Fitting Solar Panels on our narrowboat


Why did we decide to fit solar panels? Basically it was a decision taken because we have had a spate of problems with our 3.5 Kva travel power / charger combination. It was time to have a second line backup. The first line of backup is to rewire the electrics to use the starter battery alternator to charge the domestic bank as well. That takes about 2 hours, is fraught with danger (playing with over 800 amps of power is always fraught with danger) and one day that poor little alternator is going to complain of overwork. It also costs a fortune in diesel as the engine has to run 4-5 hours a day.

How did we expect the solar panels to work? We want to cruise between 2 and 4 hours a day which will bring the batteries up and then have them “topped off” by the solar panels. There was never any intention, or expectation, that solar cells would replace other forms of charging. Just reduce the amount of engine running solely for charging.

Finding out about Solar Cells

As with everything some minor research is required to understand the concepts.

Put simply there are 3 types of solar cells produced commercially

Monocrystaline. These are supposed to be more efficient per surface area, but more significantly are the cheapest on ebay.

Polycrystaline. These are supposed to be slightly less efficient than monocrystaline per surface area but, in my opinion don't look as nice and are more expensive on ebay.

Thin film panels. These are for the rich. I ignored them.

An 80 watt panel is, however an 80 watt panel, regardless of the type. We didn't buy from ebay in the end but it's a good place to compare prices. Ours are monocrystaline cells.

Then there is the charge controller, they come in three basic types:

A simple charge controller stops your solar panels from overcharging and “boiling” the batteries and does little else.

Then there are those that do “staged charging”, adjusting the electricity sent to the batteries depending on how full the batteries are.

And then there is the MPPT controller that plays with the input voltage as well. Theoretically the most efficient of the solar controller types. Naturally the most costly but they claim 10-30% better use of the power coming out of the panels especially in low light and winter conditions.

Nobody (except the Australian sites) ever mentions the cabling required.

Dreaming up the system

So it's all dead simple, one or more panels, a cable, and a charge controller, just spend your money and it will generate free power for years. If only life were that simple. So I spent some time on the potential problems.

Finding space on the roof

What is the real day to day efficiency, nobody mentions that.

Keeping them on the roof despite the vandals, thieves and muggers.

False voltage levels confusing the control of the battery by alternators/chargers

Potential cable damage when you lob a few sacks of coal or half a tree trunk on the roof.

Voltage drops due to the length of the cable.

Losses due to the panels being flat rather that being pointed at the sun; nobody ever mentions that either.

These are the sort of questions I sought to find answers to.

Roof Space

Narrow-boat roofs are a sort of inverted minefield. There are vents, flues, antennae, ropes, rings for ropes, landing planks, boat hooks and poles on the simplest of roofs. I found two spaces big enough to fit an average 80w panel. As panels get progressively wider as they get more powerful, and width was the limiting factor, I confined my thinking to a maximum of 2x80w panels. One would be at the front of the boat and one at the back because that is where the spaces were.


Modern Solar panels are supposed to be about 14-16% efficient in converting solar energy to electricity. This assumes the the Sun is directly overhead, or the panel is oriented towards the Sun, preferably in the tropics. Basically the figure is meaningless. The question is how much power do I get out of them. I cruise in the UK (approx 53 degrees north) and plan to lay them flat on the roof. I have no intention of orienting the panels a few times a day, and taking them down and re-setting them every time I move. So how do you work out the efficiency? Now I could work out the exact angle of the Sun for every day of the year to get a theoretically perfect answer, but I'm lazy. I asked a boater friend of mine who had a 400 watt array to drop it flat on an overcast day. It gave a 7% efficiency, i.e. 7 watts from each 100 watt panel, Just over half an amp per panel per hour. I used this as a baseline for my calculations.

Keeping them on the roof

Solar panels are expensive. Having someone steal them is not part of my plan. They would have to be secured not only against the elements but against the more devious elements of society.

False voltage levels confusing the control of the battery

It is part of the natural laws that when two things (or people) think they are in control of something, confusion will result. I have an “intelligent” battery charger looking after the batteries and I am about to add another “intelligent” controller, the solar controller. Both sense the voltage at the battery terminals and adjust their output voltage accordingly. A recipe for electronic anarchy. A relay to switch out the panels when the engine/generator/charger train was activated seemed essential. I did not want my 80 amp charger being “switched off” by a 1 amp charge that was half a volt higher.

