Opamp fx loop white noise

[psychepool]

I made a Bogner Ecstasy-based amplifier.

I needed an active fx loop and applied a circuit using TL072 used in Marshall 1987X.
It was chosen because it was used very satisfactorily for other amplifiers I made before.
I also liked it dosen't need to set the Send/Return level manually.

The circuit is like this.
It appears to be a structure that supplies constant voltage through a zener diode after lowering the voltage to the resistance of a high wattage.
The power source is B+ at the point connected to the PI.

01.jpg

All operations are good and there is no tone change compared to loop bypass.
The only problem is that there is a white noise.
The noise is not very serious. I can barely hear it if I set -8 dB in my THD hot plate.
Although it is not equipped yet, if it has PPIMV, it will disappear if it is lowered to a volume that will be used at home.
This is because increasing the volume does not cause the noise to grow together

However, when applied to other amplifiers in the past, I want to solve this part because it was so quiet that I could not tell which side was turned on even if I tried to turn the loop on or off.

First of all, when I played bypass, it seemed that it did not come from the input because it was quiet.
Of course, it was quiet even when the loop-in-out wiring was directly connected to each other.
To check if there is a problem on the Send/Return jack side, we removed the wiring from the Send/Return jack and connected the Send/Return points on the substrate, but there is still noise.

The ground on the amplifier was all star-grounded in the power-trans-bolt, except for the extent of the ground on the output tube cathode, bias circuit and relay power supply. The output jack ground is also at the Star Ground point.
The ground of the loop was moved to various locations, including Presencepot, the last stage of the preamplifier, and the output jack, but there was no difference.
To separate from the star-ground, the preamp ground was cut to the last stage of the preamp, and earthed to chassis near by input jack and the loop was connected to this side, but it also made no difference.

It was made with 1/4W resistor and 63V film cap when it was first made, so I thought it was a problem with parts.
But after made with PRP 1/2W resistors and Wima MKP10 250V capacitors, the results are the same.

I supposed the zener is under too much load, so I increased the resistance value, but it doesn't make a difference either.

Previously applied to other amplifiers had larger power smoothing caps. Instead of 47uF, 220uF+100uF parallelism was fitted.
So this time, I grew it to 100uF -> 220uF -> 470uF, but it made no difference except the noise comes a little later after the power is turned on. (May be because of the higher the capacity, the slower the power supply)

I moved around with the board and the wires connected to it, but nothing changes either.



I think I've done everything I can, so I can't do debugging anymore, so I'm writing this.
Where can we find the cause?
Is there anything else I can do?

The overall layout of the amplifier is as follows.
The blue box is the main circuit of the loop and the red box is the main circuit of the loop.

02_1.jpg

The picture below shows the current state of being made by putting the power supply into one board while making new parts.
The existing power supply is only fitted and not used.

03_1.jpg

A noise recording file is also attached.

[pdf64]

Has the circuit, and amp with the loop engaged, been scoped for oscillation?
Are the ICs Vdc measurements all good? eg all opamp pins should be close to Vref.
Is the hiss on the send signal, or is it originating solely from the return stage ic1b?
eg does it change if the return jack is shorted?

[Stevem]

Seems like a impeadance mismatch to me.
You may need to buffer it on one end.

[martin manning]

Have you tried a different TL072?

[pdf64]

Seems like a impeadance mismatch to me.
You may need to buffer it on one end.

Can't see what you mean, could you explain your thinking?
To me, each stage looks to connect to the ones before and after via reasonable impedance bridges, as would be expected with line type voltage signals.

[Aurora]

Off the top of my head, - try another opamp... try to scope the supply to see where the noise starts... supply, buffer/driver, input driver...
You seem to have very high gain on the neg input to the input buffer....?? How clean is the raw 30V??

[psychepool]

Has the circuit, and amp with the loop engaged, been scoped for oscillation?
Are the ICs Vdc measurements all good? eg all opamp pins should be close to Vref.
Is the hiss on the send signal, or is it originating solely from the return stage ic1b?
eg does it change if the return jack is shorted?

Oscillation does not happen.
The supply voltage is extremely normal. Vdc 30V, Ground 0V, Vref 15V...
Not all opamp pins were measured. I checked that there was a 30V flow on pin 8.
I think it's a good idea to separate Send/Return Jacks. I'll give it a try.
The return can be tested by plugging it in directly, and the send can be connected to another amplifier and tested.
It is regrettable that I have no other amplifier with a return jack, but I think it will be sufficient to connect it to the input jack.
Thank you for providing me with a good checkpoints.

Seems like a impeadance mismatch to me.
You may need to buffer it on one end.

Can you explane a little more?
Because the circuit itself worked perfectly on other amps, I don't want to change the configuration at all.

Have you tried a different TL072?

I haven't tried yet. I will find another TL072 or dual opamp and test it.

Off the top of my head, - try another opamp... try to scope the supply to see where the noise starts... supply, buffer/driver, input driver...
You seem to have very high gain on the neg input to the input buffer....?? How clean is the raw 30V??

OK, I will test it by connecting signals at various points such as Send/Return/Input/Output.
The volume going into the loop input is very small, because it is being tested at home (very low master volume setting).
The noise is in the same form and level, no matter whether the volume is raised or lowered, distorted or clean channel.
"How clean is the row 30V?" : I don’t know exactly what it means. 30 VDC is correctly applied to opamp V+.

