Record a conventional 1H spectrum. Note the o1 of the selected resonance.

Create a new dataset (edc). The pulse program for selective excitation using the SPFGE experiment is selgpse.  The default parameters to run this experiment are:

p1 (hard 90º 1H pulse at high power level pl1)
The gradient strength (gpz1=15)
The gradient shape (gpnam1=sine.100)
The gradient length (p16=1ms)
The recovery delay (d16=100u).

The most important parameters to modify are the shape, the duration and the corresponding power level of the selective inversion pulse (p12). For selective excitation, a 20 ms gaussian shape truncated to 5% is a good starting point (p12=20ms and spnam1=gauss5). The shape must be created from the xShape tool and the power level (sp1) must be calibrated.

o1 is set on-resonance of a well-isolated selected resonance.

The selectivity od the SPFGE scheme largely depends of the shape and duration of the selective pulse.

For instance, set sp1 to 60dB and start data acquisition with rga and zg.  Transform the recorded data with ef (lb=1) and correct the phase for this signal. Check that pure and clean selective excitation is achieved in a single scan (ns=1 and ds=0). In addition, there is no need for phase difference calculation between hard and soft pulses. A plot region including this resonance (using the DefPlot button) must be defined. See a SPFGE spectrum on a strychnine sample.

Execute the AU paropt program in order to calibrate the selective 180º pulse. Answer the questions as follows:
 
             Enter parameter to modify:             sp1
             Enter initial parameter value:          55
             Enter parameter increment:               1
             Enter # of experiments:                    21
 
In this case, paropt automatically acquires 21 spectra while incrementing the parameter sp1 from 55dB to 75dB.

Each spectra is automatically processed with the defined lb value (usually lb=1) and for each value of sp1, only the spectral region defined above is plotted. All spectra appear side-by-side in the corresponding procno 999. The maximum value is the selective 180º proton pulse. See the experimental result.

From this calibration, we can use other selective pulse lengths and power levels as a function of the required selectivity. Each 6 dB, the pulse duration is doubled and each 20 dB, the pulse duration is multiplied by 10. On the other hand, the duration of the corresponding selective 90º pulse is the half of the calculated selective 180º pulse at the same power level. The following table can be used as a reference:
 
 

	10 ms	20 ms	40 ms	80 ms	100 ms
selective 90º	61	67	73	79	81
selective 180º	55	61	67	73	75

 
On the other hand, the selectivity (in Hz) of the SPFGE experiment using a 5% truncated gaussian shape is
 
 

	10 ms	20 ms	40 ms	80 ms
SPFGE	140	70	40	25

 

See several experimental excitation profiles of the SPFGE scheme.

This calibration can be also performed using the non-gradient version of the selective echo experiment. The only difference is the need for a minimum two-step phase cycling.

Once the calibration has been succesfully carried out, the parameter set can be stored with the wpar command in order to be used in the future (for instance, wpar spfge all).

The use of this selective echo is widely applied as a standard selective excitation block in selective gradient-based 1D experiments (see SPFGE and DPFGE experiments)