capacitor type question

[snk]

Hello,
In order to get a bit more output volume from a Diamond Compressor (well, in fact Fred Brigg's Diamante simplified version), I would like to increase the value of one capacitor while decreasing the value of a resistor...
The capacitor is C13 on the original schematic, and its value is 1µF (On Fred Briggs circuit, the capacitor isn't labelled, but there is only one 1µF cap).
I've been told that increaseing C13 to 2.2µF and decreasing R19 to 1.5k would give me 3dB more output gain.

My question is the following : I usually use Mylar capacitors in the audio path, but I don't have any 2.2µF Mylar/polyester cap left : should I go for a monolithic ceramic capacitor instead, or for an electrolytic, or even tantalum? As it is in the audio path, in the last stage before the output, I guess the capacitor type matters (and I wouldn't want to trade a 3dB volume increase for a lesser audio quality)?
So, which would be the best option : finding a 2.2µF polyester cap, using a ceramic, a tantalum, and electro cap? Using 2 x 1µF polyester in parallel?

Thank you in advance for any advice

[ElectricDruid]

On the Fred Briggs schematic, it's marked as an electrolytic, so why not use an electrolytic?

Finding/creating a polyester cap of such a size is going to make something bulky and expensive, I'd have thought. While it does shape the tone a little bit (rolls off/ doesn't roll off the low bass) it's not exactly directly in the audio path - in fact, it's taking signals we *don't* want to ground, so I honestly wouldn't worry about it much.

The highpass frequency with 3K3/1u works out at 48Hz. If R19 is the 3K3, the mod is worked out to not change that cutoff, but to increase the potential gain. Max gain goes up from x33 to x72. So I'd say you get more like +6dB or +7dB.

Max gain, original: ((100K+6.8 )/3.3)+1 = x33.3 = +30.45dB
Min gain, original: (6.8/3.3)+1= x3 = +9.5dB

Max gain, modded: ((100+6.8 )/1.5)+1 = x72.2 = +37.2dB
Min gain, modded: (6.8/1.5)+1= x5.5 = +14.8dB

[snk]

Thank you, Electric Druid.

On the Fred Briggs schematic, it's marked as an electrolytic, so why not use an electrolytic?

I built it on veroboard, and followed this layout, which use anon polarized capacitor... So I didn't realized that the original design used an electro cap... Sorry about that!

While it does shape the tone a little bit (rolls off/ doesn't roll off the low bass) it's not exactly directly in the audio path - in fact, it's taking signals we *don't* want to ground, so I honestly wouldn't worry about it much.

Ok, it's interesting to learn.

The highpass frequency with 3K3/1u works out at 48Hz. If R19 is the 3K3, the mod is worked out to not change that cutoff, but to increase the potential gain. Max gain goes up from x33 to x72. So I'd say you get more like +6dB or +7dB.

Max gain, original: ((100+6./3.3)+1 = x33.3 = +30.45dB
Min gain, original: (6.8/3.3)+1= x3 = +9.5dB

Max gain, modded: ((100+6./1.5)+1 = x72.2 = +37.2dB
Min gain, modded: (6.8/1.5)+1= x5.5 = +14.8dB

Thank you for the figures. It's very useful to have that and to know the maths behind.
My goal is to get a bit able to better compensate for the compression volume loss occuring (with big compression amounts), using the output volume pot. I found I had it always at full volume, while sometimes struggling to get unity gain.
Yes, the mod is worked out to not change that cutoff, but to increase the potential gain.
I will use an electrolytic, and may even tweak the values a bit in order to lower the HP cutoff point a bit... [edit] By re-reading your post, I get that the 48Hz cutoff point is with the 3.3k resistor, and that using 1.5k would both increase the output gain while keeping the hpf cutoff point as low as around 20Hz.

[ElectricDruid]

I will use an electrolytic, and may even tweak the values a bit in order to lower the HP cutoff point a bit... [edit] By re-reading your post, I get that the 48Hz cutoff point is with the 3.3k resistor, and that using 1.5k would both increase the output gain while keeping the hpf cutoff point as low as around 20Hz.

The mod basically doubles the C value and halves the R value to keep the cutoff the same - roughly, since cap and resistor E-series values (mostly) don't fit neatly into "x2" and "/2" series. But 1K5 is roughly half of 3K3, and 2u2 is roughly twice 1u, and in fact because one value is slightly too small and the other is slightly too large, the RC product and resulting cutoff is actually absolutely identical (48.2Hz in both cases).

If you want to lower the cutoff, you could go up to 3u3 (half an octave further down) or 4u7 (full octave further down, 24Hz or so).

[snk]

Thank you Electric Druid.
Your explanations make perfect sense to me

So, that's the R19 resistor which defines the gain range, and we adjust the HPF cutoff point from here, by adjusting the capacitor value, that's the idea?
The lower the resistor value, the more gain, that's it?
And while the cap while change the cutoff point, it wil not influence the gain, am I right?

One last question : as there is already a 1µF (polyester) soldered, could I simply add a 2.2µF electro cap in parallel so it would make 3.2µF, or is there anything against mixing a polyester and an electrolytic in this case?

[ElectricDruid]

Thank you Electric Druid.
Your explanations make perfect sense to me

Well, I'm glad! Maybe I'm not totally incomprehensible then!

So, that's the R19 resistor which defines the gain range, and we adjust the HPF cutoff point from here, by adjusting the capacitor value, that's the idea?

Yes, exactly.

The lower the resistor value, the more gain, that's it?

Yes. It's an non-inverting op-amp stage, with a few caps added for tone-shaping, but you can just ignore the caps to look at the gain. The standard equation for non-inverting op-amp gain gives us:

Gain = ((R20+100K Pot) / R19) + 1

...so lowering R19 increases the gain.

And while the cap while change the cutoff point, it wil not influence the gain, am I right?

Exactly right. Unless you start think about the filtering as "gain at a particular frequency", which is true, but more complicated than we need for right now.

One last question : as there is already a 1µF (polyester) soldered, could I simply add a 2.2µF electro cap in parallel so it would make 3.2µF, or is there anything against mixing a polyester and an electrolytic in this case?

It's unusual, certainly, but I don't see any problem with it.
For the electrolytic cap the important thing is that the positive end stays more positive than the negative end. In this case the negative end is tied to ground, and the positive end is biased at 4.5V, so there shouldn't be any problem (this is why the original design used a electro cap in this position).
I'd say give it a try. Should be fine.

HTH

[snk]

Ok, so i refrained from being lazy, removed the stock 1µF cap, dropped a 4.7µF instead, along with the 1.5K resistor, and *bang!*, it's jsut perfect ! 
Now, I can have the Diamond compressor compressing much more, and I have enough makeup gain, louder than unity gain if needed.
Thank you again for your valuable advices and insight, it helps me to understand things better (instead of just "putting this component there without thinnking about it").

[antonis]

it helps me to understand things better (instead of just "putting this component there without thinnking about it").

That's the spirit..!! 

P.S.
Kudos to Tom..
(not only here..)