DC-3Hz low pass filter with 200ms rise time and fall time

Hmmm...interesting!

It sounds more like a power supply!

A basic low pass filter should present no problem if it is just mains or audio frequency which is the problem.

It may be more pertinent to ask what frequencies in the stop band are causing problems?

If there is clock noise, digital noise or RF noise getting in the problem is vastly different.

Hello Shyam,

Tailor made for you, with choice of 2nd or 4th order depending on your needs. I've included circuit diagram, time and frequency responses.

(If you can't read the pictures well, then right click on them and save them on your computer to display them in your favorite JPG browser.)

Neil

Hi Neil: With answers like that, who needs questions? Nice Job! I would have guessed he needed to raise his corner slightly, but I was wrong.

Dear Neil

That sure is a very good help. I will try to read the part values and construct the filter. I think two sets of filters are identical and only cascaded, so I can try few more order added to it. Can you trace the curves for 6th and 8th order also if your software permits that.

I may also have to add a 50Hz notch filter as mains frequency is a serious noise in lower band. Do you have some idea for removing the 50Hz specially and keeping the other thing as suggested now.

My input signal will be 0-50mV and have to obtain 0-10V using gain of 250. In which stage these filters will perform best? f\Before amplification or after amplification?

A final word. Thank you.

I must apologise to everyone who tried to download these JPG pictures. I did not count on CR4 doing a compression job on my JPG pictures that I uploaded. When I downloaded them from CR4 some time later, only then did I see that the values could not be seen clearly!!! So if anyone would like me to send them my original JPG files, I will gladly do this, just send me your email address.

Shyam, the two filter stages are not quite the same, as the filter routines optimise the two stages for dynamic range. You could duplicate the 4th order to get the 8th order. However, with an added gain requirement (you'll need to watch out for offset problems), as well as 50Hz noise problem, the analog route seems complex. Maybe a SCF could do it better for you.

Neil

Dear Neil J

I will greatly appreciate receiving the picture by email sst (at) sensorstechnology.com if you have the original and clear image.

Maxim suggested their standard part and that information is just that much. It looks their circuit suffers DC shift which is externally trimmed.

Did you also look into 50Hz notch filter design? I think, that is critical noise source but may also have harmonics, which will not pass through low pass filters.

I also want to look into better power source design for these amplifiers and if LM7812 and LM7912 or LM7805 and LM7905 are not a good regulator for low noise application. Have you looked into lowest noise voltage regulator also that are ready to use in low power applications with just 100mA capability.

I will like to get to the bottom of the noise to see if I can get some what better signal with greater dynamic range.

Maxim suggested MAX7480 for 8th order filter.

Hello Shyam,

you didn`t say anything about the order and attenuation requirements. Assumed you have a measurement application which requires (rather) DC-correct signal responses you could try a monolithic clock-programmable LP-Filter.

I`ve made good experiences with the LTC1065, a single supply clock-tunable 5th order LP filter which requires only a single supply and, in your application, a 300Hz clock to set the cutoff frequency. The DC-offset is only 2mV.
Regards Uwe