Which Q's need to be matched on the Ross comp build seen here?---

[rosssurf]

I was told that both halves of the rectified signal should recieve equal treatment to help this pedal perform at it's best. I have done the following build and was wondering which transistors on this layout are the ones that should be closely matched?
http://aronnelson.com/gallery/Toneys-Album/Rossv4

[Mark Hammer]

Q3 and Q4, same as on the Tonepad layout.

The matching is helpful to reduce envelope ripple with longer recovery times (stock 150k resistance).  If you use the switch to reduce that resistance down to 10k, the recovery time is fast enough that the matching may not be quite as critical because the ripple won't be as audible.

[Jay Doyle]

Personally, and this is just my opinion, but I wouldn't spend the time matching ANY of them. Use low noise, high gain transistors like 2N5087s (or whichever 2N508x's are NPNs) and you will be fine. Those rectifying transistors are cranked up to max gain so as long as the gain of them is over 100, which any current manufacture transistor will be, you will be fine.

My opinion of course,

Jay Doyle

[markm]

I've built a few of these Dyna/Ross clones using various layouts without matching transistors and haven't had any issues with them.

[rosssurf]

Well i was asking because i am hearing a slight but sharp volume boost when I turn up the sustain pot then it gets quieter for the last hour or so of rotation. Yes!.. it seems to happen at around 4 O'clock and then drop back down a little. I am not using a reverse taper, just a standard audio taper. any thoughts why this might be happening? 

[Mark Hammer]

Yeah, I've built a couple without matching and they worked fine.  I don't want to overstate the issue of matching.  In fact supertight matching  is going to be moot if the diodes and/or 1M resistors are not matched or the amplitude of the complementary signals coming off the phase-splitter (Q2) are unmatched.

That being said, there always seem to be situations where somebody's needs include having the envelope ripple be as small as possible, and having the two rectified half-waves be as equal in amplitude as possible helps in that regard.  Ripple is moot when the picking is fast and the device is set for fast recovery. As anyone who has strummed a chord into an envelope-controlled filter and let it take a while to die out can tell you, ripple only becomes noticeable (unless it is REALLY bad) when there is space between the notes, the notes are held for a while, and the device is set for very slow recovery.  If that's not what you normally do, then going to lengths to reduce ripple by matching components is unnecessary.