信号经过高通滤波之后存在直流偏置

Question

各位工程师好！ 
 本人在使用LTspice仿真时遇到一个问题，具体与高通滤波器有关，仿真电路如下：

电路包括电压跟随器、高通滤波器、反相放大器、超前校正电路。原始输入设置为频率为21kHz，幅值为10mV的正弦波，V(n208)是经过电压跟随器后的输出，可以看出其对于原始输入的跟踪效果很好，几乎没有直流偏置。反而经过高通滤波器后 V( n 211)出现了约-3mV的直流偏置，如下图所示： 
 
 后续又给原始输入添加了10mV的直流偏置，V(n208)是经过电压跟随器后的输出，可以看出其对于原始输入的跟踪效果很好，直流偏置也是10mV。经过高通滤波器后V(n211)还是出现了约-3mV的直流偏置，如下图所示： 
 
 这个问题就好像是无论原始信号有多少直流偏置，经过高通滤波器之后都会存在-3mV的直流偏置？理论上经过高通滤波器之后的直流偏置不应该很小吗？请问大家这是怎么回事呢？谢谢。

MarkThoren · Accepted Answer

Hi LANVictory, 
 I'll have to answer in English, Outlook's translation of your question pasted below. Note that there is 中文社区 as well. 
 The translation: 
 Hello engineers! I encountered a problem when using LTspice to simulate, specifically related to the high-pass filter, and the simulation circuit is as follows: The circuit includes a voltage follower, a high-pass filter, an inverting amplifier, and a leading correction circuit. The original input is set to a sine wave with a frequency of 21kHz and an amplitude of 10mV, and V(n208) is the output after passing through the voltage follower, which shows that it tracks the original input very well and has almost no DC bias. Instead, after passing through the high-pass filter, V(n211) has a DC bias of about -3mV, as shown in the figure below 
 MarkThoren: We can't see the node names that you're referrring to. But I take it that the output of U6 is correct - low offset? Indeed C18 will remove any DC offset, or more accurately, will allow an arbitrary DC voltage to appear between its terminals. If you are measuring the output of U7, then have you considered the offset of that part? It is not manufactured by Analog Devices so we cannot comment on its performance. Where did the LTspice model come from? Have you considered the datasheet offset of that part, as well as the effect of its bias current? If your signal frequency is 21 kHz, why do you need a 2 GHz amplifier? Subsequently, a DC bias of 10mV was added to the original input, and V(n208) was the output after passing through the voltage follower, which showed that it had a good tracking effect on the original input, and the DC bias was also 10mV. After passing through the high-pass filter, V(n211) still has a DC bias of about -3mV, as shown in the figure below: The question is like a DC bias of -3mV after a high-pass filter no matter how much DC bias there is in the original signal? Theoretically, shouldn't the DC bias after a high-pass filter be very small? 
 MarkThoren: If you are adding the DC bias to V6, it will have no effect on the output of U7 because as noted earlier, this offset will simply appear across C18. 
 -Mark)

Partner Forums

Partner Libraries

信号经过高通滤波之后存在直流偏置