First a small disclaimer: While this post is about the Australian NBN project, at the same time it is not. I will not post about the NBN as this has been, and is covered by a few bloggers, and my voice isn't one that needs to be added to the chorus. This isn't a political spin, nor is it meant to show my political views. This is purely a fact piece about why copper is not a good medium moving forward, especially given the condition of Telstras copper asset.
One of the biggest issues with using the existing copper network is the combination of water and copper. Many people don't understand what happens though. They just know that when it rains their internet stops, or the speed drops. The reason this occurs is easy to demonstrate though. All we need is a little bit of salt, water and a power supply. The purpose of this post is not only to show what happens, but also show you how you can do this yourself. Even take it to your local member and show them why Mr. Turnbull saying that he will just "fix that" is not a straight forward idea.
Lets look at what I used and what I did first.
- Power supply
- Copper wire
- Multimeter or equivalent
The materials I used varied slightly from what most people will have on hand but others should be just as effective. I haven't tested using a battery and I don't know if that will be able to supply enough current for long enough. My power supply was a 5VDC 500mA power pack with its connector stripped and extra wire spliced to it. I used 13 strand 0.12mm (40AWG) cable, a glass bowl and some regular iodised table salt. For measuring the voltage I used a CRO. There is no need to go this high tech, I just don't own a multimeter. If you don't have a CRO or multimeter a simple voltage divider with several LED's will do enough, especially if your results are as dramatic as mine. Also optional is a load resistor. I personally didn't. It is probably advisable to add a load resistor if you are using a multimeter. Voltage dividers will not need any form of load resistor.
The salt is mixed with the water for two reasons. The first reason is to speed up the process slightly for the demonstration. Without the salt there is not a lot in the water to attach. This is a result of how we treat our water for human consumption. The second reason is that the rain water around the phone lines is not the same as the tap water we drink, salts are absorbed by the water from the surrounding dirt, and elsewhere in the environment.
Now that you have all the materials, we need to set it up. The setup is a fairly simple circuit. In my case I stripped back the power packs wire, spliced in 2 lengths (one for positive, one for negative) of the copper wire. I then hooked it into the CRO in a reversed polarity. The purpose of this is to show that the polarity of the current is irrelevant to the experiment. Telecommunications equipment works at -48VDC, so almost 10 times the voltage of my own setup. I then placed some salt in a glass bowl, and placed the wire splice in the glass bowl, I used a small amount of tape to hold it in a constant spot. Switching all equipment on I then had a baseline without the involvement of water. In my case I set the trigger value to the voltage.
After adding the water and leaving it run for a few minutes I realised that reading the result may be slightly difficult. For those who want to though, the voltage division is 2V. My CRO has the ability to show the mean voltage, for the rest of the experiment I enabled mean voltage display to make it easier to read. Overall I let the system run for 40 minutes.
Now for the results.
To be fair on the results I have decided to use the 5 minute image as the baseline for readability with the readout. This is over 40 minutes of being in salt water. As you can see in the initial image, the CRO has the voltage inverted.
Over 40 minutes in salt water the voltage dropped from -5.385VDC to -3.615VDC. This is the problem with VDSL. While the cause of the voltage drop is different, the end result is the same. Something as serious as this with the voltage is known as a foreign battery fault.
With frequencies it is a little harder to show in a layperson way, but the easiest way to look at it across the output from the CRO (Multimeter and Voltage divider will not show this) is the noise on the line. AC power normally runs at a frequency of 50hz.
In my power pack the voltage is first stepped down from 240v, then rectified from AC to DC. This process of rectification then needs to be cleaned up as it cycles between 0VDC and +xVDC where X is the peak voltage of the AC feed. From here a smoothing capcitor is usually used, this is what is producing the little bits of noise as well as other induced currents etc. This noise is a separate problem which I may look at later due to the lack of shielding.
As you can see between the baseline and the final images the little fluctuations became more pronounced and easier to identify. The height of the wave is due to the lower voltage, the peak voltage and base voltage from the rectifier have a far greater difference. Also noticable is the frequency of the noise, there are much more gaps and the frequency appears lower. This is due to the higher frequencies that squished together now not coming through to the end.
Lets have a look in the bowl.
Salt in the bottom of the bowl - No water. As water is added, bubbles form. After a little while one side starts getting a coating. Copper disolving into water and copper chloride forming End result - wire breaks. Red colour is copper, blue is copper chloride.