Best LED's for photocouplers?

[YouAre]

i'm going to be building a lot of photocoupler based effects soon, and I plan on creating my own with LDR/LED combos. Which style LED's do you guys find work best? I think the parameter i'm looking for is linearity of brightness vs. applied voltage. So which is better for linearity, diffused, or water-clear LED's? I realize that LDR's do not operate linearly themselves on brightness vs resistance, but removing the non-linear variable of LED's may lessen that affect hopefully.

Also, I'm sure the color of the LED affects the LDR's ability to react to brightness. Which color do you guys think would work best?

I would just buy a bunch and measure them myself, but at best my methods would be inexact and would probably end up with flawed data. So what do you guys think from your experience?

[Ripdivot]

I have tested most colors both in water clear and diffused and I have found that the 3mm LED's in yellow work the best for me. I have used them for years now with great results. I'm talking the ones with the yellow lenses not clear...

[PRR]

> I think the parameter i'm looking for is linearity of brightness vs. applied voltage.

?? LEDs decide what Voltage to drop. Your input should be Current.

LED brightness versus current is VERY linear down to leakage currents (microamps for modern LEDs) and up until they get hot 10mA, 20mA, or more).

You may force semi-linearity with a series resistor dropping far more voltage than the LED voltage, or with various current source schemes. Generally an opamp and an NPN with emitter resistor. Opamp forces emitter voltage proportional to signal. An LED in the collector gets 99+% of that current. (Darlingtons, FETs, and other tricks can reduce the sub-percent error, but.....)

Your LDR non-linearity is FAR worse.

Do you even want linearity? The ear likes log responses.

> which is better for linearity, diffused, or water-clear LED's?

Why would that matter? A foggy or colored lens is foggy or colored at ALL light intensities.

> I'm sure the color of the LED affects the LDR's ability to react to brightness.

Most LDRs have a broad green response. Trying both green and red LEDs, I found no real difference: the LDR responds less to red, but red LEDs make more light for the same current... this is really a non-issue. For me it came to lower voltage drop and (at least back then) the much lower cost of red LEDs. Sure don't pay more for blue or white.

Really: for your purposes the current-driven LED is "perfect", the LDR is a mystery mix of unpredictability.

Don't over-think. Build.

[YouAre]

I have tested most colors both in water clear and diffused and I have found that the 3mm LED's in yellow work the best for me. I have used them for years now with great results. I'm talking the ones with the yellow lenses not clear...

Thank you! I'll check'em out!

> I think the parameter i'm looking for is linearity of brightness vs. applied voltage.

?? LEDs decide what Voltage to drop. Your input should be Current.

LED brightness versus current is VERY linear down to leakage currents (microamps for modern LEDs) and up until they get hot 10mA, 20mA, or more).

You may force semi-linearity with a series resistor dropping far more voltage than the LED voltage, or with various current source schemes. Generally an opamp and an NPN with emitter resistor. Opamp forces emitter voltage proportional to signal. An LED in the collector gets 99+% of that current. (Darlingtons, FETs, and other tricks can reduce the sub-percent error, but.....)

Your LDR non-linearity is FAR worse.

Do you even want linearity? The ear likes log responses.

> which is better for linearity, diffused, or water-clear LED's?

Why would that matter? A foggy or colored lens is foggy or colored at ALL light intensities.

> I'm sure the color of the LED affects the LDR's ability to react to brightness.

Most LDRs have a broad green response. Trying both green and red LEDs, I found no real difference: the LDR responds less to red, but red LEDs make more light for the same current... this is really a non-issue. For me it came to lower voltage drop and (at least back then) the much lower cost of red LEDs. Sure don't pay more for blue or white.

Really: for your purposes the current-driven LED is "perfect", the LDR is a mystery mix of unpredictability.

Don't over-think. Build.

Don't over-think? No such thing for engineers! Haha. I can appreciate/understand what you're saying, but this all came from me build a circuit where water-clear LED's seemed to not react well when used as a tempo indicator for an LFO. I'll actually be using the LEDs for Electric Druid's taplfo. So I'd like to eliminate as many variables as possible. Thanks for the advice!

[JDoyle]

If you can, check out the response of your LDRs with regards to wave length. There are two main types of LDR and they respond to different wavelengths.

This assumes you can find some sort of datasheet - no mean feat.

Regards,

Jay Doyle

[PRR]

> a circuit where water-clear LED's seemed to not react well when used as a tempo indicator for an LFO.

An engineer should be able to analyze and explain that.

My engineering background can't think why the plastic-wrap around the active device matters, unless you have some severe lighting condition baffling the eye. Even then, an LED/LDR pair generally must be in a light-tight enclosure.

Maybe the water-clear LEDs were just poor LEDs? "not react well" is not much to go on.

LEDs are plenty bright for LDRs at reasonable impedances (a few K; socking one down to 300 ohms requires REAL light, sunlight). Even at ~~1mA. Whether a bright-green focused-on-LDR LED needs 0.3mA or a fuzzy red LED needs 3mA is usually a non-issue, a set-on-test adjustment, not a make/break problem.

With the poor specs and wide tolerances on LDRs, and your uncertain coupling factor, "build it" gets answers faster than evaluating LEDs and LDRs and computing utilization factors. s