Update after 3 months
The UPS is functioning as it should, and I think the solution is viable as a simple, low cost UPS where it is needed. It was a good mind and build exercise, but there are two reasons why I am not keeping it:
First – realistically speaking, electricity fails way too seldom to actually use it. Second, unless I build one of these for each room, I will always have to rely on external torches or lights. And right now I don’t think it is worth spending the time to build multiple of these. Maybe at a later time. The original article follows below:
Why?
There are multiple LED strips in my home, all of them automated and remote controlled, as part of my home automation system. LED strips are very flexible, both literally and as usage case, easy to control and provide an interesting lighting option. While there are issues with electricity at most once or twice a year, I thought about how it might be good if some of these LED lights might work for a while once the power is down. Of course there are plenty of commercial Uninterruptable Power Supplies (UPS) devices available, but even he smallest ones are oversized and too expensive to install next to a strip controller. What follows is more of an exercise on how to make a very simple UPS for lights.
About the LED strips power
One particular thing about LED strips is that they are made of repeating modules made of 3 LEDs and a resistor (for 12V, double that for 24V). Because of that, it can be possible to reduce the consumption by reducing the voltage during the time they run on battery backup. This can be seen in the plot below, where you can see the current in 3 different strips across a wide range of supply voltage. Because the strips are very different, the current is shown normalized:
As you can see, once the LED threshold is reached around 7.5 – 8V, the strip consumption will vary almost linear with voltage. Therefore with an UPS providing a lower voltage, such as around 9V, the strips will be working at around 20% capacity. Because the human eye response is non linear, reducing the light intensity by 5X while running the strip at 20% of light will result in the perceived sensation that the light level is just half. But reducing the power consumption of the strip by 5X allows us to create an even more manageable UPS.
The UPS
Which battery
There are 3 easy options for the UPS battery: Lithium, NiMH and lead-acid. With 18650 cells being a lot more expensive now and having the preponderance to ignite, I thought it’s a good idea to stay away from them. Lead acid tend to come in big and bulky sizes, so i thought about looking into NiMH. As alternative to alkaline batteries, NiMH batteries in AAA and AA size are very common and have gone down in price immensely in the last few years. Currently, Ikea sells the Ladda AA ones for 2.5 euro each, but cheaper (and probably less quality) ones like EBL and amazon basics are easy to source for less than 2 EUR for one AA.
NiMH batteries also have another cool characteristic: they can be trickle charged, which is not possible for other types of cells. In particular for this case, the batteries can be trickle charged for years, thus a charging circuit can be as simple as a current limiting resistor, provided the battery is charged slow enough. And slow enough is quite acceptable for a small UPS.
The circuit design
The schematic of the UPS is shown in the figure below
How it works
The first thing to see is that the UPS connects in parallel to the power line so it does not operate between the power adapter and LED controller. When 12V is present on the power line, the UPS is charging the battery. If the power fails, then the DC DC converter will keep the supply line at 9V. This 9V level is enough to keep the LED strip in my office at about 20% intensity, which is perfectly fine to move around and still comfortable enough to read. I used 3 NiMH AA batteries in series.
The battery charger is a simple circuit with diode D1 for reverse protection and resistor R1 for current limitation. With the chosen value, the current in the 3 cell battery pack will be about 35-40mA, which is a bit more than 0.01C. So the battery will charge in about 100 hours, or 4 days. This is fast enough for an UPS. When the batteries are charged, they reach about 1.45V each, or about 4.3V total. With such a low trickle current, they don’t get perceivably warm, though my thermometer puts them about 3°C higher than ambient.
The UPS storage can be scaled: by using a different number of batteries (up to 6) or changing to larger/smaller batteries: AAA, C, D.
The step up DC DC converter is an easy to find adjustable module based on SB6286 which has the output connected to the 12V supply line. The output is adjusted to 9V, smaller than the 12V on the power cable when the mains power is present. Because of the topology of the boost converter circuit, it is fine to operate it in this manner, as the controller will just “think” the output is too high so it shuts down the boost. When the maine power fails, there will be no 12V present on the power cable so the boost DC DC module will try to maintain 9V.
Besides the charger and the step up converter, we have the battery together with a 1A polyfuse for protection and there is a double pole switch to turn it off. Turning OFF the UPS means disconnecting the battery and disconnecting the whole UPS from the 12V line. There is also a white LED to indicate power.
Building it
The UPS is build on a prototype board and glued to the battery back. Simple. It is powered by three AA EBL 2800 mAh NiMH batteries.
How long does it last?
Here is the UPS connected to the LED controller in my office.
So far the UPS functions as expected. When running from the UPS with the LEDs at 100% from the controller, which is still about 20% of the maximum brightness, it gets about 2h of continuous battery life. With less usage of the light, it will be 10-12 hours running only the controller.
How long will the batteries last?
This is the harder question, as there are 2 stressors for the batteries. First is that they are always charged with about 0.01C current, but I cannot find data about lifetime in this situation. The second is that when used during a blackout, if the batteries discharge completely there is no protection circuit to stop them from being over discharged or reverse polarity. People generally consider this to be bad, but I am not sure how easily they can be damaged.
So i am running some tests. Whenever possible, I will put it in discharge mode once a week and leave it for about half a day, long enough to completely drain the batteries to the point where at least one of them has reversed polarity. Then the power is back on for the rest of the week to get the battery charged. Considering the low cost of such batteries,
Data so far:
25.03.2023 – 3 cycles of discharge with the UPS. Battery capacity measured with by BC-700 charger (affiliate link): 2.15, 2.19, 2.20 Ah, while the 4th battery in the set, which is not part of the UPS, is 2.23. So far no detectable aging for the UPS batteries.
05.06.2023 – 11 cycles with UPS, with the BC-700 showing remaining capacity as 2.09, 2.08, 2.16.
Ending thoughts
This UPS is rather easy to build, with very accessible components. The NiMH batteries are much easier to source and are a lot safer than lithium ones. It can also be scalable, as the number of batteries in the pack and their size can be varied. The usage in my situation is rather limited, as I don’t expect to need it that often, but this was more of an exercise project.
Can it be made simpler: actually yes. With 7 NiMH batteries in series we get to about 8.4V nominal which means there is no more DCDC converter needed. But, I wanted a more compact design with the flexibility of adjusting the output voltage.
Can it be used for other things: yes. I believe some other 12V devices I have around, like my router, could operate with a lower supply like 11V, so this can be adapted for such an operation too.