Funnycat / Sillyfeline behaviour?

[the3secondrule]

Hi All,

I have built up the Roland Funnycat, using the Geo silly feline layout www.geofex.com/pcb_layouts/layouts/funnycat.pdf, and just had a couple of questions.

1> as stated in the build docs, there is quite a bit of treble loss in bypass (as well as some bleed through of the fuzz). i'm thinking that rather than add another switch to true bypass the whole thing, it may be simpler to a buffer on the input...

2> in harmonic mover mode "c", i'm getting some pretty heavy + painful oscillation with the blend up past halfway. i assume this isn't standard behaviour. i used 4558's for the dual opamps, the only parts sub was film caps instead of the 680n + 220n tants.

if someone who has built this could chime in, that would be hugely appreciated 

[charmonder]

Yes wow I just built this same pedal a few days ago!

It seems like the fact that it was mentioned in the build file that those 2 capacitors are tantalum, that they ought to be tantalum. I had trouble getting the 680 ( not available from small bear right now) so I used 220n in parallel with 470n which got me a 690n.You might get lucky and find a few tantalums if you have any broken dod pedals lying around.

I think you have a good idea putting a buffer on the input, I might try it myself because three 3 footswitches route seems less than convenient.

[charmonder]

Oh yeah I don't wire the rotary switch correctly, so the first time I put it all together, 2 of the 3 modes were producing feedback. If your did pcb mounted rotary this shouldn't be an issue

[the3secondrule]

yeah, i ws wondering about the tants - I didn't have any polarised caps in the right size, so I just used film instead. not sure if that would cause a problem though?

and I went with the board-mounted switch, so at least I know that's right. now to check the components *around* the switch. 

Cheers, J 

[Mark Hammer]

You really don't need a rotary switch.  Quite frankly, it will be an easier build if you simply use a pair of SPDT toggles or an additional pot and toggle.

One set of contacts in the rotary switch selects between 2 different gain presets for the envelope follower section, which you ca probably just replace with a 100k pot, wired up as variable resistance to ground.  Switch R10 from 47k to 33k for a broader range of sensitivities.

The other set of contacts in the rotary switch simply select between two different envelope follower outputs.   One of them is additionally smoothed and "tamed" by a 12k resistor and 47uf cap to ground.  The other is "Untamed", and takes its feed directly from the 1k2 resistor.  BY virtue of the reduced smoothing, it provides a rather ragged and unruly sweep, particularly in the decay portion, where you'll experience a LOT of ripple.

[the3secondrule]

One set of contacts in the rotary switch selects between 2 different gain presets for the envelope follower section, which you ca probably just replace with a 100k pot, wired up as variable resistance to ground.  Switch R10 from 47k to 33k for a broader range of sensitivities.

The other set of contacts in the rotary switch simply select between two different envelope follower outputs.   One of them is additionally smoothed and "tamed" by a 12k resistor and 47uf cap to ground.  The other is "Untamed", and takes its feed directly from the 1k2 resistor.  BY virtue of the reduced smoothing, it provides a rather ragged and unruly sweep, particularly in the decay portion, where you'll experience a LOT of ripple.

thanks Mark

I went the other way, and swapped r10 with 100k - which made the filter less sensitive, and more usable. any suggestions for taming mode C just a little bit? I'm guessing adding just a little series resistance between the 1k2, and lug C?

Cheers,

J

[Mark Hammer]

The "taming" is a function of how much the ripple inherent in the envelope is smoothed out by "averaging" capacitors.  There are two of them: C9 (1uf) and C19 (47uf).  C9 provides minimal averaging, and obviously C19 provides a lot more.  When in modes A and B , C19 is connected directly to R24, so that C19 is averaging the envelope voltage whose attack time (how long it takes to charge up C19) is dictated by R11+R23 and whose discharge time (decay) is dictated by the value of R24.

In mode C, the principal averaging going on is by C9, which is minimal.  C19 provides a wee bit of averagig, but that action is reduced by R23 now being part of a series path to ground through C19.  The resulting envelope voltage is esentially attenuated by R11/R24 acting as a voltage divider (a 221.2k "pot" turned down just a smidgen).

Now you know why mode C is as brash and untamed as it is.

If a person wanted full control over the envelope, the ideal would be to have a variable attack and decay.  Within that approach, D3 would go to something like a 100R fixed resistor followed by a 10k variable resistor (2 lugs of a pot).  That would go directly to the base of Q2.  From the base of Q2, you'd have a 10uf-22uf cap to ground, and in parallel with that you'd have a 47k fixed resistor and a 500k variable resistance in series (also going to ground). 

So, from D3, which rectifies the signal and only gives you whatever falls within a half-cycle, you would have a variable ristance, dictating how quickly the averaging cap charges up), with a minimum and maximum chargeup time set by the 100R and 10k pot.  The cap to ground from the Q2 base is the averaging cap.  The resistance in parallel with it sets the discharge/decay time, again with a minimum and maximum value.  That arrangement will provide for lingering lazy sweeps, as well as faster synth=like ones. It will not give you the "wild-dog" mode of a stock Funny Cat.

If a person simply wanted a mode C that was a little less jarring, the simplest approach is to stick a 2u2-3u3 cap to ground from the junction of R11/R23, to provide a little more averaging at the R11 output.  You could probably afford to reduce C19 to 39uf or maybe even 33uf in that case, since the envelope in mode A and B would be "triple averaged".

[the3secondrule]

Yes wow I just built this same pedal a few days ago!

It seems like the fact that it was mentioned in the build file that those 2 capacitors are tantalum, that they ought to be tantalum. I had trouble getting the 680 ( not available from small bear right now) so I used 220n in parallel with 470n which got me a 690n.You might get lucky and find a few tantalums if you have any broken dod pedals lying around.

I think you have a good idea putting a buffer on the input, I might try it myself because three 3 footswitches route seems less than convenient.

Just to confirm, I put a little Jfet buffer on the input, and it has completely cleaned up the bypass issues for me 

[charmonder]

@ Mark Hammer: these are some great ideas for mods thanks for the breakdown of the circuit! I think I'm going to add some attack decay controls, I still have the rotary switch in there I might as well do something with it.

Just to confirm, I put a little Jfet buffer on the input, and it has completely cleaned up the bypass issues for me 

nice I Just took my funny cat to a band practice I did notice the bypass was a bit muted...  I was hoping there was some way to take advantage of tat extra buffer fir the balance pot but jfet seems easy enough anyhow

[charmonder]

Oh I also have an idea for you. As mark pointed out, if you have a pot for r10, then you only really need the one side of the rotary switch acting as an spdt. Since I already had the rotary switch, I left the third leg disconnected completely. In this new third mode, the envelope follower does not sweep filter, and if youve added any knobs for rfbb rfbc and r10, you can really tweak the tone of your sound, especially useful for the SDS.

In case you can't tell i'm very pleased with the funny cat! I thought it was just going to be a novelty effect but I mean it sounded good at band practice (which I couldn't say about 99% of all the pedals I've gone through)