will this 48 V pink noise source work?

[MisterCMRR]

R16 will have only op-amp input bias current flowing in it. But the object in an inverting amplifier is to hold the + input at virtual ground (1/2 Vcc). Using such a high value opens the door to parasitic oscillation or instability due to a few pF of stray capacitance. Sometimes a resistance equal to that in the negative feedback loop is used to partially compensate for DC bias current offset in circuits requiring high DC accuracy. But, even then, a wise circuit designer will bypass it to avoid creating the aforementioned sensitive node at the + input. I'd use 1 k-ohms or eliminate the resistor altogether.

 

[Chrisfromthepast]

Sometimes a resistance equal to that in the negative feedback loop is used to partially compensate for DC bias current offset in circuits requiring high DC accuracy. But, even then, a wise circuit designer will bypass it to avoid creating the aforementioned sensitive node at the + input.

Thanks Bill, I appreciate your time. Can you clarify if you mean that I should parallel the 1M resistor in the feedback loop (R11) with like 56pf? And then R16 needs to be 1M as well? Or should I still take R16 away?
Your advice was just to take R16 and be done with it, Im just trying to make this click in my lil brain.

 

[MisterCMRR]

I would just tie the + input directly to virtual ground. R16 serves no useful purpose in the circuit.

 

[ruffrecords]

Thanks for helping me.
For clarification, why not connect pin 3 to the v+ node through a 2k2 resistor? I thought this whole impedance balance thing required matched parts on each leg so that the real and complex impedances were exactly the same.

It is. The output impedance of U1A is effectively zero so for ac it is effectively connected to gound. Hence R12 and C6 is defines the impedance driving pin2 of the XLR. C8 and R15 provide the same impedance for pin 3.

Is 2k2 so large compared to 150r to make it negligible?

From the above you can see the 2K2 is completely superfluous. Furthermore it connects the positive supply direct to pin 2 which is something the downstream equipment will NOT appreciate.

Cheers

Ian

 

[ruffrecords]

I would just tie the + input directly to virtual ground. R16 serves no useful purpose in the circuit.

Agreed. I think it is a leftover from when that stage used the non-inverting input.

Cheers

Ian

 

[MisterCMRR]

Tie pin 3 of op amp to virtual ground (+ side of C11, which is 1/2 of Vcc), not real ground, which serves as V- in circuits like this.

 

[Chrisfromthepast]

its phantom powered like a condenser mjc

 

[volker]

I still don't understand where both V+ and V- are coming from!!! I had initially thought that since J1 is receiving the 48V phantom-power and by how the OP was using J1 Pins 2 & 3 to extract the 48V from that, it made sense, although as Ian has pointed out.....that is apparently incorrect. Is there some other power source for this circuit that I have somewhere missed? I thought that the OP wanted this circuit to be powered -- BY -- the phantom-power, or have I misunderstood his intention? Can someone point this out to me? THANKS!!! (GEEZ!!! Growing old can be so confusing)!!!

This is the correct way to connect the XLR and power. V- is just ground.

 

[Chrisfromthepast]

USER=38655]@volker[/USER].
Does the placement of the 150R buildout resistor before or after the coupling capacitor make any difference? I moved it later on in the thread.

 

[volker]

USER=38655]@volker[/USER].
Does the placement of the 150R buildout resistor before or after the coupling capacitor make any difference? I moved it later on in the thread.

No, order of (passive) series elements does not matter.

 

[tinn]

An update of my drawing.
The old one would not work very well, cause 6 Volts are to low for the PNP noise source.
The Supply Voltage could be raised if you use a OP amp which draws less than the 2x4mA of the OPA 2134.
Driving the Noise source from the full Supply Voltage seems better.
Ground is Pin one of the XLR I would say.

 

[Chrisfromthepast]

I stole the output from here
https://www.opic.jp137.com/index-fet.html
@MidnightArrakis
Its kind of the same issue you had in this thread

 

[Chrisfromthepast]

Driving the Noise source from the full Supply Voltage seems better.

Yeah, thats easy enough.

Ahhh. Thanks. It feels good to be talking about the bjt again.
Bc550 still a good candidate for noise?
I only picked it because I have a bunch. Do the synth builders have a favorite type of junction for white noise?

 

[Chrisfromthepast]

I cannot properly credit this example circuit, but this is the original inspiration for using the bc550. If that helps.
Forget the phantom power/impedance balance stuff. Its different in our design.

 

[ruffrecords]

its phantom powered like a condenser mjc

Now you tell us! That makes so much difference. Forget all my previous comments about removing the 2K2 resistors.

Cheers

Ian

 

[Chrisfromthepast]

Sorry, I sometimes call it 48V nowadays because the consoles are so small, thats what the button says. Sorry, Ian, Im trying to not be as sloppy as usual.


I think everyone is happy, so next task is to pick some footprints, and start laying out the PCB.

 

[MisterCMRR]

The schematic in post 39 looks good. I did a quick calculation of V+ and would recommend changing R18 and R19 to 1 k-ohms. If supply current for the op-amps is the typical 4 mA each, V+ will be about +17 V. If at 5 mA max limit, V+ will be about +14 V. This should allow the base-emitter zener breakdown of Q1 to operate effectively. Resistors with a power rating of 0.1 W or more will be fine in every case.

 

[ruffrecords]

Just one question. What do you expect the output level of this pink noise generator to be?

Cheers

Ian

 

[Chrisfromthepast]

If the noise output of the bjt is around 30mv like it is in the example with 18v supply, that would be pretty close to mic level, and then gain it up by 11X, would put you somewhere around -7dbu. Once the noise output is measured, I'll tweak the feedback resistors to set the gain to get it to a nice neat output level. I was only intending to use this to test lines, not for any critical measurement. But, there's room for plenty of feature creep, if yall want to.

 

[ruffrecords]

So.....with me - NOT - being a "Circuit Designer" type of person, what I understand about this circuit is: Q1 creates some type of broadband noise signal that is amplified/buffered by U1B, which is then fed to U1A that has a few multi-band filters which then creates a type of "noise" that is defined and characterized as being "Pink Noise". The output of U1A is then fed to J1-3 as an "unbalanced" output. Does this sound remotely close?

That is basically right.

What is the maximum current draw allowed on a 48V phantom-powered circuit? I understand that it isn't very much.

14mA but that would give you zero volts. Basically the 48V is effectively fed to you via a 3K4 resistor so the more current you draw the lower the voltage becomes. Most designs aim for not more than 5mA which you drop the available voltage to 31 volts.

[What do you expect the output level of this pink noise generator to be?] -- Isn't that set by R17?

/

Thanks. I missed that.

Cheers

ian