3

I have some basic questions regarding the circuit shown below:

circuit

The image is from this question.

The buffer is sending signals through a twisted pair cable to an in-amp. My questions are:

  1. Does the buffer have other functions than to increase the signal strength? Maybe a transformation from high impedance to low impedance?

  2. If so, is the low impedance after the buffer important to avoid noise through external magnetic fields? Would it be wrong to have a long distance between the sensor and the buffer, because the impedance between senor and buffer is high?

  3. I have read impedance matching is important. But the output of the buffer is low and the input of the in-amp is high. Is it ok?

Kuba hasn't forgotten Monica
  • 32,734
  • 1
  • 38
  • 103
R.Hood
  • 31
  • 2
  • @R.Hood In the future, when asked for clarification in comments, please: 1. Respond by editing the question - it's deficient if someone asks for clarification, it means it should be improved. 2. Flag the comment as "no longer needed", if you have good reason to believe that you've addressed the comment. – Kuba hasn't forgotten Monica Mar 28 '22 at 18:55
  • 4
    The op amp on the left is not a buffer, it's a transimpedance amplifier – Scott Seidman Mar 28 '22 at 19:21

2 Answers2

3

Does the buffer have other functions than to increase the signal strength? Maybe a transformation from high impedance to low impedance?

It's usually a good idea to convert a high impedance to a low impedance when sending an analogue signal down a cable to a receiver amplifier. It usually cuts down on noise influencing the signal and reduces the differential noise error that would be received by the InAmp.

Is the low impedance after the buffer important to avoid noise through external magnetic fields?

The important thing here is that the impedance presented by the driver output is matched by the ground connection to the other wire in the twisted pair. In other words, the driven end of the cable is impedance matched in both wires prior to connecting to the cable.

This ensures that external magnetic and electric noise/interference fields that might "hit" the cable will produce identical voltages on both wires in the twisted pair. If the EMI voltages are the same on both wires then the InAmp (or differential amplifier or balanced receiver) can subtract them to zero.

Would it be wrong to have a long distance between the sensor and the buffer, because the impedance between senor and buffer is high?

It's not ideal.

I have read impedance matching is important. But the output of the buffer is low and the input of the in-amp is high. Is it ok?

Impedance matching can mean several things. In this scenario, you might need cable termination impedance matching (for high speed signals) but, you will also need balanced driver matching at the transmitter (as described above) and a balanced impedance at the receiver.

Regards the transmit end, it might be a good idea to put series resistors in the feed leads like this: -

enter image description here

The reason should now be clear; the output from the driver may be a few ohms whereas the other wire in the twisted pair is "hard" connected local ground. This is not ideal and the two resistors shown as "R" will "balance" driving impedances and reduce the impact of noise due to equalization.

Balancing the driver and receiver is all about making sure that whatever noise influences the wires in the twisted pair, it influences both wires identically. The InAmp will do the rest.

See also this Q and A, particularly my answer.

enter image description here

Andy aka
  • 434,556
  • 28
  • 351
  • 777
2

  1. Does the buffer have other functions than to increase the signal strength?

    The buffer doesn't merely increase the signal strength. It entirely converts the type of the signal: from a current signal to a voltage signal. It also biases the photodiode with a constant voltage, decreasing its effective capacitance and thus speeding up the sensor. The photodiode is effectively shorted, with a zero voltage across it. So, the input voltage to the buffer is almost exactly zero, but the current varies. The output voltage of the buffer varies a lot, and the current varies proportionally due to capacitive loading or line termination.

  2. Would it be wrong to have a long distance between the sensor and the buffer?

    Yes, because the signal from the photodiode to the buffer is low-level and thus much more sensitive to interference.

    The buffer amplifier essentially shorts the diode: it acts as an ideal current measurement shunt. Thus diode-to-buffer circuit is very low impedance. It is not sensitive to external electric fields, but is very sensitive to external magnetic fields. The lower the impedance of a circuit, the more sensitive it is to magnetic fields. An ideal magnetic field sensor is a superconducting short circuit loop of infinite area :) This sensitivity is proportional to the area of the loop formed by the photodiode circuit, and inversely proportional to the resistance. If the photodiode were to be far from the buffer, a nearby source of magnetic fields, such as a mains transformer or a parallel cable carrying AC currents, would easily induce interfering signal into the diode circuit.

    On the other hand, high impedance circuits are sensitive to external electrical fields, but not much sensitive to magnetic fields. As the impedance of the circuit rises from a short to an open, the sensitivity shifts from magnetic to electric fields.

  3. I have read impedance matching is important. But the output of the buffer is low and the input of the in-amp is high. Is it ok?

    It is OK as long as the highest frequency seen in the signal output in the buffer has a wavelength much longer than the length of the cable. If the wavelength of highest frequency signal is shorter than 5-10x the cable length, then the cable must be terminated on both ends. It then needs a series resistor between the buffer output and the cable, and a parallel resistor across the in-amp input. The resistance should be the same as the characteristic impedance of the cable used.

Kuba hasn't forgotten Monica
  • 32,734
  • 1
  • 38
  • 103
  • Thank you too. If the sensor would be a high impedance sensor like a hall sensor for current measurement, would it be still good to lower the impedance with the buffer before sending it through a long cable in a noisy environment with lot of magnetic fields? – R.Hood Mar 28 '22 at 19:21
  • "It also biases the photodiode with a constant voltage, decreasing its effective capacitance and thus speeding up the sensor. " Sorry, but not in this case. In fact, the photovoltaic mode being used eliminates offset voltage at the expense of speed. This configuration is the slowest possible (for a given transconductance). It forces the PD voltage to zero. – WhatRoughBeast Mar 28 '22 at 19:21
  • Kuba thank you. Wasn't aware of the different sensitivity to external electric and magnetic fields of high or low impedance circuits. – R.Hood Mar 28 '22 at 19:43
  • @WhatRoughBeast The photodiode doesn't have reverse bias - it would be slower there. But just see what happens when the op-amp open loop voltage gain goes down: the photodiode capacitance sees progressively more voltage proportional to photocurrent, and gets slower. So it's not the slowest for sure. An amplifier with a voltage (as opposed to current) input would be in the slowest class, as far as I can see it. Just replace the photodiode with an ideal capacitor and imagine what's happening. – Kuba hasn't forgotten Monica Mar 28 '22 at 19:56
  • No, that's not true. In either case, the voltage across the PD remains constant, due to negative feedback. The frequency response for a given transconductance (resistor value) is determined by the feedback capacitor. The smaller the cap the faster the TIA. And for a given resistor, the larger the input capacitance the larger the feedback cap has to be to keep the circuit stable, and the slower the response of the circuit. A PDs capacitance is a maximum at zero volts. – WhatRoughBeast Mar 29 '22 at 00:10
  • One final question. Output impedance of the PD is high, input impedance of the transimpedance amplifier TA is low. It's very sensitive to external magnetic fields. Can we generalize it with the follow statements? (A) If both (source and sink) are high impedance the circuit is sensitive to electric fields. (B) If one of them (source or sink) are low impedance the circuit is sensitive to magnetic fields. Same if both are low impedance. – R.Hood Mar 29 '22 at 06:00
  • @R.Hood A low impedance circuit (loop) is not low impedance if it contains high impedances in it. If you short the legs of a 1MOhm resistor, the resistance in the loop is still 1MOhm. So, connecting a high impedance (1MOhm) to a low impedance load (a shiort) doesn't remove the high impedance. No high current will flow in that loop, since it's broken with the high impedance. – Kuba hasn't forgotten Monica Mar 29 '22 at 20:01