LM336 -> Resistor Voltage Divider -> Transistor Buffer... problem

Status
Not open for further replies.
J

juz_ad

Full Member level 2
Full Member level 2
Joined
Dec 17, 2011
Messages
130
Helped
0
Reputation
0
Reaction score
0
Trophy points
1,296
Activity points
2,541


I've recently started using the LM336 zener voltage reference. On advice I've received here and elsewhere I've buffered the LM336 output.

The LM336 with op amp buffer shown in (1) works as expected. But (2) does not.

Is there a simple way to use a transistor buffer to follow the voltage divider formed by R2 and R3? I know an op amp will work - but I'd like to try a Transistor buffer (and I'd like to understand why this simple circuit isn't working...).

Thanks.
 

Bah - serves me right for starting stuff like this at the end of the day.

I was using a 2N3906 - not a 2N3904. User error - my bad. :roll:

So - the junction of R2 and R3 is obviously 2.5V as expected - but I'm measuring approx. 2.05V at the Transistor Emitter...

I'm assuming that's approx. one diode drop? So I guess I need to compensate for that when I'm setting the values of R2 and R3?

Thanks...

:roll:

...again.
 

You have right, I forget about base to emitter voltage drop :). You will need to compensate voltage drop but also R4 is too big for your purpose.
 

The transistor circuit is basically an emitter-follower (better without R4). So, the output will be Vbe below the base voltage. That varies, and varies quite a lot with temperature. So it won't be very stable.

(I typed this while other replies were coming in because I'm doing too many things at once - but I'll leave it.)
 

You didn't tell a specification of voltage accuracy.
If you can live with the accuracy of a transistor voltage follower (-2 mV/K temperature coefficient) and poor load regulation, you don't need a voltage reference. Dividing a regulated power supply should be sufficient.
 

FvM said:
If you can live with the accuracy of a transistor voltage follower (-2 mV/K temperature coefficient) and poor load regulation, you don't need a voltage reference. Dividing a regulated power supply should be sufficient.
Click to expand...

That's an interesting answer.

I started using the LM336 as a small step up from resistor voltage dividers with an op-amp buffer on the output. If the potential variation of an LM336 and a transistor buffer is no better than the potential variation in power supply (and for my purposes I have to assume the power supply voltage might fluctuate buy +/-5%) - then potentially there's no real advantage (and a higher parts count).

I'll have to think about that.

Thanks.
 

The load connected to the reference voltage "buffer" (or whatever you're using in this place) must be considered of course. Part count- and performance-wise, a 2.5 V shunt (e.g. LM385-2.5) or series reference (e.g. LM4120-2.5) seems preferable.
 

Status
Not open for further replies.

Similar threads