Trouble with Transistor Switch (2n3906)

Question

^{simulate this circuit – Schematic created using CircuitLab}

I'm replicating a circuit found on evilmadscientist.com: the "detecting darkness" http://www.evilmadscientist.com/2008/simple-solar-circuits/

Here is my diagram: https://i.sstatic.net/EbCcL.jpg

---

the goal is to charge a 9v batt during the day and then at night (when PV is not producing) use a PNP transistor to open the gate and allow the battery to illuminate a bank of 6 LED ( 3sets of 2 ). the batt-> LED portion does work the PV does produce a current (will illuminate an LED in dim sunlight)

When constructed, the LEDs are always on regardless (as far as I can tell) of the PVs exposure to sun.

I know this isn't so much information, but if someone could help me pin down the problem that would be very helpful. I sized the components myself so those should be considered suspect.

specs:

resistor preceding 2 LED = 100 ohm 5% resistor between PV and transistor = 1.5K ohm 5% diode after PV before batt = 100V 5A small signal Schottky diode transistor between batt - PV - LED = 2N3906 9V rechargeable batt

Answer 1

There are a number of problems here:

1. 1N4148 is not a Schottky diode, and certainly can't handle 5 A.

2. If the panel puts out 6 V max, then it's not going to charge the 9 V battery unless it is very low.

3. You are overdriving the LEDs at full battery voltage. You say the LEDs drop 3.3 V each. Figure 200 mV for the transistor when fully on, so that leaves 2.2 V across each of R4, R5, and R6, which results in 22 mA.

4. The turn on/off of the transistor will be "soft". Perhaps this is what you want, but something with a little hysteresis will cause the LEDs to snap on and off.

Answer 2

Well, the most obvious step is pretty obvious. What is the solar cell output voltage in sunlight? If it is not more than about 9 volts, you will never, ever charge your battery. So measure the PV ouput in the sun, and tell us what you get.

Here, for instance,http://www.voltaicsystems.com/2-watt-panel is a 2 watt, 6 volt solar panel. You will note that the open-circuit voltage is 7 volts. I suspect that the same applies to your panel.

To make things worse, you have misread and misapplied the circuit you linked. In that circuit, a 4.5 volt PV cell and 2 NiMH cells were used. From https://en.wikipedia.org/wiki/Nickel%E2%80%93metal_hydride_battery, you can learn that the charging voltage for NiMH cells is about 1.5 volts per cell, and

A fully charged cell supplies an average 1.25 V/cell during discharge, declining to about 1.0–1.1 V/cell

so a 4.5 volt PV unit was used to charge a (nominal) 2.5 volt battery.

On this basis, there is no reason to expect that your circuit will ever work in the sense of charging the battery from the solar cells. If you replace your 9-volt battery with 3 NiMH cells, you should get useful charging. Of course, now your LEDs won't work, so you'll have to go to single LEDs rather than 2 in series. Also, at 3.3 volts per LED, getting proper current is going to be tricky, since almost any discharge of the batteries will cause the LEDs to go dark.

If you decide to use 4 NiMH cells, you'll get better LED performance, but the PV won't provide a lot of charging current, and it will help if you use a Schottky diode between the PV and battery.