Another PT80 oscillation question to add to the masses.

[theundeadelvis]

So I built another PT80 and as usual it had a lot of runaway feedback. I did Mark Hammer's suggestion of splitting the resistor on pin 12 of the compander into 10k and 12k and running a .015uf cap from the junction to the ground. Oscillation gone. Now the problem is I only get two repeats on the max setting???

[theundeadelvis]

bump one last time for 2006.

[theundeadelvis]

I just removed the cap for the Mark Hammer mod and added a trimpot off the repeats lug. I can now adjust out the oscillation but still only get maybe 4 or 5 repeats. How many should the PT80 be able to produce?

[theundeadelvis]

I promise this is the last bump (technically I haven't bumped this since last year  ). If someone could at least let me know how many repeats they get at maximum? My Rebote seems to get 2 or 3 times more repeats than this PT80? Thanks and Happy New Year!!

[geertjacobs]

My guess:
If your Rebote gets more repeats than your PT-80, than either the schematic is different somewhere or maybe you have put in a wrong component somewhere...
My PT-80 has enough repeats but I haven't counted them.

Have you tried switching/cutting the extra capacitor you added from the circuit and see if you get more repeats again?

edit: BTW how are the three controls set when measuring the number of repeats?

[The Surgeon]

My PT acts just like analog delay: before the oscillation starts near at the end of repeat pot rotation there are sure enough repeats: ´
I can´t count them. Surely more than, say, DOD FX90.
I also built rbt25 and I don´t like the repeats keep going without oscillation, I cannot see use for that kind of feature.
Noise performance isn´t very good either.
Your PT should sound and feel as good as any delay when you get it working right.
Again, sorry for my english.

[geertjacobs]

BTW the PT-80 project page on http://generalguitargadgets.com has an audio sample with the maximum delay.
You could use that to compare.

[Mark Hammer]

Okay.  I got you into this mess.  I suppose I have an obligation to get you out of it.  Lost track of this thread over the holiday period.

Fundamentally, the challenge faced here is one of balancing off bandwidth and amplitude. 

Note that whenever one restricts bandwidth, whether at the high end, low end or both, apparent amplitude drops.  I emphasize "apparent" because that can be different than, and added to, actual amplitude.  When one filters out content, one is essentially "dumping" signal to ground.  At the same time, if it is the frequency content needed to notice a signal, then it will *seem* quieter.  The creation of more, and less, apparent amplitude is the basis of a whole lot of what we try to do here, whether it involves making effects more subtle, or increasing "cutting power", etc.

The 22k resistor to pin 12 helps to determine the overall amplitude limits of the recirculated signal.  What I suggested to you with respect to the cap to ground, was intended to reduce the bandwidth in such a way that some aspects of repeat-quality would be altered for the better.  Of course, by dumping signal to ground, this also a) reduces real amplitude of regenerated signal, and 2) further reduces apparent amplitude of regenerated signal.

Conceivably, the solution is to offset the bandwidth change with an amplitude change in the opposite direction, and to also keep just a bit more bandwidth.  So, in lieu of the original 22k resistor, try using a 10k and 8k2 resistor in series, with the 8k2 one tied directly to pin 12, and the 10k unit tied to the 1uf cap.  Rather than a .015uf cap to ground, use a .01uf, or even .0047uf cap to ground from their junction.  The .01uf cap will provide a rolloff starting around 1590hz, and the .0047uf cap will start to roll off content around 3390hz.  Obviously, intermediate cap values will roll off at points between those two limits.  The switch from combined resistance value of 22k to 18k2 will let a little more signal through; hopefully just enugh to offset the reduced bandwidth without resulting in runaway feedback of an objectionable nature.

I suppose if you want to go "deluxe", you replace the suggested 8k2 resistor with a 6k8 unit and a 5k trimpot (or a 4k7 and 10k trimpot) to nail down the "ideal" series resistance.

I hope that gets you out of the jam.

[theundeadelvis]

Thanks all! As per usual Mark swoops in like our very own stompbox Superman and hopefully saves the day! I'll let you know as soon as I have a chance to do the mods (my better half just had her wisdom teeth cut out). Thanks again to everyone especially Mark!

[zjokka]

So, in lieu of the original 22k resistor, try using a 10k and 8k2 resistor in series, with the 8k2 one tied directly to pin 12, and the 10k unit tied to the 1uf cap.  Rather than a .015uf cap to ground, use a .01uf, or even .0047uf cap to ground from their junction.  The .01uf cap will provide a rolloff starting around 1590hz, and the .0047uf cap will start to roll off content around 3390hz.  Obviously, intermediate cap values will roll off at points between those two limits.  The switch from combined resistance value of 22k to 18k2 will let a little more signal through; hopefully just enugh to offset the reduced bandwidth without resulting in runaway feedback of an objectionable nature.

I suppose if you want to go "deluxe", you replace the suggested 8k2 resistor with a 6k8 unit and a 5k trimpot (or a 4k7 and 10k trimpot) to nail down the "ideal" series resistance.

I hope that gets you out of the jam.

This is an essential note on the PT-80 that should be somewhere in the faqs or something. Discovered it in ongoing thread and built mine with the said network. Luckily it sounded good without. Was wondering though, I don't seem to be getting any oscillations although I'm using 100k pot throughout. Strange. What would be the effect of using these pots instezad of 50k (mine are in the mail).

zj