Why does the bandpass filter output such a waveform?

littleshrimp

Why does the bandpass filter output such a waveform? [Copy link]

Use a 4th order multi-feedback bandpass filter, the circuit in this post https://en.eeworld.com/bbs/forum.php?mod=viewthread&tid=1088056&extra=

If the signal amplitude of the finished board is not very large, a red waveform will appear. Is this the legendary self-excitation?

The bandwidth of the amplifier is 6.5MHz. The high frequency in the figure below is around 5MHz.

The first red is the output waveform of the first stage, and the blue is the output waveform of the second stage

A large part of the high frequency of the first stage will not appear in the second stage

maychang

"Excuse me, is this the legendary self-excitation?"

It is the "legendary self-excitation", and it is a slight and discontinuous self-excitation.

littleshrimp

maychang posted on 2019-9-8 12:24 "Excuse me, is this the legendary self-excitation?" It is the "legendary self-excitation", and it is a mild, discontinuous self-excitation...

No wonder people say that this type of filter is prone to self-excitation. Any circuit designed completely according to ADI's analog filter design wizard will be like this.

gmchen

littleshrimp posted on 2019-9-8 13:01 No wonder people say that this kind of filter likes to self-excite. The circuit designed completely according to ADI's analog filter design wizard will be like this.

Are there "two resistors"? If so, remove them and connect the op amp's non-inverting input directly to Vref.

gmchen

In addition, this circuit and this frequency have requirements for the design of the printed circuit board. I wonder if you have paid attention to the design issues of the printed circuit board, mainly in terms of grounding and decoupling.

littleshrimp

gmchen published on 2019-9-8 21:03 In addition, this circuit and this frequency have requirements for the design of the printed circuit board. I don't know if you have paid attention to the design of the printed circuit board, mainly the grounding and...

There are 4 identical channels in total. I have tried to remove the two-stage positive feedback resistors of one of the channels. The waveform has not improved. The grounding is directly connected to the ground plane through the pin. Decoupling 2.2u+100nf is close to the power supply pin. The reference level uses resistor voltage divider + ad8031 buffer + multiple capacitors above 10uf

captzs

Don't couple the two stages directly, it may be better to use a follower to isolate them. When I simulate multi-stage filtering, the output is transmitted to the next stage by resistor voltage division, or the follower is isolated, which is very effective in preventing self-excitation. Sometimes "attenuation" filtering is also used, that is, Rf is less than Re.

gmchen

littleshrimp published on 2019-9-8 21:21 There are 4 identical channels in total. I tried to remove the two-stage positive feedback resistors of one channel. The waveform did not improve. The grounding was directly drilled at the pin...

"The reference level uses resistor voltage division + AD8031 buffer + multiple capacitors above 10uf"

Are these multiple capacitors connected to the output of AD8031? Note that it is easy to self-excite when the output of the op amp is connected to a capacitive load!

littleshrimp

gmchen posted on 2019-9-9 12:38 littleshrimp posted on 2019-9-8 21:21 There are 4 identical channels in total. I have tried to remove all the two-stage positive feedback resistors of one of the channels...

The following sentence is included in the Analog Filter Wizard of ADI, which seems to guarantee that the AD8031 can drive capacitive loads and can accommodate any number of capacitors:

I have measured the reference level and there is no self-excitation, so the problem should not be with the reference level.

The filter tool adds a reference stage whenever -Vs = 0. This reference stage generates an intermediate supply voltage rail that is used as a reference voltage for the filter circuit.

The reference stage consists of an AD8031 op amp driving a 10 μF capacitor. It is generally a very bad idea to drive a capacitor directly with an op amp, but the architecture of the AD8031 ensures that this configuration remains stable as long as the driven capacitance is large enough. The advantage of this circuit is that it provides a low frequency source impedance across the entire frequency spectrum. Since the filter circuit can source and sink considerable current into the reference voltage source, a low source impedance reference source is critical to achieving optimal filter performance. When using this reference circuit, keep in mind the following points:

• Use AD8031. Using other operational amplifiers will cause the circuit to be unstable.
• The capacitance driven by the AD8031 should be at least 10 μF. Any number of bypass capacitors may be placed in parallel with this capacitance.
• No resistor should be placed between AD8031 and capacitor

The reference voltage shown in the tool is obtained through a voltage divider connected directly to the positive supply rail. However, this voltage can also be generated from a reference IC such as the ADR4520 or ADR3420. The evaluation board takes the size of these reference ICs into account for your convenience.