One of the problems that we have is knowing when we are running low on water. As mentioned earlier on the site we can go for months without needing to top up with gas or diesel, weeks without emptying Ollie's tummy (the poo tank) but only days without topping up with water.

Gas is easy to work out, we always have one bottle in use and the other a full in readiness. Whenever we change bottles I always mark on the tap when we changed it so we have quite literally months of warning before running out.

Diesel is quite simple too. We have a highly scientific, accurate and fool proof measuring system otherwise known as a stick. Granted it is marked off with graduations of 10, 20, 30, 40, 50 and 'LOTS' but let's face it a stick is well reliable.

We keep a rough calculation of 'people days' as far as the holding tank is concerned but even with a full complement of crew we can easily run to 3 weeks of average use. And there is the problem - with our water there is no such thing as 'average use'. The washing machine uses a fair bit, as does the tumble drier and showers. More mundane tasks like cooking, washing up and Ollie use hardly any at all by comparison. When moored we wait for the 'boom' - our stainless steel tank makes a fair booming noise as it gets low and we take this as our cue to fill up. Similarly when we are filling up we keep the tap on until the tank booms then overflows - no mistaking that for anything other than full.

For some months - in fact years now I have been looking into how to fit a water monitoring system. There are commercial things that need access to the water tank - something we don't have and there is of course the ultimate in low tech, the sight glass. I put the challenge to the Canal Forum I am a member of and got various responses and some good suggestions about commercial options but these all seemed very expensive and complicated.

Eventually I came upon this link on the web and decided that it wouldn't be too hard to put together so off to Maplins it was. I decided to make two of these, then fit one after the other in the same box, effectively giving me an 8 segment display but opted not to have the alarm. Three reasons for this 1) As I am running it off of AA batteries I didn't want it running needlessly. 2) I don't need a permanent display, just to be able to check when I want. 3) Running a dc current through water for a prolonged time will lead to electrolysis somewhere.

Something around £9 later including the box to put it all in I was the proud owner of a do it yourself water level kit. Just add a few hours, some solder, some bell wire that I had laying around and a test meter and away you go.

I decided that I wanted green for anything up to half a tank, two yellows to represent 1/2 down to 1/4 and red to display anything less than that. The blue press button is a momentary contact only - the circuit is only on when you hold the button in.

After much fiddling around (I built it in mid air rather than using strip board) it was finally put in its' box and the job of working out which wire went where began.

Each of the wires coming out to the left below is then connected to the probe - more of which later, the other two are going off to a battery holder that contains 4 AA batteries. It was tested by shorting the power supply lead (the unmarked one on the right) to each of the others in turn and colour coded so that I knew which wire should go where on the probe.

I made the probe from a short length of plastic water pipe and inserted into it 9 lengths of the bell wire - 1 for the power in and the other 8 for the various 'levels'. The ends of these wires were bared back around 1/4 inch and set at intervals down the inside of the pipe where they would be out of harms way. When complete I sterilised the whole lot by immersing it in a sterilising solution as this is going to live in our tank (it fits in via the filler hole which in turn is inside a locker.

When we first tested it we got sporadic results and I set off on a mental marathon to find out why. My initial suspicion was that as I was using re-chargeable batteries I only had 5 volts rather than 6 and as this is very close to the IC's operating range I thought that with more current required for more lights I was running out of power. A trip to our local supermarket and 4 new AA batteries later but we are no better off.

OK - theory number 2. The more water we have in the further the current has to travel between the supply and the appropriate probe. So out with the probe and I trimmed insulation off of the power feed in several places down its' length - taking care to ensure that these exposed ends cannot contact the probe ends. Some re-adjustment of the probe lengths to take into account the length of the tanks' filler neck and we are in business.

Just to prove a theory I then replaced the AA alkaline batteries with re-chargeable ones and it still works. Well pleased !