RF Ground - RF is not what you think "ground" is

btty038

RF Ground - RF is not what you think "ground" is [Copy link]

We learned about voltage in junior high school physics. Voltage is also called potential difference, which is the difference between any two potentials. It is generally believed that the potential at infinity is 0. The implicit condition of "voltage" is the difference relative to 0 potential, but this "0 potential at infinity" is too vague. It is impossible to detect it during actual circuit analysis. Fortunately, voltage has nothing to do with absolute potential. It is just a difference. Therefore, we can arbitrarily choose a potential in the circuit as a reference point. If we artificially define it as the 0 potential point, then this reference point is called "ground".

After the "ground" is artificially defined, the voltage at any point in the circuit has a reference. Saying "the voltage at one end of the resistor is 5V" actually means "the voltage at one end of the resistor to the ground is 5V", but if you say "the voltage at both ends of the resistor is 5V", it is clear that one end of the resistor is used as a reference, not the "ground".

According to different application scenarios, "ground" is artificially subdivided into "power ground", "digital ground", "analog ground" and "RF ground". There are many articles and theoretical analyses on how to deal with different "grounds" in the circuit. This article only briefly discusses some of my understanding of "RF ground".

In the basics of analog circuits, transistor amplifiers are discussed. When analyzing the DC operating point, the power supply voltage is taken as a given value. However, when drawing an AC equivalent circuit to analyze small signal characteristics, the power supply is treated as equivalent to the ground. Why is this?

The analog electronics teacher told you that because any signal can be decomposed into a superposition of a DC quantity and an AC quantity, that is, the "superposition theorem" , when analyzing the circuit, its DC and AC characteristics can be analyzed separately, thereby simplifying the difficulty of circuit analysis. The power supply in the circuit needs to be "source-removed", that is, when analyzing DC, only the DC power supply is retained; when analyzing AC, only the AC power supply is retained. The "Norton equivalent" and "Thevenin equivalent" theorems are mentioned in the basics of circuit analysis. The internal resistance of an ideal voltage source is 0, and the internal resistance of an ideal current source is infinite. Therefore, whether it is a DC power supply or an AC power supply, the ideal voltage source is equivalent to a short circuit, and the ideal current source is equivalent to an open circuit. For a non-ideal voltage source with internal resistance, the power supply is short-circuited but the internal resistance is retained; for a non-ideal current source with limited internal resistance, the power supply is open-circuited but the internal resistance is retained .

To sum it up, the voltage source is short-circuited and the current source is open-circuited , but it never says what does it have to do with "ground"? That's right, " removing the source" does not necessarily mean grounding! The problem is that voltage sources are usually used to power circuits. For example, the VCC mark in the figure above, although not explicitly represented by a voltage source, specifies a fixed voltage supply, so it should be understood as a voltage source. The voltage of this voltage source is defined relative to the "ground". Intuitively, the positive end of the power supply is connected to VCC and the negative end is grounded. The voltage source is short-circuited during AC analysis, so in this case it degenerates to "the power supply is equivalent to grounding . "

Let's look at an example where a 0.5V sinusoidal voltage source and a 1V DC voltage source drive the same resistor at both ends:

The DC voltage and DC current of each node are marked in the figure above. It can be seen that the DC voltage at point A is 0, which is equivalent to grounding, and the DC current is 1A, which is only related to the DC power supply and the resistance value. Let's look at the time domain waveform in the figure below. The DC component of the total current of the resistor is 1A, and the AC component is a sine wave of 0.5A.

It is precisely because all signals are defined with respect to the ground that an imperfect reference to the "ground" is equivalent to introducing noise or interference, so single-ended signals are more easily affected by the "ground"; differential transmission is to weaken the influence of the "ground" and thus support signals with much higher rates/frequencies than single-ended transmission. But please note that differential transmission only theoretically removes the influence of the ground. In real circuits, there is always a "ground", and even in differential form, there is always common-mode interference, and common-mode interference will also convert to differential mode. Therefore, the processing of the "ground" is a very important part of circuit design, especially for circuits with higher operating frequencies and faster transmission rates .

