Custom led advice?!

[Porque223]

Porque

These diode voltages you are quoting....? are they for LEDs from the same batch or are they LEDs of varying colours ?.

IIRC, red, green, and blue LEDs all have different design center voltage values.
It used to be for a red indicating LED at 20 mA and a 5v supply you assumed a Vf of 2v and so it was a 150 ohm resistor and you lived with the brightness you got.
By mistake I drove an LED with a TTL output incorrectly and got some lighting, it was 1 mA and was just visible in normal office lighting.

 

[ban-all-sheds]

I suggest that you revisit the sums above and then reconsider your serious doubts.

Point taken.

 

[JohnW2]

Porque These diode voltages you are quoting....? are they for LEDs from the same batch or are they LEDs of varying colours ?.

IIRC, red, green, and blue LEDs all have different design center voltage values.

You recall correctly. But what about the answer to Bernard's question, which is very relevent to all of this spin-off of the discussion which you started - were you really suggesting that the Vf of LEDs of a particular type (not just colour/color, also manufacturer and part number) would vary by anything like as much as from 2.5V to 4.0V??

Kind Regards, John

 

[JohnW2]

By mistake I drove an LED with a TTL output incorrectly and got some lighting, it was 1 mA and was just visible in normal office lighting.

That sounds about right. For most LEDs the Luminosity/Current curve is not far off linear for currents from zero up to normal operating current - so, if it were a '20 mA' LED you would expect about 5% (in fact a bit more)of 'normal' light output at 1 mA.

Kind Regards, John

 

[Porque223]

Verily, Sires & Miladies, this is the thread that will not die!

 

[JohnW2]

Verily, Sires & Miladies, this is the thread that will not die!

You haven't lived - 3 and a bit pages is nothing Mind you, it probably would have died at least a couple of pages back if you hadn't introduced your probabilistic comments about LED circuits - so don't blame the rest of us!!

Kind Regards, John

 

[SimonH2]

Meanwhile, back to the OPs problems ...

As the others have been discussing, LEDs are best considered in terms of being current controlled devices - for a given forward current you'll get a certain amount of light out. The exact relationship varies with both current and temperature - as well as between units in a batch.

The voltage - current curve tends to show negligible current up to a knee voltage, at which point the current rises very rapidly with voltage. In an "ideal" device, it is a sharp transition and a near vertical line. For real devices, the line isn't vertical, and the transition has a curve to it. But regardless, a very small change in voltage will produce a very large change in current and hence brightness (until, as suggested) the internals of the LED expire in a tiny puff of smoke (I believe, typically the contact wire expires like a fuse).

The best LED system designs control the current and allow the voltage to be what it ends up being for the current. Even where a system is specified as a voltage (eg 12V LED lamps), internally the chips will be current driven. Current drives also avoid issues with volt drop in cables etc. Example of current driven LEDs.

As already suggested, for "budget" designs you can approximate a current control over a small range of supply voltage with a resistor. Given that you've got a forward voltage of 4V per LED chip, then you cannot put more than two in series (for a 12V vehicle) without going to active drivers.

So the first thing you need to know is the designed forward current of the LED chips, and the minimum forward voltage at that current. Put two in series, and then work out the resistance. So if In is the nominal forward current, and Vn is the min voltage at that current, then you want R=(Vs-2Vn)/In for your resistor (minimum value). Do this with Vs = 15V and see what you get.
Then look at what the next size up standard resistor value is available, and work out the worst case current with that and see how it compares to the max limit for the device.
You can then work out what current you'll get with a supply of 14V (about what you can expect in a running vehicle), and hence how much light out. In practice, you might choose to run at a lower current, and give a bit more headroom - at the expense of light output.

What would probably be done for a "pro" design is a switch mode current drive - possibly using a "boost" mode converter to step up the voltage and allow more LEDs in series (to limit the effects of variation between LED strings). There are "lots" of driver ICs to choose from, eg http://www.linear.com/products/led_driver_ics

 

[JohnW2]

Meanwhile, back to the OPs problems ...

Whilst everything you go on to say is true and perfectly reasonable, we cannot really help the OP until we get some more information (one of the reasons the thread flew off onto side issues, based on a speculative assumption that what he has are 'bare LEDs').

All we know is that the OP "has 150 4V LEDs" which he wants to use (for an application which most of us are uncertain about, although that doesn't matter too much). Unless/until we are told exactly what he means by "4V LEDs", we really aren't in a position to even try to give him any more specific advice.

Kind Regards, John