different result under same evolution time

[Rongchun Zhang]

Hi, guys,

I am trying to simulate some signal as a function of evolution time under MAS.

However, there is some weird that I could not quite understand.

I am sorry I could not show you the sequence.

For example, I set the MAS rate=20kHz. There are three same evolution period during my sequence

If I set the increments of the evolution time as 5us, I got a signal curve as a function of the evolution time.

However, if I set the increment of the evolution time as 50us(rotor-synchronize), at a long evolution time, I found the simulated result is completely different from that when increment was 5us. It is really strange that I did not quite understand.

PS: I used n_gamma=12,  Euler_angles=rep168

Thanks for the help very much.
-- 

Best regards,

Rongchun

[Rongchun Zhang]

Hi, I just found out that the problem comes from the setting of n_gamma. Sometimes even n_gamma was set as 12, the program set it as 20( shown by the "spinev xx.in -s" commands). That could explain the difference. However, it makes me curious about how should we choose and set the value of n_gamma. Different values gave completely different simulation results.

Thanks

--

Best regards,

Rongchun

**************************************************
张荣纯 Rongchun Zhang
Room 4002, Chemistry Building, 

Biophysics and Department of Chemistry,

University of Michigan, Ann Arbor, MI, US.

Mobile: (01)734-353-9636

Email: zrc...@gmail.com
           rozh...@umich.edu

**************************************************

[Mikhail Veshtort]

Hi Rongchun,

n_gamma specifies the minimal number of steps/nodes to use for averaging over the gamma angle. The program may actually choose a larger number of steps to use if this will make the computation more efficient. This can be seen, as you have noticed, with the help of the -s option.

“Powder averaging” is essentially a numerical integration of your results (spectrum or FID) over all possible orientations of the molecule. When you choose the powder averaging set and n_gamma for the simulation, you choose the algorithm that the program will use to perform this numerical integration. If you specify a powder averaging set or n_gamma that are not sufficiently fine for your problem, the results of the integration will be inaccurate. Since it is difficult to know ahead of time if some powder averaging scheme will be good enough for the specific simulation at hand, one always has to check the results for convergence. To do this, you have to repeat the simulations using some finer powder averaging scheme and compare the results with your previous computation. If the results did not change (within some error limits acceptable to you), that means conversion is reached. However, if you see that the results do change when a larger number of gamma-steps (that can be seen with the -s option) or a larger number of {alpha, beta} nodes are used, then you must choose some finer scheme for powder averaging.

I hope this clarifies your question.

Best regards,

Mikhail

--
You received this message because you are subscribed to the Google Groups "spinev-discuss" group.
To unsubscribe from this group and stop receiving emails from it, send an email to spinev-discus...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[Rongchun Zhang]

I see。thanks。：—）