My rendering of the PUP6

Greetings all.

I have made the PUP6 phaser and I am very much pleased with the result, which is far beyond my expectations. It has the swirl of a Mu-Tron Phaser II (at least what I hear from demos) and it can do the Phase 90. It has NOT the thump of a Uni-Vibe, because the latter employs different time constants.

I have been in contact with the site owner, @bentfishbowl, and he invited me to present my rendering.

I gladly follow that, but I have decided that the original circuit is so well described and so much refined, that it probably needs no improvement.

That said, I have used a different LFO section for experimentation, intrigued by the LFO used in R.O.G.'s "Tri-Vibe", which aims to linearize the non-linear transfer function of the MOS-FETs. This works very well in conjunction with the PUP6 core, and the resulting swing appears very balanced.

This LFO rendering has the added advantage of an always constant LFO voltage (because of the anti-parallel pair of diodes, and not to say that the original design suffers from that, but several other designs DO) plus a "depth control".

Indeed some of the settings and switch combinations appear so pronounced that it is a welcome commodity to be able to reduce the depth. I think that no drawing is necessary. The Tri-Vibe is self explanatory.

Using its LFO, a small adaptation had to be made to the balancing circuit. I used R45=33k, R48=8k2, R50=47k, C14=1µF. All designators referring to the original PUP6 drawings.

I did not care too much about the size of C14, which was chosen by @bentfishbowl to avoid sea-sickness with higher LFO rates. This is being taken care of by setting the "depth" control appropriately. However, it will aid to avoid LFO bleed-through, and it does.

The R.O.G. design does not incorporate a LED annunciation that gives a visual feedback of the current LFO state, but the TH Custom kit that is provided by Musikding in Germany does. This is basically not more than a resistor/LED combo to the bypass-switch as common.

Since there is no need to re-invent the wheel, I borrowed another trick from the Tri-Vibe, namely the pre-emphasis and de-emphasis circuit, which helps to reduce noise. This works exceptionally well and comes into the bargain.

For those interested, this is achieved by nothing more than a 10k and 6.8nF series combo from U2A, pin 2 to ground, but instead of the direct wire connection from pin 1 to 2 it uses a 33k resistor. This boosts all frequencies above ca. 2kHz by a factor of 3 roughly.

The 33k resistor was chosen with purpose. It has the same value as R16 on the output, and indeed the same R/C combo is wired in parallel to R16, which yields an attenuation of exactly the same value with the same corner frequency as the R/C on the input. This counteracts the input boost, but attentuates any noise that was generated on the way with it.

This trick is totally transparent, but optional. Although it might seem redundant, since the circuit is very quiet despite the usage of OPA's that may appear mediocre by today's high-bread low noise specimen, noise amounts in situations where a high gain device is placed somewhere upstream. I have tried such OPA units, but the experiment failed catastrophically, and I strongly advise against them without applying further measures of band limiting and rail decoupling. There is so little gain involved that this effort appears entirely superfluous.

I have explored several other tangents, but those are of no importance. I feel that the combinations possible with the switch set as depicted are more than sufficient. Indeed, less is more.

Have fun.