Hi All,
I'm pretty new to electronics and I've recently taken up a project, a 32 X 8 RGB LED matrix. This will be driven by 3 LED ICs (namely the HT1632C) one for each colour with 256 RGB LEDs Common Anode.
On further experimentation I found that the IC is unable to drive the LED matrix on its own due to the common anode configuration and because I'd like to drive the LEDs with 20mA. Cue the current source/current sink issues.

Researching how to source and sink the current I found that I would need to source approximately 2 Amps to drive just one column of LEDs. 32 LEDs x 3 Channels x 20mA. I plan to use P channel MOSFETs for the 8 common lines for the matrix and I'm pretty happy my choice of MOSFET will work. However I am still a bit confused about the current sinking driver. So I decided to use the ULN2803. I understand MOSFETS would provide an advantage over darlington arrays. Due to space constraints I find the ULN2803 will be perfect.

So I've searched the internet and this forum and found examples of people working out the current requirements for sinking in the same configuration I'm doing. The problem is that I've seen people it do it different ways or arrive at an answer without explaining (maybe its very easy and I just haven't grasped it?) and I'd like just a bit of clarification

My question is

What is the max current an output of the ULN2803 will receive during operation of this matrix?

a) Any 1 output of the ULN2803 will need to sink 160mA current. 20mA x a possible 8 LEDs on at a time.
b) Any 1 output of the ULN2803 will need to sink 480mA current. 20mA x 8 LEDs x 3 Channels

If its a) then I should be OK and the thermal limitations would be within limits. (I think this is right)
If its b) the ULN2803 can handle the required 480mA but its thermal limitations would be exceeded and I would need to drop the LED current to alot less than I would like (it might still not be possible and I would be forced to use a N channel MOSFET)

I've attached a sketch of my question if it makes it any easier.
Thanks in advance for any help.

 

I can't view your image (and you should reduce the resolution so it isn't a 2MB file); but you need to do more research.

When driving an LED matrix, it's common to utilize multiplexing to reduce drive circuitry and wiring requirements. When multiplexing, you need to operate the LEDs at higher currents to have the same luminous intensity as you would direct driving at 20mA.

 

Thank you for your reply. I don't know why that image isn't working.
Ill upload a lower resolution one soon.

Ah yes, depending on the duty cycle I will need to increase the peak current to get the same brightness. Thanks for pointing that out. With that in mind, could you give any guidance to which option is closer to the truth? If I know which is right I can review if the ULN2803 is an appropriate sink for my project.

 

Can't say without more information on how you're going to drive the LEDs and the multiplexing frequency. If you multiplex, you want the same average power dissipation as in the direct drive case; but you have to follow peak current rating for the LEDs.

 

I understand. So I've re uploaded a smaller resolution image. Hopefully it works this time.

 

The 2803 can handle 480 mA per transistor as long as they all are not going at once. This lends itself well to a multiplexing scheme. Note that the 2803 is a darlington array, meaning the transistors are not saturated switches. At anything above 100 mA the chip will warm up. BTW, Maxim makes integrated constant current LED drivers.

1. What is the end application for the LED array?
2. Is it a requirement of your project to be able to address each LED individually?
3. Are you going to use PWM (or some other method) to vary the brightness of the LEDs, or are they going to be strictly full on/full off?

ak

 

Hey thanks for the reply. I understand that it can handle up to 500mA per transistor but if per say it was 20mA per LED would it ever come close to 480mA?

1.The LED array will be a display capable of producing text, patterns etc. As a conservative approach I would like to be able to run each RGB LED at full white (the whole board white) without worrying that the chip will fry (not too worried if it gets close to the frying temperature)
2. Yes and I can achieve this with the HT1632C.
3. The HT1632C can perform PWM - unfortunately only for the whole chip. So I know I'm limited to 7 colours at anyone time.


I've attached a schematic of my whole circuit for reference