Potential cable damage

Whatever cable I used would have to be tough and weatherproof. I enquired about UV resistant cable at various solar web sites. None sold UV resistant cable, all assured me that their cable was great and would cause no problems. In fact it seems not to exist in normal use and is only really used for spacecraft. I was amazed, I would have thought that any form of solar panel set-up would have used UV resistant cable. I also wanted a lot of abrasion resistance. The best thing on offer at a reasonable price was armoured cable. That had to be UV resistant.

Voltage drops due to the length of the cable

Long cables with low DC voltage are subject to fairly serious voltage drops, the larger the cable the less voltage is “dropped”. Calculations would have to be tight, not easy when you have no real clue as to what voltage and amperage would actually be traversing the cable. I did know that it would be about 13 metres long. I spent some considerable time playing with a voltage drop calculator.


I now had a reasonably clear idea of what it was I was talking about, I knew what the components were, and what they did. It is basically an expensive, complicated trickle charger.

Would it be worth it as an investment? Time to start dragging money into the equation.

I calculated the cost of a kilowatt hour taking into account the cost of diesel, engine maintenance, battery replacement cost and battery maintenance. Each Kwh is costing me £1.29. I then worked out three scenarios, cheap and cheerful, middle of the road, and top quality to give some idea of price and return on investment.

I also tried to look at the fringe benefits/problems. It was obvious that putting 2 large areas of glass on the roof would make it more “delicate”, we would have to be careful with logs, sacks of coal and ropes. I could also see a case that my batteries might well last a bit longer due to less deep discharge and generally being fuller. As my battery bank cost in the region of £1800 it is an investment worth looking after.

I spoke to many boaters who had panels on the roof, They were all, without exception, enthusiastic about them. One however shook me, and my plan. He said that the highest output from his 225 watt panel that he had seen was 16 amps. It was time to think of a 20 amp solar controller rather than the 10 amp one I was planning.

Decision time

We were about to leave Ely, where we have both friends and relatives who take in parcels for us. Solar panels are large objects. It was the only place we could reasonably get them aboard. However it would be a terrible place to fit them. It has no real shops for bits (screws, connectors, that sort of thing).

Was I sure it was a good idea. No is the quick answer. Spending £600 for less than an amp per hour when I have an 80 amp charger was a dubious adventure. However the price of diesel isn't going down and the price of decent batteries is going up.

We decided to bite the bullet. It was time to paint the roof.

I ordered 2x100 watt panels as one company did 100 watt panels the same size as their 80 watt panels. I guess they build them and see what the output is and grade them accordingly (at least I hoped so). An MPPT controller (20 amp) and 11 metres of armoured cable. The rest would wait for Peterborough. We were going solar.

We cruised the five days to Peterborough in glorious sunshine totally unable to use it.

At Peterborough I bought the internal cabling. This was 6mm (the same as the armoured cable) but being from an automotive elecricians it has lots of little cores which means it bends easier. I also bought a relay and the necessary crimp fittings. Next was a visit to an expert nut and bolt supplier for tamper proof screws. Then a switch from the high street electronics shop.

Building a solar system

The first item on the 46 point agenda was to install the switch. In my opinion you need a switch to isolate the panels in case of any sort of problem and you need it from the beginning, not after the fire. Then the solar controller was mounted, mostly to get it out of the way so it wouldn't be trodden on or dropped. It is an expensive item.

Switch and controller

Running out of room for electric gizmos

Next was the main cable that goes along the roof. It had arrived neatly wound and I had unwound it and tied it to the handrail while we travelled up from Ely to allow it to straighten in the sun. It was partially successful.

Firstly I had to experiment. I had never dealt with armoured cable before. I had to find a method of cutting it and stripping it.

Here is how you do it without leaving blood everywhere.

cut PVC outer with any old sharp knife.

use a dremel with a cutting disc and gently cut each armour strand

strip inner core with same sharp knife

seal end with epoxy putty (to stop water getting in and rusting the armouring)

optionally paint end with epoxy resin glue as well just to be sure

Not a test piece, this is the main cable

Ready to go I prepared the cable, and roof and brought the cable through roof, note just the inner wires, the complete cable would not bend to such a radius. It took me 3 attempts to find the right place but I'm used to filling holes on the roof. You can see my failures in the following picture. I then covered the entry point with epoxy putty and painted it with epoxy resin. I didn't want water leaking in to my electrics cupboard. The grommet round the cable was to ensure that the exit of the cable was “downhill” to let gravity assist me in keeping the water out.

I had decided to use 6mm cable throughout. Firstly because any voltage drop would be minimal, secondly because standard automotive crimp fittings could be used. Anything larger and I needed special fittings, and special tools to crimp said special fittings. The selection of cable size is critical to an efficient installation. I would have used thicker wire (with all it's problems) to prevent any foreseeable voltage drop. Similarly every joint causes a voltage drop so joints should be both minimised and well made. There is no point in paying for power and not getting it.

