Figure 1. Two bipolar transistors and two resistors added to this LM358 circuit to allow the output signal to swing hundreds of millivolts above and below ground. A single transistor can enable the output of an LM358 to swing below ground.
Introduction to the phenomenon of avalanche phtoemission driven photovoltaic current in bipolar transistors
Figure 2. Getting negative current from a positive power supply.
Briefly, when the emitter-base junction or a bipolar transistor is reverse biased to a high enough voltage it will avalanche. One of the bi-products of the avalanche is the emission of photons, some of which find their way into the collector-base junction
You can do the same thing with a optocoupler with a photodiode output or a transistor output coupler that gives access to the base and emitter of the transistor and probably get higher current output plus the advantage that optocouplers are designed to be optoelectronic devices, as opposed to simple bipolar transistors are are incidentally optoelectronic devices. Advantages of using a transistor are size and cost.
Avalanche Photocurrent Sources can be qualitatively characterized as follows:
• The input terminal looks electrically like a Zener diode, usually in the range of 6 to 8 volts.
• The polarity of the output voltage is opposite that of the input current.
• The output is a high impedance current source capable of supplying hundreds of microamps.
• The maximum open circuit output voltage is approximately 350 millivolts.
• Typical ratios of output current to input current range from 1x100E-6 to 1E-3.
• The phenomenon is neither specified nor guaranteed by transistor manufacturers.
As an example of how the phenomenon can be used is that to increasing the output voltage range of amplifiers. It is possible to make a Through-The-Rails amplifier (TTR, as opposed to Rail-to-Rail or RTR amplifer -note that R-T-R amplifiers do not swing all the way to the rails if loaded at all). A TTR amplifier can provide a signal that exceeds the power supply rails. This can be especially helpful in single supply applications in which the output signal will be near ground or the power supply rail, particularly those in which the signal, be it noise or the intended signal, occasionally swings below ground.
Figure 3. Simplified circuit of the output stage of the LM324 and LM358.This style of output stage is used in several other opamps and comparators.
As is apparent in Figure 3, the output of the amplifier was designed to be "helped" down toward ground by a current source. Figure 4 shows the most negative output output voltage an LM358 at room temperature is capable of when loaded only by a 10 Meg Ohm scope probe. The minimum of 15 millivolts is probably the saturation voltage of the 50 uA current source. A resistor to ground helps the output get even closer to ground. To go all the way through ground, one needs something to help the output get below ground.
Since the base of Q5 can be pulled close to ground by the collector of Q12 the emitter of Q6 should allow the output voltage to go about a diode drop below ground. It is not a surprise to me that National Semiconductor (Now Texas Instruments) did not specify the most negative output voltage under "Absolute Maximum Ratings".
Figure 4. Output of an LM358 can swing down to 15 millivolts above ground if no external circuitry is added to pull it below the saturation voltage of the current source on the output.
Figure 5 shows a Fairchild KSP-10 connected as an avalanche photocurrent source connected to the output of an LM324 (the quad version of the LM358). You can see in Figure 6 that with the help of the KSP-10 the output of the LM324 can swing 350 millivolts below ground.
Figure 6. Output of an LM358 opamp operating from a single +10 volt power supply, swinging to 350 millivolts below ground. The arrow on the left with a "1" next to it indicates 0 volts.
Swinging above and below ground can be achieved by using two avalanche photocurrent sources coupled to one-another -one to pull the output above the positive power supply and another to pull the output below ground. This is shown in Figure 7.
Figure 7. The op-amp is an LM358. A Fairchild KSP10 (made in 2014) and a National Semiconductor NT3906 (made in the mid-1970's) connected as avalanche photocurrent sources are used as an output stage with and output swing that includes voltages higher than the positive power supply and below ground.
Q2 is configured as a constant current source while current of opposite polarity from Q1 is controlled by the output of the LM358. The sum of the currents times whatever the total impedance at the collectors sets the output volltage.
The gain of the output circuit is dominated by whatever impedance it happens to be driving, so I was not surprised to see the circuit oscillate. I slapped a large ceramic capacitor between the output of the LM358 and its inverting input, putting a pole at 0.17 Hz. That allowed me to continue with this contrivance. While at it I apologize for my poor taste in using such large value resistors on the input to the LM358, but with only tens of microamps available on the output I had to keep R3 and R4 in the feedback network large. In choosing the value for the 2 Meg Ohm R5 I boldly made the simplifying assumption that the output impedance of the output stage is infinite. Of course it is not.
The output swing capability of the circuit in Figure 7 as the 50k pot is adjusted from one extreme to the other and back again is shown in Figure 8 below.
First posted in October, 2015
You can send email to me at projects(at)cappels.org. Replace "(at)" with "@" before mailing.