reverse breakdown voltage

Oh no didn't you lose credit in school for too many digits! :^)

I stuck a red LED backwards with a 100k ohm resistor is series... at about 24V I got all these 'beautiful' RC spikes. ~ 1 volt peak, rc time ~20us. Cranking up the voltage more and it was noise everywhere, on a DC background. (I was monitoring the voltage across the resistor.)

Any reason you need to know this?

George H.

I don't remember. I never gave too many digits when I was in school :). I thought about trimming it, but nah, not worth the trouble :).

RC? What C? RC spikes? 1V at the resistor and 23V at the diode doesn't look like a huge breakdown. Maybe I'll really connect a rectifier to

230V and do:

325V - -LED+ - 10MOhm (input resistance of multimeter) - 0V

Mostly curiosity. I have various LEDs here, including 3 Cree XP-G R5 cool white, which are specified as 5V reverse voltage.

Bernhard

Did you also subject the LED to various light sources concentrated on the package while doing this? Might be interesting on the results differences.

Jamie

--
Generally, it's defined as the minimum voltage required to force a
certain reverse current through the LED.

On Wed, 25 Apr 2012 17:39:42 -0400, Jamie Bernhard

--
Hardly germane.

c

on

Yeah I did, When you biased it right to the breakdown region you could definitely see the effect of sticking an light up to it. Try it. Everyone's got a red 5mm LED, 100k ohm and ~30V power supply.

George H.

,

The voltage across the 100k R looked like

Oh, I've already done this not to long ago when I was pondering on an idea.

How ever, I am trying to see where you are getting the noise from? When I did it, all I saw on the scope was a PV effect at 1 uA, it could've been more, been a while.

I used my over head light which is nothing more than a incandescent

52 watt. I did see the 120 hz in the signal, too, when getting the exposure to where I was getting 50% forward biasing voltage for that LED.

But I was able to get over 1 V as a PV with out any biasing, just the LED on the leads. It would make for a nice photo detector in a pinch though if you have the hi-z circuit for it.

I guess if one was to have a few LED's in series with HV applied and wait for a few photons to stop by for visit it would then act as a photo multiplier and push things a long, thus giving you a very sensitive detector. Or at least I think it would :)

I'll have to set that experiment up again at work, this time using the Lecroy maybe and see if I can see anything I may have missed before.

Jamie

Hmm, the DMM showed the full input voltage. No voltage drop due to the LED. IIRC before that I tried the LED in forward direction and got some voltage drop. However, after trying the reverse direction the behaviour was the same in either direction. No light, except for a short flash when I closed the circuit. The DMM showed ca. 330 mV diode drop and 445 Ohm resistance, both in each direction. I was convinced the LED was dead, but when hooked up normally it was still shining without appearant loss in brightness.

Ohh, I also had a 10uF film capacitor across the input voltage.

Bernhard

y

ic

h e

on

a.

t

the

nd

e

Hi Jamie, Here=92s some quick =91scope shots. (Still set up on my bench.)

The first is just when the PD starts to reverse conduct, with no light.

And then at same voltage with the light shinning.

The shot of a single pulse.

And then with a higher reverse voltage... conducting all the time.

(Voltage scale changed for the last.)

George H.

The currents he is seeing are in the range of possible photodiode operation.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

Cool. I wonder if you're approaching some sort of single-photon avalanche/Geiger mode. The dark behavior looks like a typical small zener diode at low current, a really noisy sawtooth oscillation.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

ng:

ally

istic

on

ly

his

k

to

ugh

n

ide

rgy

ough

hole

g.

ted on

y

idea.

cent

ED.

st the

d and

to

the

ed text -

Yeah the low current looks pretty much like a ~20V zener biased to the corner. Where I assume it's the whole diode capacitance discharging..

Here I've got 100k ohm and ~10-20 us TC so 100- 200pF. That seems like a lot for a little LED. .. ohh.. I just had some coax with 'grabber clips'... ~170pF...

A x10 probe tomorrow.

Hey for those digital type guys, you could stick a comparator on the output and count pulses. As you keep increasing the voltage you hit a quiet region, where it looks resistive.

Well I've only looked at one red LED. (I hope it's not a golden part.)

George H.

--
The currents he reported, I suspect, were measured under a single - or
not widely variable level of illumination - so it's likely that the
changes in current were due to the changes in reverse voltage across
the diode and not due to changes in illumination.

ping:

er

sually

eristic

mA on

ctly

t.

this

ink

y to

rough

ron

side

nergy

hrough

whole

ing.

rated on

ts

u

Try

y.

n idea.

m?

escent

LED.

just the

ied and

hoto

ng the

y

oted text -

HEY! I think I may be counting single photons! (At extremely low efficiency) This is amazing. (So just try it yourself.)

