Why my S21 is different in momentum simulation?

EDAboard.com Forum Index -> RF, Microwave, Antennas and Optics -> Why my S21 is different in momentum simulation?

Author

Message

TerryADS

Joined: 31 Oct 2008
Posts: 116
Location: USA

06 Apr 2009 2:26 mesh option in momentum of ads

Hi friends,
I am simulating bandstop filter(5.5G~6.5GHz) in momentum.When i set the stop frequency as 7.5GHz,i got:

And if i set the stop frequency is 8GHz,I can get:

My question,why can i get the different result as i set the different stop frequency?which one is correct?
Thanks,

kspalla

Joined: 19 Aug 2008
Posts: 848
Helped: 140

06 Apr 2009 6:11 ads momentum mesh frequency

I guess this is problem of mesh frequency and adaptive samples of simulation frequency.
Is it possible to share the design?
PM me if not desired for broadcast.

Google
AdSense
Google Adsense

06 Apr 2009 6:11 Ads

ahphs1983

Joined: 06 Apr 2009
Posts: 4
Helped: 2

06 Apr 2009 10:47 my\|phone s21

Not surprising with ADS momentum simulation, the mesh size is directly related to the stop frequency. You could manually set the mesh solving frequency, thereafter you could get the same result.

kspalla

Joined: 19 Aug 2008
Posts: 848
Helped: 140

06 Apr 2009 10:55 momentum mesh filter

I observed the following violations.
1) W33 must be less than L33, design is not flattened once exported to layout.
2) Mesh frequency is set to 3GHZ only but has to be 8GHz
3) no of freq sampling points are limited to 10 only but for filters kind designs requires nearly 40 samples to cover the entire frequency range.

Sorry, but you need login in to view this attachment

TerryADS

Joined: 31 Oct 2008
Posts: 116
Location: USA

07 Apr 2009 14:45 mesh frequency,momentum

ahphs1983 wrote:

Not surprising with ADS momentum simulation, the mesh size is directly related to the stop frequency.

Thanks.Why the mesh size is directly related to the stop frequency?and how can i set solution frequency in HFSS,it is the operation carrier or stop frequency in sweep frequency plan?

ahphs1983

Joined: 06 Apr 2009
Posts: 4
Helped: 2

08 Apr 2009 3:37 mesh density ads momentum

TerryADS wrote:

ahphs1983 wrote:

You are welcome!
You may find the answers of all your questions regarding ADS in "helps".
Under Topic "Defining a Mesh "
"The wavelength of this frequency will be used to determine the density of the mesh. In general, set the value of the mesh frequency to the highest frequency that will be simulated. "

But simulation results are meaningless unless you compare them with the practical measurement results.

Based on my experience, ADS is only useful for conversional planar structure. It is hard to simulate those structures which are more complicated.

HFSS is better in 3D enclosure structure EM simulation, but you also need to go through some "helps" before starting to use it.

TerryADS

Joined: 31 Oct 2008
Posts: 116
Location: USA

08 Apr 2009 15:19 different momentum simulation

I see now.Thanks kspalla and ahphs1983.

anil_21p

Joined: 14 Jul 2005
Posts: 33
Helped: 2

08 Apr 2009 17:07 Why my S21 is different in momentum simulation?

Momentum automatically takes highest frequency of simulation as mesh frequency you keep mesh frequency box blank. It is better keep this box blank

biff44

Joined: 24 Dec 2004
Posts: 1833
Helped: 244
Location: New England, USA

08 Apr 2009 18:30 Re: Why my S21 is different in momentum simulation?

The obvious question was already asked: which one is correct?

I think the answer might be: Neither one!

You want your program to use a mesh size that is small enough so that making it smaller does not effect the results. I think that since the program is making two different predictions based apparently mesh size, you are maybe not there yet. Try another run with a still higher frequency and see how it matches the second highest frequency run.

I think if there is a manual mesh size control, you might want to use it! At the least, you want to be able to build the filter, see the actual response, and come back and use the same program for the respin (possibly forcing the predicted data to match the measured data by screwing with the substrate εr). You want the exact same mesh size for each of such calculations.

Obviously, the smaller the mesh size, the longer the computation time (it could take days!), so start off with a large mesh, get the filter close, and shrink the mesh size at the very end of the design.

Rich
Maguffin Microwave
www.MaguffinMicrowave.com