I have a very strange problem. I wish to design a bandpass filter with a lower cutoff of .001 Hz and an upper cutoff of 5 kHz.

To satisfy this large bandwidth requirement and extremely small low frequency cutoff, the circuit needs to be built for a very low center frequency. As a result, the quality factor, and hence gain at the center frequency is vanishingly small. To restore the signal passed by such a filter would require something absurd like ~1e24 subsequency amplification.

Any recommendations on filter topologies, filter order, what else I could try would be appreciated.

 

My first thoughts are to use two separate filters, a 5khz low pass and a 0.001 hz high pass.

But what on earth do you want a cut off frequency of 0.001 hz for?

JimB

 

I agree with JimB, a high-pass filter at 0.001Hz sounds rather stupid, why on earth would you possibly want such a device?. It looks more likely that a 5KHz low-pass filter would do all you required?.

Although you 'could' design a filter for such a low frequency, the capacitors involved would be huge, and the cost would probably be prohibitive?.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

Yeah, I know its very strange, and I'm not sure it will work. My friend and I are working on a project and he suggested that we build a bandpass with such a low cutoff in order to reduce low frequency drift in a sensor circuit. I'm not an electronics expert, but it seems to me that it should in theory work, and why I made this post. Is it really that unrealistic?

 

I would suspect the filter would probably drift more than the sensor?, and you will require massive coupling capacitors - as by definition it can't be DC coupled!.

What kind of sensor is it anyway?.

If you want to explore the filter possiblity, you might have a look at the free software from MicroChip at http://www.microchip.com/stellent/id...cName=en010007, although it says it only goes down to 0.1Hz - but making the capacitors 100 times larger will give you an idea of the capacitor sizes required.

__________________
PIC programmer software, and PIC Tutorials at:
http://www.winpicprog.co.uk

 

You can buy chopper stabilized op amps that have very low offset voltage - might cut down on your drift.

You also might try using a temperature sensor to compensate for any drift due to temeprature.

You could implement your filter using DSP techniques but you will need a really long FIR filter - long time window to see the low frequencies. If you sampled slowly you might be able to do it without a ton of processing and memory.

 

This circuit will maintain zero volts out within the range of the op amp, but signals above .001 hz will pass. You can follow it with a low pass to limit the bandwidth. The gain is (1+Rgain/10K)/2.

Attached Thumbnails

 

__________________
see my website: www.geocities.com/russlk