Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Design an active MFB bandpass filter

Status
Not open for further replies.

nickagian

Member
Hi to all!

I'm designing a bandpass active filter stage using the MFB topology. You can see the schematic attached. The intended center frequency is 1.4kHz and the BW is between 1kHz and 2kHz.

The filter will have as input signal the output of an ADW22035 accelerometer.
Since it is my first active filter design and I'm generally new to electronics, I would like some help with the following:

1) Is the Cin AC coupling capacitor necessary? The signal coming from the accelerometer is already DC biased at 1.65V, however I was thinking that maybe the C1 and C2 capacitors are providing the necessary coupling.

2) How possible is the DC bias circuit (providing Vdd/2 to the non-inverting input) to interfere with filter and change its desired characteristics? I was thinking maybe to intercept an OpAmp buffer after the RC network. Just prefer to omit to save PCB space.

3) Is the polarity of the AC coupling capacitor correct or should be inverted? In fact, what is the criteria for determining the polarity?

4) Lastly, is a buffer between the accelerometer output and the filter necessary? Unfortunately, there is no information in the datasheet about the output impedance of the accelerometer. There is a 32kΩ series resistor at the output. Is this the output impedance? In that case the output impedance of the accelerometer and the input impedance of the filter are at the same magnitude. Is that ok?

As you can see, I have some pretty basic questions... :eek:

I would appreciate any help! Thanks in advance!
 

Attachments

  • Bandpass_active_f&.JPG
    Bandpass_active_f&.JPG
    145.3 KB · Views: 504
Hi to all!

I'm designing a bandpass active filter stage using the MFB topology. You can see the schematic attached. The intended center frequency is 1.4kHz and the BW is between 1kHz and 2kHz.

The filter will have as input signal the output of an ADW22035 accelerometer.
Since it is my first active filter design and I'm generally new to electronics, I would like some help with the following:

1) Is the Cin AC coupling capacitor necessary? The signal coming from the accelerometer is already DC biased at 1.65V, however I was thinking that maybe the C1 and C2 capacitors are providing the necessary coupling.

2) How possible is the DC bias circuit (providing Vdd/2 to the non-inverting input) to interfere with filter and change its desired characteristics? I was thinking maybe to intercept an OpAmp buffer after the RC network. Just prefer to omit to save PCB space.

3) Is the polarity of the AC coupling capacitor correct or should be inverted? In fact, what is the criteria for determining the polarity?

4) Lastly, is a buffer between the accelerometer output and the filter necessary? Unfortunately, there is no information in the datasheet about the output impedance of the accelerometer. There is a 32kΩ series resistor at the output. Is this the output impedance? In that case the output impedance of the accelerometer and the input impedance of the filter are at the same magnitude. Is that ok?

As you can see, I have some pretty basic questions... :eek:

I would appreciate any help! Thanks in advance!

hi nick,
As you say the 'acc' output is via a 32K resistor, so would consider that as the output impedance.

The cap Cin is correct as you have +1.65V offset from the 'acc' and the junction of R1 and R3 will be at zero Vdc.

Will you please post your LTS *.asc file, save me having to redraw it.
 
hi ericgibbs,

yes of course I can post my .asc file. You can find it attached. I've also included the OPA340 model.

Ok, then we have an output impedance from 'acc' of 32kΩ and an input impedance at the filter of 11.3k // 6.34k ? Should I lower the impedance seen at the input of the filter, by inserting a buffer?

So, you think that the Cin is required. Does the polarity play any importance?
 

Attachments

  • Bandpass_filter_&#118.asc
    2.4 KB · Views: 175
  • OPA340..txt
    2.8 KB · Views: 259
Last edited:
hi ericgibbs,

yes of course I can post my .asc file. You can find it attached. I've also included the OPA340 model.

Ok, then we have an output impedance from 'acc' of 32kΩ and an input impedance at the filter of 11.3k // 6.34k ? Should I lower the impedance seen at the input of the filter, by inserting a buffer?

So, you think that the Cin is required. Does the polarity play any importance?

hi,
Look at these images. With 32k source impedance and without.

I would not use Cin.

I have run some tests with a dummy 10K load on the filter output, it looks OK.

Also linked in thru to the absolute rectifier.



BPacc1..gifBPfreq1..gif


EDIT:
It will require to be 'ac' coupled to the rectifier circuit, ideally using a buffer OPA!
 
Last edited:
hi nick,
Over what range of 'g' will this accelerometer be used, also what rate of change of 'g' are you measuring.?

I have been looking at the BPF filter and the precision rectifier in a combined circuit and I am concerned that the response of the full system is low.!
 
hi eric,

well first of all, I've also made some tests with 32k resistor in series with the source (I haven't thought of doing that myself, thanks) and I've seen differences in the amplitude of the output signal of the filter. I think that a buffer will be needed in the end.

Why would you not use the Cin. Isn't AC coupling necessary?

Here is the complete schematic of my circuit that I use for simulations (filtering, AC coupling after the filter, full-wave rectification and peak detection). In the end I have changed the topology of the rectifier to the one that you've proposed to the other post. The output will be driven to an ADC. All this processing could be equally done digitally, but the processor is not powerful enough. I'm still considering some details. Perhaps instead of the peak detection, a LPF should be used. Don't know which is better.

Regarding the 'acc' : The only frequencies that concern me are between 1k and 2kHz. As for the g's, I don't have a clue yet for the range. The project intends to measure vibrations. Not exactly full monitoring, just detecting when a threshold is violated. But I do not have real data of the g's that will be measured. I'm not so interested in the rate the g is changing. Measurements will be taken every approximately 1min or so.

What do you mean by saying that the response of the full system is low? Is amplification needed?
 

Attachments

  • Full_path..asc
    4.9 KB · Views: 180
hi eric,

well first of all, I've also made some tests with 32k resistor in series with the source (I haven't thought of doing that myself, thanks) and I've seen differences in the amplitude of the output signal of the filter. I think that a buffer will be needed in the end.

Why would you not use the Cin. Isn't AC coupling necessary?

Here is the complete schematic of my circuit that I use for simulations (filtering, AC coupling after the filter, full-wave rectification and peak detection). In the end I have changed the topology of the rectifier to the one that you've proposed to the other post. The output will be driven to an ADC. All this processing could be equally done digitally, but the processor is not powerful enough. I'm still considering some details. Perhaps instead of the peak detection, a LPF should be used. Don't know which is better.

Regarding the 'acc' : The only frequencies that concern me are between 1k and 2kHz. As for the g's, I don't have a clue yet for the range. The project intends to measure vibrations. Not exactly full monitoring, just detecting when a threshold is violated. But I do not have real data of the g's that will be measured. I'm not so interested in the rate the g is changing. Measurements will be taken every approximately 1min or so.

What do you mean by saying that the response of the full system is low? Is amplification needed?

hi,
Ref Cin, there is no 'dc' path into the filter OPA input.

You can restore the Vadc 'dc' back to it peak 'ac' input level by doing this small change.

EDIT: I use the CA3140 model for convenience only.

AAesp07..gif
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top