Low Current Source

[gingertube]

Yeh - I know this is a TUBE Forum BUT....

I like using cascode BJT current sources with a cheap Red LED reference on tube diff amps..

For very low current sources - that is say for a few milliamps only the LED current is very wastful.

An EE associate has suggested using a reverse connected transistor with the base left open as a very low leakage zener which is supposedly self compensating with temperature.

Can anyone point me to reference material, examples or explain what this is about?

Thanks,
Ian

[Miles Prower]

I like using cascode BJT current sources with a cheap Red LED reference on tube diff amps..

For very low current sources - that is say for a few milliamps only the LED current is very wastful.

Yes, the LED current is quite wasteful. That's why I never use one for a voltage reference. See attached for how I solve the problem.

[SY]

It looks like you're running slightly over a milliamp through the reference diodes. That's OK, it will not be a very quiet reference, but all the noise is common mode. You could also use a green LED with the same current running through it and retain the advantage of the visual indication.

[ilimzn]

How about hight efficiency LEDs? These can work at below 1mA. There are also blue LEDs of that kind with a forward voltage drop of about 3V but some materials usedin those exhibit rather odd negative resistance traits, at least that was my experience - and they are a lot more expensive than the 'regular' ones covering the spectrum between red and green.

[anatech]

Hmmm, green for better bass?

I have to admit that I generally use red or green LEDs for current sinks and sources. I may stick a capacitor across the led sometimes, 0.1uF to 1uF.

-Chris

[Miles Prower]

An EE associate has suggested using a reverse connected transistor with the base left open as a very low leakage zener which is supposedly self compensating with temperature.

Can anyone point me to reference material, examples or explain what this is about?

(See attached)

The idea here is to use the CB junction as a forward biased diode. The EB junction typically has a reverse breakdown voltage of around 6.0Vdc (it does vary with actual type, but this is a nominal value). The EB junction in breakdown will act like a Zener. The main difference is that the actual voltage it establishes may not be exact since this junction isn't really a Zener, so you will have to experiment with this.

Since a reverse current responds in the opposite manner (increase with temp) as a forward current, the forward biased junction will compensate for temperature changes, just as adding a small signal, forward biased, diode does. Since the two junctions are on the same die, the temp comp should be much better.

SY: "That's OK, it will not be a very quiet reference, but all the noise is common mode."

Of course it's OK: I built it, tested it, and it works great. Both signals are balanced, as both hit the same level on the o'scope reticule, and there isn't any more noise coming from the stage than what's already there on the screen. (Due to living about 30 miles from a 50KW, clear channel AM station, my QTH is plenty noisey as it is.