SpinCore CPMG package for RadioProcessor and iSpin-NMR

Sample CPMG Data

The figures below are of a CPMG experiment conducted on household cooking oil using a 10.8 MHz permanent magnet.

Figure 1. Signal captured with data acquisition on for the entire CPMG experiment. The larg spikes between echoes are the 180 degree pulses being applied.

Figure 2. Signal captured with data acquisition set for only the tops of the echoes. 32 points are captured per echo here.

Both figures above show a CPMG experiment with the same parameters, however instead of capturing data continuously throughout the experiment, only 32 points at the peak of each echo is captured for the second figure. This allows more echoes to be captured by not storing unwanted data. The number of points captured per peak can be reduced to as few as 1 point.

CPMG Pulse Program Parameters

Figure 3 below shows the pulse sequence used by the RadioProcessor to perform this experiment. Each vertical line is a new instruction, and the names of the delays used are shown with arrows. The CPMG package includes an executable that will accept command line parameters which can be used to specify the values of each of the pulse program parameters.

Figure 3. Pulse timing sequence for CPMG experiment.

The top two lines in figure 3 show what the RF excitation pulse will look like, as well as what the acquired baseband data will look like. The Tx_Enable and Scan_Trigger signals are both internal signals used to control the timing of the experiment. Note that when the CPMG program is configured to acquire data in a continuous scan, the Scan_Trigger signal is only enabled after the 90 degree pulse and not after the 180 degree pulses.

The CPMG executable takes the following parameters as inputs:

BOARD_NUMBER - Should be set to zero unless multiple SpinCore boards are connected to the computer. If multiple boards are connected to the computer, BOARD_NUMBER should correspond to the number of the board. (Note SpinCore boards are enumerated beginning with 0. All PCI boards are enumerated before USB boards).
SPECTROMETER_FREQUENCY (MHz) - The spectrometer frequency can be between 0.0 and 100.0 MHz and should be set exactly on resonance.
SPECTRAL_WIDTH (kHz) - The desired baseband spectral width must be less than 10,000 kHz and can be as small as 0.072 kHz for certain designs.. This value is equal to the number of points captured per millisecond. Narrower spectral width values can improve the apparent signal-to-noise ratio of the acquired data.
P2_TIME (us) - The pulse length of the initial 90 degree pulse must be greater than or equal to 0.065 us.. This should be set exactly to the 90 degree pulse of the system being used.
RINGDOWN_TIME (us) - Ringdown time (or dead time) after the 90 degree pulse. This value is specific to the probe and preamplifiers being used in the system and must be greater than or equal to 0.065 us.
P2_PHASE (degrees)- Phase of 90 degree pulse. This value should be adjusted to achieve maximum amplitude for the receivers real channel.
P1_PHASE (degrees) - Phase of 180 degree pulses should be offset by +90 degrees from the P2_phase value.
TAU (us) - 2*tau is the time between 180 degree pulses
NUMBER_OF_POINTS_PER_ECHO - Number of points to capture at the top of each echo. Set to 0 to do a continuous scan (see Figure. 1)
NUMBER_OF_ECHOES - Number of echoes to produce.
NUMBER_OF_SCANS - Number of times to repeat the scan (and average the data)
FILENAME - Filename to store data to. The CPMG.exe program will generte Felix (.fid) and ASCII (.txt) output files.
BYPASS_FIR - A 1 bypasses the signal around the FIR filter, or 0 to use it. In general the FIR filter should be bypassed. See the RadioProcessor manual for details.
ADC_FREQUENCY (MHz) - The Internal clock frequency of the board. This value should always be 75 MHz unless a custom design with a different ADC frequency is being used.
REPETITION_DELAY (s) - Time to allow sample to relax after each scan.
AMPLITUDE - The transmitter output amplitude scaling factor. This value should be between 0.0 and 1.0. Note that the scaling factor is not linear.
BLANK_BIT - Number of the TTL output bit on the RadioProcessor to use as a blanking signal for the power amplifier.
BLANK_DELAY (ms) - The amount of time to assert the blanking signal bit before sending the excitation signal so that the power amplifier may warm up.
DEBUG - Set to 1 to generate SpinAPI debug output, otherwise, set to 0.

There are a few parameters in Figure 3 not specified on the command line. These are calculated as:

P1_time: 2*P2_time (us). This is the time for the 180 degree pulse and is automatically calculated as twice the 90 degree pulse length.
top_time: The top time is equal to the NUMBER_OF_POINTS_PER_ECHO divided by the SPECTRAL_WIDTH (kHz) times 1000.
P2_after1: tau - top_time/2
P2_after2: tau + top_time/2
echo_loop: This corresponds to the three instruction loop used to repeat multiple echoes. The echo_loop pattern is repeated NUMBER_OF_ECHOES times.

Tips for acquiring CPMG Data

Several suggestions for acquiring CPMG data are listed below:

Start with a simple single-pulse NMR experiment. Find the resonance frequency and adjust the transmitter output phase to achieve the largest received signal amplitude for the real channel (the imaginary channel amplitude should be approximately zero).
Find the 90 degree pulse length and verify that the 180 degree pulse length is twice that of the 90 degree pulse length.
Copy the relevant parameters from the single-pulse NMR experiment over to the CPMG program. The relevant parameters are:

BOARD_NUMBER

SPECTROMETER_FREQUENCY
SPECTRAL_WIDTH
P2_TIME (This will be the 90 degree pulse time from the single-pulse test)
RINGDOWN_TIME (Also known as transient time or dead time)
P2_PHASE (This will be the transmitter output phase from the single-pulse test. Add 90 degrees to get the P1_PHASE.)
ADC_FREQUENCY
REPETITION_DELAY
AMPLITUDE

Estimate the TAU time. This should be approximately equal to the time it takes for your single-pulse test FID to decay to zero.
Run the CPMG test initially with NUMBER_OF_POINTS_PER_ECHO equal to zero and NUMBER_OF_ECHOES equal to one. Make sure your SPECTROMETER_FREQUENCY is set to resonance. You should see the inital FID from the 90 degree pulse time, followed by a large spike (the 180 degree pulse), and then an echo. When scaled correctly, the FID from the 90 degree pulse should look nearly identical to the signal acquired using the single-pulse NMR test.
Adjust the TAU and SPECTRAL_WIDTH parameters as necessary to produce the desired echo.
After setting all of the above parameters you can adjust the NUMBER_OF_ECHOES to produce more echoes, and the NUMBER_OF_POINTS_PER_ECHO to only acquire the echo tops.
Use the amplifier blanking feature of the CPMG program. This will prevent unnecessary use of power and creation of heat. Be aware that if the blanking delay is longer than the time between acquisition and the P1 pulse, the program will leave the power amplifier on for the entire duration of a scan.

Here are a few other tips:

Narrowing the SPECTRAL_WIDTH parameter may improve the apparent signal-to-noise ratio. Keep in mind that reducing the spectral width will reduce the number of points sampled per echo which may require you to reduce the NUMBER_OF_POINTS_PER_ECHO parameter.
Try running multiple scans to perform signal averaging. This should also improve the signal-to-noise ratio. If the magnetice field drifts quickly you will need to use a small REPETITION_DELAY and be sure to set the SPECTROMETER_FREQUENCY to the resonance frequency shorlty before running the experiment.