HackSpace magazine

How I Made: a Portable Solar Generator

By Jo Hinchliffe. Posted

Solar panels for ‘off-grid’ power isn’t a new concept, but it’s true to say that it’s an area where technology is consistently improving and becoming more efficient and reducing in cost. I often consider while working in my shed that the only power requirement needed in there is for lighting, so it would be a suitable candidate for some 12 V lighting paired up to a solar power system.

The other area I think a solar power system would be useful is when I attend rocketry events where I’m camping for numerous days without an electrical hookup, but with a vehicle to carry my kit. Recently, numerous ‘solar generators’ have been hitting the market, often around the £500 mark – they consist of a solar panel and a smart-looking case containing some charge control equipment, a battery, and some 12 V or possibly inverter-fed 240 V outputs.

I decided to build my own semi-portable solar generator – it could live and work in the shed, but also be carried short distances to be set up in a tent or van. My use cases for it when at events are to be able to charge a phone and laptop, and also to be able to supply 12 V to my LiPo battery charger, as well as some small amounts of lighting. I would also like to be able to partially recharge my larger 48 V electric bike battery, potentially using a 12 V to 240 V inverter as a quick and simple approach whilst I research a more efficient 12 V to 48 V conversion system.

With my rough use case in hand, I started to specify components. Essentially we need a solar panel, a battery, and a charge controller, plus a few other items to make the system safer and easier to manage. Looking at solar panels first, it seems that the minimum entry point to this type of system would be a 100 W panel. A 100 W panel should, on a sunny day in a good position, be able to supply a fair amount of energy into the system. Perhaps more importantly though, it will also produce a small amount of energy on short, cloudy winter days. This is useful as the 200 mA or   so that the panel might produce on a cloudy day enables the battery to maintain its charge, and whilst the charge controller may be drawing a small amount of current from the battery, it’s still a net gain, and the battery is not discharging.

100 W monocrystalline solar panels are a reasonably common product – I went with a Renogy panel because it had good reviews and was priced at £77 (Figure 1). It’s pretty large at 1076 mm by 509 mm, but the aluminium frame that supports the unit is around 40 mm deep. For its size, it’s pretty lightweight and can easily be carried under one arm; if your arm is long enough to reach over it!

Whilst I’m still deciding where to place it (the current option is on a small, secure frame on my shed roof), it’s worth noting that the Renogy panel frame has lots of mount holes built in – there is nothing to stop you drilling and bolting to any area of the frame, avoiding the thin photovoltaic panel itself (Figure 2).

Battery choice is an interesting area to research as there are many different battery technologies to choose from. From traditional lead-acid batteries to lithium-ion, silicon gel electrolyte batteries, and more. Whilst larger and heavier lead-acid batteries are cheaper and have a long heritage of use in solar power systems, there are some drawbacks: the traditional ‘flooded’ lead-acid batteries, similar in appearance and structure to car batteries, require some ventilation and can only be used and stored upright.

Sealed lead-acid batteries mitigate this issue, but if you want to build a system where the battery may be moved and rotated, you might want to look at gel-type batteries. Lithium-ion and gel-type batteries tend to be smaller and lighter than their lead-acid counterparts, so if you are planning a large storage system with lots of battery capacity, these would perhaps be preferable.

The choice of battery was important in combination with deciding on a charge controller. At the budget end of charge controllers, you’ll find they only support lead-acid batteries, so it’s worth taking your time and looking over lots of options.

PWM or MPPT?

A few years ago, the only real budget charge controller options were cheap PWM-style controllers. These are still available and offer an amazing amount of technology, often for less than £20. One thing to note with PWM-style charge controllers is that they are less efficient at power conversion than other technologies like MPPT (maximum power point tracking). PWM (pulse-width modulation), at best, is probably converting around 70–75% of the power from the panel. However, if you are interested in setting up the cheapest system you can to get going, then, at the prices they are, they are well worth considering.

MPPT optimises the charge current supplied to the battery based on the solar panel voltage, current, and the monitored state of the battery. In a nutshell, it chooses the best charge rate conversion at any given time. When you set up a system using MPPT, you can usually see that the solar panel side is generating a higher voltage than the battery needs and a certain number of amps, but monitoring the battery, you will see it is receiving a lower voltage but with a higher current. For example, in Figure 3, the solar panel side (labelled PV for photovoltaic) is generating 18 V at 1.7 A, whereas, in Figure 4, which was taken at the same time, the charge controller is DC-DC converting the 18 V to give the battery 14.5 V at 2.1 A.

Having decided I wanted an MPPT charge controller, I looked at what was available. I’d read that my Renogy 100 W panel was unlikely to produce more than 6 A when in the UK, so I could have gone for a 10 A-rated charge controller, which is fine for this panel with a safety margin; however, I decided to spend a little more and buy a 20 A-rated MPPT charge controller so that, if I decided to add an extra panel, I have the capacity to safely add it.

The model I went for was a well-reviewed EPEVER Tracer 2210AN. Another advantage of this particular controller is that it can be reconfigured to work with either lead-acid batteries or lithium-ion, which again gives me options if I add or change the system in the future. Having made this decision, I paired this controller with a lead-acid battery designed to be used in solar power systems. The battery is designed to last three to eight years being periodically cycled (discharged and charged) and is rated at 110 A hours at C100. The C ratings give an indication of how the battery would perform under certain loads; C100 means that if put under a constant 1.1 A load, it would deliver that current for 100 hours. At C5, the battery is put under a 14 A load and delivers it for five hours – a total of 70 A hours.

