INTRO
The sband cable loss had been measured back
on
18mar22
to get the cable loss from the lna to the postamp chassis.
This used s11 with a 50 ohm load at the lna end.
Felix remeasured the sband cable loss on
13apr23 using S21 with the two cables (polA,B) connected at the
lna with a barrel. Assuming the barrel loss was
negligible, the single cable would be 1/2 the value (in db's).
We wanted a more accurate measurement since this cables were
going to be used to bring the wideband signal down from
the filterbank output down to the pedestal (until we move the
fiber optic cables up to the filter bank output).
The measurements
The field fox network analyzer was
used to do the measurement.
The cable configuration for each pol is:
- Starting in the pedestal
- 5' cable (ao made)
- 12ft cable (mini-circuits smsm+) precision test cable.
- 25ft cable (mini-circuits smsm+)
- 25ft cable (mini-circuits smsm+)
PolA was connected to port1, polB was connected to port
2 of the network analyzer. The cables were swept 2 to 14
GHz using 4001 points.
The
plots show the results of the measurements (.ps) (
.pdf)
- Page 1:S12,S21 cable loss and acf of loss
- black in S21 (out port 1, in port2), red is S12(out
port2, in port1)
- Top: S12,S21.. loss with both cables connected.
- There is so much loss in the combined cables, that it
looks like we are hitting the network analyzer noise
floor around 7GHz.
- Bottom: acf of S21,S12
- I'm taking abs(fft(s21)) as the acf. it's been
normalized by the integral of channels 0,nchan/2-1
- we don't see any reflections because the loss is very
large.
- Page 2: S11, and S22.
- Top: reflections looking into cableA and cableB
- The ringing is a reflection partway in each cable.
- The reflection seems to get worse above 9GHz, but it
always stays below -12db.{
- Bottom: ACF of S11,S22
- the vertical scale in the fraction of reflected power
in S11 or S22
- There is a spike around 14nanoseconds.
- Page 3: cable loss , acf in a single cable.
- i first averaged s12,s21 (linear)
- the single cable loss is then the square root of the
measured value.
- Top: loss single cable (in db's)
- above 7 GHz, the measurement hit the noise floor.
- The colored lines are the cable losses from different
length mini-circuits cables:
- Blue: 25 ft cable
- green:50 foot cable
- red: my estimate for a 67 foot cable.
- i added the 25,50 foot losses (in db) and
then used .75 db offset so the line sat on the
measured data 2 to 6 GHz
- This line can be used for the cable loss up to 14
GHz
- Bottom:acf of single cable
- The horizontal axis is in meters.
- I've assumed that the velocity in the cable it .7*c
- the largest reflection is at 1.5 meters or 4.9 Feet.
This is probably a reflection where the 5ft ao cable
connects to the first 25Ft cable.
The table below lists the losses at a few freq.
Cable loss vs Freq for Single cable
freq GHz
|
loss (db)
|
3.0
|
15
|
5.0
|
20.3
|
8.0
|
27.
|
12.0
|
32.2
|
Summary
- a 2 port s-parameter measurement was done with the sband
cables (lna to pedestal) connected together at the lna side
- These cables will be used for the wideband receiver.
- the cable loss above 5GHz will increase Tsys
appreciably (more info)
- We probably want to put the fiber optics up close to the
lna.
processing: x101/230413/sbcable_s2p.pro