Battery-operated instrument benefited from auto-power shutdown
A handheld, battery-operated “fiber finder” measures the light escaping from an optical fiber when clamped in a V-shaped block under slight pressure. A pair of photodiodes on each side of the resulting bend compares analog levels to indicate the presence and direction of light travel, and PLL tone decoders indicate the presence of as many as three optical-modulation tones. The idea is to “tag” a fiber with a central-office signal so that an operator in a pole tent or a manhole can search for and positively identify the correct fiber before cutting and splicing, thus avoiding accidental outages.
With no room on the front panel for a power switch, the design requires a sliding clamp mechanism to turn the unit on at its extreme travel position when an operator inserts a fiber. The unit must remain on each time the operator inserts another fiber and then automatically turn off when the operator finishes and no longer activates the clamp slider. This design has no room for a bulky, multipole switch; the feature works with just one pole. The design uses a PCB-mounted phosphor-bronze-plated wire against a gold post, an almost no-cost and no-real-estate switch. Without a processor or a digital clock, the function uses a spare op amp and a handful of components (Figure 1).
 |
|
| Figure 1. | Employing an RC timer, this circuit holds power on for a preset time after each momentary-switch-contact closure. |
S1’s contacts are normally open. With the power off, any residual charge on C1 drains through R5 and D1, a low-leakage switching diode, such as the MMBD2836, in a common-anode package with D2 to prevent input current through IC1 to the power rail. PNP transistor Q1, sized for the current draw to the unit, is held in cutoff; the voltage drop across R1 due to current in R4 is too low to bias Q1 on. Q2 is biased off due to zero output at the nonpowered IC1.
Closing S1 biases Q1 into conduction, providing power to the voltage regulators for the remainder of the unit’s circuitry. S1’s closure also ensures the complete discharge of C1 through R5 and D2. IC1 is now functional; its positive input is biased at 60% of the battery voltage through R6 and R7 – the voltage after approximately one RC time constant. IC1 is a single-supply CMOS device, such as an LMC6482 rail-to-rail op amp, with low-leakage inputs. You can also use a CMOS comparator with low-leakage inputs; if it uses an open-collector output, you must add the R10 pullup resistor.
With C1 remaining discharged, IC1’s output is near the upper rail and turns on Q2, which can be any low-leakage, general-purpose NPN transistor, such as the MMBT3904, or N-channel enhancement-logic-level FET. Q2 maintains base current through Q1 after S1 is open to hold the power on.
With S1 open, C1 begins to charge through R3, R4, and R5 toward the base voltage of Q1, a base-to-emitter-voltage-junction drop below the battery. Subsequent closures of S1 discharge C1 to restart the timer. When S1 remains open for longer than the RC time constant of C1R4 – approximately 10 sec with the values in the Figure 1 (the values of R3 and R5 are negligible) – the voltage at IC1’s input rises above the positive input, and IC1’s output drops nearly to ground. This action turns off Q2, which turns off Q1 and powers off the unit. As the rail voltage falls, C1 discharges through D1 and R5 to avoid clamping-diode damage to the negative input of IC1 but remains close to the power rail. The positive IC1 input is always 60% of the power rail, which ensures that IC1’s output will remain low all the way down to the power rail. Adjusting R8 and R9 to limit Q2’s base voltage below its turn-on threshold prevents any slight, dying output glitches that might exist with op amps or comparators.
