Jung Super Regs & LEDs

[underwurlde]

I see the Jung Super Reg uses the LED technique for temp compensated LED-biased current source.

This is a new one for me! I thought the Vfd of an LED varied quite a bit... Learn something new every day!

Q) How does the PhotoVoltaic properties of the LED interfere with this circuit? i.e. should the LED be blacked out?

Andy

[jackinnj]

Originally posted by underwurlde
I see the Jung Super Reg uses the LED technique for temp compensated LED-biased current source.

This is a new one for me! I thought the Vfd of an LED varied quite a bit... Learn something new every day!

Q) How does the PhotoVoltaic properties of the LED interfere with this circuit? i.e. should the LED be blacked out?

Andy

The LED is emitting when the regulator is "ON" -- so I don't know if a fluorescent light source would modulate it in this condition.

There was an amusing story in EDN quite a few years ago about the peculiar behaviour of some guy's precision amplifier -- and it turns out that a fluorescent lamp was modulating a PN junction somewhere along the path.

[underwurlde]

I ask this question based on something else 'amusing' that I came across in EDN.

An LED can act as its own ambient light detector and dim itself given the right circuit. PWM it, in ithe off phase, measure the photovoltaic voltage to get ambient light and adjust the PWM to suit the brightness (how cool is that).

Now, when you say 'ON' one would assume say 220 ohms for a 5V supply. In the Jung circuit, the series resistor is something like 10K.

Try this people (if you're bored). connect 9V (+) to an LED (as normal) 1K to (-). For 1K it should be off. Now, using a DVM, measure the voltage across the LED.

For flourecents: Hundreds of mA. For direct sunlight, 3 to 4Volts!!!

I mean, not an insignifant amount!!! So when I saw an LED configured as it is in Jung based SuperRegs, my head starts scratching!

Andy

[martin clark]

For 1K it should be off. Now, using a DVM, measure the voltage across the LED.

For flourecents: Hundreds of mA. For direct sunlight, 3 to 4Volts!!!

Eh? with a 1K series resistor, most LEDs will see around 5-7mA - enough to be rather/very bright.

Interesting point though. In working-up the pfm 'flea' (url=http://www.acoustica.org.uk/t/naim/35clockreg.html) it was brought home to me just how much of a thermal tail you can get with LEDs as a reference; it far outweighed any PV effect by 3-4 orders of magnitude.

For the 'flea', where absolute voltage doesn't matter it's not an issue - once settled LEDs do make very-low noise references. Try a search here - 'Christer' did a fantastic series of measurements on LEDs for voltage-noise properties and posted extensive results a couple of years ago.

[peranders]

Originally posted by underwurlde
I see the Jung Super Reg uses the LED technique for temp compensated LED-biased current source.

This is a new one for me! I thought the Vfd of an LED varied quite a bit... Learn something new every day!

Q) How does the PhotoVoltaic properties of the LED interfere with this circuit? i.e. should the LED be blacked out?

Andy

If you are worried you should cover the LED but normally you will have the regulator in a case, pretty dark I'd prosume. But you are right though. A LED is also a photo diode.

[SY]

No, a forward biased LED is not at all a photodiode. It has to be off or even reverse biased for the photocurrents to be significant. I did some measurements of LED-biased tube stages a few year back where I put a high intensity lamp near the LED with no effect.

[peranders]

I have used LED's as limiting low leakage diodes in a DC-servo (long time ago) and the result got better with LED's in black skrink tubes. The light had some influence.

[jackinnj]

I think it will have an effect, but you probably need one of those Keithley electrometers to measure it.

Edit: we know that heat (IR) will influence LED lamp output -- this was an early problem when folks started using LED's in outdoor applications. When it was HP they had a white-paper on this on their website -- it has now passed form Agilent to Avago so it's somewhere inthe basement out there in San Jose.

[SY]

I have used LED's as limiting low leakage diodes in a DC-servo (long time ago) and the result got better with LED's in black skrink tubes.

That was exactly my point- you're not forward-biasing them. Different physics becomes dominant.

[peranders]

Originally posted by underwurlde
I see the Jung Super Reg uses the LED technique for temp compensated LED-biased current source.

This is a new one for me! I thought the Vfd of an LED varied quite a bit... Learn something new every day!

When the absolute value is of minor importance using a LED as a voltage reference it is quite good and not that uncommon.