Discussion about birdcage coil and losses

[Tim.Herrmann]

From: Dr. Antonino M. Cassarà
Physikalisch-Technische Bundesanstalt (PTB)
Date: 20. Feb. 2013


Dear Tim Herrmann,

I recently am working with some very detailed simulations of birdcage coil
and am facing the problem to accurately reproducing the S-parameters and
the
resonant curve at 3T. It happened that simulating the coil without
including the coil losses, the linewidth of the resonance is too narrow,
and already when I move
the load within the coil of a few cm, am out of the resonance. I think the
problem is the too high Q-factor of my coil, that is not including losses.

How can I include them into the simulations? I am thinking by adding
resistances in series to the lumped elements, but how to decide their
values? I am trying
to acquire information about the Q-factor of our coil, and from it trying
to estimate the resistance to include in the simulation.

Have you some ideas? Have u never faced this problem and have suggestions?

Thank you so much,
Antonino

p.s. There is a way to pose this kind of questions to the MRI-discussion
group?


From: Tim Herrmann
Department of Biometry and Medical Informatics
OvG University Magdeburg
Date: 20. Feb. 2013


Dear Antonino, I see your point.

If you remember the talk of Jens Hoffmann at MR-EM meeting in Essen, he described the power loss by the used caps.

Every capacitor has a series resistance ESR. (Equivalent Series Resistance).

Normally ESR = 0,01-0,05 Ohm at 100MHz for Ceramic Non Magnetic Caps. ( See product manual)


Coil Losses:

use copper as material instead PEC (Perfect Electrical Conductor)
use the ESR value of our real caps in the Lumped Elements with Serial RLC ( See figure)
you can use also a variable in contrast of a value for all Lumped Elements for a better control




Please become member of the MR-EM-Forum: http://www.mr-em-forum.org/

I will publish our Email conversation at our Forum.


Best regards

Tim

From: Dr. Antonino M. Cassarà
Physikalisch-Technische Bundesanstalt (PTB)
Date: 21. Feb. 2013


Dear Tim,
Thank you for the answer. Please could you send me the link to the
discussion?
I think there is another major factor to include for the losses. Is the AC
resistance of the coil due to the skin effect.
Already at 1 MHz, the resistivity of a 1mm diametere copper wire increase
of a factor 3.8, and at 140 MHz it becomes 11 times higher.
For a whole body birdcage coil having radius 35 cm, the total lenght of
each endring is about 2.2m. This would rougly correspond to a DC resistance
of just the endring of about 0.23 Ohm. If we consider the factor 11 more or
less reasonable, that would meant about 2.3 Ohm for only one endring.

This is already larger than the ESR of all the capacitors in the coil.
Don't u think it is an important factor to be included in the simulation?

Many thanks,
Antonino


From: Tim Herrmann
Department of Biometry and Medical Informatics
OvG University Magdeburg
Date: 20. Feb. 2013



Dear Antonino, I will sent you the link.

The skin effect and the resistance of the copper will generate losses in the Birdcage coil.

If you choose cooper as material the simulation time for GPU calculation for each port will increase by 5 - 8% because the loss calculation.

By including those parameters in your simulation you will get closer to real values and the coil matching will not so strong change by load movement.


Best Tim.

It might be interesting to include the solderings as well. I would suggest do fabricate a small test board and solder a capacitor used in the design of the coil and measure the whole setup. From this setup one could derive the s-parameters and put those s-parameter blocks in the cosimulation instead of the ideal lumped elements. I suggest also to simulate the bore of the magnet and check if there is an influence.

[Tim.Herrmann]

This is a good idea. Antonino could implement this in his setup.