Ross comp distortion upon note decay

[DMS]

Hi my Ross comp (RG Keen R comp schematic) works, with sustain and level controls responding; you can hear the compression coming on pretty hard, and sounds fairly normal, however, as the notes get near the end of their decay, they get  really distorted, and if you brush chords with low strings, you immediately get some garble.  It's the kind of garble that you get as you hear when you shut down a tube amp.   In all other respects, the circuit seems to be working.  Any ideas where to focus troubleshooting?  All component values as per the Keen schemo, the IC is a random CA30080 Harris from Samll Bear.

I have an oscilloscope and signal generator, but I'm not sure what I'd be looking for and where.  I hope the fact that the behavior occurs during the note decay may give someone some ideas... Thanks in advance! DMS

[markm]

Typical problems in this circuit are solder bridges.
Also, did you use a 2K trimmer?

[MartyMart]

...... also double check the cap orientation, I had one reversed with the EXACT same symptom !!!

MM.

[Mark Hammer]

That's envelope ripple you hear. 

Normally, full-wave rectification and a suitably chosen smoothing capacitor make the little teeny blips between successive peaks in the envelope signal inaudible, the same way it eliminates hum from properly rectified/smoothed AC power.  The blips are MOST inaudible when what gets rectified is higher notes and chock full of harmonics, such that the little gaps between peaks of the "folded-over" signal are very subtle and short.  Unfortunately, for low notes, especially during the decay portion of the note, when all that harmonic content has disappeared and you are dealing with a rectified signal consisting primarily of fundamental, those blips can become quite audible, and that's what you are hearing.  In actuality, it is superfast tremolo, but people hear it as "distortion".

How to cure it?  One way is to increase the value of the smoothing capacitor.  In this case, that would mean upping the value of the 10uf cap to ground from the junction of the two transistor collectors in the rectifier.  Typically, the next standard value up from 10uf is 22uf.  Bit of a leap in my view, but that's what you'll have.  It will likely eradicate the ripple but will also change the attack/recovery parameters of the compressor, and maybe not to your liking.  Certainly, with the stock 150k resistor between V+ and that cap, it will take a lot longer for the unit to recover from peaks.  You'll get longer "sustain" I suppose, but at the cost of not being able to hear additional notes when picking fast.  Installing a lower resistance value (or some sort of switching or pot arrangement to achieve lower resistances) will help to offset this, but he attack will still not be as fast as it once was.

Another approach to curing it is to aim for "matching" the two rectifier paths from the phase splitter after the 3080.  Each of those cap, diode, 1M resistor plus tranny paths delivers one half cycle to the summing node at the 150k/10uf junction.  When they are well-matched, ripple is minimized.  Mismatching can occur for a variety of reasons.  E.g., the two diodes may have different forward voltages.  The 5% resistors may, just by chance, be 5% above nominal value and 5% below.  The caps may be 20% different in actual value.  Increasing the 10uf cap will smooth over those differences to some extent, but as I say, at a cost of attack time.

[DMS]

Thanks for the great replies - I will recheck all cap polarities and, nope, no 2K trimmer.  I used the junction of a 1k & 1k in series (I think this was point of discussion in an earlier thread on the Keeley).  I will also look closer at component tolerances in the rectifier paths.  By the way, very cool circuit explanation by Mark H... would be great to have that understanding of the whole circuit.  Wonder if any block diagrams exist, since this is such a commonly built/studied circuit.  Anyway - thanks for helping out... more debugging tonight!

[markm]

Thanks for the great replies - I will recheck all cap polarities and, nope, no 2K trimmer.  I used the junction of a 1k & 1k in series (I think this was point of discussion in an earlier thread on the Keeley).

Am I the only one that still uses a trimmer in this build? 

[Papa_lazerous]

Thanks for the great replies - I will recheck all cap polarities and, nope, no 2K trimmer.  I used the junction of a 1k & 1k in series (I think this was point of discussion in an earlier thread on the Keeley).

Am I the only one that still uses a trimmer in this build? 

No I used a trimmer also, that way you can get to the sweet spot.  who knows it may be slightly off with 2 1k's that was my thinking

[markm]

Thanks for the great replies - I will recheck all cap polarities and, nope, no 2K trimmer.  I used the junction of a 1k & 1k in series (I think this was point of discussion in an earlier thread on the Keeley).

Am I the only one that still uses a trimmer in this build? 

No I used a trimmer also, that way you can get to the sweet spot.  who knows it may be slightly off with 2 1k's that was my thinking

Mine as well.
Either my trimmer pots are off or there is a "sweet-spot" to be found thats NOT 2K split down the middle?
Maybe the trimmer is more "exact" than using resistors that are 5% or even 1% tolerance....I don't know.