Hmm. That's interesting. Let's think about how this thing is doing the phase shift and see if we can't figure out something.
The output of a common emitter amp is 180* out of phase with the input (because it is an inverting voltage amplifier). And the output of a common collector amp (like a buffer) is in phase with the input. So the amplifier in the Wobbletron is a very clever way to alternate the output between those two extremes. When the LED is bright, the LDR's resistance falls to a couple hundred ohms, and more of the output comes from the emitter (making it more like an in-phase buffer). And when the LED is dark, the resistance rises to greater than 1M, which effectively makes a short circuit, and all of the output comes out of the collector (which would make it out of phase). So by using the LFO to drive the LED, we get an output that is sometimes out of phase, sometimes in phase, and some of the time it's a little of both. Escobedo was a genius! And that's why I always steal his stuff.
So then, you are saying the following is happening:
• If you decrease the max resistance of the LDR with a parallel resistance, you get less effect. So that means we need to make sure the resistance is high to get the transistor to more fully output from the collector on part of the LFO cycle. Okay, that makes sense.
• If you move the LED further away from the LDR, you get less effect. I assume this means that the LDR is not getting enough light to make the transistor's output enough on the emitter side during the other part of the LFO cycle. Okay, that also makes sense.
So, we need to find a way to make the LED bright enough (or close enough), and also make sure it goes dark enough to get that LDR to max out its resistance.
I would suggest first seeing if two LEDs in series will work. You could probably move them back a little bit and get the same brightness, but also the distance would mean a darker dark-half of the LFO cycle. That might make the effect more pronounced. Or maybe use a higher voltage drop LED (like blue water-clear)?
The problem with LDRs is that they do not have the same response time for both their rise and fall. In this case "rise" means when they go from dark to light, and "fall" means when they go from light to dark. They usually rise quickly, and fall much slower. So it's hard to get a symmetric result from an LFO.
BUT, the Magnatone Vibe uses the same Wobbletron base, but with an LED/LDR combo driven by the EA Trem LFO, and it gets pitch bendy as hell. So I know it can be done. It might come down to the LDR itself and it's min/max resistance and rise/fall specs.
Oh, and see what happens if you change the value of C5 to something much lower. I'd rig up a switch so you can hear the change in real time.