antenna-theory.com

[gelunmak]

hi everyone,

i'm doing some rf coil, with tuning capacitors, simulations using HFSS. The impedance bandwidth is very very narrow (S11), which is expected. I found that when i changed some minor values in the simulations, say just move the coil to the left or right by delta or increased the air-box size a little, the s11 curve chaged a lot!!!

It means that the HFSS doens't help in my case, . Actually, i'm looking for a simulations software that's good for RF coil design, coil array designs, coil to coil isolations, for MRI applications! Does anyone have any suggestions on good simulations software for MRI applications?

Thanks

[admin]

I think you are probably screwing something up with HFSS if that is the case. Since it is FEM, it is full wave and should be able to handle an RF coil.

As for MRI, I believe if you are modeling body tissue and that stuff the most common tool is FDTD. I wouldn't recommend a commercial version though, especially if you have time to write one yourself.

[gelunmak]

Actually, the simulations is very simple, i'm quite sure nothing screwed up because the simulated result are pretty much agree with experiment in term of the resonant frequency (S11 dip).

I maybe misleading you by saying "a lot" in the first post. The S11 did changed but only say less than 0.1%. To be more correct, the smithchart curve looks the same when i chaged some minor values in the simulations. Since i was looking at very narrow sweep bandwidth, a little change of S11 dip looks "a lot" in my screen.

My actual problem is for two surface RF coils, partially overlapped, to find a situation that their coupling is weak at the S11 dip at simulations. (i.e. Good matching S11 and Good isolation S21 at the resonant freq). It's possible experimentally but things have to be very precise (ovelapping distance and the tuning caps value). Therefore, i've to adjust the simulation parameter little by little to look for the case. However, things are keep changing and very difficult to achieve the desire situation in HFSS.

Up to now, situation of single-good matching (single-S11 dip) but poor isolation (S21~0dB) was found in the simulations. The problem was i cannot fine-tune the things in the HFSS to get the desire situation like the real prototype (Good matching and Good isolation).

And after series of trails and tests in HFSS, i'm highly believed that HFSS cannot simulate correctly for the case.

Seeking for supporters/advices if anyone got similiar experience.

THANKS

[chaicustard]

This seems an unusual situation ( for typical RF circuits). My first thought is this: are you designing to somehow absorb the energy sent into the coil (S11) ? If S11 is low, then power flows in, but not out. So it sounds like you are actually designing 2 adjacent small antennas, that have a very low coupling coefficient.
In this concept, adding other focussing or resonating materials in/ nearby the coil may help to ease the highly sensitive un-desired coupling problem. In any case, the energy must be radiated / coupled somewhere, or absorbed as heat.
Trying to acheive better than 40 or 50 dB isolation will be very tricky without careful design to manage changes in the physical environment. ( EG temperature, aging effects, vibration) . Good luck!