OS Comp: Why the lag when switching between two bias settings?

Started by lowvolt, March 14, 2013, 12:23:39 PM

Previous topic - Next topic

lowvolt

Circuit = buck standard OS comp (GGG to be specific).

Problem = I've installed two bias controls that I may switch between so that I have (in effect) two "presets" to toggle between.  When doing so, there is a signal lag present when going from some bias to lower bias, but not when going from lower bias settings to more bias settings.  To clarify, "lower bias" (MORE resistance) and "higher bias" (LESS resistance).

Said another way, when I have one bias trimmer/pot set to produce a reduced compression sound (which is actually a higher resistance setting on the control trimmer/pot), and I switch to another bias trimmer/pot that is set to produce a more compressed sound (which is actually a lower reistance setting on the control trimmer/pot) there is a ~lag~.  But the lag does NOT occur when switching from higher comp settings (lower bias resistance) to lower comp settings (higher bias resistance).

Guesses = It sounds as though when going from less compressed sounds to higher compressed sounds that there is a capacitor being charged, and that *charge up time* is what is producing the perceived audio lag.

Question(s) = Is there a simple fix for this?  Is there just one of the circuit's caps that I can move to each ~side~ of the switch that I might use to prevent this?  My thinking is that if I were to somehow move the ~proper~ caps to each side of the switch (along with the bias trimmers/pots) that I could set up my two "presets", give the guitar a few strums on each side of the switch to properly charge each cap, then go about my business.

In other words, I'm thinking that I need to have more of the circuit on each side of the switch than just the bias control trimmer/pot.  Perhaps the bias control and maybe one of the circuit's caps along with it ... or ... erm  .... uh ... sumthin.

So how much circuitry would have to be added to each side of the "preset switch" to eliminate the lag issue?  The overly obvious answer is to simply use two complete OS circuits and set them to the desired settings and then A/B between them.  But if there is a more space-efficient manner of doing what I wish to do I'd prefer it.

Thanks for tolerating my limited electronics knowledge.  I really am trying to learn, though!  I hope to one day be a contributor rather than simply a consumer here at DIYStompboxes.
I didn't say it was your fault, I said I was going to blame you.

jasonsguild

Edit - "Maybe it's" the 4.7Uf cap in between Q1 & Q2 that's causing the lag when switching to the lower resistance bias, as more voltage gets let through.  For a quick test, I think you can safely disconnect the cathode side of that 4.7uf cap from ground to see if the lag goes away.  Not a permanent solution though, as it would likely add noise to the circuit without that filter cap.
Jason D

Mark Hammer

If the intention is to change between more and less compression, there is an easier way to do it.

1) The JFET and 82k input resistor form a virtual attenuator, equivalent to a volume pot, with the wiper being their junction.  Like any attenuator, the amount of attenuation is a function of the ratio between the resistance of the input leg (in this case the 82k fixed resistor), and the ground leg (in this case the JFET).  You are presently tringto alter the attenuation via the ground leg, but it will be easier to do so via the input leg.  Make that input resistor 100k or 120k, rather than 82k, and changes in the JFET's drain-source resistance will have a greater impact on the attenuation.

2) Alternatively, more squish can be gotten by driving the gate of the JFET harder.  The op-amp is set for a fixed gain of 23x, but there is no requirement that it BE exactly that gain.  Make the 10k resistor a 25k pot with a 3k3 resistor in series, and you'll be able to get both lighter and harder squish by feeding a hotter or cooler signal to the rectifier.

neither of these approaches will require any caps to "settle down" in order to work.

lowvolt

Thanks to both of you.  It all makes sense to me now that more of the elementry workings have been explained.  I am certain to try the suggestions mentioned here, they sound like much more viable solutions to my wants than the path I took.

Thank you!  :)
I didn't say it was your fault, I said I was going to blame you.

lowvolt

Quote from: Mark Hammer on March 14, 2013, 02:03:58 PM
If the intention is to change between more and less compression, there is an easier way to do it.

1) The JFET and 82k input resistor form a virtual attenuator, equivalent to a volume pot, with the wiper being their junction.  Like any attenuator, the amount of attenuation is a function of the ratio between the resistance of the input leg (in this case the 82k fixed resistor), and the ground leg (in this case the JFET).  You are presently tringto alter the attenuation via the ground leg, but it will be easier to do so via the input leg.  Make that input resistor 100k or 120k, rather than 82k, and changes in the JFET's drain-source resistance will have a greater impact on the attenuation.

2) Alternatively, more squish can be gotten by driving the gate of the JFET harder.  The op-amp is set for a fixed gain of 23x, but there is no requirement that it BE exactly that gain.  Make the 10k resistor a 25k pot with a 3k3 resistor in series, and you'll be able to get both lighter and harder squish by feeding a hotter or cooler signal to the rectifier.

neither of these approaches will require any caps to "settle down" in order to work.
Ok, for #2, when you say the "10k resistor" are you speaking of R10 (the one connected between pin 2/neg of the IC and ground)?  Just making sure I have a handle on your suggestions.  :)
I didn't say it was your fault, I said I was going to blame you.

Mark Hammer


lowvolt

I didn't say it was your fault, I said I was going to blame you.