This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

OPA340: Sallen Key Filter oscillation on OPA340

Part Number: OPA340

We implemented a low pass Sallen Key filter with a cutoff frequency of 4.8 Mhz, with balance R=330 and C=100pF.  With a bandwidth of  5.5Mhz, significant filtering was not expected, however the OPA340 seem to make signal worse the input noise.

Blue = input to Sallen Key,  Yellow = Output of Sallen Key

We changed to frequency cutoff to 33.9kHz

Yellow = input to Sallen Key,  Blue = Output of Sallen Key

While lower the cutoff frequency improved the response a lot it was still not as good a simple single order buffered RC low pass filter

RC low pass filter with cutoff frequency set to 159 kHz

 Yellow = input to filter,  Blue = Output of filter 

Our testing of the OPA340 Sallen Key filter resulted in unexpected disappointing results any thoughts on how we can improve the results? 

Thank you,

Scott 

  • Hi Scott,

    you cannot build a 4.8MHz cutoff frequency Sallen Key low pass filter with an OPAmp that has a bandwidth of only 5.5MHz. That's absurd. At these frequencies the OPAmp has no gain reserve and can not make the feedback loop work properly, if at all. This is very bad design practise!

    Kai

  • Hi Scott,

    An op amp used in an active filter  must have  an open-loop gain (Aol) much higher than the closed-loop gain (Acl) in the filter passband.  Normally, about 40 dB of Aol above the Acl is adequate for an accurate filter response. For example, if the filter is supposed to have  20 dB of gain (Acl) in the passband, the Aol at the cutoff frequency needs to be about 60 dB. The point being that the bandwidth of the op amp must be high enough to support the needed Aol requirement at the filter's cutoff frequency.

    You indicate that the intended cutoff frequency of the Sallen-Key low-pass filter is intended to be 4.8 MHz.  If we look at the Aol vs. frequency gain plot from the datasheet and find where the gain intersects the 4.8 MHz frequency line we see the Aol is around 0 dB. Remember, the unity gain bandwidth of the OPA340 is rated at 5.5 MHz; just above 4.8 MHz. Therefore, the OPA340 does not have enough bandwidth to properly function as an active filter, or even a simple 4.8 MHz amplifier. Since the filter is not functioning correctly unexpected responses are going to be had.

    datasheet 

    In reality, a simple Sallen-Key filter having an Acl of 1 V/V, a 4.8 MHz cutoff frequency and a Butterworth response (Q = 0.71) requires an op amp having a unity-gain bandwidth of about 341 MHz. That is a wide band, or high-speed operational amplifier. these are available from TI's High-Speed Amplifiers group.

    If you want to experiment with different active filter scenarios and find their corresponding unity-gain bandwidth requirement I suggest downloading TI's FilterPro program. It is very easy to use and will provide you with a filter design, and the op amp's gain-bandwidth requirement. The latter changes with required passband gain and the type of filter response your filter needs to have; Butterworth, Bessel, Chebyshev, etc.

    You can download FilterPro from the TI website:

    Scroll down and select Dowlnload FilterPro v3.1. You may have to set up an account but very little information is required.

    Once the active filter is synthesized and the filter schematic appears, scroll down the Filter Stage information. The Min GBW reqd. is at the end of the list.

    Regards, Thomas

    Precision Amplifiers Applications Engineering

  • Hello Thomas,

    Thank you very much for the detailed response.  I will be going through TI's "Analog Engineer’s Pocket Reference" and "The Signal" to identify good filter practices, any other references you can recommend?

  • Hi Scott,

    can you show the schematic of your Sallen Key filter? Maybe we can find a simple fix...

    Kai