|
RadioProcessor NMR Interface for MATLAB
Contents
- Legal
- Installation
- Single Pulse NMR
- Hahn Echo
- CPMG NMR
- 90 Degree Pulse Width Finder
- Included Functions
- How to Recompile the NMR Interface
Legal
Copyright © 2008 -
SpinCore Technologies, Inc.
This software is provided 'as-is', without any
expressed or implied warranty. In no event will the
authors be held liable for any damages arising from the
use of this software.
Permission is granted to anyone to use this software
for any purpose, including commercial applications.
Installation
To install and ready the console for use, simply
extract the zip file to the desired location. In MATLAB,
change the current directory to the location where the
files were unzipped.
After changing the current directory in MATLAB, execute
Single Pulse NMR m-file (spincore_spnmr_gui.m) by either
typing "spincore_spnmr_gui" in the MATLAB command
window, right-clicking the file from the command
interface and choosing Run,
or running the file from MATLAB's m-file editor.
Other interfaces may be executed in a similar manner.
Single Pulse NMR
This program is first and foremost an easy-to-use
version of the single_pulse_nmr outlined in the
RadioProcessor manual. Since operation of this interface
relies heavily on the original single_pulse_nmr code,
the section of the RadioProcessor manual relating to NMR
is still relevant here. Operation of the Single Pulse
NMR program is dominated by the use of a set of
easily-modified input parameters and two buttons.
Performing a Scan
Upon selecting the desired input parameters and
pressing the Perform Scan button, the
parameters will be passed through MATLAB to the
RadioProcessor board to initiate the scan. Since the
board can perform the scan independent of any further
user interaction, control is returned to MATLAB and the
user immediately after sending the command to begin
scanning.
It is worth noting that before beginning a scan, the
user can choose between 2 pre-programmed pulse shapes in
addition to being able to change a wide range of other
input parameters. The shape of these pulses is available
in Appendix II of the RadioProcessor manual. Note that
choosing an amplitude value of 1 produces a 4 V
peak-to-peak pulse in the desired shape.
Stopping a Scan
In the event a scan has been initiated and the user does
not wish for it to continue on in its entirety, the scan
may be stopped prematurely. To stop a scan in
progress, the user should use the Stop Scan button.
Downloading Data
To view the data acquired by the board, the Download
Data button must be pressed. Because of the
RadioProcessor's design, data can be downloaded at any
time, even while the RadioProcessor is performing a scan
and acquiring data. Once the button is pressed, all data
stored in the RadioProcessor's RAM will be downloaded
into MATLAB and plotted in the desired format.
In addition to downloading the data to MATLAB, the Download
Data button also creates ascii and jcamp-dx
5.0 files containing the data. These files are saved
under the name specified in the "filename" field at the
time the button is pressed and stored in a named
subfolder in the LogFiles
subdirectory of SPNMR11a.
If the Download Data button is
pressed while a scan is in progress, the data will still
be downloaded and displayed. In addition, a status
update will appear notifying the user to the current
progress of the scan. It is important to note that this
does not interrupt or stop an active scan.
Changing the values between running the scan and
downloading the data will not impact the proper display or
storage of the data.
Retrieving Data Later
Each time data is downloaded from the RadioProcessor, a
MAT-file is created and placed in a named folder in the LogFiles subdirectory
of SPNMR11a.
To accommodate for a variety of programs, the data is
saved as a variety of filetypes including Ascii, J-Camp Dx, and MAT-File.
The MAT-file is designed for use in MATLAB and contains
all relevant parameters and data acquired during the
scan. The MAT-File can be viewed later in the Single
Pulse NMR Interface by selecting the Load Data File option
under the File menu.
Upon selecting this option, the user will be prompted to
select a Mat-File associated
with the Single Pulse NMR Interface. If the user
navigates to the LogFiles
subdirectory and then to a MAT-File created in a previous
session, all the relevant data will be loaded into the
interface as though the user had just downloaded the data
from the RadioProcessor.
