What Does the Foxx Phase?

[armdnrdy]

Nick,

The switch board pads are a bit different than I originally drew.

The pads are Z, X, Y from the right side of the PCB when looking at the top of the board. Top being toe side.

The above drawing has been revised.

Edit:
I guess it doesn't matter on your board....the XYZ pads are laid out differently then the original.

[armdnrdy]

I have a question about the voltage divider coming out of the LFO. (R21, R25, C12)

What exactly is C12 doing? and is the value critical?

The answer to question #1 probably answers question #2.  

[anotherjim]

To answer the above, it was dealt with here...
http://www.diystompboxes.com/smfforum/index.php?topic=110857.0
In short, C12 controls a voltage divider R21/25 in a frequency dependant manner, so the LFO signal from Z3 pin8 is progressively cut by up to half amplitude  as it's frequency increases.

I now have Nicks MkI on my bench (Big thanks to Nick). My earlier question about the LFO wave shape is answered -  it's a Sine. Does get a bit bent at extreme settings of the speed pot, but I don't think that's a defect, just a consequence of adjusting speed with a single control.
I'm still struggling to find an ancestor of this circuit (But I'm sure I've seen it's like before). I'm now leaning towards the idea that it's a twin-T oscillator of some sort, even though the "T's" are disguised somewhat.

[nickbungus]

I used a polarised cap for c12 as that's all I had.  I take it that shouldn't affect the tone in any way?

[armdnrdy]

I've finished the schematic, component value overlay, (of original board) and am working on a designation overlay.

The latter is a bit difficult without trace side images and/or a correct schematic.

Any word from Musicparts?

So far...there are quite a few differences between the Sears and the Foxx models.

The cap that I was inquiring about (C12) along with R25, seem to be missing from the Foxx. I see empty holes where R25 should be.

[anotherjim]

Yes, Nick used an 0.47uF electro. Shouldn't be a difference in itself to the tone.

I've just done some measurement of the LFO, frequency and peak-peak level. The supplied battery was somewhat depleted and my earlier comment about the sine shape distorting is wrong. Powered by a 9v PSU, the wave shape is a pretty good sine at all settings.

Measured from Z3 pin8 (so not influenced by the aforementioned C12).
Slow range. Min=0.24Hz@1.24vpp. Max=0.94Hz@0.8vpp.
Fast range. Min=2.7Hz@1.52vpp. Max=11.36Hz@0.92vpp.

Note that there is a gap between the ranges. Deliberate or timing cap tolerances?

Also, despite what R21,25 & C12 are intended to do, the amplitude drops as speed increases anyway.

On listening, to my ears the sweep sounds as though it's all at one end of the LFO cycle. It does sound like the full phase sweep, but only happening in about 3/4 of the cycle. This gives it a pulsating character with slow to moderate sweep rates, although it isn't so noticeable at fast rates. So the peak to peak of the LFO leaving Z3 pin7 probably needs reducing. That said, it may be something the production pedals did anyway.

R39, the sweep bias pot, is quite sensitive, and it's just as well Nick used a multi-turn type.

[armdnrdy]

Thanks for the info Jim!

You mentioned the LFO output. (Z3 pin 7)

I have the input resistor (R34) verified from images of a few different Foxx boards as a 6.8M, instead of the Sears 2.2M.

[anotherjim]

I have the input resistor (R34) verified from images of a few different Foxx boards as a 6.8M, instead of the Sears 2.2M.

Ah-ha! Larry, do you know if the JFET's used differ between the various productions? Is R38 still 1M whatever?

[armdnrdy]

Hey Jim,

Most of the component board shots that I've collected were taken from a 90 degree vantage point so...I don't have a very good view of the JFETs.

I can make out PN4302 from one shot, and the last two numbers (02) from a different board.

The JFETs differ from board to board...standard TO-92 (PN) and glob-top transistors (2N)

Yes, R38 is verified on several boards as 1M.

[nickbungus]

On listening, to my ears the sweep sounds as though it's all at one end of the LFO cycle. It does sound like the full phase sweep, but only happening in about 3/4 of the cycle. This gives it a pulsating character with slow to moderate sweep rates, although it isn't so noticeable at fast rates. So the peak to peak of the LFO leaving Z3 pin7 probably needs reducing.

Could that be because I've made a mistake somewhere along the line?  I've selected the wrong component or I've misread the schematic, some of the decimal points are a bit ropey?

Jim, also the depth control doesn't seem to affect the overall sound to my ears (except in volume really), what do you reckon?

While you have it Jim, if there's anything you want to try, please feel free to take the soldering iron to it.  Like I've said, its mk I so I wouldn't be too bothered.  

Armdnrdy, if the main differences between the 1482 and the Studio are a few components that can directly swapped and some offboard stuff, I'd like to make a PCB layout that can do both.  If thats possible, then I'll get some boards professionally manufactured.  My wife is going mad about all the ferric chloride stains in the garage and around the garden

[armdnrdy]

With the info that I've "uncovered" so far...it seems as if the Sears model either came first...or was introduced as the "castrated" version....not to "out do" the Foxx version.

The Foxx version shows signs of improvement, better power filtering, larger signal path caps,  (C16, C21) among other things.

The Foxx phaser was made "famous" because Brian May played through one. Brian didn't play through the Sears model.

