Phaser capacitor value question

[frequencycentral]

I recently built a Phase 45 with JC Malliet's mods, one of the mods is switched cap ratios, ie the 'Univibe' mod. The stock P45 caps are 0.047uf, JC's mod chages these to 0.01 in the first stage and 0.1 in the second stage I ended up adding two extra stages both with non-switched 0.047uf caps. I noted during my experiments with this circuit that the phasing seemed more pronounced with bigger cap values.

Now, I've just completed a layout for R.G.'s Phase 180 Plus, which is an 8 stage version of the Phase 90 with some extra control over bias, width, regen and blend. Both the stock Phase 90 and R.G.'s varient call for 0.047uf caps. I'm building mine at the moment, but I find that I don't have enough 0.047uf caps. What I do have is hundreds of 0.1uf caps. Is there any reason why I shouldn't sub these in place of the 0.047uf? My experiments with my P45 build tell me that bigger caps = better. I'd just like some opinions and clarification, I'd really like to get on with the build using what I've got, but if 0.1uf caps in the phasing stages is going to detrimental to the sound I'll hold off and order some.

[R.G.]

Keen's Fourth Law says to use what you have to get it running and then do the perfect thing when you get the perfect parts.

[frequencycentral]

frequencycentral's first law says listen to R.G., but my second law says use the right parts in the first place or a suitable sub. I hate to pull parts from a completed build. I guess what I'm asking is will it sound good (and how different?) with 0.1uf caps instead of 0.047uf?

[R O Tiree]

If you have tons of 0.1µF caps, why not wire pairs in series? That would give you 0.05µF caps, which would be as near to 0.047 as makes no difference, especially as tolerances are likely to be 10% at best?

[R.G.]

Many phasers use caps even smaller than 0.05.

But yes, series/parallel is always something to remember.

[frequencycentral]

If you have tons of 0.1µF caps, why not wire pairs in series? That would give you 0.05µF caps, which would be as near to 0.047 as makes no difference, especially as tolerances are likely to be 10% at best?

Nice idea, but no room on the layout! Thanks though.

[Eb7+9]

since the jFETs are linearized by 22k resistors consider the MEAN resistance value at 11k ohm as jFETs go up and down in resistance ... you can use that value along with the cap value to establish the MEAN frequency within its sweep range using F=1/2piRC ... you can either do more MATH or Spice, or ball-park it by rembering that a doubling in cap value will correspond to an octave reduction (translation) in frequency sweep - the whole sweep range simply gets shifted down ... maybe that can be useful to you in establishing a quick educated guess in relation to what you already have ...

I understand that's not what you want here but another route you can take is keep with whatever caps you have and adjust the linearizing resistors accordingly ... in fact you could make each phasor stages continuously variable by using 100k pots instead of 22k resistors - with some linearization loss incured but that would give you continously adjustable sweep range so you could set it just where you want it ... just an idea

[frequencycentral]

Hmm, I guess the simplest thing to do is just use the correct parts to start with! Thank you!

[R.G.]

Keen's Fourth Law says to use what you have to get it running and then do the perfect thing when you get the perfect parts.

[petemoore]

  you can easily series 2 caps.."2x .1uf = .05uf", jumper [or not] 1 of them, like this for each stage.
  with a leetle wire at the crotch of 1 of the 2 seriesed caps extending longer than where the two seriesed cap leads meet, jumper/non jumper can be a simple wire twist-over/or lift-off around the joined cap leads...then you can cut or solder the jumper as you see fit...careful with the wire that starts at the cap lead at the crotch where the leg goes in the board hole, and you can de-solder/remove it ~completely, very easily.

[Andi]

Keen's Fourth Law says to use what you have to get it running and then do the perfect thing when you get the perfect parts.

I thought Keen's Fourth Law was "If you can't measure it, can't even think of a way it hyphthetically *could* be measured, it's probably not real. You certainly don't know enough about it to depend on it by superstition alone."?

[Mark Hammer]

With respect to cap values, you'll want the notches to be situated not too high and not too low.  The cap values play a part in that location of range, but then so does the resistance in parallel with the FET (as JC noted) and so does the biasing (which plays a role in determining how the fixed+JFET resistance convert to a specific parallel resistance).  Indeed, it is possible to move the overall range around (within limits), using the bias trimpot.

T'wer I, I'd opt for a smaller cap value, rather than a larger one.  So, if you have some .022, or even .01uf caps sitting around twiddling their thumbs, use them.  You may even find you prefer situating the notches a little higher up.

[R.G.]

Keen's Fourth Law says to use what you have to get it running and then do the perfect thing when you get the perfect parts.

I thought Keen's Fourth Law was "If you can't measure it, can't even think of a way it hyphthetically *could* be measured, it's probably not real. You certainly don't know enough about it to depend on it by superstition alone."?

That's Keen's 3.14159... Law.

[frequencycentral]

Where can I find a comlete list of Keen's Laws? Paperback? Hardback? Two stone tablets? 

[frequencycentral]

OK, I decided to go for the two 0.1 caps in series approach, rather than put in the wrong value and pull them later. It's working out pretty cool actually, as my caps are really small. Even soldering two together means they still only take up 3 perf holes space, which my layout requires. Cool! Thanks for everyone's help once more.

[Cliff Schecht]

Those look like ceramic caps....

*Smacks frequencycentral's hand*

[frequencycentral]

Those look like ceramic caps....

*Smacks frequencycentral's hand*

Ouch!