Receiver interview question

[yolande_yj]

for receiver input matching, do you want to step up the impedance or step down, why?

my answer: step up. because 1. higher impedance lower current and thus less noise; 2. higher impedance can do filtering easier

I am not sure, am I right? what is the correct answer?

 

[G4BCH]

Too vague to answer, depends on the input impedance of the first device. Assuming 50 ohms source a common source FET, step up, common base bipolar might need stepping down or leave at 50 ohms.

 

[yolande_yj]

it was asked in this way, no more details. Initially I said if the source is 50 ohm, and the receiver input is a CMOS LNA, then step up. but that seems not satisfying the interviewer. so I guess this is a general question and gave the answers in my first post.

 

[G4BCH]

If it is to an LNA you match to what ever impedance the device needs to see for best noise match.

 

[yolande_yj]

yes, I totally understand this as I design LNA before. but I felt like this question is to test other general knowledge on receiver not only LNA, so I gave my above answer. looking back now I think this is a trick question, something like there are 10 birds on the tree, you shot one, how many left……

 

[vfone]

If resume the situation only at two cases, which is better in a receiver, low or high impedance, the answer is low impedance, because by definition low impedances are desirable in low noise circuits. And noise is one of the most important characteristics of a receiver (and its stages).
Lower impedances can be applied in filter design also.

 

[yolande_yj]

Thanks vfone, can you explain why low impedance is desirable for low noise circuits?

--- Updated Feb 13, 2022 ---

My understanding is for the same power, higher impedance has lower current . CMOS thermal noise is proportional to gm, and thus square rote of Id. so lower current is better for noise

 

[vfone]

The most influential and limiting minimum noise at a given temperature in a circuit is the Thermal Noise.
The thermal noise is dependent only on the resistive component and is independent of any reactance in the circuit.

According to Nyquist equation, for a fixed temperature, the thermal noise voltage in a circuit can be reduced by minimizing the resistance (and the bandwidth).
Always: lower the resistance, lower the noise.

For example, a 100k resistor placed in front of a CMOS transistor will generate more noise in the circuit than the transistor itself, whatever its gm is.

Thermal Noise in Electronic Circuits Explained »Electronics Notes

Understand how thermal noise occurs in electronic circuits, how it is related to resistance and how to calculate it for various resistor values and temperatures.

www.electronics-notes.com

 

[biff44]

the correct answer was probably "neither".

Since some amplifiers need a specific input impedance, that is often something other than 50 ohms, you might have commented that if LOW NOISE FIGURE was of paramount importance, you would have tried to match the antenna to the reciever input using the optimal noise figure match impedance.

I would have mentioned that such a noise figure match might not be advised if it is a broadbandwidth receiver, where phase and amplitude ripple vs frequency was an issue, since a noise figure match is often a single frequency point match

you might have mentioned that a broadband 50 ohm match would be preferred if you needed to add a bandpass filter between the antenna and reciever front end, as often you need to do for interference reasons. for filters to work properly, they need a fixed source and load impedance, and are often designed for that impedance to be 50 ohms

In other words, answer in a way to shows you understand some of the potential issues with designing a good modern receiver front end.