And in true spirit of overdoing it, after playing with our phaser components, let’s look into LFO. For previous posts have a look here.
As usual you can skip to the video and come back later for diagrams. In this article I cover some modding options for LFO, I do some SPICE simulation and then I try it out. There’s no breadboarding diagram since this change is just adding a couple of pots.
Why Change LFO?
LFO stands for Low Frequency Oscillator, and this is what is controlling our Phaser. If we look at just the LFO part of the circuit:
There are few things we might consider changing. But from our previous analysis we came to conclusion that some of the JFETs have a really narrow control voltage swing that would be of interest. It stands to reason that controlling this voltage swing might be beneficial.
Also, if you notice on the diagram above, biasing of JFETs is intrinsically linked to the LFO too. If we keep bias pot as trim-pot, any changes will require opening up the pedal to do adjustments. If we use a proper pot, then we have very small margin of useful change.
In the previous post we saw that even as little as 25mV difference in biasing might affect our sound. The bias trim-pot can control bias voltage in a range of over 4V. Back of the envelop math (i.e. very imprecise) tells me that 25mV is 0.6% of the rotation of the pot. This is less than 2 degrees if we take that standard pot has 270 degree rotation.
That is very hard to control, this is why I opted to add additional biasing pot that should give me control of about 200mV. Now, this might be useless or not, but we’ll see that shortly.
I ended up with this schematic for the biasing and LFO:
How do these changes work?
Well, RV4 there is for bias, to get this to work, set RV4 pot to, say 10 or 12 o’clock. Then adjust bias trim-pot (RV2) until you get JFETs properly biased as much as you can. Now you can play around with RV4 without touching the trim-pot.
The original resistor value there was 1 Mega ohm. I replaced that with 770K resistors and 500K pot. This means we can vary this resistance between 770K and 1270K and this affects our biasing point a bit due to interaction with R15.
RV3 affects our swing by changing reference point for our Schmitt trigger. This affects amplitude of our triangular LFO output so the maximum swing goes from 500mV peek-to-peek all the way down to 115mV.
Note though that smaller the swing faster the LFO rate, so changing RV3 might require corrections of speed (RV1) and some adjustments of bias (RV4) too. If you peek at the diagram of SPICE simulation below you can notice clearly that smaller swing makes LFO frequency higher.
This is my full schematic so far:
Other Things to Try
There are several other things I could try, but I feel I’m overstaying my welcome with this series of posts so I’ll just skip those 😀. For example, adding a mixer pot at output might give us more pronounced or less pronounced effect of the notches:
The effect is subtle anyway so this might not do much for you.
Or we could also add ramp up/down (sawtooth) option – so instead of triangular wave – something like this:
I hope you can see above that the output has 3 different shapes.
With three position switch that is disconnected in the middle position you get 3 wave options.
I tried them out and two stage phaser is too subtle for making a huge difference I think. Still, definitely something to be aware of if thinking of trying it in different scenarios.
I stripped out all unrelated circuitry to play with my LFO:
You can see here on the diagram what to expect from the output when swing pot is manipulated:
And here’s the output when swing is fixed and we manipulate bias:
You can see above that amplitude stays nearly the same (there is a slight drop in it so some adjustments may be needed using all pots) but the voltage of our midway point varies by about 200mV.
Here’s the LTSpice file if you want to play with this yourself:
Subtle changes, but very much worth pursuing. I kind of loved this effect from the get go. It just makes me want to continue playing and this adds another layer to it. For the next time, I have to complete this circuit and assemble it, that is going to be the goal anyway.