Saving and Loading Settings
The Single Pulse NMR comes pre-loaded with a set of
default parameters that appear every time the program is
run. In the event the user changes these parameters and
wishes to have quick access to them later, a Save
Settings option is included under the File menu.
Choosing this option will allow the user to save the
current settings as a MAT file.
Similar to the Save Settings option
is the Load Settings option which is
also located under the File
menu. This option will prompt the user to select a MAT
file in which previously selected settings have already
been stored.
This feature differs from the Load Data File
option in that these functions save and load only the
input parameters. The Load Data File option loads not only the
input parameters but also the generated NMR data as
well.
Detecting Resonance Frequency
One of the more valuable features of the Single Pulse
NMR Interface is its ability to automatically detect the
resonant spectrometer frequency. Available under
the Edit menu,
Set to Resonance
will run a scan, download and analyze the data to
automatically determine the resonant spectrometer
frequency. Upon completion, the Spectrometer
Frequency box in the interface will automatically be set
to this new frequency.
The algorithm for detecting resonant frequency is not
particularly robust and therefore will not work in every
scenario. Primarily, the algorithm works by
observing the power spectral density of a given FID and
observing the frequency offset of the largest
peak. The algorithm then adjusts the spectrometer
frequency such that this peak is centered at 0 Hz.
In particularly noisy systems or systems with some types
of systematic errors, the resonant frequency detection
may not function properly.
Calculating Signal-to-Noise Ratio (SNR)
After performing a scan and downloading the data to the
interface, the Calculate
SNR option becomes available under the Edit menu.
Choosing this option opens a new window which displays a
semilog plot of the FFT of the data.
To use this interface, the user must click the Select signal and noise
button. The user should then select the signal
region of the data containing the signal peak of
interest. Next the user will be prompted to select
a representative section of noise from the plot.
After these steps, the plot will have a signal portion
highlighted in green and a noise portion highlighted in
red. The user can then calculate the
Signal-to-Noise Ratio of the data by pressing the Calculate SNR button
or can select new regions of the plot by again clicking
the Select signal and
noise button.
Currently the Signal-to-Noise Ratio is calculated as
the ratio of the maximal peak detected in the signal
region to the RMS value of the selected noise
region. These values are currently meant for
comparison purposes only and are not intended to be used
as absolute data.
Scanning Continuously
SpinCore's Single Pulse NMR Interface also has the ability
to perform continuous scans. To scan continuously,
the user should set the input parameters then simply press
the Scan Continuously button.
This will then cause the RadioProcessor board to read the
input parameters and begin scanning. In this mode,
the RadioProcessor performs a scan, downloads the data,
performs another scan and so on until the user stops the
scan.
The input parameters become locked in place while the
continuous scan is running. This requires the user
to temporarily stop the scan in order to change the
parameters.
Utilizing Undersampling Functionality
As the RadioProcessor user's manual indicates, the
RadioProcessor is capable of acquiring input signals with
frequencies of interest above the Nyquist frequency.
Although aliasing may occur, the RadioProcessor is still
capable of adequately detecting and acquiring these
signals. The MATLAB interface has been coded to
reflect these capabilities.
Explanation of input parameters:
Number of Points: Specify the number
of complex data points to capture, this should be an
integer between 1 and 16384.
Spectrometer Frequency: The frequency
of the spectrometer, in MHz
Spectral Width: The desired baseband
spectral width, in kHz.
Pulse Duration: Duration of TX
excitation pulse, in µseconds.
Ringdown Time: Transient, or ringdown
time, in µseconds.
Repetition Delay: Delay between scans,
in seconds.
Number of Scans: Number of times to
scan and average the data.
Output Phase: Phase of the TX output
channel, in degrees.
Clock Frequency: Clock frequency of
the board, in MHz.
Bypass FIR: Check to bypass the FIR
filter, leave unchecked to use it. If this is left
unchecked, the maximum useable Spectral Width is 135
kHz.
