I would like to explore the ground plane just a little. It seems the ground is considered as a fixed reference at all points. In low frequency circuits we can ignore the ground plane but in RF we must consider strays effects including how the ground acts. As the frequency increase the wavelength decreases and the circuit layout becomes more critical. First let's look at a couple of datasheet references.
Note pins 4 and 8 go to ground. Someone ask why didn't they just tie them together inside the IC? That was a very good question. Now a look at the recommended hookup.
My friend connected 4, 7, and 8 at the chip. The manufacturer has pin 4 going way over there and connecting to the input circuit while 7 and 8 are connected with the output. Why? Consider a length of wire suspended in space will have inductance and capacitance. This is where to sage advice "keep your leads short and close to the ground plane." came from. The ground plane connecting the input and outpoint grounds will effect the circuit. A signal flowing between the two points will have a phase shift (time delay) simply because the distance between the two points represents a fraction of a wave length. So as frequency increases the strays become more prominent. More on this later. (don't want to start chasing rabbits.)
What if I was wiring this circuit and one component lead was to short but the other was long enough. Why not do this? As it says we have a common current path for two signals.
If C2 is bypassing R1 and we make that connection we add L1 into our circuit. C1 is not bypassing R2 it is forming a series circuit with L1 that could introduce instability. At some frequency C1 and L1 will be at resonance and possibly cause oscillations.
So when I connect the circuit as shown doesn't it introduce two coils into the circuit? Even if they couple together the effect is weaker. Maybe more on that later.
It's a lot in a short time but I hope this is enough to show the ground plane is not universal.
In a future post we may look into a circuit that uses this info to inject inductors into the design.
Note pins 4 and 8 go to ground. Someone ask why didn't they just tie them together inside the IC? That was a very good question. Now a look at the recommended hookup.
My friend connected 4, 7, and 8 at the chip. The manufacturer has pin 4 going way over there and connecting to the input circuit while 7 and 8 are connected with the output. Why? Consider a length of wire suspended in space will have inductance and capacitance. This is where to sage advice "keep your leads short and close to the ground plane." came from. The ground plane connecting the input and outpoint grounds will effect the circuit. A signal flowing between the two points will have a phase shift (time delay) simply because the distance between the two points represents a fraction of a wave length. So as frequency increases the strays become more prominent. More on this later. (don't want to start chasing rabbits.)
What if I was wiring this circuit and one component lead was to short but the other was long enough. Why not do this? As it says we have a common current path for two signals.
If C2 is bypassing R1 and we make that connection we add L1 into our circuit. C1 is not bypassing R2 it is forming a series circuit with L1 that could introduce instability. At some frequency C1 and L1 will be at resonance and possibly cause oscillations.
So when I connect the circuit as shown doesn't it introduce two coils into the circuit? Even if they couple together the effect is weaker. Maybe more on that later.
It's a lot in a short time but I hope this is enough to show the ground plane is not universal.
In a future post we may look into a circuit that uses this info to inject inductors into the design.
No comments:
Post a Comment