[psychepool]

I tried somethings.

I reconnect send/return jack and test again.
First, I turned on the amplifier and connected the guitar to the return jack.
There's the same noise. I think it's a little smaller than when it was connected to the input, but there's no meaningless difference, so I think it's probably an illusion.
Then, after connecting to the amplifier input, and connect a cable from the send jack and connect it to another amplifier input (for safety, the Bogner amp's speaker out is connected to the dummy load).
There was almost no noise.
But the send side is attenuated and the amplification stage is a return stage, so it can be a natural result.

I changed two different TL072s, but there was no difference.

How should I interpret it?
Could it be a lack of electricity?
The circuit diagram shows the power being drawn from the PI B+ point.
I connected it from the screen grids of the output tube to see if I could get more current, but it didn't make any difference either.

Is there any other reason to infer?
Is there any other attempt I can try?

[pdf64]

Does the setting of SW1 alter the noise?

[psychepool]

Does the setting of SW1 alter the noise?

I didn't use sw1.
It's +4dB/-10dB mod switch.
I don't need +4dB mod so I made it -10dB setting.

Connect 47n on the send stage after 270K.
And omit 2.7K on the return stage, then connect 47uF to Vref directly.

I thought it wouldn't be a problem to omit the 2.7K resistance.
Is it nessary to install 2.7K even if it would be jumped?

[pdf64]

...Is it nessary to install 2.7K even if it would be jumped?

No. Any other deviations from the schematic to report?
What happens to the hiss if the + input of IC1b is shorted to Vref?

Are you sure you need -10dB? If the send signal level was higher then the return stage could be adjusted for gain gain and hence lower hiss.

[psychepool]

No. Any other deviations from the schematic to report?
What happens to the hiss if the + input of IC1b is shorted to Vref?

Are you sure you need -10dB? If the send signal level was higher then the return stage could be adjusted for gain gain and hence lower hiss.

I don't think there's anything in particular except what's been written so far.
In order to secure higher voltage, it was produced using MC34072 and 20V zener (18V zener in the Signal section), but the same noise was generated.(but other operations are well)
Now it's made exactly like the original circuitry.

I connect the +input of IC1b. There's no hiss noise. Of course there's no guitar sound too.
What was the purpose of this work to check?

Each +4 dB/-10 dB switch is known as a balanced/unbalanced switch.
+4dB is a spec used for racks and we know that -10dB mode should be used to connect typical stompbox effectors.
When I used it previously made amp, stompboxes are distorted in +4dB mode.
Since only stompboxes will be used anyway, I tried to fix them at -10dB.
For the same reason, there are concerns that raising the send signal may cause distortion in the Stompboxes.

[pdf64]

How about if the node of c3 and r9 is shorted to Vref? If the hiss is there, are you sure r9 isn’t 470k?
Regarding the fx loop signal level, have you actually measured it? The signal division and multiplication ratios, and hence the +4 / -10 label accuracy, are only relevant to the amp it was designed for.

[psychepool]

How about if the node of c3 and r9 is shorted to Vref? If the hiss is there, are you sure r9 isn’t 470k?
Regarding the fx loop signal level, have you actually measured it? The signal division and multiplication ratios, and hence the +4 / -10 label accuracy, are only relevant to the amp it was designed for.

The loop board has been removed from the chassis for a while. So I'll try what you said later.

I make amps just for a hobby and I don't have professional electrical knowledge.
Signal level in +4dB/-10dB mode has not been specially measured.

You seem to say that hiss noise occurs because the current -10dB mode is excessively amplified in the recovery section.
I'll put the mode switch back on and test it.
If the hiss disappears in +4dB mode, may I say what you say is the cause?
So how do I adjust the attenuation/recovery stage for each?

[psychepool]

Loop refited to the amp.
Now +4dB/-10dB mode switch is installed.

Noise disappeared in +4dB mode!
I'm not sure if it's gone to perfectly zero, but I can be sure it's normal level.
Unfortunately, in -10dB mode it is the same as before.

It has yet to test whether stompbox effectors operate normally in +4dB mode.
There was definitely distortion in the amplifier used before.

If there is no distortion in +4dB mode, then it can be used as it is, but otherwise it will be necessary to adjust the amount of attenuation/recovery appropriately.
Which part should I touch to adjust?

The attenuation stage (IC1a) appears to attenuate as if it were a potentiometer through the resistance value of the input stage.
Given that the recovery stage (IC1b) is a configuration that jumps R1 resistor on a mode switch, it appears that the more the R1 value increases, the less amplification, the more amplification increases.
Am I right to guess?

So my debugging plan is :
-Reduce the value of R1 as much as possible within the noise-free range.
-In that state, adjust the values in R4/R11/R12 to find the values that do not differ in level from the bypass.
-Then, connect the Stompbox to check if any distractions occur.
Do you think it's a good idea?


I found the following circuit on the web.(There is some error, but I don't think there's a problem understanding the content.)

opamploop2.jpg

It is the same circuit as the currently worked loop, but the above mentioned attenuation/recovery seems to have been adjusted.
If +4dB clearly causes distortion, I think it would be better to test the values specified here.