Output Amplitude Scaling: Amplitude of
excitation pulse, a value between 0 and 1.
File Name: Name of the file the
results of the scan will be saved under.
Pulse Shape: Choose either to use a
hard pulse with the scan or a sinc pulse.
CYCLOPS: Single Pulse NMR GUI can run CYCLOPS
by checking the CYCLOPS checkbox.
Timing Diagram
In this experiment, a single scan is performed as
follows: a single RF pulse is applied to the sample for
a specified amount of time (Pulse time). The high
voltage induced on the sample coil is allowed to ring
down (Transient time), after which the Free Induction
Decay (FID) signal is acquired (for length Acquisition
time). The sample is then allowed to relax for time
Repetition delay. This scan procedure is then repeated
an arbitrary number of times as desired to improve the
signal to noise ratio.
In the timing diagram below, tx_enable, phase_reset,
and trigger_scan are the control lines used by the
internal PulseBlaster timing core to control the
experiment.
Hahn Echo
The Hahn-Echo GUI is an interface introduced in Release
12a that allows users to easily perform the Hahn-Echo
experiment using MATLAB. The Hahn-Echo interface can be
executed by either typing in "spincore_hahnecho_gui" in
the Command Window or run the spincore_hahnecho_gui.m file
in our Spnmr12a package.
The Hahn-Echo experiment applies a 90 degree pulse, waits
a period of Tau, and then applies a 180 degree pulse. This
causes an echo in the FID response of the sample, which
peaks at a duration of Tau after the end of the 180 degree
pulse.
After the execution of "Perform Scan," the
experiments starts and the status bar will display the
string "scanning." The Command Window returns the current
status of the experiment. The status bar will display the
string "DONE" indicating the experiment is finished. The
interface then will generate two data files - ASCII and
jcamp.
The Hahn Echo interface provides a ''Set to Resonance"
feature which sets the the spectrometer frequency to a
resonant spectrometer frequency value. However, due to the
noise of the system, this feature is not very robust due
to the noise of the system.
The Hahn Echo interface provides the option of processing
ASCII data, which allows users to process existing ASCII
data generated by RadioProcessor and plot the data on to
the interface. The default address of the ASCII data is
./LogFiles/hahn_echo.
SpinCore Hahn Echo Interface:
(an example of an off-resonance data)
Timing Diagram:
Explanation of Input Parameters
File Name: the name of the output file containing
the experiment data
Board Number: the board number in your system to be
used
Blank Bit: which TTL Flag to use for the power
amplifier blanking signal
Debug: enables the debug output log
ADC Frequency (MHz): the ADC frequency of the
RadioProcessor board
Spectrometer Frequency (MHz): Must be between 0 and
100 MHz
Spectral Width (kHz): Must be between 0.15 and
10000 kHz
Amplitude: the amplitude of the excitation signal,
must be between 0.0 and 1.0
Pulse 90 Time (us): 90 degree pulse time, must be
at least 0.065 us
Pulse 90 Phase (degrees): 90 degree phase, must be
greater or equal to 0 degree
Pulse 180 Time (us): 180 degree pulse time, must be
at least 0.065 us
Include_90: If checked, the 90 degree pulse
response will be included in the acquisition.
Bypass FIR: If checked, the FIR filter will be
disabled
Number of Scans: the number of consecutive scans to
run (Scan count may not be consecutive due to latency)
TAU (us): the wait time between 90 degree pulse and
the 180 degree pulse, must be at least 0.065 us
Blanking Delay (us): TTL blanking delay, must be at
least 0.065 us
Transient Delay (us): transient delay, must be at
least 0.065 us
Repetition Delay (s): the time between each
consecutive scan, must be greater than 0
CPMG NMR
The CPMG NMR GUI is an interface new to Release 2 that
allows users to quickly and easily perform CPMG NMR
experiments. Modeled after the original Single Pulse
NMR interface, users will note that the two interfaces
operate in a very similar manner. Although some of
the parameters differ and the frequency domain plot as
well as the SNR calculator have been removed, the two
interfaces otherwise operate identically.
