Phase 45 (w/ Univibe mods) Troubleshoot

[DJPsychic]

Thank you for the rundown!

[DJPsychic]

Concerning  "depth"

If I replace the singular "R9" with a 10k pot, can I get more "depth" then the stock, or does this only act to reduce the wet signal?

Versus

JC Maillet "Mix", replacing both 10k's (R9 & R10)

[Rob Strand]

Concerning  "depth"

If I replace the singular "R9" with a 10k pot, can I get more "depth" then the stock, or does this only act to reduce the wet signal?

Versus

JC Maillet "Mix", replacing both 10k's (R9 & R10)

There's a difference between adjustable depth and more depth.

For more depth try reducing the 3M9 resistor to 3M3.   That increases the sweep depth.  Even when everything is working fine this value can need some tweaking as technically the right value depends on the JFETs.   You might need to tweak the bias again.

The phase can lack depth if the bias is not set correctly, or, something is wrong.

Keep in mind the  Phase 45 only has two all-pass filters so you can't expect too much.  Listen to some youtube videos to get an idea where your unit stands.   You can also compare against the Phase 90 to get an idea of the effect of adding more all-pass stages.   (and keep in mind units vary and the later Phase 90s don't sound like the vintage ones so perhaps find a video with a vintage unit.)

[Rob Strand]

FWIW,

Unfortunately I can't remember what the correct zener is for the Phase 45 and I don't have all my notes and original PCB pics.

The 1N5230 shown on the Phase 45 layout is a 4.7v zener which ends up at 3.6V (to 4.4V) or so due to the low operating current.

The Phase 90 uses a 1N5231 which is a 5.1V zener that ends up around 4.4V to 4.8V.    A few schematics incorrectly had 1N5230.

The zener will affect the sound because it changes the way the sweep works.

The web isn't a good place to resolve this issue you really need original pics.   There might be a post on the forum about it but most of the traffic is about the Phase 90.

EDIT:
What zener did you actually use?  500mW?, 1W?  part number?
Perhaps even 3.6V is low, maybe that's for a 1W.

[DJPsychic]

Thanks man. I realized  after I asked the question how obvious the answer was lol

I ended up testing  the "R9" 10k pot, and it might be a useful to clean up some of the tone but not really doing what I hoped. .

Would putting a 5m pot in place of the 3M9 be overkill?

- And one last issue I need to address is I'm barely getting much "headroom" before I get breakup. My signal wants to distort pretty quickly. It's usable if I'm playing quieter but muddy if hit a chord or play heavier.

Maybe try a different IC?

[Rob Strand]

I ended up testing  the "R9" 10k pot, and it might be a useful to clean up some of the tone but not really doing what I hoped. .

Would putting a 5m pot in place of the 3M9 be overkill?

It would work but a 5M pot is hard to get.

Another way is to put a 500k or 1M pot in series with the 1M resistor that goes to the bias trimpot wiper.  1M is easier to get.

The bias will probably need to be adjusted when you change that pot setting.

There's about 4 ways to mod the unit to add a depth trimpot.   One of those ways will require less re-adjustment of the bias when the depth is changed.

- And one last issue I need to address is I'm barely getting much "headroom" before I get breakup. My signal wants to distort pretty quickly. It's usable if I'm playing quieter but muddy if hit a chord or play heavier.

Maybe try a different IC?

The problem is the low (actual) zener voltage.    Ideally you want the zener around 4.5V measured across the zener.

Typically that works out when you use a 5.1V 500mW zener.

If you use a 1W zener the actual zener voltage could be lower than 4.5V.  If you can only get 1W zeners then it might be a good idea to get a 5.6V zener (as well as a 5.1V) then it will end up at 4.5V.     This stuff is quite imprecise.

And who knows, perhaps after fixing the zener it could sound better.


At the moment you might have a 4.7V 1W and it's ending up way down at 3.3V to 3.6V.

[DJPsychic]

I ended up testing  the "R9" 10k pot, and it might be a useful to clean up some of the tone but not really doing what I hoped. .

Would putting a 5m pot in place of the 3M9 be overkill?

It would work but a 5M pot is hard to get.

Another way is to put a 500k or 1M pot in series with the 1M resistor that goes to the bias trimpot wiper.  1M is easier to get.

The bias will probably need to be adjusted when you change that pot setting.

There's about 4 ways to mod the unit to add a depth trimpot.   One of those ways will require less re-adjustment of the bias when the depth is changed.

- And one last issue I need to address is I'm barely getting much "headroom" before I get breakup. My signal wants to distort pretty quickly. It's usable if I'm playing quieter but muddy if hit a chord or play heavier.

Maybe try a different IC?

The problem is the low (actual) zener voltage.    Ideally you want the zener around 4.5V measured across the zener.

Typically that works out when you use a 5.1V 500mW zener.

