Designer’s Guide Community

Lilylily Junior Member Offline Posts: 15	some questions on the amplifier in the RC Active filter Aug 19^th, 2009, 2:15am Hi,everyone,I have to design two filters having cutoff frequencies of 100kHz and 1.2Mhz. And I don't know how to select the amplifiers in the filters. And what is the requirement on the gain, bandwidthand so on?Thanks!! �
Back to top	IP Logged

buddypoor

Community Fellow

Offline

Posts: 529
Bremen, Germany

Re: some questions on the amplifier in the RC Active filter
Reply #1 - Aug 19^th, 2009, 6:53am

Lilylily wrote on Aug 19^th, 2009, 2:15am:

Hi,everyone,I have to design two filters having cutoff frequencies of 100kHz and 1.2Mhz. And I don't know how to select the amplifiers in the filters. And what is the requirement on the gain, bandwidthand so on?Thanks!!

In most cases, gain is not a problem as it is high enough.
But the GBW should be app. 50...100 times larger than the pole frequency of the filter stage. By the way: 2nd order ?
But in general it depends on the filter topology. Some are more suszeptible to opamp gain/phase errors than other structures.

Sorry: read "suszeptible" as "sensible" .

Back to top

« Last Edit: Aug 19^th, 2009, 8:50am by buddypoor »

LvW (buddypoor: In memory of the great late Buddy Rich)

IP Logged

vivkr

Community Fellow

Offline

Posts: 780

Re: some questions on the amplifier in the RC Active filter
Reply #2 - Aug 20^th, 2009, 2:57am

Lilylily wrote on Aug 19^th, 2009, 2:15am:

Hi,everyone,I have to design two filters having cutoff frequencies of 100kHz and 1.2Mhz. And I don't know how to select the amplifiers in the filters. And what is the requirement on the gain, bandwidthand so on?Thanks!!

This really depends on your requirements. I would advise using a simple (but reasonably realistic) opamp model and playing around withthe gain/bandwidth/slewrate etc. to see what the minimum requirements are, and then setting your spec to be a bit higher than that (safety margin).

For continuous-time filters, it is quite easy to do this exercise. There is also something called "sensitivity analysis" which you can use. This is available with the Cadence Analog Design Environment for Spectre, but I have never used it yet. You can try it out with a schematic of your filter containing some opamp macromodels to see if you get something useful out of it.

Best regards,

Vivek

Back to top

IP Logged

buddypoor

Community Fellow

Offline

Posts: 529
Bremen, Germany

Re: some questions on the amplifier in the RC Active filter
Reply #3 - Aug 20^th, 2009, 5:21am

Hi vivek, I am not familiar with SPECTRE, However, I doubt that the sensitivity analysis can give valuable information about the dependence of the transfer function to the parasitic opamp phase (and that is the dominant parasitic parameter determining the upper freq limit).
As far as I know, sensitivity analysis gives information about the amount of influence (in percent for example) which a part and its tolerance has on the transfer function.
But your other advice - to play a little with different opamp makros - is a very good one. For my opinion, this exactly is one of the fundamental tasks of a simulation program.

Back to top

LvW (buddypoor: In memory of the great late Buddy Rich)

IP Logged

Lilylily

Junior Member

Offline

Posts: 15

Re: some questions on the amplifier in the RC Active filter
Reply #4 - Aug 20^th, 2009, 8:26pm

Hi, thanks to all of you! And the fact is that I want to design a filter tunable to 5 bandwidths ranging from 100K to 1.2MHz? But I don't know what the first step is. At the beginning, I think I should determine the amplifier structure firstly. But now I think It's better to determine the structure of the filter (sallen-key, MBF, inverting) and the order,�then get the corresponding requirement to the amplifier. Is that right? Do you have the experience to design a tunable filter. May you tell me how you deal with that? Thanks!

Back to top

IP Logged

Mayank

Community Fellow

Offline

Posts: 334

Re: some questions on the amplifier in the RC Active filter
Reply #5 - Oct 8^th, 2009, 3:36am

Hi Lilylily,
Hey it's a late reply..But plz dnt mind. You should proceed like :--
1. First step is to decide the Filter approximation and its order required based on your specs such as Roll-Off rate, PB ripple, SB attenuation etc.
2. After deciding upon type of filter(ie filter approximation), you design the top-level of the fitler, which basically means work out the math and calculate component values for any one of the numerous implementations you choose (eg: Active RC, Sallen-Key/Tow-Thomas/Acken-Moserberg Biquad/ LC Ladder / MFB etc.).
3. Design an opamp such with a GBW in accordance with your filter cutoff( or the effective TF cutoff in case of cascading type of structure in Higher Order Filters). Generally, we keep Opamp GBW atleast 10-50 times greater than the Filter cut-off.

thanx,
Mayank.

Back to top

IP Logged