Boss PH-1R --- TOO BRIGHT!

[Slowpoke101]

It looks more like a 0,5 Watt zener diode but that shouldn't be causing you problems at this moment. Anyway I have a silly question as usual. Having looked at your pictures I'm wondering about about the orientation of the electrolytic cap C2. It's difficult to see properly but it appears that it may be in backwards. I may be wrong, but that is to be expected.

[ElectricDruid]

Q5      D-5v
        S-2.9v
        G-5v

Q6      D-5v
        S-0v
        G-5v

Why is Q6 Source so different from all the others? (2.9V)

Q6 is well separated in the circuit from Q2-Q5.  I think it is performing a different function entirely than those 4 others.

Ah, yes, you're quite right. Q6 is one of the switching FETs, not one of the phaser FETs. Sorry, my bad. I should have checked.

[rundgrenrules]

It looks more like a 0,5 Watt zener diode but that shouldn't be causing you problems at this moment. Anyway I have a silly question as usual. Having looked at your pictures I'm wondering about about the orientation of the electrolytic cap C2. It's difficult to see properly but it appears that it may be in backwards. I may be wrong, but that is to be expected.

I had some bipolar 1uf caps on hand so all those replacements are bipolar!

[rundgrenrules]

I wonder if I need to just build a little JFET matcher circuit and pull these JFETs and test them.  If one was wired in backwards I wonder what else might be off about them.

[ElectricDruid]

I wonder if I need to just build a little JFET matcher circuit and pull these JFETs and test them.  If one was wired in backwards I wonder what else might be off about them.

What's the evidence that leads you to think that might be the problem?

Have you audio-probed the output of each phase stage to see if there's one particular stage where the signal gets bright? Have you lifted the dry path so you can hear the phase stages on their own?

I'd try that first to try and narrow things down, rather than desoldering a load of parts just to test them. What would it even prove? Say one was bad - you wouldn't know if that was the cause of the problems, or whether you damaged it getting it out off the board.

[duck_arse]

check the datasheet for the part at hand. does it show centre gate pinout? does it state drain and source are interchangable? if yes, there is no backwards with those parts. effectively.

[rundgrenrules]

Have you audio-probed the output of each phase stage to see if there's one particular stage where the signal gets bright? Have you lifted the dry path so you can hear the phase stages on their own?

I just made a probe and traced out the audio path from input to output and probed that from board input to board output with particular focus on the opamp inputs and outputs.  It all sounded pretty good to me with the exception of a big volume leap between IC2 Pin 2 (Output) and IC1 Pin 1 (Output).  This is where the clipping seems to come in and the harsher harmonics.  I assume there is some makeup gain thats supposed to be occuring right here, but when I probe directly back and forth between the Board In (Header 2) and Board Out (Header 5), there is just this big volume difference between the two.  Clean at the board input, phased and clipped/harsher/significantly louder at the board output. 

When I lifted R31, three things happened:
1. I could hear around the effect a little better for probing
2. There was a less drastic volume jump between the output of IC2 and the output of IC1
3. The volume balance between the board IN + OUT was much less and there was no clipping occuring at the output

So I am assuming that there is a bit of an over-accumulation occuring between the wet and dry signal.  I measured R31 and it measures correctly at 47K.  I ALSO assume that tweaking the value of R31 might not net me any less clipping or harshness and might actually just mess the the balance needed for proper phasing.

I'm a beginner, so I apologize for my lack of knowledge.  But I feel like I am learning some stuff here.  I haven't really delved into tracing out a schematic/audio path/probing/etc before.  I like having to dig deeper to figure this one out.

[Rob Strand]

When I lifted R31, three things happened:
1. I could hear around the effect a little better for probing
2. There was a less drastic volume jump between the output of IC2 and the output of IC1
3. The volume balance between the board IN + OUT was much less and there was no clipping occuring at the output

So I am assuming that there is a bit of an over-accumulation occuring between the wet and dry signal.  I measured R31 and it measures correctly at 47K.  I ALSO assume that tweaking the value of R31 might not net me any less clipping or harshness and might actually just mess the the balance needed for proper phasing.