Capturing Data
in a Continuous Scan:
To capture data in a continuous scan, the Points per Echo
parameter must be set to 0. This will cause the
RadioProcessor to trigger and begin capturing data
immediately after the transient time on the 90 degree
pulse. Once data has begun to be captured in this
manner, it will continue until the desired number of
echoes have been performed.
Capturing Only the Peaks of the Data:
As an alternative to capturing data continuously, it is
possible to capture only the peaks of each echo. For
this, the user must specify some value for Points per Echo and Number of Echoes other
than 0. In this case, the RadioProcessor will capture the number
of points specified by Points
per Echo around the peak of each baseband signal
peak. Also, the total number of data points to be
downloaded in this case is equal to the value specified by
Points per Echo multiplied
by 1 plus the value specified in Number of Echoes.
If this results in a value greater than 16k, the user will
receive an error message.
Downloading Data
Similar to the Single Pulse NMR Interface, the scan and
download functions have been separated for most function
of the CPMG NMR Interface. As such, the user must
press the Download Data
button after performing a scan if they wish to view the
results of the scan. This action not only
immediately downloads and displays the data in the current
interface, but also stores all relevant information in
various log files. Note that if the Download Data button
is pressed before the entire experiment has complete you
may obtain unwanted data. If you are performing
multiple scans, the status bar will notify the user if the
experiment has completed and if it has not completed, what
percentage of the experiment has completed when the Download Data button
is pressed.
As with the Single Pulse NMR Interface, there are 4 log
files created each time the Download Data button is pressed.
These files are saved under the LogFiles subdirectory in an automatically
generated folder with the name specified in the File Name box
associated with the last scan.
Calculating T2:
For NMR samples that produce data with singular
exponential decay during a CPMG experiment, the CPMG NMR
interface can attempt to calculate the appropriate value
of T2. This process is done by first observing a
series of single maximal points collected from the peak of
each echo. The number of echos to observe and thus
the number of datapoints can be specified by changing the
Number of Echoes
parameters. Users may find that their calculations
for T2 vary greatly depending on the number of echoes they
specify.
Once this data has been collected, T2 is calculated by
finding the slope of the line of best fit on a semilog
plot of the peaks. After T2 has been calculated, its
value will be displayed in the original CPMG Interface
window. Also, a new figure window will be generated
containing Cartesian and semilog plots of the first 150
peaks with the calculated exponential overlayed.
Note that the T2 calculation requires that only one point
per echo is captured. If the Number of Echoes is
large, noise may dominate the signal for later echoes.
This will cause the decay to be non-exponential and the
line of best fit will not properly fit the exponential
decay. For better results, modify the number of
echoes so that the acquired data, when plotted on the
semilog scale appears linear.
Explanation of Input Parameters:
Spectrometer Frequency:
The frequency of the spectrometer, in MHz
Spectral Width: The desired baseband
spectral width, in kHz (this value is equal to the
number of points sampled per millisecond).
Pulse Duration: Duration of 90 degree
excitation pulse, in µseconds. The 180 degree
pulse is calculated to be twice this value.
P2 Phase: Output phase of the 90
degree pulse. The 180 degree pulses will have
a phase offset of 90 degrees from this value.
Ringdown Time: Transient, or ringdown
time, in µseconds.
Tau: One half
the time between 180 degree pulses. Also, the time
between the 90 degree pulse transient time and the first
180 degree pulse.
Points per Echo:
Number of points to capture around peak of each
echo. Set to 0 for continuous scan.
Number of Echoes: Total
number of echos to capture data from. Parameter is unused
during continuous scan.
Repetition Delay: Delay between scans,
in seconds.
Number of Scans: Number of times to
scan and average the data.
Clock Frequency: Clock frequency of
the board, in MHz.
File Name: Name of the file the
results of the scan will be saved under.
