I need help with tweaking low pass filtering on a chorus

[tokenofdevotion]

So I'm currently trying to figure out how can I reduce bbd hiss on a Guyatone PS-017 Distortion & Chorus.
The pedal is perfect as is. It's straight up 80's, but I wanted to double the delay time just to experiment with it and learn more. There is no schematic available and these are literally the only internal pictures of the PCB you'll find on the internet as of right now.




It uses the MN3207 and MN3102.








I thought I could just swap out the 47pF capacitor next to the 3207  for a higher value like 100pF. I immediately noticed and read about that the increased noise coming from the BBD's is a result of the mismatch between the clock rate and the high pass filter on the delay path: I confirmed this by reducing the stock cap value to 20 and the swirly hiss was greatly reduced. The chorus is noisy even with the stock cap value. So I was wondering what can I do to keep the higher delay time while getting rid of the higher noise besides just sticking to lower values. I read somewhere some units use bias trimpots to balance everything out. However, very cheap pedals like the Aria ACH-1 I got don't suffer from loud hiss. I just want to understand how can I solve this, even if it means sacrificing high frequency for a quick fix

[Mark Hammer]

I only see one trimpot, so I assume it is for biasing the BBD, and not for balancing the two outputs.  You are correct that, while it is recommended to do so, many chorus pedals do NOT use a balance trimmer on the dual outputs of the BBD, opting for simply two value-matched fixed resistors.  But "better" BBD-based pedals DO include a balance trimmer to cancel clock noise more effectively.

In many respects, you would have fewer issues if you were able to find an MN3208 (2048 stages), and whip up a little retrofit board that could plug into the space currently occupied by the MN3207.  This would double the delay time without changing anything about the clock or filtering.  Of course, the MN3208 is not pin-for-pin replaceable with the MN3207, the way that, say, an MN3209 or MN3204 could.  It has all the same 8 pins.  They're just not in the same place as an MN3207, so that requires making a board that reroutes everything appropriately.  The good news is that, with the MN3102 right beside the BBD, there is room for such a daughterboard to fit in that space without interfering with anything.  The other good news is that several companies are still making 3208 replacements, so they can be gotten inexpensively (well, relative to some BBDs)

[tokenofdevotion]

Hey Mark, thanks for the idea, I really appreciate it. What if I were to alter the filter? What exactly is it that makes it annoying or undesirable to mod in a pedal like this? Aside, from no schematic being available of course. I'm just curious how to modify it without needing to buy another IC so maybe more people could try to do it if they don't have a spare MN3208

[Rob Strand]

Here's the parts you need.  Please check it.
No idea what the original values are.
Even the 10k resistors could be different.











FWIW,  if you lower the filters too much it will affect the sound.   You probably don't want the filters cut-off much below 3kHz.  I think the Boss CE-2 values are actually close to the mark.   

[tokenofdevotion]

Hey Rob, thank you so much for the markups, I guess I just need to swap out the components for the ones on the CH-1 schematic? Sorry for the dumb question, but it's the first time I modify a filter on a chorus

[Rob Strand]

Hey Rob, thank you so much for the markups, I guess I just need to swap out the components for the ones on the CH-1 schematic? Sorry for the dumb question, but it's the first time I modify a filter on a chorus

If you want to lower the filter frequency, one approach is:
- markup the schematic I posted with the correct values according to the PCB.
- increase the six filter capacitor values by some factor,
  that will decrease the filter frequency by the same factor
- replace the six filter caps with the new values

You have to decide how low you want to make the filter.  It's a bit of an experiment at the end of the day.
You could approach the problem a little differently and simulate the filter with the old and new values and see if the new values are OK.

I found this schematic for the Guyatone PS-013 Chorus:
https://www.freestompboxes.org/viewtopic.php?t=16517

Interestingly the values are the same as the Boss CE-2 Chorus.   IMHO you don't want to drop the filter frequency too much below those values because the chorus will get very muddy.

You definitely don't want to start with the CH-1 values as the filter cut-off is at a very high frequency and are nowhere near what you intend to do.

The unit could have an issue, or ,changing the VCO cap has done something unexpected.   You could confirm the clock frequency before and after the VCO cap change to see if it is in fact half the clock frequency.