I'm a beginner, so I apologize for my lack of knowledge.  But I feel like I am learning some stuff here.  I haven't really delved into tracing out a schematic/audio path/probing/etc before.  I like having to dig deeper to figure this one out.

I was going to suggest that the other day.  I'm in the middle of some technology upgrades so I haven't been around lately.

As a sanity check I think you should bypass Q6, like I mentioned earlier.   At this point I thinking less that Q6 is a problem.

I really think the only thing you are hearing is the volume boost when the pedal is engaged and you just happen to be hearing it in the higher frequency spectra.

There's no easy way to fix that issue since it needs another JFET to knock the gain down in effect mode.  One way to tackle the issue is the way CE3 chorus does it.   It might be possible to do it without adding a JFET with a devious mod but I haven't had a chance to look into it ; only had a hint of an idea.   The other option is to put a low-pass filter somewhere in the IC2/Q6/IC1 path.  However, your experiment with a cap across R9 is kind of indicating it's not enough and maybe you are hearing some volume boost in the upper-mids to treble, which are harder to filter without being radical - not impossible with some finely tuned shelving filter tweaked by ear.   The fact you increased R32 and it wasn't enough is also pointing to the fact simple mods are not going far enough.  So ... that only leaves the gain drop mod as an option.

[rundgrenrules]

I was going to suggest that the other day.  I'm in the middle of some technology upgrades so I haven't been around lately.

As a sanity check I think you should bypass Q6, like I mentioned earlier.   At this point I thinking less that Q6 is a problem.

I really think the only thing you are hearing is the volume boost when the pedal is engaged and you just happen to be hearing it in the higher frequency spectra.

There's no easy way to fix that issue since it needs another JFET to knock the gain down in effect mode.  One way to tackle the issue is the way CE3 chorus does it.   It might be possible to do it without adding a JFET with a devious mod but I haven't had a chance to look into it ; only had a hint of an idea.   The other option is to put a low-pass filter somewhere in the IC2/Q6/IC1 path.  However, your experiment with a cap across R9 is kind of indicating it's not enough and maybe you are hearing some volume boost in the upper-mids to treble, which are harder to filter without being radical - not impossible with some finely tuned shelving filter tweaked by ear.   The fact you increased R32 and it wasn't enough is also pointing to the fact simple mods are not going far enough.  So ... that only leaves the gain drop mod as an option.

I just went back and put some alligator leads in place of R32.  I experimented with this value until I felt like I wasnt getting the gain boost that was there with the stock 22k, and arrived at 47k.  Now, isn't this resistor post the Q6 transistor that flips the effected path on/off, and wouldn't changing this value also change the bypass volume as well?  Either way it seems easier on the ears and dealt with the clipped/pushed sound.

Once this was done I was able to revisit paralleling R9 with a cap and detected a more noticeable improvement in filtering the high-highs.  I played with that value until I arrived at 47nf.  That seemed just enough to take the harshness out of the upper range without robbing too much up there.

I think the order of how I did things before didn't allow for me to land on the combination of tweaks in it now.  I'm pretty satisfied with the overall sound but I AM curious about R32 and if/how it does/doesn't affect the bypassed volume.  I understand that it is setting the gain of IC1, coming into the Inv Input of side A from the effected path, and that the clean signal is coming into pin 3 on the non inv input of side A and thats where the two are combining.  It is my understanding that Q6 is acting to turn the wet signal of and on, so I would think R32 being downstream of this transistor would affect the bypassed volume?

If someone else has already stated as such I apologize.  The phaser does sound quite good now, very happy with it!

EDIT:  Also would like to add that the noise floor is now lower.  Definitely low enough that I dont see the need to do anything more to address it.  To summarize, the only change from stock is 47k @ R32 and a 47nf cap paralleling R9.

