Guitar preamp op amp tl072 input noise

[Audioguru]

On your new schematic, why do you have R3 that throws away half your signal level?

[HoracioDos]

On your new schematic, why do you have R3 that throws away half your signal level?

Yes, I was trying to flatten frequency response towards 100Hz. Does it make sense?

[HoracioDos]

I'm sorry for hijacking the thread. Rollback to V2 with proposed capacitors values.

[Zero999]

Lots of those AC coupling capacitors are unnecessary.

[Audioguru]

Hero: "Lots of those AC coupling capacitors are unnecessary."
Without those capacitors then the two opamps are a low frequency oscillator with positive feedback from the output to the input.

HoracioDos: "Yes, I was trying to flatten frequency response towards 100Hz. Does it make sense?"
You reduced the input resistance of the preamp to 1.5M and kept the same 1nF coupling capacitor but you attenuated the signal with a voltage divider. The 5kHz peak is probably needed when the input resistance of the preamp is 3M and the value of the coupling capacitor can be changed to produce any cutoff frequency you want. It was crazy to use a voltage divider to reduce the signal from the pickup.

Now I am guessing because I cannot make quotes on this website and I cannot see the thumbnails while I am replying.

[Zero999]

Hero: "Lots of those AC coupling capacitors are unnecessary."
Without those capacitors then the two opamps are a low frequency oscillator with positive feedback from the output to the input.

Positive feedback from where? C3 and C4 firmly bypass R3, R4 and R6 to 0V, attenuating any feedback. The Bode plot shows no signs of anything which would cause oscillation, between 10mHz and 10MHz.

By the way, I've had no problems with making quotes on this site.

[Richard Crowley]

Why is the output capacitor value so high...? Calculate a suitable value for the load resistance it will have (100k? Then the capacitor will be 100nF).

It is trationally pretty large because we don't know what is the load impedance.  It could be as low as 100-300 ohms, but more typically 600-1000 ohms, or even 10K.  If we set the lowest expected load impedance at 300 ohms and the lowest frequency of interest at 70 Hz, we would need a 35uF capacitor to maintain a 3dB rolloff at 35Hz.

that it will take weeks to charge

And that is why a "drain resistor" of perhaps 1~10K from the output side of the capacitor to ground is commonly seen.  And/or using the switched node of the output connector to short the output capacitor to ground when there is no load applied.

High frequency resonance is used and is needed because guitar speakers have poor high frequency response like an old telephone or AM radio.

And it would appear that even modern guitar speakers continue to have terrible frequency response (by modern hi-fi standards).
But the lousy frequency response and peaky frequency response is all part of the sound of the instrument.
So "improvement" it to modern standards is apparently not very desirable in this context.   

[Richard Crowley]

Now I am guessing because I cannot make quotes on this website and I cannot see the thumbnails while I am replying.

Yes, that is one of the unfortunate things about this forum.
Dunno whether is is a fundamental design of SMF, or just how this forum is configured?
In either case, pretty annoying.  I have to open a second copy of the thread so that I can see the graphics in the previous posts.

[Zero999]

Oh, I see the issue with seeing schematics whilst replying. That's one of the reasons why I usually make my attachments inline, using [img] tags.

HoracioDos: "Yes, I was trying to flatten frequency response towards 100Hz. Does it make sense?"
You reduced the input resistance of the preamp to 1.5M and kept the same 1nF coupling capacitor but you attenuated the signal with a voltage divider. The 5kHz peak is probably needed when the input resistance of the preamp is 3M and the value of the coupling capacitor can be changed to produce any cutoff frequency you want. It was crazy to use a voltage divider to reduce the signal from the pickup.

Yes, C5 and C8 are too small. I can't believe I missed that, especially with the Bode plot staring me in the face.

[Audioguru]

Positive feedback from where? C3 and C4 firmly bypass R3, R4 and R6 to 0V, attenuating any feedback. The Bode plot shows no signs of anything which would cause oscillation, between 10mHz and 10MHz.

Without seeing the schematic during my reply I forgot that the gain of this preamp is only 7 times and much less at low frequencies so of course the filter of the virtual ground prevents positive feedback.

By the way, I've had no problems with making quotes on this site.

This quote looks different than before. I think the quote I am making in this reply will work.

[Zero999]

Now the gain is much lower, there's no point in having the gain split across two amplifiers, as the bandwidth of one, is more than enough. The circuit can be simplified greatly.

[HoracioDos]

Amazing shape!

