High gain vactrol amplifier

Introduction

This is about making an own analog optocoupler (vactrol) and using it to make a single stage high gain amplifier. 

In this circuit the vactrol is used as feedback system to stabilize the amplifier by means of negative feedback.

Schematic of vactrol single stage amplifier

Bill of material

R1 470K

R2 2K2 

R3 5K trimmer

C1 330n (see text)

C2 330n

LDR GL5516

LED see comment

TR1 2N2222 / BC549C

TR2 J113

The LED and the LRD are joined to form a vactrol (see below). You can practically use almost any typical 5 mm lamp. White LEDs of this kind are fully on at some voltage around 3 V drawing 20 mA. In my trials I reached a quarter of the maximum current - but it all depends on the efficiency of the LED. I used an Avago white LED HLMP-CW11-X1000 which delivers over 7000 mcd at just above 3 V / 20 mA, 15 deg angle meaning that most light reaches the LDR. 

The thing you need to do is adjusting the value of the resistor R2 to find a good compromise between gain and thermal stability. This is further explained as follows.

Principle of operation

How does it work. The FET is used as a high impedance current source. The current depending on Vgs - the higher VGS, the lower the current. Vgs is determined by the current itself and by the resistance of the LDR. If the current tends to increase, e.g. because of the BJT thermal shift, the voltage drop across the LDR increases, reducing the voltage on the point of connection of the LED, which in turn becomes brighter. This causes the resistance of the LDR to decrease, reducing Vgs. The whole loop, if the proportion among the key electrical parameters is correct, stabilizes the collector voltage while allowing for the current to change in function of the temperature - which is not ideal, but preserves the voltage swing capability. It is a sort of servo bias circuit otherwise realized by other means - by the way, I find this article, or better said the whole site, really excellent in my view: https://sound-au.com/ism.htm. 

One may wonder why the vactrol + FET circuit does not provide negative feedback on the signal. The thing is, the LDR has slow response, so, unless you are after frequencies below 100-200 Hz, you do not see noticeable effect. In any case, if you are after signals of frequency below 1000 Hz, then increase the capacitance of C1. 

It is interesting to note that the voltage gain can be adjusted by changing the supply voltage. This is because the gain is proportional to the impedance of the collector network, which increases when the supply voltage is increased.

The selected FET J113 has Idss 2 mA or slightly higher. 

In the trials with the 2N2222, the current through the collector was about 8 mA

This may seem a bad match for the 2N2222, which features higher current at larger collector current, and was selected to test the achievable speed of the circuit, since the initial idea was to develop an amplifier for a magnetic coil picking up signals in the range of 20-200 KHz.

But in practice,even when tested at higher collector current, the gain was about the same. The reason being that the gain depends mostly on the impedance as seen from the collector.

I did some trials with a BC549C, higher beta, and in terms of collector current vs. current gain a much better match, which confirmed the above statement. 

The poor repeatability of the FET characteristics adds to other sources of variability in such a circuit - the home-made vactrol is certainly a main one. This is why some experiment with R2-R3 is required.

Trials 

I used a signal generator delivering 110 mV peak at 21 KHz, passing the signal through a 1:100 resistor divider. 

For the following two tests, R2 + R3 are set to provide in collector voltage 2.7 V with supply 24 V and 3.8 VDC with supply 14 VDC. This working point gives thermal stability enough for use at room temperature but far from optimal: by blowing the circuit...