Re: [odin-devel] sinusoidal gradients

[Lazar F. <laz...@gm...>]

Hello everybody,

I think the same problem exists with SeqGradTrapezParallel() even if
trapezoidal ramps are used. When I set gradeint integral to some high value,
the gradient amplitude does not reach the maximum gradient strength
set in the constructor, making horizontal part of the gradient longer
than it could have been.

GxyzOscillator2 = SeqGradTrapezParallel("Oscillat2",0.01, 0.01, 0.01,
0.8*systemInfo->get_max_grad());
produces a gradient with 0.021 mT/mm instead of 0.032 mT/mm
( I have the system limit at 0.04 mT/mm for the gradeint strength)

Can some one else observe the same result?

Lazar

On Fri, Apr 3, 2009 at 8:50 PM, Lazar Fleysher <laz...@gm...> wrote:
> Dear Thies,
>
> On Fri, Apr 3, 2009 at 12:30 PM, Thies Jochimsen <th...@jo...> wrote:
>> Dear Lazar,
>>
>> can you please post the relevant part of your code?
>>
>> Thank you, Thies
>>
>
> Here is the line of code which I am using.
>
> I am trying to set a sinusoidal gradient with amplitude equal to the 0.8 of max
> and which has a period (full sinus) of 1 ms.
>
> The relationship between the gradient integral (under one sinus lobe) with the
> maximum amlitude is: intergral = gmax*period/pi., the ramp duration
> is, obviously,
> period/4. Of course, this is limited by the slew rate, so basically, I
> can not have
> very high gradient amplitudes with very short periods.
>
> However, when this gradient is simulated, it has relatively large
> horizontal component.
>
> // integral = gmax/pi * period of oscillation, and the ramps are equal
> to the quater period, try for oscillation period of 1 ms
> // slew rate limit sets limit on max oscillation frequency to use
> sinusoidal gradients and reach maximum amplitude:
> // gmax*omega < slewrate limit
>     GxyzOscillator1 =
> SeqGradTrapezParallel("Oscillat1",0.8*systemInfo->get_max_grad()/M_PI*1.0,
>
> 0.8*systemInfo->get_max_grad()/M_PI*1.0,
>
> 0.8*systemInfo->get_max_grad()/M_PI*1.0,
>
> 0.8*systemInfo->get_max_grad(),
>
> 0.01,half_sinusoidal,1.0/4.0);
>
>
> Am I doing something wrong?
>
> Thanks
>
> Lazar
>>
>> On Freitag, 3. April 2009, Lazar Fleysher wrote:
>>> Hello everybody,
>>>
>>> I can not seem to be able to make sinusoidal-shaped gradients.
>>> I was thinking about using trapesoidal gradients and set the rapmtype
>>> to "half_sinusoidal".
>>> Then I made a calculation for the integral, max gradient and
>>> rampduration so that the
>>> shape of the gradient will be a sinus (with horizontal part of the
>>> trapezoid equal to zero, only ramps).
>>> Instead, I seem to observe that some horizontal part of the gradient
>>> is kept and the gradient amplitude
>>> is reduced to produce correct gradient intergral. I know that if I set
>>> the total length of the gradient to
>>> a very short one, I will exceed the allowed gradient slewrate and this
>>> would happen. But I have done
>>> checking and the slewrate is 1/3 of the maximum. This indicates that
>>> there is some sort of hard limit
>>> on the minimum duration of the horizontal part of the gradient which
>>> is substantial. I could not find where
>>> it was set in the code.
>>>
>>> Am I trying to do something illegal? maybe it is a design feature
>>> based on some limits which I do not understand,
>>> but then, how to make sinusoidal gradients?
>>>
>>> Thanks a lot for your help
>>>
>>> Lazar
>>>
>>> ---------------------------------------------------------------------------
>>>--- _______________________________________________
>>> od1n-devel mailing list
>>> od1...@li...
>>> https://lists.sourceforge.net/lists/listinfo/od1n-devel
>>
>>
>>
>> ------------------------------------------------------------------------------
>> _______________________________________________
>> od1n-devel mailing list
>> od1...@li...
>> https://lists.sourceforge.net/lists/listinfo/od1n-devel
>>
>