Introduction

I work regularally with large files so my work demands a high speed internet connection. Currently i’m expierencing   bad service with my 4G router. Download is often at 40 mbps and regularly  falls below 1 Mbps especially during rush hour. I'd also preffer to have a static public IP adress so i can host services for development in my office at low cost.

I’ve considerd a 5G modem with an directional antenna pointed to the cellular tower 850 meters down the road. In theory I would be able to get broadband internet but i still wouldnt have a static IP. To overcome that i could rent a VPS  with a static IP adress and route all traffic to the VPS and from there on into the world.

Considering the €500 cost for a UniFi 5G Max Outdoor modem, the montly cost for an unlimited 5G high speed subscription and the additional 5-10 euro’s a month for renting a VPS plus cellular network pressure from a highway and road in proximity during rushhour and the influence of the weather on the 5G connection health i deem this a last resort. There is a FTTO fiber optic comming into our office but its 400+ €/m for a gigabit connection.

Condering all options the advantages of a wired connection simply vastly outweigh of a mobile modem. 

ISP

I observed 3 Coax cables Coax cables ascending from the ground into the meter cabinet. So i called Ziggo for a subscription but they said the adress is not connected and there is no neightbourhood cabinet near, service was not possible. I couldn’t believe it so i did a KLIC to figure out of the cables where registered witht he Dutch Kadastral institutue “Het Kadaster”

And it showed a cable with a run of aproximatly 250 meters owned by Ziggo B.V. to a neighbourhood cabinet. I  visually confirmed there was a cabinet at location from the KLIC. This was undeniable proof there is a COAX cable present an intentionally put in the ground and registered by Zigoo. I already determined the cable was “cold” and there was no real transmission by Ziggo on the line. Now the next step would be to check if the cable was really cconnected to the neighbourhood cabinet. 

VNA TDR 

Figure 1: VNA Calibration

I wanted to  use my Vector Network Analyzer for a long time to do TDR but I didn't have a project untill now to try this feature out. I wanted to test VNA TDR first in a controlled environement. So i cut a piece is RG178 and soldered an SMA connector to it. Calibrated my VNA with 1-port SOL after a 20 min warmup. I set my S11 to TDR timedomain, with a lowpass hamming window and a velocity factor 0.69

Figure 1: VNA Calibration

This method determined the cable length was aproximatly 1.66m. I verified it using a taped measure and measured 1.69m. This method is considered pretty accurate and has just a 1.8% error enough for my goal to determine the coax cable length of 250m where a accuarcy of 5m would still be acceptable 

The tiny error could be explained by that i damaged the RG178 by soldering the SMA connector with a two high temperature, there is a wrong calibrarion plane. Futhermore i notice the coax shows 42Ω impedance in the plot which  potentially has its roots is the same fault, using another pre-fabricated coax and it plotted  51Ω.

Field Measurement

Figure 2: VNA connected

Connected the 50Ω Vector Network Analyzer to the 75Ω Ziggo Coax 3 cable.  Unfortunately the plot started 50Ω with a negative distance of -60m and went all the way up to +55m so a span of about 115m. I'm aware there is an impedance mismatch between the Ziggo cable and the 50Ω input impedance of the VNA but i guessed it wouldnt matter that much.

Under various settings (lower frequency sweep range) different signal stength settings, and more measurment points i kept getting negative length measruments. There are quite a lot connectors/adapters in between to go from the VNA to the massive C3 cable and an impedance mismatch. I considered the measurments incredibly untrustworthy and unreliable to build conclusions on. Instead of pouring more time in this method i decided Pulse TDR might yield better results.

Pulse TDR

Figure 3: Pulsing

A signal generator feeds a  square wave into the coax and the scope is “tapping” the signal to measure the step response time. My signal analyzer doesnt have the fastest rise time but for long cable  length it doesnt really matter. 

Figure 4: Step Response

In figure 4 the inital response at X1= 6 [ns] and then the reflected step response at X2=776 [ns]. This brings the step response to about 770 [ns].  The formula to caluclate the length is 0.5•Δt•VF•c = L [m] so  0.5•(770•10⁻⁹)•0.89•(3•10⁸) = 103 m. I didnt place the cursors completely accurate and the velocity factor could a bit be off.