Nuts and Volts - June 2014

When you’re ready, connect the main power supply and the battery pack to the UPS, and use a suitable USB cable to connect the UPS to your Pi — which should immediately begin its boot-up process. When your Pi has completed booting, measure the voltage between the two ends of the battery pack and make a note of it. Next, unplug the main power supply. The Pi should continue to run without interruption. Again, measure the voltage between the two ends of the battery pack; it should be significantly lower when it’s providing power.

My battery voltage dropped from 11.3V to 10.5V, but I had already been doing a considerable amount of UPS testing so your voltage levels will probably be somewhat higher. This is an important point to keep in mind because we will be measuring the battery voltage when it is not powering the Pi. Consequently, we need to be sure to replace the batteries well before they reach the “spent” level of 6.3V. (I plan to replace my batteries when they drop below 7.5V.)

Experimenting With the UPS Voltage-Sensing Circuitry

Now that our UPS is functioning correctly, we’re ready to conduct a couple of simple voltage-sensing experiments. Even though we won’t be communicating with the Pi this month (that’s still on my “To Do” list), I used our Pi interface board (from the August 2013 Primer) for these experiments because we need to power everything at 3.3V. Our Pi interface board is a convenient way of doing that. If you have a 3.3V

breadboard power supply, that would work just as well.

Figure 8 is a photo of my breadboard circuit for the following two experiments. I haven’t included a schematic because the necessary connections are very simple. The six-wire ribbon cable is inserted into the breadboard with a double-ended three-pin male header. (You can also use a 2x3 pin header, but it isn’t necessary because each pair of pins — 1 and 2; 3 and 4; and 5 and 6 — carries the same signal.)

The other end of the ribbon cable is inserted into the 2x3 male header on the UPS, with pin 1 (the red stripe on the cable) inserted into the “B” (battery voltage) line. Therefore, on the breadboard, the three connections (in order from left to right) are main power voltage (divided down), ground, and battery voltage (not divided). The ground connection is made underneath the cable, so it isn’t visible in the photo.

At this point, I need to mention an important caution before we proceed. The battery voltage at pin 1 (red stripe) of the ribbon cable is not divided, so it can be as high as

11.375V. Therefore, directly connecting pin 1 of the ribbon cable to a PICAXE (or Pi) I/O pin is likely to damage or destroy the I/O pin!

Experiment 1: Battery Pack Voltage Sensing We’re going to use a simple two-resistor voltage divider circuit, execute a readadc10 statement to measure the voltage at the junction of the two resistors, and then compute the battery voltage. The Microchip documentation for PIC micro devices (on which all PICAXE processors are based) indicates that a total resistance of about 10K will produce the most accurate ADC reading. However, the two resistors will always be in the circuit, so they will produce a continuous current-drain on the batteries. At the maximum battery voltage of 11.375V, the current drain of a 10K (total) voltage divider would be more than 1 mA. That might not seem very large, but I was concerned that over time it would significantly reduce the life span of the batteries.

In order to determine how much accuracy would be lost by using larger resistors, I decided to experiment with three different pairs of resistors: 10K and 1K; 100K and 10K; and 1M and 100K. I chose those values because a ratio of 10-to- 1 simplifies the programming calculations that need to be made. Also, in order to maintain as much accuracy as possible, I measured the actual value of a dozen or more resistors for each of the above nominal values and paired them so the measured ratios were as close as possible to 10-to-1.

The program that I used for this experiment ( Pi_PS_Test1.bas) is available at the article link. It’s fairly straightforward and well commented, so we won’t discuss it in detail. It