As our system is primarily going to be supporting less than an amp’s worth of lighting with an occasional bit of around 5/6 A to charge RC batteries plus other experiments, it seems a fine capacity for a starter system. The battery arrived well-packed and it had some travel plugs inserted to stop any leakage of the acid in transit. It’s important that these are removed straight away – don’t use the battery with them in. However, I kept the plugs and have taped them in a small bag to the side of the battery – this means that if I’m driving with the system disconnected, I could add the travel plugs back in if I felt it necessary.  

Get connected

On arrival, the solar panel has a short set of cables around 60 cm in length marked as positive and negative and terminated in some IP67-rated connectors. These types of connectors are called MC4 – they’re robust and weatherproof (Figure 5). I wanted to create some extension cables to run between the solar panel and the rest of the generator. I could have used slightly cheaper 10 AWG wire, but again, in case I wanted to extend the system, I bought two metres of 8 AWG flexible silicon wire, which is fine for up to 40 A.

I don’t plan to ever need that capacity, but it’s good to oversize these wires for low resistance to minimise losses from the solar panel. MC4 connectors are easy to add: strip a small amount of cable and crimp a male or female connector onto the end and slide this into the case (Figure 6). Whilst they are designed to be left once connected, it’s simple to connect and disconnect them as needed, simply squeeze in a couple of clips and they slide apart. I used a cheap crimp tool with a slot marked 5.5 mm, which worked well for the crimp assembly.

The tool also arrived with a set of plastic spanners, which you can use to tighten up the MC4 casings to make them as watertight as possible. Whilst I’d recommend getting a crimp tool for the MC4 connectors, we’ve seen examples of people crimping them carefully with diagonal plier cutters.

The charge controller arrived preconfigured to work with a 12 V lead-acid battery, so I didn’t need to change any settings. It has three different charge modes – these automatically kick in depending on the amount of available power from the panel, the battery state, and the load. Wiring up the charge controller is simple with an input from the panel, a battery connection, and a ‘load’ connector. The manual asks you to connect the battery first and then attach the solar panel. On the EPEVER unit, there is plenty of capacity to connect the 8 AWG-thick wires. At the battery end I found some battery clamps that had busbars built on, which allow you to make further connections direct to the battery (Figure 7).

As a test, I used the battery clamps to connect and disconnect to power the system. Later I added a fused circuit-breaker, which adds both safety and a simple way to power down the battery side of the system. With the battery connected, the unit springs to life and then you can connect the panel. The interface for this is simple: you press the select button and it scrolls through the screens that give you information on the solar panel, the battery, and the load. To attach something you want to supply power to, it’s common for people to use the battery connectors connecting into the system with a suitably rated fuse.

Additionally, the EPEVER charge controllers also have a ‘load’ output – this isn’t useful for my intended use, but the ‘load’ connection can be set up in different modes and allows you to control the connected device, which is probably aimed at lighting. For example, you can set the load channel to only be powered for a certain amount of time after the panel has stopped charging – very useful if you want to have automated lighting that runs for a few hours after nightfall. You can also configure the load channel to be always on or always off. Of course, using the load channel also means that the load can be switched depending on the battery condition. This means, for example, if lighting is attached and is accidentally left switched on when the battery reaches the safe discharge threshold, it will power down the load channel output. The charge controller is backlit, but only for a period after a button is pressed.

As an experiment, I used a multimeter, and even when the back-light was on and I was switching between screens, I read a current consumption of only 26 mA, which is excellent. The charge unit also came with a thermistor, which you can attach directly to a small port on the unit, or you can run a wire from this port to place the temperature sensor near the battery. This again protects the battery, and if the battery reports a high temperature, it will shut down the charging system.

With everything in place and tested, I began to make a wooden case to hold everything together (Figure 8). I wanted to keep things simple and strong and also well-ventilated. The charge system and battery need to be ventilated, and the charge controller states it requires a minimum of 15 cm in each direction when mounted to a wall. I decided to mount the charge controller to the top of the case so that air can freely circulate around it. I made the lower deck long so that it has enough room to stow cables and accessories alongside the battery. The battery is quite heavy, so it’s always going to be a two-handed lift and carry, so I wasn’t too concerned about adding weight with the timber.

I made the frame from 9 mm plywood, using 20 mm square pine to make internal parts to screw into. The base, back, and sides are all screwed and glued to make it solid. The top/shelf is screwed into position but not glued, this means I can remove the top if I need to access the battery, make changes, or reroute cables. As well as the top shelf having the charge controller mounted to it, there are a few other additions. For safety and ease of powering down the system, I added a 20 A-rated circuit-breaker. These circuit-breakers are designed for high-power car audio but are often used in solar power systems (Figure 9).

You can quickly switch the breaker between the open and closed position, which is mounted on the positive cable between the battery and the controller. To disconnect the solar panel side of the system, I unclip one of the MC4 connectors, but I think it would be a good idea to add another of the breaker switches on the solar panel lines, meaning I can isolate all parts of the system. I’ve also added a small automotive fuse-box to the top shelf (Figure 10). This is used to connect loads to the battery safely with different in-line fuses. We wouldn’t use anything on this system that would require over 10 A currently, but we can swap in lower or higher value car-type fuses as needed. It’s also possible to buy single in-line automotive fuse holders rated up to 10 A – these are useful as they can be wired in-line to devices that we might use less frequently.

With everything laid out correctly in the case, we removed everything and painted the case to seal the wood, as it will spend most of its time in a damp shed! To finish off the case, I added a couple of cast iron handles to the sides, which make it much easier to lift.

This project has already started to earn its keep as it’s running the low-power 12 V lighting I had in my shed. I’ve also been using it to charge LiPo batteries for my RC and robotics projects. It gets you thinking about what you could run off a 12 V system. I’m looking at converting one of my smaller 3D printers to work off the power of the sun!

From HackSpace magazine store

Subscribe