[luvdunhill]

Hi! I am in the need of an LED with a V_f of 1.58v. This led sets parameters for a current source, and I've calculated this value directly from the actual circuit. I was wondering if it's possible to get something in this range by buying 20 or so LEDs with a V_f_typical of 1.6v, and V_f_max of 1.8v, as I haven't been able to find any LEDs with anything less than 1.6v. The reason I'm a bit skeptical is that I bought 10 LEDs with a V_f_typical of 1.7v and they only varied from 1.712-1.723, so it may not even be possible to get a V_f below the published V_f_typical. Also, I could just use a diode here, but my searches on digikey for something that looks like it would work have failed, other than perhaps sticking two diodes in parallel. Anyways, I'm definitely at the limit of my knowledge here, so thanks for the help!

 

[tangent]

If I were in that situation, I'd first decide how sensitive the downstream circuit is to the exact current. If it's really so sensitive that 20 mV is going to break it, I'd be very worried. That's down in the noise with some circuits, so who knows what else will break it?

If you absolutely positively must have a CCS with that tight an accuracy, I'd choose a different design that doesn't require LED matching. The hard part of designing a CCS is choosing among the trade-offs in the many CCSes available, not in finding CCS design alternatives. There are plenty of CCS designs to choose from.

Finally, you may find that you can adjust the CCS circuit you've already chosen another way. In one LED-based CCS design I know of, it'd be incorrect to think of the LED as setting the current, but rather one of the two resistors in the design. Maybe if you show your circuit, we can find a simple solution like this for you.

 

[bhjazz]

I would agree with Tangent. You may want to adjust your design to accomodate available parts and/or downstream parts of the circuit.

If you want to peruse around, you might also check Super Bright LEDs and see what they have to offer.

 

[kipman725]

if you do need such an exact reference an LED is not a suitable device to provide it.

For good circuit design the circuit should largely be immune to parameters that change/are inexact like forward voltage drops and HFE/Hfe. Instead the circuit should relie on the values on passives which can have values set to a high degree of precision like resistors.

Consider using a different current source such as the willson current source:
http://en.wikipedia.org/wiki/Wilson_current_source
the book "the art of electronics" has info that's relatively easy to understand on current sources/sinks. (good book to get hold of for everything else aswell

)

 

[luvdunhill]

tangent, bhjazz, and kipman725:

Thank you so much for your offers to look at the circuit. I'm not sure whether I should post all the details here, but I do in fact have a post in another forum describing my dilemma and including a link to the schematic with some voltages and currents listed. I have in fact taken all of the marked voltages down for my version so I can provide those numbers if need be. A link to my message is here:

http://www.diyaudio.com/forums/showt...&pagenumber=11

Unfortunately, I haven't had much response ... basically, all I'm looking for is a "nah, that doesn't matter one lick" and I'll move on

When your source input is 0.24mV and your looking at a gain of 76dB+ I tend to think that the small voltage matters, hence my concern

Feel free to assuage said concerns though!

 

[dip16amp]

Quote:

Originally Posted by luvdunhill /img/forum/go_quote.gif Hi! I am in the need of an LED with a V_f of 1.58v. This led sets parameters for a current source, and I've calculated this value directly from the actual circuit. I was wondering if it's possible to get something in this range by buying 20 or so LEDs with a V_f_typical of 1.6v, and V_f_max of 1.8v, as I haven't been able to find any LEDs with anything less than 1.6v. The reason I'm a bit skeptical is that I bought 10 LEDs with a V_f_typical of 1.7v and they only varied from 1.712-1.723, so it may not even be possible to get a V_f below the published V_f_typical. Also, I could just use a diode here, but my searches on digikey for something that looks like it would work have failed, other than perhaps sticking two diodes in parallel. Anyways, I'm definitely at the limit of my knowledge here, so thanks for the help!

I have 20 red LEDs (TLUR4400) from mouser (part # 78-TLUR4400) and I measured them with 5.0195 vdc and a 1003 ohm resistor. Here is how they read.
1.587
1.573
1.603
1.614
1.611
1.602
1.593
1.603
1.578
1.580
1.604
1.581
1.583
1.608
1.591
1.601
1.588
1.609
1.593
1.592