When I simulated a filter with a cutoff frequency around 200 kHz in LTspice, everything looked fine. However, when I tested it in the real world using a signal generator, I noticed that the signal amplitude dropped significantly right at the input of the filter.


Is there a special method or consideration for testing filters with such low cutoff frequencies?

 

Hi emir,
Welcome to AAC.
What is the impedance of your signal generator?
The LTSpice SigGen is zero ohms.
E

 

Hi emir,
Welcome to AAC.
What is the impedance of your signal generator?
The LTSpice SigGen is zero ohms.
E

Hi, I use HighZ and 50 output. I also tried to add 50 ohm series resistor in LTSPICE and it's normal to dropping amplitude of signal at filter input. I think i have to use voltage buffer for test it but is there any other test methods?

 

hi emir,
Please post your LTSpice asc file.
E

 

Here it is

 

Put a 50Ω series resistance in your LTSPICE simulation and see what happens.

 

Put a 50Ω series resistance in your LTSPICE simulation and see what happens.

Hi Bob,


Yes, when I add a 50-ohm series resistor, the signal at the input of the filter almost completely disappears, even at very low frequencies. So now I’m wondering:
Is the actual cutoff frequency of my filter the one I observe with the 50-ohm series resistor, or the one I see without it?

If the real cutoff frequency is the one without the 50-ohm series resistor, then how am I supposed to test this filter in practice?

Even when I use an OPA548F/500 buffer, the input signal to the filter still weakens as the frequency increases. This makes me think something else might be affecting the test setup.

 

If the real cutoff frequency is the one without the 50-ohm series resistor, then how am I supposed to test this filter in practice?

Drive it from a low impedance source, eg an opamp voltage follower, and buffer the output with same, for its high input impedance.

The output impedance of the source and the input impedance of the load are critical parts of the circuit.

 

Is this just a theoretical question, or are you designing the filter for use in a real circuit? The characteristics of the filter will depend on what circuit is driving it and what the output is connected to. To get realistic answers from the simulator, you must add the input and output circuits.

 

This is a bad scheme. Increase the load resistance!

 

Drive it from a low impedance source, eg an opamp voltage follower, and buffer the output with same, for its high input impedance.

The output impedance of the source and the input impedance of the load are critical parts of the circuit.

Hi,

I'm writing to provide an update. I implemented a unity-gain buffer using the OPA548F/500 op-amp and drove the filter through it. Thanks to the buffer's 5A output capability, I did not observe any attenuation at the filter input. The filter's behavior at different frequencies is provided below.

Blue: Filter Output
Yellow: Filter Input
Purple: Buffer Input


50 KHz - As expected, no attenuation is observed at this frequency.



200KHZ - The cut-off has started, but it is not yet at the -3dB point. There is some distortion at filter input. I couldn't get it.



700KHz - At this frequency, it is clearly observed that the response is well below -3dB.

 

This is a bad scheme. Increase the load resistance!


View attachment 348499

Hi bordodynov,
Thanks a lot for your valuable comment
However, I don't have a load condition as shown in your schematic. The resistor and capacitor I added are actually part of the damping network for the filter. They were included to prevent signal amplification in case of any resonance.
Also, in the schematic you provided, the "OUT" side is incorrectly labeled — the left side of C3 is actually the output.

 

Use a 10,000 ohm resistor in series with the signal generator output as well as in the spice simulation. probably the response will be quite similar.
As already mentioned, the generator in spice is a zero resistance device.