OK, both you all. All I did was build an emitter follower amplifier and replace RE with an LED. That simple. If I drive the base with the supply, the base bias will do the current limiting.
But with no degree of control, and no idea what the current might be.
The purpose of my post was to show that (A) I personally have not seen this done before and so I wanted to demonstrate that the common transistor switch configuration (driving a Xsistor into hard saturation) is not the only method to drive an LED.
You've not seen it because it's such bad design - no problem with the LED in the emitter, but put a resistor in series with it - it then makes a working circuit.
And (B) it puts a light load on the driving circuits. I have had need of the latter in many situations such as tapping of the AC house current, power line and making a small rectifier circuit (no Xformer) to get enough DC to run some logic and ultimately driving a fairly hefty relay which could shut the AC house current off and on. In the voltage converter section I was unable to use anything but an opto-isolator which doesn't have allot of drive. So it would have been nice had I used a buffer circuit to turn on the relay Pre-driver Xsistor.
Yes, no problem at all, just add the RESISTOR - which will reduce the load even further. But I fail to see how the load is significantly less than having the LED (and resistor) in the collector? - assuming the gain of the transistor is 200, the LED current will be 0.5% more with it in the emitter, not terribly significant.