So to prepare the panels. Our narrowboat has a slightly rounded roof, feet would be needed on one side to make them nearly level. Nothing is actually level in a narrowboat but what I wanted was nearly level but with enough slope so that water didn't puddle on top of the panel. On opening the panels to test them I found a little notice inside that said “do not turn loads on and off suddenly”. I reconsidered the relay as it would do just that.

Cut-outs were required for the feet, and holes for the cable. The back panel also needed a waterproof junction box. I could not buy one that was small enough to fit in the panel so I made one and attached it to the back panel.

Then there were the legs, made from ordinary cheap brackets, a lot of cutting drilling and tapping.

Turning a bracket into a foot

The next step was to “blind” the panels, basically cut up the cardboard packing and tape it onto the active side of the panels. This was to stop stray sparks. Solar panels are always live, they have no switches. Also a connector block was applied to each end of the main cable to stop short circuits should the positive and negative say hello to each other. Even blind they were producing 6 volts.

The cable run around the roof obstacles was determined and the cable laid in it's final position.

The back panel was put on the boat and the cable split for the junction box. The electrical connections were made and everything was sealed against moisture.

I was told that the only way to join 3 6mm cables was soldering.

At this point I completed the internal cabling. I elected not to install the relay until I found that I needed it. The system was started. It produced power.

The back panel was blinded again while the front panel was attached. The system produced twice as much power. So far so good.

Over the next couple of (scorching) days I gradually fixed the panels down with a mixture of 3 different types of tamper proof screws, some so close to the handrail that I had to spend considerable time and a cranked screwdriver fixing them. Any thief is going to have to be well armed and incredibly patient.

There are always finishing up jobs, one was making the cable entry somewhat neater so I filled a “brick burst” grommet full of sealant and screwed it over the cable entry.

Cheap Grommet, £1.50, the real thing was £16 and still had to be filled with sealant

Unfortunately the sealant chemically attacked the grommet. I ordered a new one, they cost about £1.50. This time I will fill it with ordinary silicone sealant. The cable is yet to be fixed to the roof, when I can get the right screws. Getting bits in the UK is hell, they want to sell you sealed boxes of complete solutions.


It's high summer, if solar panels are going to work well they should do so now. Winter will be different.

In sunshine the peak output has been about 12 amps, average peak 10 amps. Four hours either side of that it is about 6 amps.

In overcast weather the midday we get about 4-5 amps trailing to around 2.

What does that really mean? Well, the three main items that chew power are the overhead (the inverter, and things on standby), the fridge, and my laptop.

The overhead is approximately 1.7 amps

The fridge is 6.2 amps

The laptop is 3.2 amps.

As the fridge and laptop are not on all of the time, there are more times when the batteries are gaining than losing. Very approximately, from 8 till 8 we are about even as regards power generated as against power used (it is high summer!)

The panels are, at best, 65% efficient laying flat (even the guys with the big moveable arrays don't get 100%).

It turned out that the solar controller considered the batteries full at 13.86 volts and switched off charging at that point. No relay was fitted as my “real” charger goes above that and any conflicts would be at the top end of the charging range.

Final pictures

The front panel

and the back panel

both suitably boring and silent, the way I like things


Not doing this last year, or the year before

Buying special sealant rather than using ordinary silicone.


Any conclusions are initial conclusions, under ideal conditions but here goes.

On a summers day the panels supply most of our electrical needs between 8am and 8pm. With an engine run of 2 hours (cruising) no extra running of the engine will be needed.

The system works better than I expected. Even in total overcast it produces a few amps (which covers the overhead use) although one big dark cloud did reduce it to zero for about 10 minutes.

Fitting the system was straightforward but the details, such as weatherproofing the electrics were tedious and time consuming. I will re-visit them in two weeks to see if I did a good job before fixing the cable to the roof.

Give considerable thought to fixing the panels down. Lots of people recommended various methods including various forms of bonding (it seems it is the normal method for caravans) but you will be bonding to paint which is not the strongest of substrates.

Our battery monitor computer readout is now unreliable as it shows the solar charger output voltage rather than the actual voltage of the battery until it gets dark. We will have to re-learn the management of our electrics.

Mooring the boat will be more difficult. At the moment I try to find a place which has a good telephone signal, a good mobile broadband signal, line of sight for satellite TV. Now I must add free from shadows.

This is the link to the autumn update and results