[Ale]

Hello everyone! I've been a little busy lately, but finally I made out. This is the final circuit I build, on the breadboard it works great, with very very (very) little noise if I use an external power supply (EI, I'm still waiting for the toroidal and the switch psu is still noisy), almost nothing with batteries. I'd like to thank you another time, I really learn alot and now I could experiment easilly.

Ah, I was forgetting to say... I put the 24V because I made one psu (unregulated) like that, I still don't hear any difference with or without R4 and neither with C3. I added some diodes to prevent the circuit to explode with reverse voltage. R1 and C2 are used to decoupling and lower the DC noise (I found many videos on youtube with this schematic).

[Audioguru]

The 470uF output capacitor value is way too high and will take 23.5 seconds to almost fully charge by R4. Calculate a value suitable for 50Hz and the load impedance of minimum 2k ohms parallel with R4 (1.9uF). A 2.2uF capacitor will pass 50Hz pretty well and will fully charge in 0.022 seconds. 

R4 charges the output capacitor when power is applied to the circuit. Then there is no loud POP when a load is plugged into the output jack.
You show the value for C3 as .22nF which is impossible. If it is 22nF then it cuts frequencies above 35.2kHz. It will barely reduce 7kHz.

[Ale]

Yes, it is 22nF not .22, my bad  I will reduce the C7 value and make some tests.

[HoracioDos]

Please post final schematic when it's ready. I'm very interested in what you are doing but sadly I don't enough time right now to build it.

Thanks in advance.

[grants225]

Guys, what about changing RV1 of 100k to 1M fixed resistor, volume/gain controll moving after opamp (or if we still have possibility to use second part of TL072, put potentiometer in beatween of two cascades) and drop R5 to 10k-20k rate?

Current gain of x10 is too much as some guitars may output up to 1Vrms. To avoid overload in second amp stage, RV1 should be turned down and what happens that poor weak input signal first should be divided by <~10 in order to be later on amplified by x10. Higher rate of set opamp gain adds distortion and additionally RV1 on input can became possible noice pickup source due to hard to implement proper grouding and shielding  (again, picked noise goes straight into x10 amplification)

[Ale]

What do you think about this?
Grant225, I place a buffer before the real amplification stage now the RV is between them, were you think somethink like that?

I also added a low pass filter and high pass filter in the first stage, is that a bad idea? are they correct like this?

[Ale]

Perhaps it's better like this, with the hi pass before and low pass after the buffer.

[Audioguru]

If you modify the awful buffer then might as well do it right:

[Ale]

Yes of course, but you are using the inverting configuration, I was using the non-inverting. Is it not possible to do the same with the non-inverting? I read that the non-inverting is the less noisy configuration, that's why I was using it. Is it better to use the inverting when apply filter because the bias on the non-inverting input is isolated right?

Another question... the 470nF and RV1 are forming an high pass filter right? but the cutoff frequency is 1hz at 100k and 11hz at 30k (but it also dimish the volume) but it's not important because it is too low hz value, am I wrong?

thank you, I'm ready to modify the schema with your suggestion.

[Richard Crowley]

Why would non-inverting be less noisy?  I never heard of that before.

[Audioguru]

I used an inverting opamp so that the lowpass filter is a simple integrator. I doubt that anybody cares about its 180 degrees of phase shift.
I was going to use a 100nF capacitor to feed the 100k volume control but it adds another -3dB at 16Hz and affects up to about 80Hz.

[Ale]

Don't remember where I read it, sorry... This is the new schema of the buffer.
I have a question: if I end the first stage with the 470nF cap, can I remove the .1uF cap on the second stage?

thanks Audioguru, I'm eager to try it out 

[Audioguru]

I have a question: if I end the first stage with the 470nF cap, can I remove the .1uF cap on the second stage?

The 470nF capacitor can be 100nF for a cutoff frequency of 16Hz.
Your new schematic does not have a second stage so I am guessing that its .1uF capacitor is its input coupling capacitor.
If the volume control that is connected to 0V connects directly to the input of the second opamp then the input of the second opamp will be at 0V and this opamp will rectify the audio causing severe distortion. Actually if the DC input of a TL071 opamp is within about 3V from the negative supply (which is ground in this circuit) then the output goes as high as it can and produces no sounds but maybe popping sounds on the loudest sounds that exceed 3V. The input of the second opamp must be biased at half the supply voltage so that it can swing up and down with the signal. Then the signal from the volume control (which has 0VDC) can feed it with a coupling capacitor.