The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> OTA-C versus active RC filters
https://designers-guide.org/forum/YaBB.pl?num=1200224613

Message started by aaron_do on Jan 13^th, 2008, 3:43am

Title: OTA-C versus active RC filters
Post by aaron_do on Jan 13^th, 2008, 3:43am

Hi all,

I'm trying to compare OTA-C and active-RC filters in terms of power consumption. Assume they both need to meet the same high linearity requirement, and the same noise and gain requirement. Which one can be designed with a lower power consumption?

From what i understand, active-RC filters can't reach the same frequency because of the miller multiplied feedback capacitor. However, if we assume that the frequency of operation is only a few megahertz, then the OTA-C filter would need large capacitors anyway.

Lastly, which type of filter is typically used for receivers with a low IF? I would think active-RC filters are more popular because they should be more reliable. Any input is welcome...

thanks,
Aaron

Title: Re: OTA-C versus active RC filters
Post by sheldon on Jan 13^th, 2008, 6:03am

Aaron,

Not an expert, but my understanding is that active-RC filters have better
linearity and dynamic range: IIP3, P1DB, than OTA-C filters. Even when
using techniques to extend linearity such as doublets, etc. it is difficult for
an OTA-C filter to match the linearity of a closed-loop op-amp. Not sure about
noise, my guess is that the noise of an active-RC is higher since their are
more components to contribute noise in the circuit.

Not sure what the comment about reliability means, however, active-RC
filters also have significant limitations. Compared to OTA-C filters, they require
additional circuitry for trimming and can have significantly larger area.

Best Regards,

Sheldon

Title: Re: OTA-C versus active RC filters
Post by aaron_do on Jan 13^th, 2008, 6:58am

Hi Sheldon,

thanks for the reply. I figured that op-amp based active-RC filters would be more reliable since if the op-amp's gain is high enough, then you can assume it is ideal while for a gm-C type filter you need to know the gm. Not sure how true that is though.

Aaron

Title: Re: OTA-C versus active RC filters
Post by ACWWong on Jan 13^th, 2008, 4:37pm

Hi aaron,

Yes I tend to agree with Sheldon, its easier to get linearity with activeRC... linearity in gmC filters can get tough especially for high-order/Q structures. But if linearity isn't a tough specification and power is critical, then gmC has advantages. So as to which to choose it ... it depends on circuit specifications and requirements. For any specification given, it wouldn't be time wasted if you were to spend a couple of days investigating both approaches before deciding... this decision is probably best made in conjunction with filter architecture (cascaded biquads or ladder/SFG) choice.

As to reliablity... i don't see your point... sure having masses of open-loop gain/bandwidth in your opamp will mean it won't come into play when designing your active RC filter stages (are you over designing it ?), but you'll still have to cope with R and C variation which is usually trimmed (e.g by algorithm). For gmC both gm and C vary but its often quite easy to tune gm to compensate.

As to low-IF receivers, I've worked on both types of complex bandpass: activeRC and gmC... the activeRC ones were more linear.. when linearity could be achieved in gmC, then gmC should be lower power.

cheers
aw

Title: Re: OTA-C versus active RC filters
Post by buddypoor on Jan 15^th, 2008, 4:33am

Hi Aaron,

in general one can state, that OTA-C filters are preferable over RC-opamp filters only if
- you are designing a fully intergrated filter structure, or
- you have the requirement to change resp. tune some parameters (pole frequency or pole Q).

For a design with lumped elements an opamp-based RC structure has certainly advantages (better linearity, greater dynamic range, more structure alternatives, low output impedance).
You should not forget that OTA-structures require output buffers, which have to be acounted for in the power budget.

In addition, all design formulas assume ideal parameters for the active device; and this requirement can be fullfilled by far better in case of opamps. Thus, you have to live with more deviations from the ideal filter response in case of OTA-filters.

Regards
Lutz