In practical application circuits, almost every power pin of an independent device and different power pins of the same device will be connected to several capacitors nearby. What is the reason for this? According to the different focuses of these capacitors, people have given them different names: filter capacitors, energy storage capacitors, bypass capacitors, decoupling capacitors, and AC grounding capacitors. Some articles specifically distinguish different meanings under different names. In the author's opinion, no matter what name is used, the essence of the capacitor has not changed, but the angle of looking at the problem is different. For example, the filter capacitor is to filter out the power ripple or the noise leaked from the device. From another perspective, it is to allow the AC ripple and high-frequency noise on the power supply to be bypassed to the ground through the capacitor . When the power pins of different devices are each connected to appropriate capacitors nearby, the noise on the power supply will not easily interfere with each other through the power supply network, the so-called "decoupling"; and while the energy storage capacitor meets the instantaneous large current of the load, can't it be understood as stabilizing the power supply voltage? Can't it be understood as filtering out instantaneous pulse interference for the power supply? Can't it be understood as allowing the pulse interference to be bypassed to the ground through the capacitor? No matter which statement is made, isn't it essentially to keep the power supply voltage as an ideal DC invariant as much as possible? It is precisely because the power supply in reality is not ideal, the internal resistance is not 0/∞, and there are various parasitic resistance and capacitance in the circuit routing that we use capacitors of appropriate values to help give AC components of different frequencies an ideal "ground" that is as close as possible; once the AC components are handled well, the DC will be "pure".

However, the capacitors in reality are not ideal either, and there are also various parasitic parameters. These parasitic parameters are not obvious at DC or low frequency, but they will become more prominent as the frequency increases, or even lose the properties of capacitors. Therefore, in order to make the capacitor still play the role of "capacitor" at different frequencies, it is necessary to select suitable capacitors. The parasitic parameters of capacitors are related to the manufacturing materials, package type, and package size. Generally, the larger the capacitance and the larger the size of the capacitor, the lower the applicable working frequency. In the power supply network, multiple capacitors of different sizes must be selected in parallel for AC grounding of different frequency bands. The characteristics of the circuit are almost always closely related to the signal frequency/wavelength. The wavelength of high-frequency signals is shorter, and it is easier to reach the physical size of the circuit and leave the lumped parameter approximation and enter the category of distributed parameters. Therefore, the "ground" with the highest frequency should be given priority. Therefore, the capacitor with smaller capacitance and higher operating frequency is placed closer to the device power pin or the position where AC grounding is required. The following examples are all true.

The application circuits of several amplifiers here remind me of an experience I had when I first came into contact with MMIC design: a designer with many years of experience told me to connect 100pF, 10000pF, and 10uF capacitors to the power pins in accordance with the actual application situation when simulating the amplifier circuit. I certainly agree with this point of view, but his simulation circuit is like this, which I dare not agree with:

If you don't find any problem with the simulation circuit diagram above, you can find a random transistor model to verify whether adding or not adding the capacitors C3~C8 has no effect on the circuit performance. However, after inserting the ideal DC_Feed (only DC, no AC, equivalent to infinite ideal inductance) between the power supply and the capacitors in the following circuit, the circuit performance is completely different with or without C3~C8. Why is this? It has been explained in the previous article.

It is precisely because of the AC grounding of the power supply in theory and practice that the status of power supply and ground in high-frequency circuits is equal and should both be treated according to "AC ground" .

So remember: RF is not what you think it is !

What do you think about RF? Welcome to leave a message!

MartinFowler

Thanks to the host. After all these years, I still don't really understand the meaning of "land". I just know that the land should be separated.

In fact, these basic concepts should be made clear.

btty038

MartinFowler posted on 2020-5-29 11:19 Thanks to the host. After so many years, I still don’t really understand the “land”. I just know that the land should be separated. In fact, these basic concepts...

All it takes is a little professional guidance. This kind of thing also requires experienced people, professional knowledge, popular metaphors and pictures. Only we rookies can understand it well.

led2015

I have always known some basic knowledge, but every time I do PCB routing, I always forget the importance of ground.

btty038

led2015 Published on 2020-5-29 23:36 I have always known some basic knowledge, but every time I do PCB wiring, I always forget the importance of ground

It is still quite particular, especially for RF. If it is not done well, the impedance will change.