[Rob Strand]

I just went back and put some alligator leads in place of R32.  I experimented with this value until I felt like I wasnt getting the gain boost that was there with the stock 22k, and arrived at 47k.  Now, isn't this resistor post the Q6 transistor that flips the effected path on/off, and wouldn't changing this value also change the bypass volume as well?  Either way it seems easier on the ears and dealt with the clipped/pushed sound.

R32 only affects the effects mode.

With R32=47k the overall should be perceived about equal (as much as can be expected).

The key point to keep in mind it using R32 isn't really the "technically correct" way to change the level as it changes the phaser characteristic as well.   A slightly better compromise might be to use 39k as it gives deeper notches, ie. affects the phaser characterisic a little less, with only a small increase in level on 47k.   I'll admit there's not much in it.

If you were to set the gain using the technically correct method it would *only* affect the gain and not the effect.  The use of R32 is a compromise to get around changing the actual level since changing the level alone requires more extensive mods. 

Once this was done I was able to revisit paralleling R9 with a cap and detected a more noticeable improvement in filtering the high-highs.  I played with that value until I arrived at 47nf.  That seemed just enough to take the harshness out of the upper range without robbing too much up there.

All good.

I think the order of how I did things before didn't allow for me to land on the combination of tweaks in it now.  I'm pretty satisfied with the overall sound but

I'm pretty sure now your issue is mainly due to the slightly perceived volume increase.

I AM curious about R32 and if/how it does/doesn't affect the bypassed volume.  I understand that it is setting the gain of IC1, coming into the Inv Input of side A from the effected path, and that the clean signal is coming into pin 3 on the non inv input of side A and thats where the two are combining.  It is my understanding that Q6 is acting to turn the wet signal of and on, so I would think R32 being downstream of this transistor would affect the bypassed volume?

It definitely doesn't affect the bypass volume.

The opamp IC1 at the output is configured as an inverting mixer.   The dry signal input at R31 is completely independent of the effect signal input at R32.   What happens in this feedback connection the inverting input (pin 2, IC1) looks like it is shorted to ground and that prevents interaction between the two signal paths;  it's called a Virtual Ground.

[rundgrenrules]

I just went back and put some alligator leads in place of R32.  I experimented with this value until I felt like I wasnt getting the gain boost that was there with the stock 22k, and arrived at 47k.  Now, isn't this resistor post the Q6 transistor that flips the effected path on/off, and wouldn't changing this value also change the bypass volume as well?  Either way it seems easier on the ears and dealt with the clipped/pushed sound.

R32 only affects the effects mode.

With R32=47k the overall should be perceived about equal (as much as can be expected).

The key point to keep in mind it using R32 isn't really the "technically correct" way to change the level as it changes the phaser characteristic as well.   A slightly better compromise might be to use 39k as it gives deeper notches, ie. affects the phaser characterisic a little less, with only a small increase in level on 47k.   I'll admit there's not much in it.

If you were to set the gain using the technically correct method it would *only* affect the gain and not the effect.  The use of R32 is a compromise to get around changing the actual level since changing the level alone requires more extensive mods. 

Once this was done I was able to revisit paralleling R9 with a cap and detected a more noticeable improvement in filtering the high-highs.  I played with that value until I arrived at 47nf.  That seemed just enough to take the harshness out of the upper range without robbing too much up there.

All good.

I think the order of how I did things before didn't allow for me to land on the combination of tweaks in it now.  I'm pretty satisfied with the overall sound but

I'm pretty sure now your issue is mainly due to the slightly perceived volume increase.

I AM curious about R32 and if/how it does/doesn't affect the bypassed volume.  I understand that it is setting the gain of IC1, coming into the Inv Input of side A from the effected path, and that the clean signal is coming into pin 3 on the non inv input of side A and thats where the two are combining.  It is my understanding that Q6 is acting to turn the wet signal of and on, so I would think R32 being downstream of this transistor would affect the bypassed volume?

