Green LED I-V Curve

C of an MCU. However, I can't find a good chart of the I-V curve below 1 m A. The circuit supplies up to 1.15 mA which is measured by dropping it acr oss a 1.5 k resistor yielding up to around 1.7 volts. This seems like it m ight be getting into a range where the LED might be stealing a significant portion of the current being measured. But none of the references I can fi nd provide enough detail to know for sure. Most seem to be something along the lines of an "artist's concept" of a curve lying along the X axis until reaching some voltage where it takes off like a hockey stick. Others just seem to be a wag with only a very few points measured. None of them show any detail at the low end with 10 mA as the lowest line on the Y axis.

will see, but I can't lower it too much or lose resolution of the measureme nt.

was trying to keep a part off the BOM if possible. There's already a green LED in use.

At low forward current the

ought to work fine. The reverse current - Is - has to be very low, and if y ou tried to measure it, what you would probably see is the LED acting as a photodiode to ambient light.

A blob of black paint might be called for if you were to use a green LED as a voltage limiter.

--
Bill Sloman, Sydney

Make sure that no external light fall on the LED.

I did a manual measurement of various no-name LEDs 10 years ago. It included only one green 5mm LED. The current level probably doesn't go as low as you want but FWIW this is what I got:

mA V

0.44 -- 2.608 0.5 --- 2.627 1 ----- 2.711 2 ----- 2.801 3 ----- 2.864 4 ----- 2.902 5 ----- 2.939 7 ----- 3.003 10 ---- 3.077 15.5 -- 3.141

ADC of an MCU. However, I can't find a good chart of the I-V curve below 1 mA. The circuit supplies up to 1.15 mA which is measured by dropping it a cross a 1.5 k resistor yielding up to around 1.7 volts. This seems like it might be getting into a range where the LED might be stealing a significan t portion of the current being measured. But none of the references I can find provide enough detail to know for sure. Most seem to be something alo ng the lines of an "artist's concept" of a curve lying along the X axis unt il reaching some voltage where it takes off like a hockey stick. Others ju st seem to be a wag with only a very few points measured. None of them sho w any detail at the low end with 10 mA as the lowest line on the Y axis.

D will see, but I can't lower it too much or lose resolution of the measure ment.

I was trying to keep a part off the BOM if possible. There's already a gre en LED in use.

Thanks for the data. That was just what I was looking for.

From that data it looks like it will have very minimal impact on the curren t over the voltage range of interest. But if I am using this to save addin g a Zener diode but I have to have the part painted... doesn't seem worth the trouble. The case only has small air openings, but it has a large port on top covered by plexi so the workings can be monitored. So it will have indirect light inside the case falling on the LED. The case is stainless steel, so lots of reflection.

I wonder just how much the light will affect the LED. Low uA won't be a pr oblem. 100's of uA will. Sounds like I should use the Zener.

--

  Rick C. 

  + Get 1,000 miles of free Supercharging 
  + Tesla referral code - https://ts.la/richard11209

When I need to light-proof some part, I just use adhesive aluminium foil. Easy. Paint and electrical tape are not completely opaque.

I first used Al foil a long time ago when I needed to use several IR receiver modules outdoors. The modules needed to receive only radiation coming at a narrow angle from the front but the boxes housing them were not light-proof.

There's a black crepe paper tape that used to be the standby for light blockage. The crepe has the advantage of allowing it to stretch slightly to make a good seal against uneven surfaces.

C of an MCU.

In the distant past, I have NOT had much success with using diodes to clamp ADC inputs. (Unless you can afford to jettison 4 or 5 bits, or more, of r esolution?) The last time I tried it, I found that the diodes (Zener's) re ally affected the voltage much too early in the curve and I couldn't calibr ate it out with reasonable effort. I ended up just taking them out, for th e much improved response.

But maybe that's not what you're doing here. (?) And I don't think it is, so just mentioning the above.

As for green LED currents, we recently started using one that is amazingly bright at only 330 microamps. I'll send you the part# when I get to the of fice.

I have some green LEDs that are visible in average room light at 1 uA, and visible to a dark-adapted eye, up close, at 1 nA. Had to get out of bed at 2AM in Truckee to measure that.

