What is causing this step-up converter circuit to fail when introducing a transistor?

Question

I built three separate circuits with the intention of making a colored led cycling solar sun jar.

Each circuit works on its own. With just the RGB ATtiny85 circuit it works with 3x 1.2 volt AA NiMH batteries. If I reduce the AA batteries to just one, 1.218 volts on its own, and put the output of the LT1073 step-up converter circuit to the RGB ATtiny85 circuit the LED lights up as expected. At this point I see 3.218 volts at the output of the LT1073. If I take the battery + and put it to the charging circuit's 2N3906 emitter then the led no longer works. A reading of the 2N3906 collector voltage shows 1.284 volts when disconnected from the LT1073 circuit. When I connect it to the LT1073 circuit it reads 0.6 volts and the output of the LT1073 reads 0.592.

These are the circuits of the project:

1. RGB light cycling microcontroller circuit modeled very similarly to this project, but only using one common cathode RGB led.
2. LT1703 based step-up converter using a single AA 1.2v NiMh battery. The circuit schematic comes from the LT datasheet on page 11, only changing R1 and R2 to acquire the output voltage of 3.2 volts.
3. A simple solar cell charging circuit, see detecting darkness for the schematic.

I assembled the schematic for viewing this circuit by combining images of the three sources and modifying for existing conditions. Note that the LT1073 circuit is constructed with carbon film resistors where the LT schematic calls for metal film and also I'm using electrolytic capacitors.

I'm puzzled how the introduction of the 2N3906 transistor has resulted in the circuit to fail.

Answer 1

Your transistor doesn't have enough gain.

When the solar panel is dark, it looks like a resistor. That's why the transistor turns ON when the solar panel doesn't have any light on it.

Because your SMPS regulator is a boost circuit, it requires more current with its input at 1.2V than your circuit would need if it was running directly from 3 cells (nominal 3.6V).

The combination of low gain and small base current means that the transistor can't turn ON fully.

You can try a couple of things:

1) put a resistor in parallel with the solar panel. You will have to experiment with the value.

The downside of this is that now you are wasting solar energy in this resistor.

2) Use a P-channel low-threshold-voltage MOSFET instead of the transistor. Do a search for a family of MOSFET called "Trench FET" - Digikey has several that are suitable.

You want the Rds-on resistance to be a low as possible and the fully-enhanced gate voltage needs to be lower than 1.2V.

Answer 2

The transistor is permanently off since there is almost no base current possible in such a circuit. It can work only if the "NiMH battery voltage" is higher than "solar panel voltage" by more than \$ V_{BE} \$ (~0.65V). \$ I_B \$ will be then:

\$ I_B = \frac {V_{NiMH} - V_{BE} - V_{SOLAR\_PANEL}}{5k} \$

From that it is clear that one NiMH cell is practically insufficient for such a circuit to work properly.

If the "solar panel voltage" goes up (i.e. during charging the NiMH battery), the transistor will turn off.

(2015-04-27)
The solar panel internal resistance (\$ R_{i\_SOLAR\_PANEL} \$) should have also be taken into account, so the above formula will look like this:

\$ I_B = \frac {V_{NiMH} - V_{BE} - V_{SOLAR\_PANEL@I_B}}{5k+R_{i\_SOLAR\_PANEL@I_B}} \$

Both \$ V_{SOLAR\_PANEL} \$ and \$ R_{i\_SOLAR\_PANEL} \$ change according to current conditions (illumination, temperature, etc.), of course.