It definitely doesn't affect the bypass volume.

The opamp IC1 at the output is configured as an inverting mixer.   The dry signal input at R31 is completely independent of the effect signal input at R32.   What happens in this feedback connection the inverting input (pin 2, IC1) looks like it is shorted to ground and that prevents interaction between the two signal paths;  it's called a Virtual Ground.

Thanks for all the help!  I might try out 39k again. 

There's no easy way to fix that issue since it needs another JFET to knock the gain down in effect mode.  One way to tackle the issue is the way CE3 chorus does it.

Would that happen to be the job of Q4/Q5 on the CE3?

[Rob Strand]

Would that happen to be the job of Q4/Q5 on the CE3?

Q4/Q5 on the CE3 only switch in the effects mode, not unlike Q6 on the PH-1R.

The thing that tweaks the gain on the CE3 is Q2 and R18.  For the PH-1R you would place that type of circuit at R5 or R33.

FYI, on the web will see people taking about R18 on the CE-3.   You will also see people complaining about the volume *drop*.   It's all to do with what you expect and not resolvable across all people.   If you compare a CE-3 against a CE-2 the CE-3 sounds lower in level.  However, if you compare the dry and effects levels the CE-3 does seem more balanced, perhaps a tad low at times.    It's very hard to get an exact value.  I'd suggest going a tad higher in level than the CE-3.

[rundgrenrules]

Would that happen to be the job of Q4/Q5 on the CE3?

Q4/Q5 on the CE3 only switch in the effects mode, not unlike Q6 on the PH-1R.

The thing that tweaks the gain on the CE3 is Q2 and R18.  For the PH-1R you would place that type of circuit at R5 or R33.

FYI, on the web will see people taking about R18 on the CE-3.   You will also see people complaining about the volume *drop*.   It's all to do with what you expect and not resolvable across all people.   If you compare a CE-3 against a CE-2 the CE-3 sounds lower in level.  However, if you compare the dry and effects levels the CE-3 does seem more balanced, perhaps a tad low at times.    It's very hard to get an exact value.  I'd suggest going a tad higher in level than the CE-3.

If I were to modify the pedal to incorporate a small relay true bypass board and sidestep the Boss bypass, would that allow the gain to be more easily brought down by a more invasive/correct method (than R32 value tweaks)?

[Rob Strand]

If I were to modify the pedal to incorporate a small relay true bypass board and sidestep the Boss bypass, would that allow the gain to be more easily brought down by a more invasive/correct method (than R32 value tweaks)

That will work but you need to be careful not to load the Flip-Flop with the relay driver circuit.

If you added a FET switch it would require a JFET, a resistor and a diode.  The amount of modding would be comparable to a relay.  Except no clicking and less power.

I had a look at one of the devious mods I mentioned earlier.    I thought I had a nice clean mod but I overlooked the fact the RES control is still connected to the first opamp in bypass mode and that screwed everything up.   Need to ponder some more.

[BJF]

Hi there,

Very much out on a limb here

Fun little phaser and it seems to be working just fine. As for bright when engaging try putting a resistance between output and output jack in series with signal and this can be a trimmer at 50K. What this resistance will do is to form a lowpass filter with the capacitance in the output cable.
If you measure the capacitance of your cable you can calculate what this does
It will effect both bypass and effect but as you describe the increase in brightness is only in effect mode which suggests this occurs high up  and so if you get corner frequency of the filter to just glance out the added brightness chances are it won't affect dry signal at any negative degree.

Another way is to increase C16 while you might want to define at what frequency which you can by doing the trimmer experiment above.

If you need more of a staircase filter C16 in series with a resistor can do that


Bjorn Juhl
BJFElectronics
Sweden

[Mark Hammer]

Nice to see your here, Bjorn! 

[BJF]

Nice to see your here, Bjorn! 

