Phase 45 unbalanced sweep

[slashandburn]

I'm back begging for help again. I finally built a Phase 45 on the tone pad pcb and build doc, here:  It's way more of a phaser than I had expected, wish I'd built it earlier now.  Most accounts I had read (and YouTube demos I've watched) gave me the impression that this was a fairly subtle phaser but it sounds pretty huge to my ears (not a bad thing!). There's a strong chance there's nothing wrong and I'm just being dim here, but the Sweep" doesn't sound even to me. Not sure how best to describe it, unbalanced? Uneven, lopsided? The lower half of the Sweep seems to be over too quickly and the upper half of the seems to drag on for ages.  I've been adjusting the trim pot by ear as per the GGG build instructions. The sweetspot is fairly easy to find but I wonder if I'm missing something that I could only dial in with a lighter touch or a multi-turn trimmer. I also understand there's two 10k resistors that should be closely (perfectly?) matched that I didn't pay attention to when building, so I'm not sure what to look at first. There's also the possibility my jfets aren't as well matched at my jfet Matcher implied ( http://2.bp.blogspot.com/-zLpU6E6w8xs/UBRUu_pzPhI/AAAAAAAACCo/3XLcB0oReSc/s1600/Greatly+Improved+JFET+Matcher+II.png)

I'm not sure where to look first, the trim pot, the jfets, the 10k resistors or something else entirely. If I recall correctly the only parts substitutions I made were using 2n5457 as the jfets and a 200k trimmer. Also a 1m Linear Pot with a resistor across lugs 1+3 in an attempt to get an anti-log 500k.

Any nudges in the right direction would be hugely appreciated. I can provide voltages and the matched jfet values if they would help.  Cheers in advance! 

[R O Tiree]

Most of them do that.  The only ones that don't will have an LFO that pushes out a hyper-triangle waveform.

OK, the LFO you have in there is linear... triangular... smoothly ramps up, smoothly ramps down, so many volts per second rise/fall rate.  But our ears' perception of frequency is nowhere near linear - it doubles every octave.  So, say the LFO starts at 0.5V (it probably doesn't, but let's keep the maths simple) - by the time it gets to 1V, the "phase frequency", for want of a better term, has doubled.  The LFO now needs to get to 2V to double the phase frequency again, and 4V to re-double it and then 8V.  The voltage required to get the same-size "step", as our ears perceive it, has to double every time.  So...

A hyper-triangle ramps up slowly to start with, increasing its slope as time marches on.  When it hits the top stop, it ramps down sharply to begin with, gradually flattening out until it reaches the bottom stop.  There's a post from about 4 years ago where this is explored, a bit, with an interesting reply (reply 7) from Mark Hammer about the potential drawbacks and how EHX solved the problem.

There's a neat circuit, if you want to check this out, here at birthofasynth that converts a triangular waveform (which you're already getting from your existing LFO) into a sine or hyper-triangle.  The circuit as shown runs off +15/0/-15V, but you can run it off 9V/0V (bias at the LFO's Vref, of course).  The trimpot at R1 adjusts the shape of the sine wave (which you'll need to do to get a good hyper-triangle shape later).  The pot at R6 will give you a sine wave in the centre, hyper-triangle at one extreme and inverted hyper-triangle at the other.  Once you've found which way round that is, hard wire it.  Not many components and you can always install a switch for LFO source if you run into the problems Mark Hammer mentioned in my first link.

[slashandburn]

Ah cool, thanks! I understood just about enough of that to gather that the phaser is okay.  I had a quick play around again today and I'm much much happier with it knowing its meant to be a bit lopsided. Sounds good all the same!

That is a neat circuit you linked! I was bracing myself for something complex but that looks totally doable on breadboard. Something for a rainy day. Thanks!