Apply gradient correction independent of the sequence?

[schote]

The translation of the gradient voltage waveform to current by the gradient power amplifier is not ideal.
The rising edge is more like an exponential increase which saturates at the gradient amplitude.
Since the gradient amplitude may not be reached immediately but after a short delay, the position of the projection in k-space might be slightly shifted.

To compensate for that a short time delay in the order of a few hundred us is added to the readout gradient prior to the ADC gate.
Since this adds to the gradient delay it must also taken into account for the prephaser gradient of the readout dimension.

The overall gradient momentum is not affected since the falling edge suffers from an exponential decay which compensates the delay from the rising edge. Thus the phase encoding gradients must not be corrected.

[schote]

If timing corrections are applied during sequence unrolling, there is a deviation between pulseq and the replayed waveform. It is open for discussion, whether we want to apply the correction somehow automatically or require the sequence deveopers to take care for any timing corrections.