Bypass FIR: Check to bypass the FIR
filter, leave unchecked to use it. If this is left
unchecked, the maximum useable Spectral Width is 135
kHz.
Timing Diagram:
For more information about CPMG experiments, see: http://spincore.com/CPMG/
90 Degree Pulse Width Finder
The 90 Degree Pulse Width Finder is a new interface
designed to run a series of single pulse scans with
variable pulse widths that will ultimately allow automatic
detection of the 90 and 180 degree pulse widths.
Pulse Width Detection
The 90 Degree Pulse Width Finder attempts to detect both
90 and 180 degree pulse widths by observing the amplitudes
of the magnitude output of a series of single pulse
scans. These scans vary the pulse width from some
user-specified initial value to a user-specified final
value in user-specified increments while holding all other
input parameters constant.
After each NMR scan is completed and the FID data is
acquired, an amplitude for that FID is calculated by
finding the average of a pre-defined number of points
following some offset at the beginning of each FID
scan. By default, these values are defined in the
m-file code to be 15-point averaging and a 10-point
offset. Therefore, the amplitude of each scan is
taken to be the average of the 11th to 25th
time-sequential amplitude values of the FID. The
user is welcomed to change these values as they please.
These calculated amplitude values are gathered for the
real, imaginary and magnitude data and are plotted
sequentially in the interface until the scan is completed
or stopped. The FID that each value is being based
off of is displayed in real-time as well so that the user
may attempt to observe or diagnose any errors that may
occur.
Once the full range of scans has been completed, the
software attempts to automatically detect the 90 degree
and 180 degree pulse widths using the acquired data.
The software performs this detection by observing the
magnitude data amplitudes and assigns the 90 degree pulse
width to be the time index of the maximal amplitude
value.
The 180 degree pulse width is then detected as the index
of the first valley that appears in a filtered version of
the magnitude data amplitudes. This filtered version
of the magnitude data is simply the original magnitude
data that has been smoothed by a simple 4-point moving
average filter in an attempt to remove any erroneous
valleys that may return improper values for the 180 degree
pulse width.
Due to the somewhat simple method of detection, the pulse
width values may not be accurately calculated if the
system is particularly noisy or has significant errors.
Stopping the Scan
In the event a scan has been initiated and the user does
not wish for it to continue on in its entirety, the scan
may be stopped prematurely. To stop a scan in
progress, the user should use the Stop Scan
button. Once the scan is stopped the user can still
change the plot and open the figure in a new window as
though the scan had completed. The automatic pulse
width detection will not occur, however, unless a scan is
allowed to run until completion.
Explanation of Input Parameters
Number of Points:
Specify the number of complex data points to capture in each scan,
this should be an integer between 1 and 16384.
Spectrometer Frequency: The frequency of
the spectrometer, in MHz
Spectral Width: The desired baseband spectral
width, in kHz.
Ringdown Time: Transient, or ringdown time, in
µseconds.
Repetition Delay: Delay or relaxation time
between scans, in seconds.
Number of Scans: Number of times to scan and
average the data.
Output Phase: Phase of the TX output channel, in
degrees.
A/D Clock Frequency: Clock frequency of the
board, in MHz.
Output Amplitude Scaling: Amplitude of
excitation pulse, a value between 0 and 1.
Bypass FIR: Check to bypass the FIR filter,
leave unchecked to use it. If this is left unchecked, the
maximum useable Spectral Width is 135 kHz.
Pulse Shape: Choose either to use a hard pulse
with the scan or a sinc pulse.
Initial Pulse Width: The initial pulse
width to begin scanning from.
Final Pulse Width: The final pulse width
to terminate scan at.
Step Size: The step size between pulse
widths. This can be used to improve the resolution
of the detected pulse widths.
Included MATLAB Functions
The source code for several C programs that have been
written specially for use with MATLAB are included with
SPNMR11 under
the MEXSource
subdirectory.
