I have taken on a hobby project and I am trying to make a circuit which would be able to identify an acoustic signal with a frequency of 40kHz.

Here is the circuit of a band pass filter for this:

enter image description here

Here is the magnitude repsonse curve for it enter image description here

The issue is:

1) What would be the best way to detect the signal for the input?

2) This circuit will be mounted on a moving object and so it would be connected to a DC power supply (battery). This means that I would need a single supply op amp which has a GBW of 928k Hz but was also recommended 3.98M Hz.

I am new to this and would usually just try to figure this out myself but I cant for the life of me find an op amp of this specification?

I would appreciate any help.

  • 1
    \$\begingroup\$ You can search for an appropriate op amp using keywords like "single supply" and "rail-to-rail". Why do you want an op amp that will operate at 4 MHz? What do you mean by "detect the signal"? \$\endgroup\$ Jan 25 '19 at 13:24
  • \$\begingroup\$ Also, please provide citations or links for the graphics you include in your question. \$\endgroup\$ Jan 25 '19 at 13:25
  • \$\begingroup\$ @ElliotAlderson The software used to generate these listed that this was needed. What difference/ effect does GBW of the op amp have on the effect of the functionality of a circuit? \$\endgroup\$ Jan 25 '19 at 13:32
  • \$\begingroup\$ analog.com/designtools/en/filterwizard \$\endgroup\$ Jan 25 '19 at 13:33
  • \$\begingroup\$ James, that filter wizard comes with a component recommendation. \$\endgroup\$ Jan 25 '19 at 13:53

I recommend you do not use narrow band filtering, but a phase locked loop. In 1970's the following IC became popular as a replacement for narrowband filtering + level detector:


It's tone decoder LM567 and it's still available.

Some pre-filtering may be needed if there's strong other signals in other frequencies. But that filter needs not to be exactly tuned to the wanted frequency, it only must attenuate enough the unwanted.

ADD due the comments:

Tone decoder doesn't detect more reliably than a high Q filter + output level detector, but tuning and changing the detection bandwidth are radically simpler. In addition this can be a single IC solution and it's designed to handle up to 500kHz.

  • 3
    \$\begingroup\$ Sidebar note: The LM567 schematic in Figure 10.2 is a smorgasbord of functional blocks worthy of careful study. Just one corner of it is essentially an entire MC1496 (lower right corner area.) It's good stuff to read and understand. I won't speak to its applicability here, though. You should do that. \$\endgroup\$
    – jonk
    Jan 25 '19 at 15:04
  • 2
    \$\begingroup\$ Can you explain why a PLL is better than a bandpass filter? That would make your answer much more valuable. \$\endgroup\$ Jan 25 '19 at 15:04

James - I do not recommend to use a MFB topology with additional positive feedback. This additional feature is necessary only if you need a relatively large pole-Q (small bandwidth). However, in your case, the quality factor Q is - as far as I can see - only 40/10=4.

Therefore, you should make a redesign for a bandpass without positive Q-enhancement.

For single supply operation, you have nothing to do than to feed the pos. opamp input with a dc voltage which will bring the DC ouput to app. half of the supply voltage. Because the DC gain from the non-inv. input to the output is unity, it is a very simple task: Connect the non-inv. input to the midpoint of a resistive voltage divider consisting of two equal resistors and feed this divider with the supply voltage.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.