Boss FA1-ish becomes gated fuzz at max treble only?

[drdn0]

Built a FA1-ish off of this schematic (less gain, used values I had on hand, got rid of the subs switch, etc).


Works amazing and sounds perfect UNLESS I turn treble to max (even 99% of pot rotation is fine) at over about 75% volume, where it becomes a huge RF antenna and the output gets crazy modulated. At the switch over point it almost sounds like the op-amp is rebiasing, the signal gets super gated and seems to phase in and out. If I roll the treble back all the way on my guitar, it becomes a pretty damn nice sounding fuzz - but obviously it's not what this is meant to be.

Only that combination (100% treble, 75% volume and up) causes issues. If I keep volume lower, everything works perfectly with full bass/treble pot rotation sounding great. Even at max volume and up to 99% treble it's sounding good. Happens on different guitars, different pickup combinations, etc.

All components are correct, all soldering is good, there's filtering/etc off board (but not in the schematic). Decoupling caps are right next to the op-amp pins.

Thoughts? I don't really understand enough about James/Baxandall tone stacks to even know where to start.

[Rob Strand]

Try putting a 100pF (47pF to 220pF) cap across pin 7 and 6 of the opamp.
You can refine the value.

[bluelagoon]

Nice try, but No Cigar!

You know main reason you are getting the oscillation out of control at full treble is because you have incorrect capacitor values hanging off your treble control, By increasing your treble threshold higher with lower than original Boss FA-1 design component values your treble pot goes into oscillation, as the original design was not meant for that much treble output. The correct values at C3 and C6 on your schematic should be 5n6 not 4n7
Also your Bass pot caps ought to be 33nF not 10nF

Another thing mismatched from the original Boss design is the input impedance you have established with both those 1Meg resistors hanging off your input, effectively only gives you 500k input impedance, the Boss design has a much larger input impedance of 3.3 Meg ohm.

If you want the same as the boss unit, then ditch your first 1M resistor to Ground, and up the value of your second 1M resistor to 3.3 Meg then it will be correct to Boss Spec.

You are also missing a 10uF cap from the circuit, right next to R6.

Aside from that you can still sometimes with the original Boss circuit get the oscillations at the high end of the treble control, the easiest way to resolve that issue is to

Cut the connection of Treble Pot 3rd leg that goes on to C3 4n7 your circuit, then send that treble pot 3rd leg to the 3rd leg of a 5k trim pot send middle leg of the 5k trim pot to C3 4n7 where it was connected previously to treble 3rd leg, it usually takes only a slight reduction in the treble pot of only about 1k to allow for max treble pot use without oscillation, but since you are using different caps, the reduction would likely be a higher resistance reduction, you might even want to put in a 10 k pot, and adjust down to where the oscillations are no longer apparent.
Alternatively you could just stick a small value resistor in there for the reduction, something 1 to 5k should be enough.

Another way to also help tame down the oscillation at high treble is to decouple your op amps with a 100nF capacitor, ceramic will do, placed as close as possible to your op amp Positive Voltage leg input, Pin 8. connect the cap between pin 8 positive and Ground, place cap as close to op amp as possible next to the positive pin 8. This also helps.

See amended circuit.

[drdn0]

Nice try, but No Cigar!

You know main reason you are getting the oscillation out of control at full treble is because you have incorrect capacitor values hanging off your treble control, By increasing your treble threshold higher with lower than original Boss FA-1 design component values your treble pot goes into oscillation, as the original design was not meant for that much treble output. The correct values at C3 and C6 on your schematic should be 5n6 not 4n7
Also your Bass pot caps ought to be 33nF not 10nF

Another thing mismatched from the original Boss design is the input impedance you have established with both those 1Meg resistors hanging off your input, effectively only gives you 500k input impedance, the Boss design has a much larger input impedance of 3.3 Meg ohm.

If you want the same as the boss unit, then ditch your first 1M resistor to Ground, and up the value of your second 1M resistor to 3.3 Meg then it will be correct to Boss Spec.

You are also missing a 10uF cap from the circuit, right next to R6.

