I'm using an 8753C to measure WR1500 waveguide. The run of guide is divided into 80 lengths at 136" each. I'm sweeping from 540 to 710MHz with 801 points and using the time domain bandpass mode.

Looking in the time domain I can see the reflections from the flanged connections as expected. However, I also see unexpected "reflections" between some of the flanges, as if there was a dent in the middle of the 136" length causing a reflection on the order of the flange (-40dB) These unexpected reflections occur in groups. In other words, the first 8 lengths of guide show no reflection in the middle and the next 10 do, and then another approximately 17 without, and then about 10 more with the reflection.

If I remove the first 10 lengths of guide and remeasure, the location of these reflections moves out 10 lengths and keeps about the same pattern of recurrence.

I think I must have something wrong with the 8753C setup. I've increased the #of points with no change.

What could be the problem here?

Looking in the time domain I can see the reflections from the flanged connections as expected. However, I also see unexpected "reflections" between some of the flanges, as if there was a dent in the middle of the 136" length causing a reflection on the order of the flange (-40dB) These unexpected reflections occur in groups. In other words, the first 8 lengths of guide show no reflection in the middle and the next 10 do, and then another approximately 17 without, and then about 10 more with the reflection.

If I remove the first 10 lengths of guide and remeasure, the location of these reflections moves out 10 lengths and keeps about the same pattern of recurrence.

I think I must have something wrong with the 8753C setup. I've increased the #of points with no change.

What could be the problem here?

But, in time domain, there are other issues such as aliasing that can cause similar problems. I see you change number of points, and don't see any change, so it is less likely that aliassing is the problem.

You see this when the phase is changing conitually so that at some point the angles are changing with the real (or imag) part getting very small, and changing in abrupt 1/32000 radian increments. Where real and imaginary are about the same size, the delta phase jump is much smaller.