Transistors... It has been a while.

[JoeJoe]

See here:

Question is: why would you want to put a variable resistor between Z and Y, and not a fixed value one?

I could do this in my sleep years ago, honest...

[Andy Bennet]

Adjust the temperature at which the LED comes on? Presumably some kind of overheat warning device (assuming PTC sensor)?.

[Adrian Brentnall]

Well - if the component between Z and X is a thermistor (temperature-dependent resistor), the variable resistor between Y and Z lets you set the temperature at which the LED lights..... (I think).

[Another Dave]

I'd go with that. There's a thermistor below the variable resistor and an LED triggered by the transistor.

[The Natural Philosopher]

Because teh ciruit is so poorly designed obnly an adademic could have desighed it

It allows variation in gain and so on of the transistor to be adjusted for. For a single temperature anyway.

It is a circuit taht shouyld be thrown away though. Ther is no negative feedback to compensate for tolerance, whatsoever.

[Brian Gaff]

Well, cannot see your circuit, but from others comments surely if you wanted it to trigger at a temperature you would need it to suddenly come on at full brightness not just come on slowly as the bias is altered. I'd have expected some feedback in the system to make it happen like when you need to make a relay go on hard.Schmidt Trigger or an IC that does similar job.

Sorry I cannot spell today. Brian

[Dave Plowman (News)]

Easiest way is a comparator circuit using something like an LM311

[Dave Liquorice]

The biggest snag is that unless the temperature changes rapidly the will be quite a band where the transistor is operating linearly and the LED at variable brightness. Don't you need positive feed back to make it switch cleanly from one state to the other? You then need to add some hysterisis to suppress any tendancy ot oscilate just at the switching point.

Recntly been round that loop with a circuit to swich a bank of IR LEDs on when it gets dark and off when it's light. Ended up with something similar, two transistors in darlington arrangement, LDR/resistor potential divider on T1 base, 10K bias on the T1 collector T2 base connection, emmiters commoned then grounded via a

3R3 resistor, LED and current limit resistor and relay coil between T2 collector to +V. Certain amonut of "adjust on test" value selection to get the on and off points set. B-)

[Fredxxx]

The circuit relies on transistor gain to make the LED transition from off to on as fast as possible but is not 'triggered' as such.

There is no feedback, just a very simple circuit that compares the Base-Emitter threshold with a potential divider made from a thermistor and variable resistor.

[The Natural Philosopher]

schmitt trigger with two transistors is probably optimal.

absolultely defined switchover points.