Biasing Stompflo Modulation Output

[KarenColumbo]

Gentlemen, I have a question. I breadboarded a double PT2399 chorus (based on the cool "Jenny Greenteeth" chorus by Valve Wizard) using Tom's excellent StompLFO ("Stompflo").
Works fine, but I think the LFO output isn't biased correctly. I think it "hangs" a bit too long on one side of the sweep, because I am overloading the PT2399 or feeding it with a voltage (?) beyond Pin 6's scope.

Can you tell me which part I have to wiggle so that it makes a full curve? I'm using the filter network Tom posted on his site ("option two: passive filter"), whereas R10 is 22k instead of 100k, because it seems to have made it better a bit. Any further reduction of R10's value seems to make it worse. Also I left out C10 because it made no audible difference. I guess I have to make it swing around a certain artificial zero-point, what where is it? What does the PT2399 need here?

Here's the "Jenny" schematic. R10's output (from the schemo above) goes directly into the base of the lower BC337.

[ElectricDruid]

Without analysing anything closely, I'd guess that the LFO output is going down so low (it goes to 0V) that there isn't the 0.6V on the base of the BC337 that you need to turn it on. So it's being turned hard off and you're clipping the bottom of the LFO waveform.

One simple way to fix this would be to limit the LFO's Depth CV to only provide 4.4V (e.g. keeping us 0.6V away from the bottom) by adding a 1K2 resistor to 5V above the 10K pot. Then you can tie the Offset CV to 5V with a 10K so the modulation starts from the top down and it won't ever reach the part where it cuts off.

Alternatively, the reason the original Jenny Greenteeth works is that the LFO that's there operates around a 4.5V bias and swings to probably about 2V and 7V either side - it depends on the op-amp and the actual values, but that's fairly typical. For this situation, Option 1 (the active filter version) provides the same offset and should be a drop-in replacement.

Third possibility is to try adding some biasing resistors direct to the base of that BC337 and then feed the LFO to the base via another resistor. I can't work out how that will go off the top of my head (need more tea) but it's worth a shot.

[patrick398]

Would it be possible to use the biasing configuration from the phase 90? Springs to mind as it's on my breadboard at the moment and allows you to tune the lfo niceley to the fets...

[Kipper4]

Did you try a series cap (1uf) and a large R (100k work your way down)from StompLFO pin3 straight to pin6 of pt2339?
Do not omit your min R to gnd off of pin6 too. (1k) or it might latch up.

> ---][--/\/\/\-------->
                       /                 
                       \
                       /
                       gnd

[KarenColumbo]

I looked at the Phase 90 thing - maybe not hard enough. Will give this a try, thanks!

Hey Rich, I have the "Jenny Greenteeth" on the breadboard. This one has a anti-latch-up-circuit (those BC337ies), so I guess there's no chance of them going down - that's why I "omitted" the pin 6 resistor because there's a failsafe mechanism at work already.

[patrick398]

I think i'm actually having the same trouble as you with my lfo on the phase 90. Sounds like the sweeps are uneven, kind of like it's got an irregular swing to it. On slower sweeps it stops doing anything after its done the up sweep (for quite a while), then starts doing the down sweep, then stops doing anything again (for a slightly shorter time than after the up sweep). It would be nice to have regular and constant phasing without all the gaps in between cycles.

Not trying to hijack your thread here, just thinking our problems are very closely related!
I'm thinking it's to do with the biasing of the fets/bjts. Maybe a larger resistance across source and drain/collector emitter?

[patrick398]

So adding more series resistance from the output of the stomp lfo seems to have worked for me. I felt like the lfo's output was too hot for my fets even though it's already going through a 3M3, then 1M and into 250k pot. I added another 1M8 and that seems to have lowered the output enough for the fets to register the whole sweep.
This is largely hunches and guesswork so point me in the right direction if i'm wrong here people.

Maybe you could try playing the resistances into the bias pot before your BJT?
Sorry i can't be more practical help here

[Kipper4]

Ref jenny green. Only 1 of the bjts is anti latch the other feeds the lfo signal to pin6.
Traditionally a phasers is used with tri or sine wave iirc.
So don't be surprised if it ain't as phasers like as you wish with all waves and speeds.
You can set the lfo to a set wave with a voltage divider.

You could also back off the depth on some waves.
I found a series R (large) did help
Plus some lpf on lfo output helped.
I think I used a fc of around 350hz.

[ElectricDruid]

Plus some lpf on lfo output helped.
I think I used a fc of around 350hz.

I'd have to recommend this with the StompLFO, especially when used with a digital chip like the PT2399. The "raw" output from the StompLFO is actually a series of narrow 5V pulses at 2MHz ("Pulse Density Modulation", if you want to look it up) so that might well scramble the clock on the PT2399. Hence the need for a bit of lowpass, but since the frequency is so high, a simple RC rolloff seems to work perfectly well without more complicated active filters being necessary.
If you've got a series resistor already to reduce the level a bit, just hang a little cap off the end down to ground and you're good. The cap's value obviously depends on your resistor, but with 3 or 4 Meg in series, you'd be around 100pF.

HTH,
Tom