相同频率范围内的线性扫描和分段扫描之间的全局相位偏移

使用N5230c VNA，我在1 MHz范围内执行线性和分段S21测量，以6 GHz为中心。
起始和终止频率相同;
但是，分段扫描有几个段，每个段具有不同的间距。
存在全局相移，使得线性扫描的相位比分段扫描低0.05度。
注意，采用分段扫描使得频率从一个段到另一个段增加。
通过采取一系列线性扫描获取“分段扫描”时，这种小偏移也是可见的，每次线性扫描频率更高，并且在软件中将它们拼接在一起（效率不高，但在我进行分段清扫工作之前使用简单代码进行快速测试
）。
相位从一个段到另一个段随机移动。
这是校准错误吗？
在同一范围内似乎不太可能。
这次全球相移的起源是什么？



     以下为原文

  Using a N5230c VNA, I perform both a linear and a segmented S21 measurement over a 1 MHz span, centered near 6 GHz. The start and stop frequencies are the same; however, the segmented sweep has several segments, each with different spacing.  There is a global phase shift such that the linear sweep is 0.05 degrees lower in phase than the segmented sweep.  Note that the segmented sweep is taken such that the frequencies increase from segment to segment.  

This small shift was also visible when acquiring a "segmented sweep" by taking a series of linear sweeps, each successively higher in frequency, and stitching them together in software (not efficient, but a quick test using simple code before I had segmented sweeping working).  The phase shifted randomly from segment to segment.

Is this a calibration error.  It seems unlikely over the same span.  What is the origin of this global phase shift?