[HotBluePlates]

Morgan Jones pointed out the same issue in his books, but he recommended an added cap 1/100th the value of the original. Forexample, a 10uF filter cap gets a 0.1uF bypass cap.

The stated issue is that electrolytic caps deviate from the notion of an "ideal capacitor" both in their ESR (an issue for low frequency/d.c.) and in their apparent inductance at very high frequencies (way above audio, in the range of rectifier transients). The idea is that if you have noise issues caused by this high frequency mess, a more perfect film capacitor can bypass the big electrolytic. The small required value of the film cap keeps the size of the bypass reasonable (try checking out the sizes of polypropylene filter caps).

[Fresh_Start]

IRRC Kevin O'Connor claims that there is a benefit to having 2 or more caps of equal (or similar) value in parallel for the "reservoir" cap.  Don't know if that's true but sometimes it's cheaper to buy two smaller caps.

Also, adding an "extra" filter cap to serve as the reservoir with a small resistor before the filter cap for the plate(s) does a LOT in terms of ripple reduction.  Well worth it in single-ended amps IMHO and my limited experience.

Cheers,

Chip