Positive feedback in high gain preamp

Thread Starter

hrs

Joined Jun 13, 2014
321
Hi,

I'm working on a guitar amplifier with a high gain preamp. Some photos and the preamp schematic are attached. When I open up the preamp volume control RV1 beyond a certain point I get squealing at ~1.5 kHz. I think this is due to positive feedback. As you can see from the photos most of the wiring is longer than needed. These will be shortened as I settle on the configuration of the circuits.

Before I start randomly shortening wires and changing things, which wires are most sensitive to introducing feedback (if this is indeed the problem)? I have some shielded wire. Where does this make most sense, on the input or output of the preamp? How important is it to shorten the power wires to the power amp and the speaker output? Other considerations?
 

Attachments

crutschow

Joined Mar 14, 2008
27,198
Yes, it sounds like feedback.

The input wires that have the lowest level signal are the most sensitive.
You might try shielding them.

The power and speaker output wires are not likely the problem.
 

Audioguru again

Joined Oct 21, 2019
3,177
Does the guitar have a magnetic pickup? Why is R15 feeding DC current to the guitar?
Why does the circuit use an input transistor with an input impedance too low instead of one of the Jfet-input opamps that have the needed very high input impedance?
Why does it have gyrators to cut frequencies around 87Hz, around 900Hz and around 1300Hz? The 1500Hz feedback frequency is reduced a lot.
The polarity of C32 is backwards.
The circuit has many distortion-producing diodes so it must sound awful or maybe you like the "fuzz".
 

Thread Starter

hrs

Joined Jun 13, 2014
321
Does the guitar have a magnetic pickup? Why is R15 feeding DC current to the guitar?
:eek: Yes, magnetic. R15 is a remnant from the stompbox this is based on. I guess I can just cut it out right? Since Q6 has its own bias network.
Why does the circuit use an input transistor with an input impedance too low instead of one of the Jfet-input opamps that have the needed very high input impedance?
In the next step there will appear a TL07x on the input configured as a buffer. That will feed into this and and another preamp.
Why does it have gyrators to cut frequencies around 87Hz, around 900Hz and around 1300Hz? The 1500Hz feedback frequency is reduced a lot.
Because it produces that Swedish sound we all know and love so much.
The polarity of C32 is backwards.
Indeed! I'll add it to the list.

Thanks!
 

Audioguru again

Joined Oct 21, 2019
3,177
I looked up "Swedish Death Metal" sound videos and saw and heard guys playing high pitched buzzers that looked like guitars.
One guy showed that all his level and distortion knobs were at maximum!
There was too much distortion for me. Maybe people who played with guns and therefore have a high frequency hearing loss like it.
 

Thread Starter

hrs

Joined Jun 13, 2014
321
The input wires that have the lowest level signal are the most sensitive.
You might try shielding them.
Perfect, thanks! This fixes the issue. Now I need to turn it really loud before the normal feedback sets on. I also removed R15 based on AGs review. I don't know if DC on the guitar could contribute to positive feedback.

I looked up "Swedish Death Metal" sound videos and saw and heard guys playing high pitched buzzers that looked like guitars.
The chainsaw type sound I'm referring to is this amongst others:
One guy showed that all his level and distortion knobs were at maximum!
Yes. Those gyrators have pots on them in the original Boss HM-2 circuit. To get the desired sound I fixed them to 11.
There was too much distortion for me. Maybe people who played with guns and therefore have a high frequency hearing loss like it.
Still it could be worse:
 

Ian0

Joined Aug 7, 2020
2,196
How long is your lead? 1500Hz resonance sounds like the LC parallel resonance of the pickup and the lead capacitance. Does it change if you use a shorter or longer lead?
 

Audioguru again

Joined Oct 21, 2019
3,177
Of course the unshielded input cable was picking up the output causing feedback and also picking up other interference.
The sound effects and other noises are not music.
 

Ian0

Joined Aug 7, 2020
2,196
Of course the unshielded input cable was picking up the output causing feedback and also picking up other interference.
and the interaction of the pickup inductance and cable capacitance will put a peak in the gain at about that frequency.
Another place worth looking is the two input preamp stages. They share one decoupling network. Is there coupling between them in the shared decoupling network? Why not have a single op-amp stage, and you can fix the low impedance input at the same time as @Audioguru again mentioned in post #3
 

Thread Starter

hrs

Joined Jun 13, 2014
321
How long is your lead? 1500Hz resonance sounds like the LC parallel resonance of the pickup and the lead capacitance. Does it change if you use a shorter or longer lead?
Before fixing the problem with shielded input wiring I used one 1.5 m cable and a 4 m cable. It made no difference.
 

Ian0

Joined Aug 7, 2020
2,196
It was a bit of a longshot, as the ridiculously low input impedance should damp any resonance. If you use a more conventional 1M input impedance, it might show up.
 