The point I'm trying to make is...I believe that the Foxx is the improved version of the Sears and...people are familiar with the Foxx Phaser because of Brian May's use.

Common sense tells me that since no one has heard of the Sears Phaser....why would anyone build the lesser version?

All three versions including the Regal are made with the same board...it would be easy to do but...I guess I don't understand why  you would want to make the lesser Sears version. 

[anotherjim]

Nick, there can always be component errors - even on production boards. An official schematic can fail to reflect any late changes that were found necessary once production got under way. Sometimes, mods were made that can only be seen on the copper side.
I will see what I can do to improve the sweep AND the volume drop when engaged without adding anything.

Sorry Nick, what do you call the depth control? Do you mean the pedal only with LFO switched out? I haven't looked at that aspect yet.

[nickbungus]

Armdnrdy:True.  It's just good to keep the options available.  What I don't get is how the 1482 schematic is on the net and the studio isn't.  Especially coming from Greg Covington, as it seems he definitely did his homework.  His site is down now but he make all the Brian May stuff.

I'll be making your version next regardless.

[nickbungus]

Hi Jim. R2 and R7 I believe are the depth control.  But remember I'm an idiot.

[anotherjim]

Gotcha.
It's a bit strange really. It's the Q control. Originally, "Q" represents the quality factor or efficiency of a resonant circuit or filter. It's often used to describe the boosting of effected frequencies by providing positive feedback from output back to input. The thing here is that the feedback is from the output wet/dry mixer Z2 pin8. So it includes the un-effected dry signal - which is fed back and mixed with the incoming signal Z2 pin14. The strange part (to me) is that it's more usual to only send the wet signal (as from Z2 pin7) back to the input.

I haven't played with it that much, but it seems to work, changing the density of the effect and makes it more "nasal" sounding.

[nickbungus]

According to this, its the depth control, but even I could work out that's a wet/dry mixer.  But I also read the following on Wiki.

When signals from the two paths are mixed, the frequencies that are out of phase will cancel each other out, creating the phaser's characteristic notches. Changing the mix ratio changes the depth of the notches; the deepest notches occur when the mix ratio is 50%.

[anotherjim]

I can see where you might come to that conclusion, but in fact what you're thinking is a feed from input to output via the Q control works in the opposite direction. Z2 pin 13 is the (-) inverting amp input. Operational amplifiers with a feedback path (R3) and resistors feeding the (-) input will maintain the voltage at the (-) pin at null. By Null, I mean something we call "virtual ground" or "virtual earth". The amp will drive it's output so that  the feedback resistor path balances the input paths, equal and opposite so the null in between them persists whatever the signal is doing. So pin 13 can only be an input to the amp, it cannot send signal out even though there are other paths connected there.
There are 2 paths into Z2 pin 13 (-) input. The input jack and the Q control. Whenever you see something like this, you have a "virtual earth inverting mixer". However many input paths there are, the amp output will be a mix of the signals from each path and each input does not affect the others because where they meet at the (-) pin is "virtual ground".

The situation isn't quite the same with the (+) input pin, so mixing is done using the inverting input. The cost is the output signal polarity is reversed, so another inverting stage should appear somewhere to correct that. And here is does with the output wet/dry mixer...

The same thing again. The wet/dry mix is set by fixed resistors in each input path, R45 & 48. Z2 (-) pin9 has these 2 paths going in and a feedback path from R49. Another inverting mixer.

[anotherjim]

Tried increasing R34. Largest pot I had was 2.2M so put that in series with the 2.2M R34. Increasing the pot to max I still had plenty of effect without the impression that it was over-driving the phaser sweep.

However, it changes the DC level on the output Z3 pin7 (expected), so a readjustment of the bias pot was required. Also, the variation in amplitude of the LFO signal with frequency means that any change here must be checked at all rates - and no doubt end up having to be a compromise. So, I have R34 effectively totalling 4.4M. I have a feeling it could go higher yet, and of course, Larry found 6.8M in some units.

Other observations...
The pedal sounded very distorted today, but I traced it to RF interference in my workshed (especially from my soldering station). Nick used an unscreened plastic shell, but even then, the circuit has zero RF protection at the input. I think that a 47pF cap across R3 might do some good here, without affecting tone.

Incidentally Nick, if you have to use a plastic shell for MkII, you could screen it with conductive paint inside and make contact to the circuit ground with a tag washer screwed on + one more trapped between the base plate and shell.

Volume drop while engaged... The Q control can boost the volume, so without the pedal having an output volume control, it would be difficult to pick a fixed change to the design. As it stands, it must be a compromise.

[nickbungus]

All sounds really interesting.  Don't change it back, I want to hear it.

I'm keeping my eyes open for an aluminium wah shell, but if I don't get one, I'm happy to discard the treadle and just put it into a normal aluminium enclosure.

Feel free to try the 47pf cap across R3 if you want.  I'll definitely give it a shot if not.

I'll try the conductive paint on this version, nothing to lose and all to gain.

I'm thinking Mk II should definitely have an LPB-1 for volume to combat any drop

[nickbungus]

It seems to me putting a pot in replacement for R34 (like you mentioned earlier Jim) for a Range control is a worthy mod too. Especially as it can be up to 6.8M in some units.