Issues with original design distortion

[Voltzy]

Hello all,

I have been playing around with my own design that is essentially a Sound 2 Light merged with the Thunder Alley:



I have been trying out three different input buffers, while trying each one the other two are completely disconnected from power and ground and the audio path. Eventually I will add a tone stack.

I feel the JFET has the best sound, the problem is when I activate the Bass Boost I get a high pitched squealing sound while not playing and a sine-wave like sound when the guitar is unplugged. While I am playing with Bass Boost the sound is an incredible full on distortion so I would like to keep that if possible.

I have tried adding a 100nF cap to gnd between the Bass Boost switch and R6 and it does reduce the squealing a bit but it feels more like I am treating the symptoms and not the problem.

The TL072 buffer works fine with the Bass Boost however with it on or off it always sounds gated and no amount of re-biasing the final 4N26 transistor will fix it.

The LM741 works the best but the sound is kind of lacking in my opinion.

I also have a very strange problem with using any of the buffers, or even not using any, where once I turn on the power it will sit around 5V and then slowly raise itself to around 6.5V, you can also hear this change while playing something through it, it takes around 10 seconds to fully "turn on". I have double checked the polarity of the power filter caps and have tried multiple batteries and power supplies all with the same slow-power issue.

Really at this point I am clutching at straws, I have come across other issues in the build and have always been able to fix the problems myself, however these three issues (JFET bass boost squealing, TL027 gating, and slow-power build) have got me really stumped.

[antonis]

Buffer2 (JFET) isn't a buffer (strictly speaking) rather than a Booster (with a gain of 3)..!! 

If you like to use it as inverting buffer (like the other two) make R10/R11 ratio close to unity..
(but you'll keep to have R11 value output impedance - not a problem in case of Op-Amp input..)

The above should "ruin" Bass Boost <incredible full on distortion> so it might be better to keep it as it is and place a xxpF feedback cap from Drain to Gate to prevent squealing..

[ElectricDruid]

Does LM386 need bias on its +ve input? At the moment that's just floating there at some undefined DC level behind that 47u cap.

I don't know this chip - can you do that on LM386?

Tom

[dschwartz]

I would filter the Vref of the tl072 buffer . And make the 47uF coupling cap smaller and /or put a 47k resistor to ground at the input of the lm386.
The 47uF is too big and not referenced to anything, so the bias can float around.

[duck_arse]

the output of the LM386 will sit at 1/2 supply. I think you need a cap between the output and the led CLR. and probably a pulldown resistor after C7.

[Voltzy]

Buffer2 (JFET) isn't a buffer (strictly speaking) rather than a Booster (with a gain of 3)..!! 

If you like to use it as inverting buffer (like the other two) make R10/R11 ratio close to unity..
(but you'll keep to have R11 value output impedance - not a problem in case of Op-Amp input..)

The above should "ruin" Bass Boost <incredible full on distortion> so it might be better to keep it as it is and place a xxpF feedback cap from Drain to Gate to prevent squealing..

Yes I was originally going to use it with a gain of 1 but then I read up on the Tillman and Fetzer preamp designs so I decided to try those instead, now that I think about it the FET "buffer" probably sounds the best of the three because it is already boosting the signal a little before the LM386.... anyways I will try out a feedback cap thank you.

I would filter the Vref of the tl072 buffer . And make the 47uF coupling cap smaller and /or put a 47k resistor to ground at the input of the lm386.
The 47uF is too big and not referenced to anything, so the bias can float around.

OK thank you, I had planned to filter Vref but had completely forgotten. What size cap would you recommend to change the 47uF to? I put that in simply because it was on hand, perhaps a 10uF instead?

Does LM386 need bias on its +ve input? At the moment that's just floating there at some undefined DC level behind that 47u cap.

I don't know this chip - can you do that on LM386?

Tom

You don't need to bias the inputs of the LM386 which is great, plus it has super high gain which makes it my favorite op/power amp

the output of the LM386 will sit at 1/2 supply. I think you need a cap between the output and the led CLR. and probably a pulldown resistor after C7.

OK I will give these a go, but I am unsure of the purpose of them. I'm guessing the CLR cap will reduce any DC noise going through the LED? And the pulldown resistor will provide a path for C7 to drain when then effect is off? I suppose it will also create a high-pass filter...

[duck_arse]

the CLR will stop the 386 from burning your led w/ too much current, same as any led. the cap between output and CLR will block the DC, so the led won't fall-back to 4V5 all the time, and will go from OFF to whatever ON-ishness the 386 drives it to. as for the pulldown, it anchors the free end of that cap, refers it to ground so's the next thing you plug in doesn't go WHUMP.

a(n additional) 10uF ~ 47uF cap to ground at the junction of R3 and R4 will be plenty to filter your Vbias.

