• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Answers 1 answer
  • Subscribers 39 subscribers
  • Views 1281 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Multiple feedback filter not matching filterpro design recommendation - please help.

Edward Polzin
Prodigy 40 points

Other Parts Discussed in Thread: OPA4820, TL974, OPA4830

I used the filterpro desktop to design a Bandpass filter. the parameters that i entered that deviated from the default values were: center freq = 125kHz Passband bandwidth = 10kHz Stopband bandwidth = 80kHz stopband attenuation = -40dB.  The result was a multiple feedback 6th order bessel filter. I built up the second stage of the circuit using an OPA4820 (waiting for resistor values to come in to build remaining stages). however, the center frequency was about 140kHz and the gain was only about half. I've had a lot of trouble trying to get these styles of filters to work. Can anyone shed some light as to why the real circuits dont match the simulations? Attached is the filter pro design file that i used.  any help would be appreciated.  thanks.

  • 0
    TI__Expert 5515 points

    Edward,

    The problem is likely parasitics working in tandem with the large feedback resistor values. As noted in the OPA4820 datasheet, the part really doesn't like impedances around it higher than a couple of kΩs - because at higher values, the parasitics really start to affect the bandwidth of the op amp. Your design requires a fairly high GBW for the op amps - around 150MHz. Tweak that slightly and center frequencies and Qs can be skewed.

    Also as noted in the OPA4820 datasheet, the Sallen-Key implementation may be better suited for these types of designs using these high-frequency op amps. Using a Sallen-Key approach, the resistors around the op amp can be kept to 1kΩ or under, and the other resistors in the circuit become smaller as well, which will give improved noise performance anyway.

    Hope that helps - give the Sallen-Key approach a try and see if it works better for you.

  • 0 in reply to Rick Downs - WEBENCH Design Center
    Prodigy 40 points

    Thanks for the info.  Will try that.

  • 0 in reply to Edward Polzin
    Prodigy 40 points

    Went back and selected the Sallen-Key type.  Built the circuit and all it does is oscillate at about 36MHz.  Not sure how to start trouble-shooting this one.  Please help.

  • 0 in reply to Rick Downs - WEBENCH Design Center
    Prodigy 40 points

    After failure with the Salen-Key implementation with the OPA4820, i dropped a TL974 in the same circuit and it works well.  Center frequency and roll-off looks acceptable.  However, i do not have unity gain, but more an attenuation of about 4 to 5.  i assume this is because the TL974 has a smaller GBW.  Can someone suggest a better op-amp choice that's available in a TSSOP-14 package (same as the TL974) for this implementation?  Thanks.

  • 0 in reply to Edward Polzin
    TI__Expert 5515 points

    Edward,

    I tried to respond to your earlier post and was having some trouble with the forum, so now I'll address both:

    I suspect the problem you may be having with the OPA4820 is mostly a matter of how you're building the prototype. Remember I cautioned you on parasitics, and if you're not using a PCB then that may be the cause of the instability.

    The TL974 only has a GBW of 12MHz, which is about 10x less than what you need for this design to work accurately. The only other part I could find in the same footprint that would be adequate might be the OPA4830 - it's a lot like the OPA4820, but with less GBW, so may be a little less touchy.

    If you give it or the OPA4820 another try, I'd suggest laying out a PCB and building it up on that. Suggest you follow the guidelines for layout of the board that you can find in High Speed Amplfier PCB Layout Tips.

  • 0 in reply to Rick Downs - WEBENCH Design Center
    Prodigy 10 points

    i have to construct an active bandpass filter for 2Mhz to 12MHZ.the passband should be at 0db and the stopband should be at least at -40db.....i have tried many circuits but can not get accurate results....plz guide me