Ultradim LED's (ncd @ nA) / Electromagnetic Radiation / Forums

Forums

4hv.org :: Forums :: Electromagnetic Radiation

« Previous topic | Next topic »

Ultradim LED's (ncd @ nA)

Move Thread

LAN_403

klugesmith

Fri Jun 10 2011, 06:46AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1714

Worked on 2 hobby projects in one brief session tonight. This was to console myself after leaving a dentist's office with one less bicuspid than I went in with.

1. Initial testing of my newly arrived Keithley picoammeter (from Ebay)
2. Initial experiments with LED's at the threshold of visibility.
The pA-meter is for ionization chamber experiments. Digital readout with push-button selection of decade ranges from +/- 1.999 mA to +/- 1.999 nA.
The LED work is for a new project to measure the luminance (in ncd/mm^2) of an old radium clock dial, barely readable with dark-adapted eyes. But that's another story.

With clip leads I connected in series: a 3.6-V NiCd battery pack, 100K resistor, LED (from a mixed box with unknown part numbers), and the pA-meter. Set this up in a dark room, under a red safelight.

The indicated current was about 5 uA with white LEDs, 12 uA with an amber, and 14 uA with reds.
One white LED (in traditional 5mm package) stood out as the brightest of the lot.
At 5.50 uA I could barely read a newspaper column, one word at a time (with safelight off).

With series resistor of 1 megohm, the current was 0.702 uA and I could still see the spot of light projected by LED onto paper 4 inches away.

With 100 megohms, the current was 12.50 nA and I could see the difference between on and off. Thought that was spectacular (as do amateur astronomers looking at a 6th magnitude comet!)
I think the LED's apparent luminance, viewed on-axis, was of a magnitude similar to that of the clock lume under test (which was not here). Caveat: I would not be surprised to learn that the LED's luminous efficiency at this operating point is very different from that in normal service.

To check the low range on the meter, I replaced the LED with a 5000 megohm glass-body resistor.
Indicated current (with & without battery reversal) was +/- 00.6 on the 100 nA range,
+0.65/-0.59 on the 10 nA range, and +.464 to +.484 / -.444 to -.496 on the 1 nA range.
Time to RTFM.

[edit] Would any readers like to share observations of LED's at very low currents? If not equipped to measure nA, you could use a voltage source with more headroom (e.g. 9 V) and calculate current from resistance. 1 ncd/mm^2 is the same as a millinit. Anyone care to find the formula relating visual star magnitudes to candelas?

radiotech

Mon Jun 13 2011, 06:57AM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Just a thought. Do you think that the led's might have close, but distinctly
measurable groups of spectra at the very feeble currents you are using, and,
may be regenerating some of the dim light at close-to-noise levels?

klugesmith

Mon Jun 13 2011, 07:00PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1714

radiotech wrote ...

Just a thought. Do you think that the led's might have close, but distinctly
measurable groups of spectra at the very feeble currents you are using, and,
may be regenerating some of the dim light at close-to-noise levels?

Good question.
For now it's enough of a challenge to measure the total luminance or radiance.
If that goes smoothly, with good S/N ratio,
then we might tackle the much greater challenge of measuring the spectrum.
Or (harder yet) doing that with high spectral resolution.

It might be easier to put your question to someone very knowledgeable
about the physics of LED's. It might be unexplored territory, because it has little
practical value. Like the question of how much the luminous falltime of white
LED's is greater than that of their blue cousins, because of phosphor persistence.
(once I did a crude measurement that established an upper bound of 1 microsecond).

radiotech

Mon Jun 13 2011, 08:01PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

The pAmeter, what is the input device for the electrometer?

Is it a FET or a thermionic device?

You are sort of reverse approaching a spectrophotometer sans prism.
A good photomultiplier has a rating in A/lumen , and they are wholly
capable of detecting faint scintilations. This might be
a way to get an energy figure to calibrate the pA reading differential
from dark to illuminated of your LED

The magnitude of stars in candelas cd might be harder since it is
a comparative scale.

In the LED, if you knew the area of the active photocell surface
this could be calibrated in cd/ m^2.

I will come down to some (candle) at some distance yields so many
joules...

Antares is 3470 times as bright as the sun. I think they used a wedge
photometer, but they still don't know how bright either is,

Perhaps the LED's will need to be conditioned before low level
experiments, and may have memory. Does the pA meter work with AC
and could a little light chopper wheel be placed in front the LED?

Bored Chemist

Tue Jun 14 2011, 07:09PM

Registered Member #193 Joined: Fri Feb 17 2006, 07:04AM
Location: sheffield
Posts: 1022

I had a great plan to test this until I realised that my 1980s vintage picoammeter has a display that is roughly 8 orders of magnitude brighter than the LED I'd be trying to look at. Of course, if I could look at the meter through a telescope from a few Km away...

A cheap and cheerful experiment indicates that a 1980s vintage yellow led is visible with a current of 250nA.
I got tired of waiting for my eyes to get used to the dark. I also found that you can read an LED display through a layer of black cloth and that my cellar isn't very dark on a Summer evening.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob