How can i obtain a faster attack and a longer sustain for an Orange Squeezer?

[gigimarga]

Hello,

I've built an Orange Squeezer using Tonepad's project: http://www.tonepad.com/getFile.asp?id=88.
The only mod that i made was to replace the 4.7uF after the Ge diode with 1uF for a faster attack (i hope that's right).

My questions is how can i obtain a longer sustain and if i can to improve the attack more?

Thx a lot!

[Mark Hammer]

That 4.7uf cap determines the entire timing relationship of the compressor's rectifier, start to finish.

What happens is that the cap stores charge coming via the 1k5 resistor and diode.  The cap charges up to its peak capacity depending on how much current is permitted to flow into it, and also depending on what "full" is.  "Full" is larger for a 10uf than for a 4.7uf than for a 1uf.

The decay time is a function of how fast that cap discharges or drains.  Once again, like a bottle of water, how fast it drains depends on how big the "opening" is and what "full" is.  You will note that there is a 100k resistor in parallel with the 4.7uf cap.  Think of this like the opening to the bottle that lets the cap drain to ground.  If that resistance is smaller, it is like a larger opening in the bottle and current drains out faster.  If the resistance is larger, then the "opening" is smaller and it takes longer for that current to drain off.

The resistance between the diode and the cap also plays a role.  Again, think of the cap as a bottle that we wish to fill up.  If there is a resistance between the diode and the cap, then there will be less current fed to the cap and it will take longer to charge up.  This is why you will often see the "attack" control on many circuits in the form of a variable resistance between the diode and cap.  Make the resistance larger, and less current passes so the cap/bottle takes longer to fill up.

So while a smaller cap will give you a faster attack (because that "bottle" fills up faster), it cannot give you a longer decay because it drains faster than a larger cap would.  To achieve a longer decay/sustain/"drain time", you would need to use a larger cap (go back to 4u7) and increase the value of the 100k parallel resistor.  It would not be unreasonable to replace it with a 47k resistor in series with a 500k pot (variable resistance), so that your decay time goes from fast (47k) to slow (547k).

Just one caveat I will mention, though.  The OS uses a half-wave rectifier.  This type of rectifier tends to introduce what is called "envelope ripple".  In theory, it is like a very subtle audio-frequency tremolo.  In practice, it sounds like uneven bad distortion.  The MOST ripple happens during the point where the strings are vibrating freely and slowly dying out.  When the timing cap is small and the decay fast, you will not hear the ripple because things happens too quickly.  When the decay is very long, the ripple is smoothed out by the cap, and you don't hear iot either.  When the decay time is in-between, however, sometimes that ripple can be unpleasant.  happily, it is fixed by simply increasing the decay time with the suggested pot.  Just be aware that it is a natural occurrence and not a technical problem with the circuit.

[Headshot]

Thanks for that great explanation Mr Hammer!

[gigimarga]

Thx for this exceptional answer, as usually, Mark!

If i understood well, if i will use a 10uF cap instead of the 1uF it could better for what i want (because the bottle will have more water in)?

[Mark Hammer]

10uf will certainly give you a longer decay.  Will it give you a fast enough attack?  I will leave that to your ears to say.

As a general question to others here:  Is it possible to increase the amount of current being fed to that storage/averaging cap in a way that is independent of the series resistance ahead of it?  In other words, any series resistor may limit the current charging up the cap, but that current has to come from somewhere, right?  Can I assume that there are "somewheres" that supply lesser and greater amounts of current "at source", such that even a larger cap like 10uf or 22uf can be charged up quickly so as to offer fast attack in addition to long smooth decay?

And if that is a reasonable assumption, how complicated an addition to the basic HWR circuit are we talking?

[gigimarga]

After i used a pot as Mark Hammer suggested below and i made some tests, I replaced the 100K resistor with a 470K one and i left the 4.7uF cap in his place.

Thx a lot Mark!

[freeride]

I have a question about the "sustain".  If we increase the resistance in parallel to the cap, that means the cap will drain slower, so the JFET gate voltage will be higher longer and the JFET will be on longer and the compression will last longer (i.e. notes stay volume limited [quieter] longer).  I'm looking for something that adds sustain, not just something that cuts the loud notes for longer.  Anybody have any ideas for how to modify the orange squeezer to make notes last longer - a sustainer? 

[Mark Hammer]

Compressors never make the notes last longer.  They just make them seem like they do by evening out the difference between the loudest and softest segments of the note over time.  Best you can do is lengthen the decay/recovery period, as described.....and stand near an amp turned way up.

After that, you have to turn to a device that mechanically or electromechanically keeps the string vibrating, or electronically synthesizes a note and holds it to a constant amplitude.