Aside from that you can still sometimes with the original Boss circuit get the oscillations at the high end of the treble control, the easiest way to resolve that issue is to

Cut the connection of Treble Pot 3rd leg that goes on to C3 4n7 your circuit, then send that treble pot 3rd leg to the 3rd leg of a 5k trim pot send middle leg of the 5k trim pot to C3 4n7 where it was connected previously to treble 3rd leg, it usually takes only a slight reduction in the treble pot of only about 1k to allow for max treble pot use without oscillation, but since you are using different caps, the reduction would likely be a higher resistance reduction, you might even want to put in a 10 k pot, and adjust down to where the oscillations are no longer apparent.
Alternatively you could just stick a small value resistor in there for the reduction, something 1 to 5k should be enough.

Another way to also help tame down the oscillation at high treble is to decouple your op amps with a 100nF capacitor, ceramic will do, placed as close as possible to your op amp Positive Voltage leg input, Pin 8. connect the cap between pin 8 positive and Ground, place cap as close to op amp as possible next to the positive pin 8. This also helps.

See amended circuit.

Bass pot caps are 33n, I made a mistake on the schematic originally as I copied components over, but they're right on the board. I didn't have 5n6 caps on hand, but I can order some and see if it makes a difference. I didn't think the circuit would be that close to oscillation from factory.

Are the 10uf caps even necessary? There's no real need to block DC there from what I can see, and there is no scratchiness over the volume pot.

There is plenty of decoupling on the offboard switching PCB, it's within 1" of the IC. I have already tried adding it in on the back of the board and it made no difference whatsoever.

[Rob Strand]

Try the cap I suggested.

(If that's not enough try a small cap across the same terminals of the other op amp.  Much smaller though, maybe 10pf to 22pF.  Only increase if problem still exists.)

Cut the connection of Treble Pot 3rd leg that goes on to C3 4n7 your circuit, then send that treble pot 3rd leg to the 3rd leg of a 5k trim pot send middle leg of the 5k trim pot to C3 4n7 where it was connected previously to treble 3rd leg, it usually takes only a slight reduction in the treble pot of only about 1k to allow for max treble pot use without oscillation, but since you are using different caps, the reduction would likely be a higher resistance reduction, you might even want to put in a 10 k pot, and adjust down to where the oscillations are no longer apparent.
Alternatively you could just stick a small value resistor in there for the reduction, something 1 to 5k should be enough.

That added resistor isn't much different to increasing the existing 10k (R9).   Also it changes the response of the circuit.

You should be able to put a very wide range of cap values in the treble circuit.   That's how you set the treble frequency.    Some values may be more susceptible to oscillation than other values but it's not the root cause.

[bluelagoon]

Hi Rob, haven't tried upping the R9 10k resistor, you might be right, But I have used the limiting trim pot and or small resistor to reduce the upper end of the treble control to stop oscillation, and it does work, and the only response change to the circuit is it limits the over emphasis of the treble control to stop the oscillation. no different to turning the treble control down to stop it, only with the limiting resistor or trim, it allows the treble control to go full on, with no more unwanted oscillation.

[bluelagoon]

As for the unnecessariness of the 10uF at C4 and C5 on traditional FA-1 schematic, you are correct, a lot say they are unnecessary, but then there is also the argument, that if there could be an issue there, better to be safe than sorry, for a few extra cents worth of components it makes redundant any potential issues, that is likely why Boss put them there from the start

[bluelagoon]

Here's an image of the original Boss FA-1 circuit. You will also note that R2 3M3 resistor has its own Vref voltage bias point at a lower voltage than the op amp vrefs. All component values here match the original aside from substitute op amps and jfet

[antonis]

So, how about that oscillation prevention cap across output and inverting input..??
(is it so difficult to put it there..??)

[bluelagoon]

Well how does that proposed anti oscillation cap across the op amp actually work? Perhaps its just 6 of 1 and half dozen of the other, where both methods may get the desired result. I can only speak from personal experience that a small resistor to the size difference from where the oscillation starts on the treble pot resistance to actual full pot resistance value. It works, and doesn't take out any more treble response than entirely necessary

[drdn0]

So, how about that oscillation prevention cap across output and inverting input..??
(is it so difficult to put it there..??)

Haha not difficult, wife was in meetings all arvo (works from home) so I haven't been able to test anything out

[antonis]

Well how does that proposed anti oscillation cap across the op amp actually work?

In brief: It shunts (bypasses) high frequencies out of NFB loop..