These programs have also been precompiled with the
MATLAB mex function and can be called just like any
other function in MATLAB. It is important to note
that for these functions to work properly, MATLAB must
be made aware of the location of the function's
associated mexw64 file if it is not located in MATLAB's
current directory. The precompiled version
of these programs have the .mexw64 extension and are
located in the EssentialGuiFiles
subdirectory of SPNMR11.
These functions, in their present form, provide a
powerful group of tools for interaction between the
RadioProcessor and MATLAB. Particularly, they
allow for scripting and construction of more advanced
programs for the simplistic acquisition and importing of
data.
These included functions are outlined in the table
below:
| Name |
Prototype |
Description |
| pbGetData |
[REAL, IMAG]=pbGetData() |
This basic function accesses the
RadioProcessor, downloads the data from RAM
and stores the real and imaginary data in the
matrices REAL and IMAG, respectively. While
this is a quick and easy way to import
RadioProcessor data into MATLAB, the data
lacks any frequency information and is simply
a list of magnitudes.
Using this function does not clear the data
from the RAM on the RadioProcessor, nor does
it interfere with a scan in progress. As
such, data may be acquired and displayed many
times before a longer scan has completed.
|
| pbGetDataPlus |
[REAL, IMAG, STATUS, RESONANCE_FREQ]=
pbGetDataPlus(num_points, SF, SW, 'filename',
num_scans, adc_freq, pulse_time, trans_time,
repetition_delay, tx_phase, bypass_fir, amp_val,
shape); |
A more advanced version of pbGetData, this
not only stores the data from RAM into REAL
and IMAG but also creates ascii, felix, and
jcamp-dx files in the active directory using
the specified filename.
This function requires additional inputs for
creating the additional files. These inputs
are the number of points, spectrometer
frequency, spectral width, filename, number of
scans ,adc frequency, pulse_time, trans_time,
repetition_delay, tx_phase, bypass_fir,
amp_val and shape respectively.
The value returned to STATUS specifies
whether or not the scan was successful. A 0
indicates that the data was successfully read
from the RAM. For further information
regarding this function, consult the SpinAPI
documentation concerning the pbGetData
function.
The latest version of spinapi now supports
calculating resonance frequency if available
from the data stored in the RadioProcessor
board. However this value does not guarantee
the correct resonance frequency if there exist
number of different peaks in the data.
This function's capabilities regarding the
ability to acquire data during a scan match
those of pbGetData in that neither the RAM is
cleared nor is a scan in progress stopped when
this function is called.
|
| pbReadStatus |
[STATUS, SCANS] = pbReadStatus() |
This function, when called returns the
current status of the RadioProcessor Board as
well as the current number of scans.
For information concerning the value
returned to STATUS, consult the SpinAPI
documentation concerning the pbReadStatus
function.
|
| pbStop |
pbStop() |
This function stops any scan in progress. It
is not necessary to use this function first if
the user wishes to begin a new scan. Moreover,
this function is intended for users that wish
to stop a scan without beginning a new one.
|
| SpnmrScan |
SpnmrScan(num_points, SF, SW, pulse_time,
trans_time, repetition_delay, num_scans,
tx_phase, bypass_fir, adc_freq, amp_val) |
This function, when called, causes the
RadioProcessor Board to perform a single pulse
NMR experiment using the user-specified input
parameters. This function returns control to
MATLAB immediately after sending instructions
to the RadioProcessor board and does not wait
for the scan to complete before returning
control.
Calling this function does not return any
outputs to MATLAB. Though status updates are
displayed, no actual data is stored as a
result of the function call.
Data acquired during the scan is stored in
RAM on the RadioProcessor board and may be
retrieved with pbGetData or pbGetDataPlus.
|
| SpnmrWithDelay |
SpnmrWithDelay(num_points, SF, SW, pulse_time,
trans_time, repetition_delay, num_scans,
tx_phase, bypass_fir, adc_freq, amp_val) |
This function operates identically to
SinglePulseShapeScan with the exception that
the function does not return control to MATLAB
until the scan has completed.