A fun experiment would be to use a PMT and some signal averaging to see if any photons are emitted at, say, 1 pA.

One of my most common causes of writing ECOs against new designs is that the LEDs are too bright. Gotta do that again today.

There are many better ways to clamp an ADC input than using one diode to ground.

I wonder why Ricky doesn't measure an LED himself.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard

It is even more true for blue leds than for green. When a device has a blue indicator LED, it usually is far too bright. Both my PC and my router have indicator lights that, when not taped over, cause a large blue spot to appear on the opposite wall.

My BeagleBone Black has as least 5 of them. One of them is a heartbeat indicator. I'm unable to operate this thing laying on the table. I always put something on it.

Cheers, Gerhard

Blue seems to annoy people. As LEDs got better, we got complaints from users, and had to reduce currents. The first blue LEDs, the SiC Crees, were about right for a panel indicator at 50 mA. Now 1 mA is a bit bright.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard

The best use of this I've seen is Bosch dishwashers, the models that have no visible controls or indicators when the door is closed:

They use a blue LED to project a spot on the floor - if the spot is steady, the diswasher is happily progressing through its cycle. If blinking, go figure out why.

Joe Gwinn

must be an early version,

Ours is red! It's a really good machine. As is the Dyson vacuum cleaner.

It's a pleasure to, once in a while, see some really good engineering.

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard

I think you will find a zener to be worse than a LED

Low voltage zeners are very soft. Above about 7 volts, they leak just nanoamps right below the zener voltage.

(Yes, we call all those things "zener diodes")

--

John Larkin         Highland Technology, Inc 

Science teaches us to doubt. 

  Claude Bernard

I used a blue LED projection as an indicator in a custom-built product I made last year. The central unit is housed in a white plastic box which is completely closed except for an antenna and a USB port. Going by a last minute inspiration, I projected the blue LED at the inside of the box from about 2cm away instead of having it poke through the wall. The result is a diffused circle of blue light. The customer loved it. The LED current is 1.5mA.

ADC of an MCU.

mp ADC inputs. (Unless you can afford to jettison 4 or 5 bits, or more, of resolution?) The last time I tried it, I found that the diodes (Zener's) really affected the voltage much too early in the curve and I couldn't cali brate it out with reasonable effort. I ended up just taking them out, for the much improved response.

y bright at only 330 microamps. I'll send you the part# when I get to the office.

That is exactly what I am doing, limiting the input to the ADC for protecti on reasons. We don't have any small Schottky diodes in the design nor a su itable Zener. Thought I'd try this to save adding to the BoM which is grow ing.

I do not want to add any unusual steps to manufacturing, so if the light cu rrent is at all significant, the LED is out!

This is to measure motor current via a tap in the controller with a very lo ose spec on the ratio. At lower currents the ratio varies from 4670 to 101

10 (3 amps) and at higher current it is still 6060 to 8330 (8 amps) or ? ?15%. So I can lose some precision from 10 bits and never know it, but i t's only 1 mA at 8 amps, so maybe 100 uA could be tolerated at a max. If c loser to 10 uA it could be ignored. I don't have a part number for the green LED in question. We also have the motor selection in the air since the earlier chosen unit didn't have enoug h umph, but still managed to strip gears! Meanwhile we seem to be pushing to define the schematic... hurry up and do the wrong thing.

Partly at my urging they are finally starting to take the requirements a bi t more seriously, but I think they still don't get that requirements are su pposed to be done first. lol

Long road ahead.

--

  Rick C. 

  -- Get 1,000 miles of free Supercharging 
  -- Tesla referral code - https://ts.la/richard11209

Worse in what regard exactly??? My only real complaint with the LED is the light sensitivity. At the voltages I need to measure (up to 1.7 volts) th e dark current would seem to be very minimal.

If the current at low voltages is truly logarithmic, does that mean there i s a current at zero or even negative voltages? So a diode will pump out ve ry low levels of power? Perhaps the level of power is below the uncertaint y principle level of detection like the quantum foam?

--

  Rick C. 

  -+ Get 1,000 miles of free Supercharging 
  -+ Tesla referral code - https://ts.la/richard11209