Excess low frequency noise in LEDs

I usually use two poles, passive, with about a factor of 10 difference in impedance. Calculating the noise bandwidth isn't difficult, and the existence of very accurate resistors and pretty accurate capacitors makes it easy to get a good result--'good' being a percent or two (~0.1 dB).

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs
Loading thread data ...

On a sunny day (Tue, 2 Feb 2016 06:12:01 -0800 (PST)) it happened George Herold wrote in :

Yes, after posting that I remembered 'e' was in that equation, but still..

So e = (Inoise * Inoise) / (2 * Idc * bandwidth )

I found this:

formatting link

Analog is cool.

Na, lots of stuff to think about now :-)

Soundcard? Lots of cool high pass filters,

Probably .. maybe electron charge will have to be adjusted ;-)

Reply to
Jan Panteltje

I have a more basic suggestion than that: if you want to rule out (or at least fail to rule in) room vibrations, then do what you can to induce them while monitoring the output: tap the thing, jump up and down, bring your cane to work and thump it on the table or floor, etc.

If you can tap it, thump it, jump up and down next to it, etc., and see it not change its output at all, that's a pretty good indication that it's not responding to vibration.

--
Tim Wescott 
Wescott Design Services 
 Click to see the full signature
Reply to
Tim Wescott

Grin well I know that's not going to happen.. at least for the case of the light bulb. I can suspend the thing in air.. but if I talk loud enough I can see that on the output.

What's hard to grasp, is how really tiny the shot noise is (at higher currents) and a little bit of variation in the light source, just ruins your whole day.

George H.

(Oh dang I meant to post this data on Jan's reply... anyway here's some old data. (The units on the y- axis could be better.)

formatting link

If you look close at the high current end (100uA) you'll see that my data is getting a bit of excess noise. (data was taken before I had learned to vibration isolate... it's even worse at higher currents.)

Reply to
George Herold

Chuckle.. the world awaits your results. You'll have to read about ENBW (equivalent noise bandwidth.)

Reply to
George Herold

On a sunny day (Tue, 2 Feb 2016 13:57:18 -0800 (PST)) it happened George Herold wrote in :

I have a theory about electrons having female and male personalities.

Reply to
Jan Panteltje

Spin up and spin down.

--
Grizzly H.
Reply to
mixed nuts

from

nment a bit.

and

ffect.

LED

quency

Hi all, Well I got a chance to look at the excess noise from an LED. Rather than retype, I'm just going to copy parts of the email I sent to the grad student.

*** Hi Michael. I had a chance to look at the shot noise from the red LED toda y. You are totally right! This looks like real 1/f noise. It certainly w as not vibration! I was careful to tighten the LED and photodiode firmly i nto the holder. And then when I banged on the low level electronics box I could only get the a small response on the meter. Small was at most a doubl ing of the V^2 noise. If I tried the same with the light bulb it would pin the meter..

(I set the output time constant to 0.01 sec. this gives a fluctuating nois e display. But lets me see fast changes.)

I've taken some data, but I've made no attempt to plot it yet. However her e is some sample data. Parameters Red LED, led current = 18 mA TIA photo diode, R = 100 k ohm I dc = 10V/100k = 100uA AC coupling everywhere, LP filter = 10 kHz High level gain = 300 V^2 out put = 7.70 (0.770 volts on meter) (this number changed with time. .. At some point the 1/f noise seemed To double and then it reduced and bec ame smaller.)

formatting link

TEk009 is the Quaf. The spike is the white noise to 10 kHz, The broad bump is the excess 1/f noise Tek010 is the FFt from the qacf Tek010 is an expanded scale showing the 1/f section.

Finally I took a picture of the output of the SRS spectrum analyzer and the 'scope spectrum so to give a comparison.

This is interesting and I'd like to write it up to put in a newsletter some time. I would like to give you credit for "discovering" the excess noise . As long as that is all right with you. I'll refer to you as Michael xxx xx a graduate student in the physics department of the University of Innsbr uck. Is all of that correct? If you'd rather I didn't use your name that is fine too... just let me know.

Finally I wanted to thank you for bringing this to our attention.

Well done! George

Reply to
George Herold

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.