If you use a 1W zener the actual zener voltage could be lower than 4.5V.  If you can only get 1W zeners then it might be a good idea to get a 5.6V zener (as well as a 5.1V) then it will end up at 4.5V.     This stuff is quite imprecise.

And who knows, perhaps after fixing the zener it could sound better.


At the moment you might have a 4.7V 1W and it's ending up way down at 3.3V to 3.6V.

Thank you for the thorough breakdown!

I think I have a 5.1V zener on hand I'll give that a shot.

Ideally I'm  going for a pot that acts as close to the "intensity" (depth) pot of a Uni-Vibe type clone.

[Rob Strand]

Ideally I'm  going for a pot that acts as close to the "intensity" (depth) pot of a Uni-Vibe type clone.

It's do able but the best way to do it would need some design time and also analysis of exactly what happens with the univibe.

The intensity on the univibe goes from full down to nothing.    When intensity is on minimum the unit isn't "bypassed" as such, there is a notch.   The notch is located at a particular frequency "somewhere" in the middle of the normal sweep range.

Upfront, the tweaks to the depth on the phase 45 so far aren't going mimic the univibe when the depth needs to work over a large range to nothing.   Over a small range it's OK, mainly for tweaking but not for a user settable control.

The depth control "mod" as it is now, when you back depth off the moves the notch to a low frequency.  It doesn't move the notch to "somewhere" in the middle like the univibe.     By restricting the range of the depth control to a small range for tweaking that effect doesn't get out of hand.  However having the depth control go to zero depth isn't going to work well at all and it won't be like the univibe.

The whole design effort for a univibe depth control would focus on getting the notches to be located at the correct frequencies over the range.

There are some depth mods for the Phase 90.  They basically work over a narrow range.   In fact I'm pretty sure they are just like the first mod I suggested, putting a pot in series with the 1M resistor to the bias trimpot wiper.

The biasing on these things is touchy so I'm reluctant to recommend untried "forum hacks" towards a univibe version.

[DJPsychic]

Awesome info thank you again Rob.

I had a pack of these and tried one out. Seems to "clean" the signal up a tad, but feel I lost a a little "throb"

Do you or anyone recommend a better 5V zener for this application?

[Rob Strand]

I had a pack of these and tried one out. Seems to "clean" the signal up a tad, but feel I lost a a little "throb"

You might be able to tweak the LFO duty cycle  by adding a large value resistor across the 10uF cap.

I'd say it would be possible to get what you want the problem is someone has to spend the time to do it.   Lots of playing around tweaking the circuit.   Adding stuff which isn't there now - that's probably the biggest stumbling block.

Do you or anyone recommend a better 5V zener for this application?

Any 5.1V 500mW zener will put you in the right zone.   The original units averaged out at about 4.7V to 4.8V across the zener.  You might see +/- 0.3V on that.

Some people can only get 1W zeners.    The first thing to try would be a 5.6V 1W zener.   And if the voltage is still too low you can drop the 10k resistor between the 9V rail and the zener to raise the zener voltage.   Ideally you don't want to go below about 2.2k.  If you can get 4.4V or above with 2k2 or above leave it.  If it's still lower than 4.4V go for broke and put in a 1k.

Some 5.1V 1W zeners might work with the trick of lowering the 10k.  No guarantees.

If you use a 4.7V 1W zener I doubt you will make 4.4V.     If you have already bought one of these you might  be able to add a 1N4148/1N914 silicon diode in series with the zener (pointing downward, the opposite direction to the zener).    If that still doesn't make the voltage try dropping the 10k resistor.  I suspect with a bit of stuffing around you can this to work - no problems.

A 4.7V 500mW zener might work by reducing the 10k.

A 5.6V 500mW zener will work but the voltage will be a little high.  Better off not going there unless you are desperate.

The 5.1V 5W zeners you have are too large and the voltage will be low.
Perhaps try the series diode trick with the existing 4.7V 1W.

That pretty's much the DIY'es guide to zener tweaking!

[DJPsychic]

Dude thank you for the master class!

I ordered a set of 1/2w zeners last night. I'm def going to dig deeper in this circuit and shape it the best I can. May keep it as a one-knob, but I think having the some sort of depth may be valuable. I'll update with my findings. Thank you again Rob!

[DJPsychic]

So I got a 5.1V .5W zener in there. Im reading 4.8v at the diode.

Still  not a very "clean" signal. Kind of muddy and easily distorted. Usable with a fuzz but alone not a very clean phase.

[Rob Strand]

So I got a 5.1V .5W zener in there. Im reading 4.8v at the diode.

Still  not a very "clean" signal. Kind of muddy and easily distorted. Usable with a fuzz but alone not a very clean phase.

That's a bit weird.   Would you say it's cleaner than before?   We could be dealing with multiple problems.    The first is clipping due to the Vref being off.  At 4.8V it's about as good as you will get.   The second could be distortion from the JFETs themselves.   Beyond that you could have a minor build error, or even a faulty opamp.