More detailed: https://www.allaboutcircuits.com/technical-articles/miller-frequency-compensation/
                    https://www.analog.com/media/en/training-seminars/tutorials/MT-058.pdf
                    https://www.analog.com/media/en/training-seminars/tutorials/MT-045.pdf
                 

[ElectricDruid]

Try putting a 100pF (47pF to 220pF) cap across pin 7 and 6 of the opamp.
You can refine the value.

+1 agree. Helps tame the op-amp at the high end, which is especially important when you start boosting the treble. You need something to roll-off the gain as you get into ultrasonics.

[Rob Strand]

Hi Rob, haven't tried upping the R9 10k resistor, you might be right, But I have used the limiting trim pot and or small resistor to reduce the upper end of the treble control to stop oscillation, and it does work, and the only response change to the circuit is it limits the over emphasis of the treble control to stop the oscillation. no different to turning the treble control down to stop it, only with the limiting resistor or trim, it allows the treble control to go full on, with no more unwanted oscillation.

Yes, I get what you are doing.  The advantage of raising R9 is the pot mid position remains flat whereas the added 5k trimpot skews the response a bit towards cut at the mid position.  (No offense but) Even though reducing the amount of treble might stop the oscillation it's solving the wrong problem (with either method).

Well how does that proposed anti oscillation cap across the op amp actually work?

It should be possible to just shave off some of the HF *outside* of the audio band so the unit is much more robust to layout and wiring ; same as ElectricDruid said.  That way it performs like the original unit.

As you go up in frequency stray capacitance has more of an effect thus making it more prone to oscillation, especially outside of the audio band.   So it's best to reduce the gain outside of the audio-band.

The unknown is whether the tone control section itself is the cause, or,  it's the whole unit.  Whatever the cause you can try the small caps in either or both positions.

Here's an image of the original Boss FA-1 circuit. You will also note that R2 3M3 resistor has its own Vref voltage bias point at a lower voltage than the op amp vrefs. All component values here match the original aside from substitute op amps and jfet

The original designer did try to do their best to preserve headroom.   The original JFET biases nicely with the original circuit.   Just plugging a 2N5457 in lose a bit of swing and the trimpot for VREF in the OP's schematic kind of trades off JFET swing and opamp swing.   The best bias would be probably be circuit #4 in my examples and take the opamp off the Vcc/2 point.



Something else which might be an issue for oscillation is the OP's circuit uses a small cap on Vref whereas the original circuit use a large cap.   [On closer look at the schem it shouldn't be a big deal.]

[Rob Strand]

Here's a sim of the roll-off from backing off the treble control vs the added HF roll-off cap.
Note further roll-off above the audio band is achieved with a small cap on the gain stage opamp.

[Treble control backed off with 5k left (4 O'clock position) - sort of like 5k trimpot at max.]



With controls flat and HF cap in place the high frequency -3dB point is about 60kHz, and the roll-off is -0.6dB @ 20kHz.

[bluelagoon]

Hi Rob, I;ll keep in mind the cap over the opamp trick for future reference, just figured that would potentially roll off more treble than you really want to lose, may be a matter of getting the cap size just right to not do so, as for adding in that restor or trim pot, the amount that was ever needed when the oscillation problem was encountered was so small usually less than 1k from a 5ok pot, making it pretty negligible to throwing the circuit off balance, I have made over 100 of the boss FA-1 effects and variants plus combination with other effects, and to me placing the limitation on the pot from opening full was the best easiest and less intrusive on the original circuit. I did notice, that the Boss units although very same components never had the issue, So I put it down to being a PCB board layout and wiring issue being different to the Boss layout. It also helped and often eliminated the problem by placing a small MLCC cap right next to the opamp positive leg and to ground. I didn't get the oscillation from every build that I had done, it was quite a random issue that only popped up  like 1 in 5 or so builds. I have even heard of the complaint for the FA-1 kit sold by AION Effects to have the same issue, not always, but occasionally with his FA-1 clone kit.

One thing about your pick for bias setting as number 4. That matches identically to the original boss circuit, so there you go.