This function is especially useful for use
in writing scripts that require multiple scans
to be run consecutively.
|
CpmgScan
|
CpmgScan(SF,
SW, P2_time, P2_phase, P2_after, tau,
echo_points, echo_loops, repetition_delay,
num_scans, bypass_fir, adc_freq)
|
This function,
when called, signals the RadioProcessor Board to
perform a CPMG NMR experiment using the provided
input parameters. Similar to
SinglePulseShapeScan, control is returned to
MATLAB immediately after the instructions are
sent to the RadioProcessor Board.
Also similar to SinglePulseShapeScan, CpmgScan
does not return any data to MATLAB. Data
must be acquired using either pbGetData or
pbGetDataPlus.
|
CyclopsScan
|
[ERROR] =
CyclopsScan(num_points, SF, SW, pulse_time,
trans_time, repetition_delay, num_scans,
tx_phase, bypass_fir, adc_freq, amp_val, shape,
blanking_Bits, num_Boards);
|
This function,
when called, causes the RadioProcessor Board to
perform number of CYCLOPS NMR experiments
specified in 'num_scans' using the
user-specified input parameters. This function
returns control to MATLAB immediately after
sending instructions to the RadioProcessor board
and does not wait for the scan to complete
before returning control.
Returns:
0 is returned on success. otherwise the
value has following meanings
-1 : The board does not support cyclops.
Only RadioProcessor with firmware ID later than
12-15 supports CYCLOPS.
-2 : Wrong number of scans entered.
CYCLOPS requires multiple of 4 number of scans.
-3 : No boards found.
|
How to Recompile the NMR Interface
The functions in the above section were all originally
written in C code, the source files of which are available
under the MexSource subdirectory
of SPNMR11.
The source was modified from its original form in order
that it might be compiled properly in MATLAB. This
modification is as simple as adding a gateway Mex function
that handles the passing of data between the C code and
the MATLAB workspace.
When attempting to compile Mex-files that make use of
SpinAPI, the spinapi.h, spinapi.dll and spinapi.lib must
all be present in the current directory for the file to
compile properly.
For example, assume I've created a C file called example.c
that makes use of SpinApi that I have formatted to be
compiled in MATLAB as a MEX-file. The command line
to compile this file might looks something as follows:
mex example.c spinapi.lib
Assuming that the example.c and the various spinapi files
are located in the current directory and the compiler has
been configured properly, this should generate a MEX-file
with some extension appropriate to the platform that
MATLAB is currently running on.
In order to recompile the MATLAB Interface files (or to
compile new MEX-files) the compiler must first be
configured properly. The following steps explain how
to setup the compiler. This only needs to be done the
first time you compile unless there are settings you want
changed:
- Download MinGW and install it to C:\MinGW. You
can download MinGW as part of our
pre-configured compiler package.
- Download Cygwin
and install it to C:\cygwin. You will need
CygwinDevel.exe
- Download gnumex
and extract all files to C:\gnumex.
- Configure the mexopt.bat file by running gnumex.m in
MATLAB. Follow the instructions
available on the gnumex page.
The following steps need to be done every time you
recompile:
- If you wish to download the latest version of the
SpinCore MATLAB interface, you can find it at the main SPNMR page. After it is
downloaded, unzip all the files to your working
directory.
- Copy the latest versions of spinapi.dll, spinapi.lib
and spinapi.h (and registers.h if necessary) into the
MexSource directory of the unzipped MATLAB files.
- Navigate to that directory in MATLAB and type
"mexCompile_64.m" under a 64-bit system or
"mexCompile.m" under a 32-bit system.
- Now your compilation is complete.
In addition, the information located at http://gnumex.sourceforge.net/
may prove to be helpful as well.
All Mex-compiled functions were placed under
/EssentialGUIFiles of the Spnmr11b directory. Please note
that if you were to use a different version of MATLAB
under a 64-bit system, it may be necessary to recompile
the source code by running mexCompile_64.m file under the
mexFile folder.
|
