OPA657: PSRR decreasing when use resisotr divider as voltage reference

Hi Follin,

take care, voltage references can generate lots of noise and can totally ruin the noise performance of your TIA. The LM4120 shows 20µVpp noise in the 0.1Hz...10Hz band. A 1k resistor, on the other hand, will only show a noise of 85nVpp in the same band.

A two stage low pass filter may help and/or a way larger filtering time constant. If the input bias current of TIA OPAmp is too high, a low noise, low input bias current buffer can be added, best followed by an additional low pass filter.

If the exact bias voltage plays no role, for instance because the bias voltage is also taken into consideration by an ADC, a simple capacitance multiplier arround a NPN transistor could do, as discussed in this thread:

https://e2e.ti.com/support/amplifiers-group/amplifiers/f/amplifiers-forum/1173311/gras-47ac-preamp-circuit

(In order to provide a voltage division, you may need to add a resistor or Z-diode in parallel to C1.)

Capacitance multipliers are known to be ultra low noise:

https://www.tnt-audio.com/clinica/regulators_noise3_e.html

Kai

Hi Kai,

Capacitance multiplier is really a great idea. I'm going to try it on my board.

I have several questions:

1) Is the "exact bias voltage" you mentioned related to the voltage drop of the current block diode and NPN since the Vbe of T1 is not a accurate/constant voltage? And I think the diode maybe removed.

2) You said the 100uF capacitors shall be Al-cap in the thread you posted. But I have limited space so I might use Tantalum or just Ceramic (1210) capacitors. Will it affect the result a lot?  (Voltage is below 5V here)

3)  Currently I have 8 reference voltage on one board. So  if I want to be able to adjust them individually, may I just use several voltage devider after the multiplier? For instance:

4) I also made a negative one, is there somthing else I need to care?

Thank you so much.

Hello Follin,

For your 2nd question: you should be fine to use alternative capacitor types to Al-Electrolytic; you should be aware of the various specifications and aspects of different capacitor types.  Kai likely had good reason to recommend a specific capacitor type; my thoughts are that electrolytic makes more sense for higher capacitances due to cost.  If you would use a 1206 ceramic with the desired capacitance and rating, that might be just fine. Please investigate this for yourself if you do not have a good understanding of capacitor types & their differences.  

Please see these additional resources also:

Project with 'negative' capacitance multiplier for negative supply rail: https://sound-au.com/project15.htm

Audio Xpress Article on Capacitance Multipliers: https://audioxpress.com/article/the-capacitance-multiplier

These articles might help answer your 3rd & 4th questions.  

I will let Kai take a try at your 1st question, as it pertains to his response.

Best,

Alec

Hi Kai,

Thank you! I learn new things. I haven't noticed the microphonic effect. I will do some test later.

And for the protect of the input of the OPAmp, I modified the circuit a bit and so far I have the final one below.

It seems using more components now...but I still need the opamp to avoid voltage division with the resistors around the FDA.

Hi  Alec,

Yes I will try with different capacitor types. And thanks for your additional resource! I modified my circuit a bit.

Hi Follin,

I should explain why my earlier capacitance multiplier was designed that way:

The original circuit is a microphone amplifier and the first stage of a microphone amplifier and the microphone itself is often powered by a battery. When disconnecting or connecting a battery or replacing an old battery by a fresh one in the field a short-circuit or even polarity reversal can easily happen. That's why I have added D4. Because the supply voltage is with 18V rather high, the additional voltage drop across D4 would not harm.

But D4 isolates the input of capacitance multiplier from the low source impedance of battery now. Because this may cause stability issues I have added C8. And to protect the battery from high inrush currents into C8 I have added R12.

In your circuit there's no need for D4, I guess. And because of that you would not need R12 and C8 either. And because the low pass filtering time constant of R12 and C8 is much smaller than R1 and C1, you would not need the low pass filter formed by R12 and C8 either. Remember, R12 and C8 weren't introduced for low pass filtering but for stabilizing T1 and limiting the inrush current.

So, you could omit D4, R12 and the big aluminium electrolytic cap C8. A simple 100...470nF for C8 should do.

Omitting D4 would also help to decrease the voltage fluctuations at the output of the capacitance multiplier. Only the base emitter voltage of T1 would contribute to the voltage changes, but no longer D4. The voltage fluctuations are usually caused by varying loads (which is not the case in your circuit ?) and temperature changes. The expected voltage fluctuation is about -2mV/°C.

To be continued...

Kai

Hi Kai,

Thank you so much for your detailed explanation, it's really helpful!