Good book reading together - "Hardware Design Guide: From Device Cognition to Mobile Baseband Design"_Part 2 and 3

chrisrh

Good book reading together - "Hardware Design Guide: From Device Cognition to Mobile Baseband Design"_Part 2 and 3 [Copy link]

The second part mainly explains the power supply in the circuit

2.1. BUCK step-down power supply

Selection of inductance value: According to the output current I, the current ripple coefficient a () can be constrained to 30%, the switching frequency of the BUCK can be calculated, and a margin of 20~30% is left, and then a relatively close inductance is selected.

Saturation current of the inductor: 20~30% higher than the peak current flowing through the inductor (= output current + peak-to-peak current/2)

2.2 Linear power supply LDO

LDO must work in the saturation region (only in very rare cases will it work in the variable resistance region, at which time the regulation ability will be greatly reduced)

Use TINA to simulate the LDO of NMOS and the LDO of PMOS

2.3. Causes of negative pressure at BUCK switch node

The negative voltage generated by the body diode is characterized by low amplitude and no overshoot. If the BUCK routing and layout are not good, the loop inductance will increase, which will cause ringing at the SW and cause EMI.

Therefore, when designing the power supply, you can refer to the official layout suggestions.

2.4 Why is a resistor connected in parallel with the LDO output?

In order to further reduce power consumption, LDO will enter low power mode when it detects that the current is lower than a certain threshold. If it suddenly switches from low power mode to normal mode, abnormalities may occur if the response cannot keep up. The mitigation method is to connect a large resistor in parallel to the ground at the output end of the LDO. This resistor consumes current all the time. The current value consumed is just lower than the power consumption of the LDO and slightly higher than the threshold current of the normal mode. Then the LDO will stay in the normal mode instead of entering the low power mode, thus avoiding the problem.

2.5 PWM and PFM of Switching Power Supply

PWM: Pulse Width Modulation: The switching frequency is fixed or the switching period is fixed, rather than the duty cycle. Continuous switching

PFM: Pulse Frequency Modulation: The switching frequency is varied, or the switching period varies. Intermittent switching, non-continuous operation mode.

PFM is applicable: the load has different current consumption in different working states, and the difference is large, ranging from milliamperes to microamperes. The switching power supply has low efficiency at low load current due to its fixed frequency, so it is appropriate to use PFM mode at this time.

PFM ripple is larger than PWM mode

The third part is about: analog signal processing

3.1 ADC

The number of ADC bits affects resolution rather than accuracy

Least significant bit LSB: The higher the ADC bit number, the smaller the voltage signal that can be resolved: LSB = FS/(2^n-1), FS is the full-scale input voltage

3.2 Signal Analysis Basics

The signal-to-noise ratio (SNR) is the power ratio of signal to noise (effective value ratio) and is one of the important indicators for measuring system performance.

The noise floor of the FFT waveform will be different if the number of FFT points is different, because FFT has processing gain. This is the impact of the number of sampling points on FFT. When doing FFT analysis, it is best to keep the number of sampling points unchanged and then perform comparative analysis.

3.3 Modulation and Demodulation

Modulation is divided into amplitude modulation (AM) and frequency modulation (FM): spectrum shifting

3.4 Why is the system bandwidth defined as -3dB?

3dB is a doubling relationship, so the -3dB point is the half-power point. When G=-3dB, the power amplification factor Ap is exactly equal to 0.5. From the power perspective, 50% is used as the watershed to distinguish between the passband and the stopband to define the system bandwidth.

3.5. Filter first or amplify first?

Amplify first and then filter: After the signal is superimposed with noise, the signal-to-noise ratio is reduced, and it is difficult to distinguish the waveform when processing it again.

When processing weak signals, you can choose to filter first and then amplify to ensure the needs of subsequent circuits, which can meet most applications. Or the first-stage amplifier has a very small amplification factor, and its main function is to collect noisy signals with a very high input impedance, and then through signal processing, multi-stage amplification, filtering to ADC, convert analog signals into digital signals to achieve high-precision acquisition.

The first step in analyzing a problem is always to reproduce the phenomenon.

See the essence through the phenomenon~