The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Question about OTA-C filter ^^
https://designers-guide.org/forum/YaBB.pl?num=1211808963

Message started by yixiusky on May 26^th, 2008, 6:36am

Title: Question about OTA-C filter ^^
Post by yixiusky on May 26^th, 2008, 6:36am

Hello everyone, i have some questions about OTA-C filter design, I would like to share with all of u, thank u very much ! ^^

Design object : tunable bandpass filter
Tuning frequency range : 50M-860MHz
Bandwidth of filter : 15MHz

a . Consider all of requirements, i decide to use OTA-C filter. However I don't know is there other choice exept gm-c filter? How about your opinion? ^^

b. about the requirement of OTA design

1) the phase marge of OTA shoulbe 90 degree

2) How to decide the bandwidth and gain of OTA ? I have no idea

3) The linearity of OTA also should be improved. By using source degeneration... I still can not get good linearity

4) Do you think it is possible design one filter with this range tuning ability?

Sorry to ask so many questions :(

thank you very much ^^

Title: Re: Question about OTA-C filter ^^
Post by buddypoor on May 27^th, 2008, 1:13am

Hi Yixiusky,

I´ll try to give some answers:

Design object : tunable bandpass filter
Tuning frequency range : 50M-860MHz
Bandwidth of filter : 15MHz

a . Consider all of requirements, i decide to use OTA-C filter. However I don't know is there other choice exept gm-c filter? How about your opinion? ^^
In general, OTA-C structures are appropriate for tunable filters,however, an upper limit of about 900 MHz seems to be problematic. Monolithic integration iis necessary.

1) the phase marge of OTA shoulbe 90 degree
No problem as the OTA normally has no feedback.

2) How to decide the bandwidth and gain of OTA ? I have no idea
OTA has no "gain", but a transconductance instead. Frequency range could be a problem.

3) The linearity of OTA also should be improved. By using source degeneration... I still can not get good linearity
What means "good" ? What are your requirements regarding dynamic range and input/output levels ?

4) Do you think it is possible design one filter with this range tuning ability?

No, I don´t think so. More than that, a tunable BP structure with a constant bandwidth (instead of constant Q) is hard to design. So, it looks not too good for your task.
Regards

Title: Re: Question about OTA-C filter ^^
Post by yixiusky on May 27^th, 2008, 4:37am

buddypoor^^ Thank you very much.

I would like to discuss with u more, since i am not totally get your idea :(

a . In general, OTA-C structures are appropriate for tunable filters,however, an upper limit of about 900 MHz seems to be problematic. Monolithic integration is necessary.

Question : I am not clear about "Monolithic integration is necessary". I read some papers, people can design the filter up to GHz with gm-c filter. i would like to know your meaning, thank you!

2) How to decide the bandwidth and gain of OTA ? I have no idea
OTA has no "gain", but a transconductance instead. Frequency range could be a problem.

Question: OTA has Iout/Vin (transconductance). But when we design OTA for OTA-C filter, how can we set the specification for OTA? And what's the relationship of OTA's bandwidth and bandpass filter's bandwidth?

3) The linearity of OTA also should be improved. By using source degeneration... I still can not get good linearity
What means "good" ? What are your requirements regarding dynamic range and input/output levels ?

Question: "Good" means i would like to get IIP3=15dBm for OTA-C filte. But now i only get -9dBm. so i modify the OTA, add source degeneration. but still can not get 15dBm IIP3 :(

4) Do you think it is possible design one filter with this range tuning ability?

No, I don´t think so. More than that, a tunable BP structure with a constant bandwidth (instead of constant Q) is hard to design. So, it looks not too good for your task.

Question: I would like to know what's the reason you feel one filter is impossible to cover whole frequency range? based on analysis or ..

5) I would like to build the 4th order biquad OTA-C filter, but i don't know the exact steps to build OTA-C filter. It is not like

active-RC, we can use "filter solution" to get coefficient...I am really confuse and not clear :(

sorry to ask so many questions :)

sincerely yours

yixiusky

Title: Re: Question about OTA-C filter ^^
Post by buddypoor on May 28^th, 2008, 1:02am

Hi, Yixiusky,
I´ll try to clarify something.

Question : I am not clear about "Monolithic integration is necessary". I read some papers, people can design the filter up to GHz with gm-c filter. i would like to know your meaning, thank you!

It is not recommended to design a filter near 1 GHz from lumped elements. Therefore, monolithic integration (IC) is necessary. Is this a problem ?

Question: OTA has Iout/Vin (transconductance). But when we design OTA for OTA-C filter, how can we set the specification for OTA? And what's the relationship of OTA's bandwidth and bandpass filter's bandwidth?

Are you sure that bandwidth has to be constant for different mid frequenncies ? This means the filter Q will rise up to 57 ! More than that, a filter structure wich can be tuned with constant bandwidth is hard to find.

Question: "Good" means i would like to get IIP3=15dBm for OTA-C filte. But now i only get -9dBm. so i modify the OTA, add source degeneration. but still can not get 15dBm IIP3 Sad

I don´t understand - does this mean you have already chosen an OTA type ? Or do you design an OTA by yourself ?

Question: I would like to know what's the reason you feel one filter is impossible to cover whole frequency range? based on analysis or ..

I understand that your filter has to be TUNABLE . Is this correct ? Or are you to design several filters with different frequencies within the band you have mentioned ? As already said, tuning with constant BW (and rising Q) is a big big problem within this band.

I would like to build the 4th order biquad OTA-C filter, but i don't know the exact steps to build OTA-C filter. It is not like active-RC, we can use "filter solution" to get coefficient...I am really confuse and not clear Sad

Why 4th order ? Why biquad type ? Another problem: 4 th order and tunable ??? (nearly impossible).

Recommendation: You should check all requirements again and sort them in a systematic manner with the aim to select one concept (filter structure and realization). At the moment your information about requirements are confusing and partly contradictory: biquad type bandpass of 4th order, mid frequency tunable with constant BW between 60 MHz and nearly 900 MHz.
good luck

Title: Re: Question about OTA-C filter ^^
Post by yixiusky on Jun 1^st, 2008, 5:54am

Thank you very much, buddy :)

I will follow your recommendation and make my requirements in a systematic manner:

4th order biquad type gm-c bandpass filter, mid frequency tunable with constant BW between 60 MHz and 860MHz.

As u said ""I don´t understand - does this mean you have already chosen an OTA type ? Or do you design an OTA by yourself ?Why 4th order ? Why biquad type ? Another problem: 4 th order and tunable ??? (nearly impossible). "

Answer : (1) yes, i am designing ota-c filter now, and design OTA by myself. But i am not sure the phase margin of OTA

must be 90 degree or not.

(2) I design 4th order, because i need 3rd order harmonic rejection ration larger than 30dB. With 4th order i
can get this 30dB rejection ^^

Thank u very much ^^

Title: Re: Question about OTA-C filter ^^
Post by buddypoor on Jun 1^st, 2008, 6:41am

Don´t care about phase margin or other details which are not important during the first design phase.
The main question is: Did you find any 4th order filter topology which is tunable and has constant BW ?
I would be surprised !

Title: Re: Question about OTA-C filter ^^
Post by yixiusky on Jun 1^st, 2008, 6:21pm

:) Tunable bandpass filter with constant BW is my design goal.

So i should try to achieve this goal. The normal biquad can not satisry this goal so i should add bandwidth tuning circuit to the filter also ^^ i will try to make it ^^ thank you very much for your instruction ^^