A variable power supply reverse biased red LED in 5mm package (Vr~24 volts). 100 kohm current limiting (and sense) resistor, look at current through the 100k ohm with a x10 scope probe. Adjust voltage till you see some RC spikes. Now cover the LED with something black. The pulses stop! I turned out the lights in my lab and can still 'see' the light that's coming in from the room next door. I need a bit of buffering so I can send the pulses into a counter.

The count rate is highly dependent on the bias level.

24 V no counts (at 200mV threshold) 24.5 V 500 counts/sec in room light 25 V 4.5k counts/ s in same light. 25.5 V 19kk/sec.

Total Friday Fun! (I've gotta put some final changes into a pcb too.)

I should hit up google 'cause it seems like someone must have done this before.

George H.

opping:

ower

Usually

cteristic

f mA on

xactly

out.

nd this

think

rgy to

through

ctron

er side

energy

through

he whole

eling.

ntrated on

ults

you

=A0Try

ply.

an idea.

rom?

ndescent

at LED.

, just the

ch

plied and

photo

e

sing the

e.

my

m

quoted text -

Just a quick addition, At the higher voltages it's not as light sensitive.

The diode is just avalanching away on it's own... with or without any light. Geo

Usually

characteristic

through

whole

on

idea.

text -

You may be in possession of a Geiger mode photon detector at about

1/1000 the regular price. But expect a high rate of dark counts.

I wonder how you could demonstrate single photon counting. Probably put the detector in a box with an LED and a lot of optical attenuation, and drive the LED on/off, and measure something.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

dropping:

Usually

characteristic

on

this

through

side

energy

through

whole

on

idea.

LED.

text -

That _is_ fun. Do you have enough ND filters lying around that you can make a 1 photon per millisecond per die area light source for it? You can figure out the threshold number of photons by looking at how the count rate goes up with light level-- a few ND1 filters in a row would allow a simple measurement of that.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

dropping:

power

s. Usually

racteristic

of mA on

wn

exactly

r out.

and this

r think

nergy to

e through

lectron

ther side

gh energy

it through

nd

the whole

nneling.

centrated on

esults

n you

=A0Try

upply.

on an idea.

from?

t

candescent

he

that LED.

ng, just the

inch

applied and

a photo

ive

using the

ore.

on my

com

e quoted text -

e

t -

So I tried comparing light levels with a photodiode. In the room lights this was hopeless. (vastly different wavelength sensitivity.)

But then I turned off the room lights and stuck a red LED onto a PD and then onto the reversed biased (RB)LED. Trying for maximum signal each time.

The numbers for the RBLED are a bit scattered. The maiximum light input was 'tweaky' wrt angle and position. Here's some data

Vf red led photodiode RBLED

1 kohm in voltage counts/sec. series 100k ohm gain

0 0.004 0-3 some light in the room.

2.05 0.25 25 3.03 1.22 72 4.06 2.14 120 5.33 2.8 160 10.0 4.4 280

Then the same but with a yellow (high brightness) LED as the light source.

Vf PD RBLED

2.03 0.303 260 2.6 1.35 1.4k 3.04 2.57 2.4k 4.14 4.4 4.3k 5.02 7.7 5.6k 6.38 11.5 7.2k

So the RBLED is a lot more sensitive to the yellow light. (as you would expect.) The count rate is at least roughly linear with the light intensity. (I think that means it can't be some weird two photon process that gives the pulses.) At the highest count rates for the yellow LED I was getting some overlapped pulses on the scope... those might not have been counted. So the apparent fall off in sensitivity for the high light intensity yellow LED is expected. I've got a piece of coax with clip leads connecting it to a little amp and the pulse width is still about 10-20 us. I need a little one or two transistor buffer to hang on it.

The photon generated count rate is a strong function of the bias voltage. All the above data was taken with Vb=3D 24.3V. This gives counts in the room light of ~450 per sec. At 24.9 V the room light count rate is 4800. (And this is still all photon generated, I can trun off the lights and cover it with black plastic and get zero counts per second.

I'm not sure if I answered your threshold question or not Phil.

Well I tried another brand of red LED. (Avago HLMP-EG30-NR000)

635nm. And this showed none of the pulsing avalanche breakdown behaviour.

The LEDs we purchased were from Newark part number 92N5331. No longer carried. :^(

So I don't know if these are 'special' in some way. I tried three of them from the drawer and all avalanched at ~24 volts.

I've got a few other different color LEDS that I'll try.

George H.