Darlington driver output higher than expected...

A

Andrew Cowan

Senior Member
I'm using an 18X (running off 5V) to control two cascaded 74HC595 shift registers (also running of 5v).

These then feed into two ULN2803 darlington drivers, which are connected to 16 LEDs. I'm running the ULN2803 drivers off 3.1v, so I don't have to use limiting resistors on all the LEDs.

However, there seems to be a problem - the brightness of the LEDs. The positive side of the LEDs are getting the full 3.1v, which shows the problem is with the darlington's negative switching.

Looking at the voltage of the LEDs with a scope, the negative line seems to be around 1v when the LED should be on - not 0v as it should be.

I wired the circuit up quickly at school today and didn't have long to troubleshoot. I'm sure the problem is something small I've done wrong, but I can't get back to the circuit before tomorrow.

Does anyone have any suggestions in the mean time? The PICAXE/shift register stage is fine and runs direct from 5v.

Andrew

Edit - just thought I'd add that I've measured 0V directly on the ground leg of the ULN2803.
 

Attachments

Last edited:
fernando_g

fernando_g

Senior Member
11th commandment: Thou shalt not run an LED without some resistive or current-limiting element, otherwise thou shall receive the eternal wrath of the semiconductor gods.
Now seriously. An LED is a semiconductor. Its forward voltage drop varies with temperature. Also from device to device, it may be very close but NOT IDENTICAL.

What is happening is that your poor darlington has to absorb those variations by operating in its linear range -another word to indicate that it will dissipate heat- lots of it.

Lastly, the light output and sometimes the actual wavelength shifts with temperature.

Google is your best friend, unless you live in China. So please google "led temperature coefficient".
 
A

Andrew Cowan

Senior Member
The LEDs are SMD 3-chip LEDs (three LEDs in one package). Due to size constraints, I am not able to make individual connections to each LED in tha package - I have to parallel them.

Each package of LEDs draws 90mA at 3.6v. They are rated for 90mA max. At 3.1v they draw about 50mA.

That means I cannot use a constant current method, because that would cause a cascade effect if one chip blows. Instead, I am running them at a reduced voltage - and they seem happy enough so far. It is not for an industrial/consumer product, so I am happy with how it is working.

A
 
B

BeanieBots

Moderator
I think fernando_g put that very well.:cool:

But they're your LEDs to blow up, so I'll leave you to do that.
(BTW, that 50mA @ 3.1v will be VERY temperature sensitive and vary considerably from LED to LED)

The reason you were not successful in blowing them up is because you trying it with a darlington. A darlington DOES NOT pull down to 0v. (have a look at the circuit of a darlington!)
It pulls down to 2X base voltage drops (~1.2v @ full current).
 
A

Andrew Cowan

Senior Member
That's the problem!

Doh! Lucky I used an adjustable voltage regulator - it won't be much of a problem to sort it out.

Thanks for solving my problem.

A
 
B

BeanieBots

Moderator
Don't thank me too soon!
I may well have just started your problems now that you know how to drive the LEDs from a low impedance voltage source.
It really isn't a good idea.
 
You must log in or register to reply here.
Top