[tokenofdevotion]

I just checked the resistors on the PS-017 and they're all 10k. Cap values are as follows:

C1 3.3nF
C2 8.2nF
C3 0.47 nF

C4 3.3nF
C5 8.2nF
C6 0.47 nF

All same values from the PS-013 so the filter is exactly the same as in the PS-017. Very interesting. Now I don't have a clue what's causing the hiss, which only shows up when the chorus section is active, and is decreased when I lower the clock capacitor value

[Rob Strand]

I just checked the resistors on the PS-017 and they're all 10k. Cap values are as follows:

C1 3.3nF
C2 8.2nF
C3 0.47 nF

C4 3.3nF
C5 8.2nF
C6 0.47 nF

All same values from the PS-013 so the filter is exactly the same as in the PS-017. Very interesting. Now I don't have a clue what's causing the hiss, which only shows up when the chorus section is active, and is decreased when I lower the clock capacitor value

I had a feeling it might be like that.

I'm not familiar with that unit.  I was wondering if the gain of the distortion part of the circuit is boosting the noise level of the chorus.    Another thing that can happen is the noise from the chorus is getting into the distortion or other parts of the circuit.  When the clock frequency is high you don't notice it but when you drop the VCO frequency the noise starts getting into the audio path.

If there is a distortion only mode it would interesting see if changing the chorus VCO cap makes the distortion noisier.   That would point to a possible issue with the power supply or Vref (4.5V) circuit.   Dry electrolytic caps could cause such a problem.

If there is a chorus only mode and only that gets noisier then that makes it more difficult to determine if it's normal BBD noise or not.   You always expect some increase in noise when lowering the clock frequency.

Something which can occur on these pedals is the input or output buffer transistors get damaged and go noisy.   If the distortion only mode isn't noisy then the input transistor is probably OK.

If you can measure the clock frequency it might help.   If the delay is going out to 20ms then it's pretty normal to get a lot of noise.

[tokenofdevotion]

If there is a distortion only mode it would interesting see if changing the chorus VCO cap makes the distortion noisier.   That would point to a possible issue with the power supply or Vref (4.5V) circuit.   Dry electrolytic caps could cause such a problem.

If there is a chorus only mode and only that gets noisier then that makes it more difficult to determine if it's normal BBD noise or not.   You always expect some increase in noise when lowering the clock frequency.

So after some addititional testing I double-checked and the noise only comes up whenever the pedal is set as chorus only and distortion + chorus. I changed the clock cap from the stock value 47pF to 200pF while the unit is set to distortion only and no noise comes up. Switched back to chorus + distortion and noise is still there (increased by the 200pF cap).

I'm starting to think this might just be a design fault

[ElectricDruid]

I'm starting to think this might just be a design fault

No, noise in a BBD-based pedal is definitely a design *feature*, not a fault!!

It's pretty typical, unfortunately. There's a reason everything delay-based went digital as soon as DACs and RAMs were cheap enough.

[tokenofdevotion]

Of course, there's always some residual noise, but I was just trying to understand why in this specific unit, compared to my other chorus pedals it almost has 4 or 5 times the amount of noise you usually have on a chorus pedal. The delay time on the PS-017 is very short and sounds flangy. I'd say it's probably 5-6ms. While on the Aria ACH-1 I got the delay time is longer around 10ms and there is no insane amount of noise, while using a MN3209

[ElectricDruid]

Ok, I see what you mean.

I suppose the next question is whether the problem is specific to *this particular unit* (which might suggest some component failure) or whether it's common to all pedals of this type (in which case it could genuinely be a design fault).

What power supply are you using with the pedal?  There have been more than enough "My pedal is noisy!" threads around here where the eventual answer is "My pedal is fine, but my power supply is rubbish"!
In a similar vein, I'd certainly check electrolytics across the board. Often pedals use cheap electrolytics very close to their maximum voltage rating (I've seen 10V caps on a 9V supply enough times...) which is asking for trouble in the long term...and a pedal from the 80's is defnitely "in the long term" now!
The most important ones would be power supply filtering, and bias supply filtering. There may well be more than one bias supply, separated for op-amps and for the BBD.

HTH.

[tokenofdevotion]

I only see 4 electrolytic capacitors rated barely 10v and only one is within the chorus section so I'll try to recap all of those and see if I can get some improvement.

By the way, is there any additional step I'm missing when swapping BBD's? I just popped in a MN3208 from an old yamaha ch 100 I had laying around just for testing out the double delay time without changing anything. I made a daughterboard and once I soldered it to the main board I noticed there's no chorus on any of the 3 pedal modes. I tried tweaking the only bias trimpot and no changes, also tried changing the clock cap value with no results. I desoldered the daughterboard, made another one to double-check nothing was wrong and still no chorus. Pretty weird or did I just goofed?

