Difference in distance measurements between encoding directions

[Xen]

Thanks for your reply.

Actually we did these scans to calibrate the gradients. It's an animal scanner thus QA is pretty scarce (once a year, and not that effective!). Even after we calibrated the coils this difference remained; each of the three gradient coils seems to do the same thing, make the phantom look larger in any of the two directions in the imaging plane where the frequency encoding is set. You might think we have a small bandwidth, but we set a large one (100kHz) to avoid artifacts in frequency encoding but the problem persists.

We will have a look to the other possible causes you mentioned and see how we can get through this.

Xen