You do remember what VP values you got when you matched the JFETs?    Having said that a few people have build phasers with surprisingly low VP values and didn't complain about it not being clean.

When you switch between effect and bypass you notice any boost in signal in effects mode?  Any gain would point to too much gain on the first opamp.

Perhaps the next step is to listen to the clean path and the phased path separately.   The circuit has two 10k's at the output of each signal path which mix the clean and phased paths.    Lift the 10k from the phase path so you are listening only to  clean signal.   When you lift the 10k resistor the gain of the unit will double, just be aware of that.  You should get a clean signal.     If the clean signal isn't clean then stop there.  Something is wrong.   Remeasure the voltages on the output of each opamp.

Next is to restore the 10k on the phased signal and lift the 10k for the clean signal.   That will give you a vibrato effect.   The signal gain will be double normal like the previous test.    Again the signal should be clean.

Perhaps going through those tests you will find something, or at least find something that narrows down where the problem is.

[DJPsychic]

Thanks Rob for the response

I bought 2 sets of "quad matched" JFETS from "pedal parts and kits" to be used in P45/P90. I didn't check voltages before putting in the circuit. Probably should have!

As far testing the clean paths. I believe I tested this (circled) resistor using a 10k pot as sort of "mix". If I remember correctly I got a vibrato effect when pot was turned down and signal was clean.

Which is the 2nd 10k I should be testing?

[Rob Strand]

hanks Rob for the response

I bought 2 sets of "quad matched" JFETS from "pedal parts and kits" to be used in P45/P90. I didn't check voltages before putting in the circuit. Probably should have!

Not to worry.  We should able to narrow things down anyway.

As far testing the clean paths. I believe I tested this (circled) resistor using a 10k pot as sort of "mix". If I remember correctly I got a vibrato effect when pot was turned down and signal was clean.

Which is the 2nd 10k I should be testing?

That resistor was the correct one for vibrato.  The key thing is not that you got vibrato but that you got a clean sounding vibrato.   We are trying to find the source of the distorted sound.

The top lead of the 10k for the clean signal is three holes to left of the lower lead of the 10k you circled.  It has long leads.  The two 10ks connects to a track which is the whole width of the PCB and it goes to a 47n at the far left.
   

[DJPsychic]

The top lead of the 10k for the clean signal is three holes to left of the lower lead of the 10k you circled.  It has long leads.  The two 10ks connects to a track which is the whole width of the PCB and it goes to a 47n at the far left.

Just to confirm:

[Rob Strand]

That it.

[DJPsychic]

Ok so I lifted the first "clean signal" 10k and everything sounds fine. More gain but clean.

Tested the 2nd "vibrato" resistor, it is clean but there is a slight "clipping" or breakup on part of the modulation cycle. It's only if you pick a note or play a chord hard that you get the "distortion". Useable but a bit muddy at certain points in the modulation.

Having trouble defining the "distortion" sorry, maybe wrong word. It's like when a chorus pedal or vibe distorts when you use a wah. It's hitting the ceiling or threshold  of the effect. (I'm not an engineer obvi)

[Rob Strand]

Ok so I lifted the first "clean signal" 10k and everything sounds fine. More gain but clean.

Tested the 2nd "vibrato" resistor, it is clean but there is a slight "clipping" or breakup on part of the modulation cycle. It's only if you pick a note or play a chord hard that you get the "distortion". Useable but a bit muddy at certain points in the modulation.

Having trouble defining the "distortion" sorry, maybe wrong word. It's like when a chorus pedal or vibe distorts when you use a wah. It's hitting the ceiling or threshold  of the effect. (I'm not an engineer obvi)

No problem.  It's hard to convey finer points with words.

From you description, the clean path seems OK but the all-pass (phased) path is where the problem is.   On face value that would imply the FETs are the cause.   That doesn't necessarily mean the JFETs are faulty as that type of circuit does produce distortion.   The difficulty is trying to work out if you circuit has an abnormal amount for that circuit.    When the JFET VP values are low the distortion increases.

For sake of sanity it's probably a good idea to swap the positions of two opamps and redo the clean/phase check.  If the opamps are OK then the distortion should remain.  If one of the opamps is faulty the distorion will move to the clean path.

Assuming the opamps are OK.   What you can do it mis-bias the JFETs.   Lift the long leaded 10k so you are listening to phased path alone.    Set the trimpot to one extreme and see if the distortion is gone or reduced.  Set the trimpot to the other extreme and repeat.   One end might sound cleaner than the other, that should be when the trimpot is set with the wiper connecting to ground.
Don't worry about losing the phase effect.  The idea is to see if cutting off the JFETs electrically remove the distortion.  That would be more evidence the issue lies with the JFETs.

While you at it, re-check the JFETs are in the correct way.

[DJPsychic]

I'm thinking it's something with the jfets. I actually completely replaced the IC's yesterday and have the same issue.

Thanks for all the help, I'll perform the trim pot test next and report back.