Hi there Mike

Thank you

Bjorn Juhl
BJF Electronics
Sweden

[rundgrenrules]

Hi there,

Very much out on a limb here

Fun little phaser and it seems to be working just fine. As for bright when engaging try putting a resistance between output and output jack in series with signal and this can be a trimmer at 50K. What this resistance will do is to form a lowpass filter with the capacitance in the output cable.
If you measure the capacitance of your cable you can calculate what this does
It will effect both bypass and effect but as you describe the increase in brightness is only in effect mode which suggests this occurs high up  and so if you get corner frequency of the filter to just glance out the added brightness chances are it won't affect dry signal at any negative degree.

Another way is to increase C16 while you might want to define at what frequency which you can by doing the trimmer experiment above.

If you need more of a staircase filter C16 in series with a resistor can do that


Bjorn Juhl
BJFElectronics
Sweden

Thanks Bjorn!  I think the brightness I originally though was the issue was indeed just a perceived increase in brightness due to the volume boost and saturation as others have said.  If there was a simple way to reduce the "ON" effected volume of the pedal while retaining the wet/dry channels mix that would be ideal.

I assume there is no way to simple add a simple passive trimpot to ground at the end of the circuit, say between C17/Pad 5 & ground?  This would potentially mess with the impedance and thats where the active vaolume reduction circuit with the JFET will be needed?

[BJF]

Hi there,

Yes that can certainly be. If you look at the circuit you can see input to the allpassfilters is attenuated roughly -6dB ( R8 and R9) and then at output mixer the output is is amplified roughly 6dB  ( relation R33 to R32). This requires precision resistors to make difference between attenuation and amplification exactly none so in practice you choose values that give a slight boost.
Also because within limits between two sounds compared and one being louder or brighter humans percieve that as better.
Also Q1 has a voltage gain that is less than 1....

Also remember when two signals are added in phase perceived volume will rise.

To experiment what increased output impedance would do there are two fairly simple ways

The added resistance shall go in series with signal

A. There is a wire from pcb to output jack. This wire can be desoldered and a trimmer wired as restate can be inserted in series with output
This must be done with the pedal open and purpose would be to define what value should R34 have with the cables used to make perceived no change

B. R34 is changed from a fixed resistor to a trimmer of say 25K

As an example let us assume a typical guitar cable and it has a capacitance of 125pF/ meter and a length of 5 meters and it can be viewed as a capacitive load of 750pF.

Output impedance is defined as the impedance looking backwards into the circuits output and here output impedance becomes roughly the
470 Ohms (R34)

To compute function of the filter created also termination impedance must be known and assuming this is another effects pedal or amplifier input we can estimate a load between 300K and 1M and ignore the influence because deviation will be less than 1%..

If output impedance would by brute force be made higher such as an inserted resistance of 30K signal loss would be about 1% which would be industrially acceptable as insignificant loss

However taking the two extremes corner frequency of 470 Ohms and 750pF yield 451KHz and corner frequency of 30K and 750pF would yield 7KHz.

Bjorn Juhl
BJF electronics
Sweden

[rundgrenrules]

I went ahead and brought R32 BACK to stock (22k) for the best phasing sound.  I made a quick little contraption with an input jack and output jack soldered directly to a ground bus bar along a 100k pot between them and connected to the tips on the respective jacks.  I put this after the phaser in a loop of a single true bypass looper.

Using this simple passive volume control circuit I was able to dial down the perceived volume (pedal engaged at all times) and blip back and forth with the true bypass looper.  I was able to get exactly the same perceived volume with the bypass loop engaged and disengaged by trimming that 100k pot to ground a little bit.

Could I not get what I am after in terms of reoving the volume boost by bypassing the built-in flip-flop using a small relay board like this one...

https://griffineffects.com/mods-kits/griffin-effects/switching/parasite-true-bypass-module

...and then adding a 100k trimpot to ground between the effect output (pad 5 of Boss PCB) and relay board return, tweaking to taste for perceived volume bypass and engaged?