Battery Monitor

This mini-project came about from building a garden shed. I decided to re-purpose an old UPS battery along with some spare halogen downlighters for illumination. But as I researched lead-acid technology, I decided that I needed something rather more sophisticated than a simple switch if I wanted the batteries to survive.

Picture of completed
      unit (45k)

One of the biggest problems is deep discharge - in other words allowing the terminal voltage to drop too low. A Yuasa data sheet (for a NP10-6 battery) specifies the expected number of cycles verses the depth-of-discharge - for 100% DOD, you get 250 cycles, for 30% DOD, you get 1200. Unfortunately, the data sheet doesn't expand on the full meaning of DOD, but it essentially translates to terminal voltage.

Why bother with an electronic switch? A circuit to prevent deep-discharge would need some sort of voltage monitor, a high-current switch and a latch. The latch is required because the battery terminal voltage will rise once the load is removed - this will cause the monitor circuitry to switch the lamp back on again. Undesirable unless you particularly want a lamp-flasher, which could be slightly surreal - imagine your garden shed flashing through the night!

Inside view (50k)

Rather than relying on the slightly confusing method of switching the manual switch off and on again to reset the latch, it seemed logical to add electronic switching. It hardly increased the component-count, and eliminated a potentially lossy and/or unreliable set of contacts. The switching scheme that I chose (press and hold, rather than the more normal toggle action) was built around the slightly unusual switch that I found in my junk stores.

Closer view of the circuitry
     (50k)

Why bother with an LED bargraph? I could have used a simple comparator to detect the low voltage condition, but as I plan to add a solar cell (or some other means of charging - such as a length of bell-wire from the garage), I wanted a simple at-a-glance indication of what the battery is doing.

Circuit description

The circuit is very simple, combining very basic circuit elements. Many electronics students will have built a two-transistor bistable back at school, and will probably have also seen the LM3914 bargraph driver IC. Click to enlarge the diagram...

Schematic of the Batt-Mon (12K)

The switch is a slightly odd device that has separate NC and NO poles. Pressing the switch closes S1b, which sends a negative-going pulse via C3 into the base of TR1, turning it on. This forces off TR2, and hence the MOS-FET TR3. At the same time, S1a opens, allowing C1 to charge slowly via R2. Once the voltage across C1 rises to a volt or just over, TR2 is turned on, providing drive for TR3, and the lights...

So, a brief press of the switch turns off the light, and to turn the light on, you hold the switch for a second or two... This is very simple and hardly consumes any power; less than a milliamp in the 'off' state. Using the two transistors rather than logic ICs means the circuit operates over a wide range of voltages, and is a useful building block in its own right...

The voltage monitor is little more than an implementation of the LM3914 data sheet - the IC operates in 'dot' mode to keep power consumption low. The only 'clever' bit is the way TR4 detects low-volts - it would've been slightly easier to look straight at pin 1 (the red LED) and use that to trigger the 'off'. However, this doesn't cover the situation when battery voltage is below 10V. So, the light can only latch on when one of the top 9 LEDs are on, when TR4 is held on by the voltage drop across R15...

Conclusion

This simple little project worked really well for many years. I decided to take it with me when I moved house, and plan to install it in a playhouse that we've built for the kids. Of course, things have moved on a bit since I first built this, and the halogen lights will probably be replaced with LEDs.