Edit: I've tested 2-3 power supplies so I'm sure that's not the issue, I swapped the guyatone for the 2 other choruses I got and no noise at all

[Rob Strand]

So after some addititional testing I double-checked and the noise only comes up whenever the pedal is set as chorus only and distortion + chorus. I changed the clock cap from the stock value 47pF to 200pF while the unit is set to distortion only and no noise comes up. Switched back to chorus + distortion and noise is still there (increased by the 200pF cap).

I'm starting to think this might just be a design fault

I agree with ElectricDruid about the power and yes it's probably a good idea to replace the Electros.   You could even tack some electros on the back of the board to see the effect.

Without a schematic, or at least tracing out part of the circuit, it's difficult to know what the electro's are doing and which ones are related to power.

The Boss pedals have a regulator feeding the BBD and you can see on the BF2B that they added an RC filter on the power rail to the BBD (R65, C41 located in the middle of the schematic).   The BF2/BF2B use MN320x BBD series like the Guyatone unit  (as opposed to the MN300x series used on the Boss CE-2 CH-1).

https://www.synfo.nl/servicemanuals/Boss/BOSS_BF-2_SERVICE_NOTES.pdf

At this point I'm not suggesting adding the RC filter as we don't know how the power is done on the Guyatone.

[PRR]

There's a reason everything delay-based went digital....

But one lost corner of delay-land is still with us:
Copper pipe made into a studio delay that runs at the speed of sound
https://www.youtube.com/watch?v=cBZKziM0vlI

[tokenofdevotion]

I was gonna trace the pcb for posterity anyway, it's actually the main reason I bought it for.

It looks like for now I'll be following Mark's suggestion at the beginning of the thread. So I swapped the MN3207 for a MN3208 and there's no sound coming from the delay path in any of the modes. I double-checked the connections and everything's re-routed properly. I even went back to the MN3207 for a moment and got the chorus working again. Am I missing something?

[tokenofdevotion]

bump?

[ElectricDruid]

So I swapped the MN3207 for a MN3208 and there's no sound coming from the delay path in any of the modes. I double-checked the connections and everything's re-routed properly. I even went back to the MN3207 for a moment and got the chorus working again. Am I missing something?

The MN3208 doesn't fit in the same 8-pin footprint as the MN3207, so if you actually used an MN3208 and did it with an adpator like Mark suggested, there must be a problem with your adaptor. There's no other reason why it shouldn't work. (The '3208 follows the same footprint as the 4096-stage '3205, so it's not *totally* illogical..)
There are modern clones like the V3208 that are in the same 8-pin package as the MN3207 and *would* be a direct drop-in replacement.

[tokenofdevotion]

So I traced down the main board and this is what I got. I have no clue of how to draw a proper schematic so had to just get all the values and that's pretty much all I can do. I did noticed when increasing the level pot I could get more headroom and lower the noise floor that way. But, there is still this odd swirling hiss that does not disappear unless only distortion is active. Any clues on how to improve this while increasing the delay time?

[Rob Strand]

So I traced down the main board and this is what I got. I have no clue of how to draw a proper schematic so had to just get all the values and that's pretty much all I can do. I did noticed when increasing the level pot I could get more headroom and lower the noise floor that way. But, there is still this odd swirling hiss that does not disappear unless only distortion is active. Any clues on how to improve this while increasing the delay time?

The hiss seems to be removed in distortion mode because of the extra filtering in the distortion circuit.  (1K, 47n, 6k8, 10n near the TL4558P IC)

I've drawn a PSU schematic and marked up your overlay.  The PSU parts for the BBD are located near the center of the PCB.





The distortion uses output clipping diodes followed by a two stage low-pass filter.

On the whole the Chorus circuit is very much like a Boss CE-2 with a few minor values part changes.  Obviously the BBD and BBD clock chip are different.  Due to the NMOS/PMOS differences the BBD output resistor goes to ground not +V.

The PSU circuit doesn't look too bad.  One criticism would be the Clock + LFO + VCO run off the same power rail as the BBD.  That could inject noise although some Boss circuits also do this.  Raising C4 to 100uF might help a little bit.

Yet unconfirmed: I haven't traced out the whole circuit but I have a suspicion where the hiss problem is:  There is no pre-emphasis and de-emphasis circuits (as used on Boss CE-2, CH-1, BF-2 etc).  That will cause hiss.    I think the mixing of the direct and delayed signals are mixed very simply to simplify the circuit.

If that's the case then the circuit details in and out of the chorus need to be traced.    Maybe pre-emphasis and de-emphasis can be shoe-horned in.  However,  a different tact is to modify the BBD filters I traced earlier.  The aim is to build in some pre-emphasis into the input filter and some de-emphasis into the output filter.  Such a redesign needs some care and the filter design goes outside of the textbook type filter designs.  It might be enough to save it.