示波器采样率设置多少比较合适？

采样原理

尼奎斯特采样定理规定，对于具有最大频率f_MAX且带宽有限的信号，等距采样频率f_S必须大于最大频率f_MAX的两倍，这样才能唯一地重建信号而不会产生混叠。

示波器带宽和采样率

什么是示波器的采样率?

采样是将一个信号（即时间或空间上的连续函数）转换成一个数值序列（即时间或空间上的离散函数）。示波器采样频率，又称采样速度或采样率，定义了每秒从连续信号中提取并形成离散信号的采样数量，它使用赫兹（Hz）来表示，采样频率倒数为采样周期或采样时间，即采样间隔，一般来说，采样频率是指计算机每秒采集多少个信号样本。

示波器带宽通常是指按3dB（-30%幅度误差）衰减输入信号正弦波的最低频率。对于示波器带宽，采样原理规定所需的采样率为f_S=2f_BW。但该原理假设频率分量都在f_MAX（在此情况下是f_BW）以下，并且需要具有理想的砖墙频率响应的系统。

理论上的砖墙频率响应

然而，数字信号的频率分量在基本频率（方波由基本频率的正弦波和数量不限的奇次谐波组成）之上，通常，对于1GHz及以下带宽，示波器具有高斯频率响应。

采样率和示波器带宽

因此，实际上，示波器的采样率应为其带宽的四倍或以上：f_S=4f_BW。这样就会减少混叠，并且混叠的频率分量会出现更大的衰减量。

混叠

当信号采样不足 (f_S<2f_MAX) 时，将发生混叠。混叠属于信号失真，是由于错误地从数量不足的采样点重建低频率而导致的。

混叠

采样率与混叠现象

• 采样率不足会导致信号的混叠失真

• 不会发生混叠的前提是：采样率>信号的最高频率的两倍(Nyquist采样定律）

• 实际应用中：采样率>所需最小示波器带宽的4倍（Sin(x)/x内插）

• 采样率会受时基和存储深度的影响会发生该变

数字示波器的三种采样方式

实时采样(Real Time)

等效时间采样(Equivalent Time)

顺序采样(Sequential Sampling)

实时采样技术(Real-time Sampling Technology)

实时采样/Real-time Sampling

In real-time sampling, the signal is digitized (on the fly) in real time; all data is acquired on a single trigger event. This fact is represented in the figure by the number "1". Therefore, there is a simple 1:1 correspondence between the samples and the times at which they were taken. Real-time sampling is needed to capture signals that absolutely do not repeat. The time between samples is called the sample interval. The sample interval is 1/ sample rate. For example, at 1 GSa the sample interval is 1ns. In the real time mode, the resolution is the sample interval.

ADVANTAGES: Pre-trigger viewing is possible as long as the scope is set up to acquire data before the trigger event is actually received.

实时采样：单次触发完成所有样点的采样。

数字带宽 = 采样速率/4

• 能够测试重复或单次（瞬态）信号

• A/D采样速率决定示波器的最大采样率

• 具有“预触发”功能，即既能观察触发点之后的信息，也能观察触发点之前的信息

• A/D转换速率决定示波器的数字带宽

如何计算你所需要的最低采样速率?

1.知道你的信号的最快上升时间, tr

2.计算信号的截止频率

Fknee = 0.4/tr (20% to 80%)

Fknee = 0.5/tr (10% to 90%)

3.计算需要的示波器带宽, BWscope

BWscope =  1.9*Fknee (高斯频响示波器)

4.计算需要的最低采样速率, SRscope

SRscope =  4*Bwscope  (高斯频响示波器)

实时采样特点：

•在观察示波器的波形的时候，示波器的实际采样率可能会远远低于标称的最大采样速率

•实际采样率会受到时基的设置和存储深度的影响（后面详细讨论）

•采样率下降会造成示波器系统带宽的下降

系统带宽 = Min（前端模拟带宽，采样速率/4）

采样率越高越好吗？

--高于4倍带宽的采样率，并不能提高测试快信号与毛刺的能力
--过高的采样率，反而会浪费宝贵的存储资源

等效时间采样技术/Equivalent-time Sampling

In equivalent-time sampling it takes multiple trigger events to build up the waveform. In the figure shown, only 3 points are taken per trigger event. Clearly, it would take longer to build up this waveform, compared to the situation in which 50 points/trigger event are acquired. The trigger event is asynchronous to the sample clock; thus each trigger yields a group of samples that are shifted in phase with respect to other groups of samples.

ADVANTAGES: Bandwidth is NOT limited by sampling frequency; BW is limited by analog path (mostly by the preamp chip); pulse performance is typically better than with real-time sampling; timing resolution (which is a function of the trigger interpolator) is typically very good.

等效时间采样特点：

•只能用来测量重复信号

•多次触发建立示波器的波形

•单次采样率并不关键

•触发电路的好坏影响时间分辨率

等效时间采样的波形重建过程

顺序采样技术/Sequential Sampling Technology

In sequential sampling, only ONE sample of the signal is digitized on each trigger. With each successive trigger, the sampling point is delayed further from the trigger point. After a full screen of samples (about 500) are acquired and digitized, the signal is moved from acquisition memory onto the display. In sequential sampling the signal MUST be repetitive. In the 54750 Series the resolution is 62.5 fs.

ADVANTAGES

Sequential sampling does NOT miss repetitive glitches unless they are less in width than the fine interpolator's time resolution. Bandwidth is NOT limited by sampling frequency, but rather by the analog front end of the scope. Because the A/D's used for sequential sampling aren't run at high speeds, they typically are higher resolution successive approximation converters. Timing resolution is better here, compared to real-time sampling.

顺序采样特点：

•只能用来测量重复信号

•一次触发采集一个样点

•多次触发建立波形

•通常用在带宽>10GHz的采样示波器上

•没有预触发信息