AnalogKid

Joined Aug 1, 2013
9,249
If you use a more conventional 1M input impedance, it might show up.
Not sure that's the "convention". The input impedance that the pickup sees has a direct effect on its tonality. The lower the input impedance, the more high frequency roll-off there is as the coils are more heavily loaded. Mic preamps often have an optional 200 ohm input impedance because some dynamic mics are designed for this impedance to give them whatever their characteristic "sound" is.

ak
 

Audioguru again

Joined Oct 21, 2019
3,177
I think that an old Shure SM-58 dynamic mic has its high frequencies peaked and resonating because that was as high as it could go in those olden days, I doubt it was designed for the peaking. Then the marketing guys said it produces "presence" in the sounds. It also adds some high frequencies to the muffled sounds of an old AM radio.
 

Attachments

Ian0

Joined Aug 7, 2020
2,196
Not sure that's the "convention". The input impedance that the pickup sees has a direct effect on its tonality. The lower the input impedance, the more high frequency roll-off there is as the coils are more heavily loaded. Mic preamps often have an optional 200 ohm input impedance because some dynamic mics are designed for this impedance to give them whatever their characteristic "sound" is.

ak
It's the value used by Fender, Marshall and many others. Blencowe says in Designing Tube Preamps for Guitar and Bass that it should be greater than 470k in order not to load the "electronics" in the guitar - the "electronics" being a seemingly random collection of potentiometers and capacitors that passes for a tone control.
A quick scan through Pittman's The Tube Amp Book reveals that 470k is quite common, but I found values up to 10M, obviously for those who like the sound of mains hum.
 

Thread Starter

hrs

Joined Jun 13, 2014
321
Before I start laying this out, do you see anything wrong with this? Output1 and output2 will go to buzzer preamp 1 and buzzer preamp 2 and then into a mixer. Should I include a small resistance in the high side near the switch? Can I expect this to work without plopping sounds? Both buzzer preamps have input coupling caps so I didn't include them in these outputs.
 

Attachments

AnalogKid

Joined Aug 1, 2013
9,249
As shown, you will get a loud pop whenever a switch closed, because the input side of an external coupling cap has to charge up to 7.5 V. When an output is off, the shunt leg of the output switch discharges the external cap to 0 V, so things pop each time the output is enabled.

If you add a high-value resistor from each output to the 7.5 V reference, these will pre-charge the external coupling capacitors when connected, eliminating a large pop when the series switch closes, and keeping them charged when the switch is open. For this to work, the shunt legs of the output switches (U2B and U2D) should switch to the 7.5 V instead of GND. Also, you will get much better linearity through the series switches with 0 Vdc across them (7.5 V on each side).

ak
 

Audioguru again

Joined Oct 21, 2019
3,177
Why do you need the opamp to make 7.5V for the other opamp's bias voltage when the other opamp's input is a Jfet that needs almost no current? You have two 100k resistors as a voltage divider and a filter capacitor feeding the input of the opamp that is not needed. Instead, the voltage divider and capacitor can feed R2 on the useful opamp instead and it can be a single TL071 opamp. At Digikey today, one TL071 costs $.60US and one TL072 costs $.73US in through hole packages.

If you switch inputs on your circuit during a sound (speech or music) then the sudden voltage change from zero will cause a loud POP sound.

I worked with a large high quality intercom system that ramped down an audio level, switched it to an audio amplifier when the signal was zero then ramped up the audio level to avoid a POP.

If you ramp levels up and down from 7.5V then you might need the "un-needed" opamp.
 

Thread Starter

hrs

Joined Jun 13, 2014
321
If you add a high-value resistor from each output to the 7.5 V reference, these will pre-charge the external coupling capacitors when connected, eliminating a large pop when the series switch closes, and keeping them charged when the switch is open. For this to work, the shunt legs of the output switches (U2B and U2D) should switch to the 7.5 V instead of GND. Also, you will get much better linearity through the series switches with 0 Vdc across them (7.5 V on each side).
Great, thanks! Does the attached circuit cover what you intended?

Why do you need the opamp to make 7.5V for the other opamp's bias voltage when the other opamp's input is a Jfet that needs almost no current? You have two 100k resistors as a voltage divider and a filter capacitor feeding the input of the opamp that is not needed. Instead, the voltage divider and capacitor can feed R2 on the useful opamp instead and it can be a single TL071 opamp. At Digikey today, one TL071 costs $.60US and one TL072 costs $.73US in through hole packages.
I had it like that because I thought I might make use of an unused opamp. All I have is TL072 and TL074 in terms of jFET opamps. Maybe I'll buy a TL071. TI says to connect unused opamps like this:
2553.Slide-0001.JPG-650x0.jpg
Do you think it's better like in the revised circuit?
 

Attachments

Top