1

This is a generic question for about frequency sensors. But as an example I use this sensor which has around 1k output impedance. So in below drawing the source in dashed box has 1k output impedance:

schematic

simulate this circuit – Schematic created using CircuitLab

The voltage signal from the source to the receiver will be transmitted by STP cable. And the receiver is a sort of differential input instrumention amplifier.

But the system is not 100% balanced because of the 1k output impedance.

I think the best way is to buffer both signal and ground of the output for balancing but this would introduce some noise.

Assuming GND has very low impeance, In Figure 1 the 1k is lowered by a buffer. And on the other hand in Figure 2 the imbalance issue is tried to be solved by adding a 1k resistor on the ground output.

My questions are:

1-) If I use a resistor as in Figure 2, then I would increase DC error but I can calibrate it out. In that case can we say that Figure 2 is a better solution than Figure 1? I'm asking because in Figure 1 the buffer circuit will introduce noise.

Which one is better than the other?

2-) Would you measure the exact value of the output impedance or would you rely on the datasheet?

user1245
  • 4,125
  • 4
  • 38
  • 95
  • 2
    You talk about ground but you don't show it in your pictures and, neither do you present the rogue effects such as extraneous noise or EMI. It's not clear what you are asking. – Andy aka May 24 '18 at 12:00
  • The sensor is floating wrt Diff ended amplifer's AI GND. Im only focusing on how to balance the sensor properly without introducing noise by using extra electronics. Imagine diff amp is an isolated daq board channel. – user1245 May 24 '18 at 12:02
  • Yes 100% balance system for common mode EMI. But for that the source need to be balanced. But in this case it has 1k output impedance. – user1245 May 24 '18 at 12:04
  • You need to be clear in your question and not rely on comments for addressing shortfalls. People don't read comments so much. – Andy aka May 24 '18 at 12:09
  • If you want to drive a differential input over STP, why wouldn't you use a proper differential line driver? They're readily available. Then it doesn't matter whether the actual signal source is balanced or not. – Dave Tweed May 24 '18 at 13:04

1 Answers1

0

The sensor is floating wrt Diff ended amplifer's AI GND

If it is truly floating and presents a high impedance to ground then there is nothing wrong with scenario 1 (buffer amp not required) because the 1 kohm can be regarded as being two 500 ohm resistors; one in each leg. If you regard the source as a black box then you cannot differentiate between one leg having 1 k, the other leg having 1 k or both having 500 ohms.

However, the devil is in the detail and nothing is as simple as this scenario BUT without further information then that is all that can be said.

Andy aka
  • 456,226
  • 28
  • 367
  • 807
  • Figure 1 will introduce noise due to buffer interface. What about Figure 2? (Adding 1k to GND line of the sensor.) Would that balance the lines? Or any side effects? – user1245 May 24 '18 at 12:24
  • Sorry I meant use figure 1 but without the buffer. I shall amend my answer. – Andy aka May 24 '18 at 12:34
  • I though the lines must be balanced 100% for all cases for CM noise. – user1245 May 24 '18 at 12:48
  • If the source is floating and presents high AC impedance to ground (and these are big "ifs") then balancing is inherent and you can regard the 1 kohm as being equally split into 2x 500 ohms either side of the internal voltage source. – Andy aka May 24 '18 at 12:49
  • Very interesting.. Would you mind if I ask another question about it and you might answer with couple of diagrams to explain why? Im really curious – user1245 May 24 '18 at 12:54
  • Please go ahead and ask. – Andy aka May 24 '18 at 13:01
  • Here I just opened: https://electronics.stackexchange.com/questions/376184/output-impedance-imbalance-and-cm-noise-in-truly-floating-sources – user1245 May 24 '18 at 13:21