There's a good article here that refers to oscillations in op amps, and the various remedies to correct the issue.
The method I chose was to limit the treble pots bandwidth right at the cutoff point where the oscillation started very near the end of the treble pot full on rotation, by using a small resistance to effectively limit the bandwidth going to oscillation, its worked for me and it has still enabled maximum treble response from the preamp pedal, no qualms.

https://tangentsoft.com/audio/hs-opamp.html

[bluelagoon]

Hi Rob, How did you do that scope analysis, is it as circuit simulation? and if it is, might be better for assessing the FA-1 circuit if in fact you had the correct value caps in place as on the original Boss FA-1. The treble side caps are meant to be 5n6, and the bass side caps should be 33nF.
Also which line corresponds to which varient, Green is ? and Blue is? might help a learner like myself. Cheers

[Rob Strand]

Hi Rob, I;ll keep in mind the cap over the opamp trick for future reference, just figured that would potentially roll off more treble than you really want to lose, may be a matter of getting the cap size just right to not do so, as for adding in that restor or trim pot, the amount that was ever needed when the oscillation problem was encountered was so small usually less than 1k from a 5ok pot, making it pretty negligible to throwing the circuit off balance,

OK I've updated the results below.  Yes, less than 1k isn't going to be a big offset.  Less offset than pot tolerance or people's ability to center the knob on panel-mount/PCB-less pots.

One thing about your pick for bias setting as number 4. That matches identically to the original boss circuit, so there you go.

Yes, done on purpose.

There's a good article here that refers to oscillations in op amps, and the various remedies to correct the issue.

There's many causes for oscillation.  That makes it tricky to come up with a solution unless you go step by step through the possible solutions.  In this case a hint is the treble is advanced and it oscillates, which points to a HF feedback issue; but in the end there's a chance this isn't the case.

Hi Rob, How did you do that scope analysis, is it as circuit simulation? and if it is, might be better for assessing the FA-1 circuit if in fact you had the correct value caps in place as on the original Boss FA-1. The treble side caps are meant to be 5n6, and the bass side caps should be 33nF.
Also which line corresponds to which varient, Green is ? and Blue is? might help a learner like myself. Cheers

It was done with LTSpice simulation software.

In the first simulation (V1.0):  blue is the HF cap (more HF roll-off), green is the backed off treble control.

Here's the new sim with the treble back off only 680 ohm, the HF roll-off cap reduced, and the correct FA1 part values.

[bluelagoon]

Nice Rob  Just looking at that latest spice test, it appears the 47pF holds more of the treble response than a backed off treble pot,
Do you think that the 47pF would ultimately rein in the oscillation on the last 680 ohm of the treble pot?, I will give it a try next time I encounter any such issues.
I probably should learn to use LTspice, I am sure it would give me a better grasp on things. Its just there's so much to do, and so little time to do it all, And like Neil Young says, "If you go down every path, you just might get lost" and then there's Housework Ahhh!
Cheers, Trevor

[Rob Strand]

Nice Rob  Just looking at that latest spice test, it appears the 47pF holds more of the treble response than a backed off treble pot,
Do you think that the 47pF would ultimately rein in the oscillation on the last 680 ohm of the treble pot?, I will give it a try next time I encounter any such issues.
I probably should learn to use LTspice, I am sure it would give me a better grasp on things. Its just there's so much to do, and so little time to do it all, And like Neil Young says, "If you go down every path, you just might get lost" and then there's Housework Ahhh!
Cheers, Trevor

Well, if you go high enough in frequency the feedback cap always wins since it eventually rolls off instead of flattening out like the shelf.   Also it's difficult to compare apples to apples since we don't know exactly what the minimum resistor value is and the minimum capacitor is to the stop oscillation - only after both of those are known can we compare the response.

Also, even though both methods roll-off the highs we don't know the roll-off is fixing it.  At the moment it's the best theory and it's *highly* likely to be case but I don't believe anything until I see it working (even my own stuff).  Maybe the cap doesn't work at all and the cause is deeper.

As for LTspice.  It definitely requires some commitment to learn it, and a lot of experience to know when to believe it or how close to expect the results to be to reality.   You can probably find 10 models for 2N3904 transistor and 10 models for 1N914/1N4148 diodes and get many different results.   Unless you compare the models against real measurements you don't know what models are good.   That leaves you 10 (or more) different results and you scratch your head about which one to believe most.  I'll admit often the differences are close enough but the point is it's not an easy task working out good models - I posted something in the simulation part of this forum about this:
https://www.diystompboxes.com/smfforum/index.php?topic=130372.msg1265474#msg1265474

For AC analysis like in this thread, where the circuit depends largely on resistors and caps, I'd believe spice more than most measurements.