NPN Phototransistor Best Practice?

[hal9001]

Is it best to connect NPN phototransistor as common emitter or common collector?



Cheers!

[Whales]

If that's a 2-pin phototransistor then both circuits will give identical (but inverted) output.

Hint: erase the base wire from those drawings. 

[Kleinstein]

Most phototransistors are just 2 pin devices and they don't care which confuguration is used. The phototransistors seem to be NPNs, but one would not notice if they would make PNP phototransitors. They could be interchangable and no need to change the circuit.

It manly depends on which way the output signal is wanted.  The usual model for the phototransistor is a current proportional to the light intensity. So the version with the resistor at the ground side has a slight advantage, as it is not that much effected by the supply voltage. The difference is not because the resistor is at the emitter, but because the resistor is towards ground.

[Whales]

So the version with the resistor at the ground side has a slight advantage, as it is not that much effected by the supply voltage. The difference is not because the resistor is at the emitter, but because the resistor is towards ground.

I think it could be the other way around.

First let's think of voltages as the difference between two points (this makes things a bit less confusing).  There is the voltage across the resistor V_R and the voltage across the phototransistor V_PT.  Let's assume we can measure both independently and choose either as the output measurement.

It doesn't matter which way around you arrange the two components (R on top or R on bottom) you will have the same current travelling through both the R and PT.  This means that V_R will be the same regardless of arrangement.  This means that V_PT will be the same regardless of arrangement.  The question comes down to whether or not V_PT or V_R is a better output in terms of rejecting noise from the main power supply, not which one is on top/bottom.

I vote V_PT being better, because if it acts anything like a diode then it will have a reasonably constantly voltage drop even if the main voltage rail changes a bit.

(If this is true then putting the PT at the bottom might make more sense, as measuring voltages in reference to ground is the norm and might (?) make opamp choice easier)

[Whales]

Whilst I'm here it's worth mentioning phototransistors vs photodiodes.

Last I looked: bog standard phototransistors are rather slow (KHz analog bandwidth), if you want speed then you can get some much faster photodiodes (but you then have to build amplifiers for them).  Faster photodiodes seem to have less output current (presumably because smaller surface area = less capacitance but also less current?) which probably means having to deal with more noise.

I'm not sure if photodiodes can be used for more linear sensing than phototransistors.  I've seen diagrams that suggest phototransistors are a bit like a photodiode being connected directly between the B and E (or C and B?) legs of an NPN transistor, suggesting they would have very nonlinear and temperature dependent outputs. 

[Kleinstein]

If it is about a small signal, the voltage over the resistor is definitely the better output. If not saturated, the differential resistance of the phototransistor is quite high. So a change of the supply voltage would mainly effect the voltage at the phototransistor and only to a small part (kind of Early effect) also the current and thus the voltage at the resistor.
With a large signal so that the phototransistor saturates and essentially turns full on, the high signal would be effected by VCC in both cases, but one would not really care about small variations anyway.

Phototransitors are a bit slow, but mainly when used all the way to the saturation or with a large resistors. With low impedance they can still work to the low MHz range.
A photodiode is also only fast with a low impedance - in the wrong circuit it can also be slow.
The phototransisistor is like a relatively slow transistor with a photodiode between collector and base. It gets a bit nonlinear as the gain depends on the current. In addion the gain is temperature dependent (going up with temperature).
A photodiode is usually linear over a large range (e.g. pA to mA) and the temperature dependence is relatively small, except for the wavelengths near the IR limit.

[Psi]

+1 for being sure you want a phototransistor vs photodiode.
For any application one of them is usually a good idea and one is a bad idea

[hal9001]

Most phototransistors are just 2 pin devices and they don't care which confuguration is used. The phototransistors seem to be NPNs, but one would not notice if they would make PNP phototransitors. They could be interchangable and no need to change the circuit.

It manly depends on which way the output signal is wanted.  The usual model for the phototransistor is a current proportional to the light intensity. So the version with the resistor at the ground side has a slight advantage, as it is not that much effected by the supply voltage. The difference is not because the resistor is at the emitter, but because the resistor is towards ground.

So the version with the resistor at the ground side has a slight advantage, as it is not that much effected by the supply voltage. The difference is not because the resistor is at the emitter, but because the resistor is towards ground.

I think it could be the other way around.

First let's think of voltages as the difference between two points (this makes things a bit less confusing).  There is the voltage across the resistor V_R and the voltage across the phototransistor V_PT.  Let's assume we can measure both independently and choose either as the output measurement.

It doesn't matter which way around you arrange the two components (R on top or R on bottom) you will have the same current travelling through both the R and PT.  This means that V_R will be the same regardless of arrangement.  This means that V_PT will be the same regardless of arrangement.  The question comes down to whether or not V_PT or V_R is a better output in terms of rejecting noise from the main power supply, not which one is on top/bottom.

I vote V_PT being better, because if it acts anything like a diode then it will have a reasonably constantly voltage drop even if the main voltage rail changes a bit.

(If this is true then putting the PT at the bottom might make more sense, as measuring voltages in reference to ground is the norm and might (?) make opamp choice easier)

Thanks for explaining!