The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> RF Design >> inductor choice for lna for 60ghz or higher freqs
https://designers-guide.org/forum/YaBB.pl?num=1364186247

Message started by kamela on Mar 24^th, 2013, 9:37pm

Title: inductor choice for lna for 60ghz or higher freqs
Post by kamela on Mar 24^th, 2013, 9:37pm

I am going to design an LNA for 60ghz or higher . I have read some texts and papers but I couldn't gain a good viewpoint about of which type of inductor ( transmission lines , active inductor , pasive spiral inductor ) I have to use in my design with respect to all things such as area , power , linearity . noise , gain , stability and other layout and manufacturing considerations. but i guess the active one isn't a good choice for these frequencies but I'm not sure . I've just a bit mixed up.
thanks in advance for any opinion

Title: Re: inductor choice for lna for 60ghz or higher freqs
Post by aaron_do on Mar 24^th, 2013, 11:48pm

Hi,

comparing transmission lines to spiral inductors, spiral inductors will be smaller and probably have better insertion loss (being shorter). I'm not sure if variation is going to be an issue, but your EM modeling will probably need to be much more accurate for spiral inductors. Unwanted coupling/radiation should be less of an issue for transmission lines as compared to inductors. Another consideration is the return path which should be much better defined for a transmission line than inductors. BTW, this all depends on the characteristic impedance of your t-line since if Z0 is high enough, it will look more or less like an inductor anyway.

In conclusion, if you do your modeling well, I believe you will get better raw performance using spiral inductors.

Aaron

Title: Re: inductor choice for lna for 60ghz or higher freqs
Post by kamela on Mar 25^th, 2013, 2:44am

Hi
thanks for your reply but I have read that spiral inductors become too complex at higher freqs( 40ghz and above ) and their models for no longer are accurate .

Title: Re: inductor choice for lna for 60ghz or higher freqs
Post by raja.cedt on Mar 25^th, 2013, 2:58am

hello,
Modeling problems are very difficult to talk like this, generally spiral inductors and transmission lines are similar from modeling point of view, but as aaron_do pointed out spiral inductors are smaller due to mutual coupling as compared to transmission lines but at the same time spiral has bad return path (i mean to say generally there is no gnd plane) but for Tlines defined ground will be there. One more thing at 60G probably the inductance required will be smaller (in the order of 200-300pH) that's why many people at this range and mm wave Tlines. Normally Tlines have higher self resonance frequency.

Have a look at 60G publications, they all preferred Tline with high Z0 rather spiral

http://www1.ee.ucla.edu/~brweb/papers/Journals/RJan06.pdf

Thanks,
Raj.

Title: Re: inductor choice for lna for 60ghz or higher freqs
Post by aaron_do on Mar 25^th, 2013, 5:27pm

Hi all,

Quote:

Have a look at 60G publications, they all preferred Tline with high Z0 rather spiral

http://www1.ee.ucla.edu/~brweb/papers/Journals/RJan06.pdf

Are you sure about that? That looks like quite an old paper. If I remember correctly, several of Prof. Razavi's later 60 GHz works used inductors. Anyway, to be safe, t-lines may be a better choice.

Quote:

thanks for your reply but I have read that spiral inductors become too complex at higher freqs( 40ghz and above ) and their models for no longer are accurate

yes you are quite right...

Quote:

your EM modeling will probably need to be much more accurate for spiral inductors

One problem which applies to both spiral inductors and t-lines is you need to make sure to properly define where the inductor starts and ends. Otherwise the parasitic inductance could be on the same order of magnitude as the designed inductance. For spiral inductors, coupling to nearby metals and other inductors may be stronger, meaning that you will need to do some kind of multi-port EM simulation to include all major structures. Equally important is the return path...

Quote:

Another consideration is the return path which should be much better defined for a transmission line than inductors.

It is important to have a well-defined return path to prevent unwanted interaction between different stages in the circuit.

regards,
Aaron

Title: Re: inductor choice for lna for 60ghz or higher freqs
Post by raja.cedt on Mar 26^th, 2013, 3:12am

Hello Arron,
Sorry, i should be careful while concluding , as far as i know many people are using Tlines even in recent publication, of course this depends on many things for example prof Sorin P. Voinigescu used spiral inductors in his 60G Chips. One good thing about Tline is less coupling from substrate due to shielding. Spiral inductors has some coupling problem, but recently published 8 shaped inductor (attached fig) has less coupling.

One imp thing parasitic inductance will play critical role in both cases, one example what observed recently is from inductor to transistor base connecting Via has ~7pH inductance.

Thanks,
Raj.