[Voltzy]

OK so I tried a 22pF feedback capacitor across the drain and gate of the JFET which has no apparent effect with bass boost off, and almost cuts the squealing with bass boost on (still get a tiny bit almost sounds like popping) so I reckon I will put in a 27pF and that will do it. I originally tried around 100-200 pF and those gave me a noticeable cut to the amount of gain on either setting.

With the TL072 buffer I changed C4 to 10uF, added a 47k to ground from pin 3 of the LM386, and added a 10uF filter cap for Vref. It does sound a bit better, still gating but a different gating sound if that makes any sense. Perhaps now adjusting the 4N26 bias will fix it I will try that next.

I had tried a DC blocking cap for the LED before its CLR but they all just completely cut the sound/light. I tried a range from 1uF down to 220nF to no avail, just realized I forgot to try out my new collection of tiny pF caps so I will do that tonight but I don't have much confidence.

I haven't added a pulldown resistor after C7 yet but I will incorporate that into the final design, I was thinking of using a 1M.

[duck_arse]

for the DC blocker before the leds, you want the value to start at 1uF, and then work your way UP, 10uF, 47uF, 100uF, maybe 470uF. see what happens then.

[Voltzy]

Oh right I have been going the wrong direction then haha, OK thank you I will give that a go.

[Voltzy]

So I have been tweaking my design quite a bit, ended up not being happy with the sound I was getting out of it. I simply tried the circuit skipping the optocoupler entirely and it sounded great! The tone control gave me all the bass I needed so I did away with the Bass Boost switch.

This is the design I have now (I had added a BMP style tone stack and a recovery stage).


So it probably isn't that original anymore, but I am loving the sound of it. I wanted to know if I should add another two zener diodes to the gate of the second JFET? I gather that any static shock generated at the input will be dealt with by the first two zeners...

Also (as this is still all new to me) is there any major things I have done wrong here? Anything anyone can suggest to make it work better/less noisy/more utility etc? To give some context I play in some doom metal bands and frequently use tuning of B-standard or lower. So far I have found it sounds great with a bass guitar with the gain turned almost all the way down.

The only things I still have on my list to try is a larger value for the gain pot, and to add a wet/dry blend pot. With the blend pot I figured I could just blend the input straight with the output, as the signal gets inverted with Q1 and then inverted again with Q2.

[Voltzy]

Can no one tell me if I need the second set of Zener diodes?

[ElectricDruid]

Do you even need the first set?

I can't think of a single example stompbox circuit that uses that sort of protection. Why do you think you need it?


Tom

[287m]

my eyes and minds trick on me when i see that Jfet drawing

and why 10u in IN and OUT? tillman-ish?

[antonis]

With a very rough estimation, you'll need about 12-13dB gain recovery after tonestack so you might have to by-pass R11 with a cap..
(I'm not familiar with J201 transconductance habits so you might need to partialy by-pass it for a Q2 voltage gain of about 4)

[GibsonGM]

Do you even need the first set?

I can't think of a single example stompbox circuit that uses that sort of protection. Why do you think you need it?


Tom

Misplaced from Mosfet designs?  *shrug*

[antonis]

<Misplaced from Mosfet designs? >
From bi-polar supply Mosfet designs, I would dare to add, Sir.. 

[ElectricDruid]

<Misplaced from Mosfet designs? >
From bi-polar supply Mosfet designs, I would dare to add, Sir.. 

So we can follow long-established hacker-engineering practice for parts reduction in a circuit design: Take some bits out. Does it sound worse? If yes, put them back and pull something else. If no, leave them out.

T.

[Tony Forestiere]

https://en.wikipedia.org/wiki/Muntzing 

[Voltzy]

Thanks for the replies!

Do you even need the first set?

I can't think of a single example stompbox circuit that uses that sort of protection. Why do you think you need it?


Tom

I got the idea from this https://www.electrosmash.com/1wamp site where it talks about the fetzer amp, and suggests the JFET is susceptible to static discharge. I am really worried about cooking any of my J201s as they are not easy to get a hold of here and I don't have that many. From what everyone is saying the first set of Zeners are just being super safe and the second set are not needed at all.

my eyes and minds trick on me when i see that Jfet drawing

and why 10u in IN and OUT? tillman-ish?

Do you mean the input and output caps for the entire circuit? I had heard that lower value caps can cause some bass to be attenuated and that 10uF was a good safe value. Is it bad using such high value DC blocking caps in the circuit?

With a very rough estimation, you'll need about 12-13dB gain recovery after tonestack so you might have to by-pass R11 with a cap..
(I'm not familiar with J201 transconductance habits so you might need to partialy by-pass it for a Q2 voltage gain of about 4)

Holy crap another 13dB would make this pedal loud, it is already very loud as it is. I will give this a go and aim for 4V of gain. When you say partially bypass, can I add a cap+resistor in parallel to R11 and adjust the new resistor value accordingly?