Can a ULN2803APG drive a 12V led strip with 440ma load?


Sorry if this has been asked somewhere else, I couldn't find it.

I plan to drive 8 pieces of a led strip, each consuming 55ma at 12V (measured with a multimeter), they will work on a 100% duty cycle most of the time. I just want to use the ULN2803APG to be able to control it from an Arduino Mega, allowing the LEDs to blink a few times for notification purposes.

According to the datasheet, the current on each channel is below the limit for a 100% duty cycle with 8 channels, but I am concerned about the thermal dissipation (I am just a curious computer scientist that doesn't know much about electronics). Would my load fry this IC, or would it be OK?

Thanks in advance for your collaboration,


There is a chart in the datasheet showing output current vs. duty cycle with all outputs active. I think that it is showing that you can get 440 mA out but only at about 5% duty cycle. At 100% duty cycle the output current drops to about 60 mA.

Perhaps some logic level power MOSFETs would handle the current better.

It is 55ma per channel, 440ma total.

My concern is the thermal dissipation, 12V 440ma.


This data sheet says:

? JA Package thermal impedance, D package: 73.14 °C/W

(3) Maximum power dissipation is a function of T J (max), ? JA , and T A . The maximum allowable power dissipation at any allowable ambient temperature is P D = (T J (max) – T A )/ ? JA . Operating at the absolute maximum T J of 150 ° C can affect reliability

So if T A is 25C, and T J is 150C, max dissipation is 1.7W.

Connected to your strips, the chip will be sinking 440mA, but not at 12V. Most of that 12V will be dropped by the led strips. Probably only a volt or a little less will be dropped by the chip, so less than 440mW.


So I am probably safe with this setup?

Yes I think so. Anyone care to second that?

I think you'll smoke it with 2 outputs at 440mA.
I think the package is only good for 500mA on one IO at a time.
All the characteristics only show data up to 350mA.
Look at VCE-SAT:
With 350mA, Vce will be up to 1.6V.
Power dissipated = IV, so 560mW. x 8 = >4W, and the part has no additional pins for cooling or an exposed pad to mount a heat sink against. Poof.

Use a N-channel MOSFET rated for higher currents, such as AOI518:

I designed & offer a board with 32 of them, can just build up with 8

Ah - just reread the top - 55mA/channel. At 100mA, Vce = 1.1V max.
P - 0.1 * 1.1 = 110mW x 8 = 880mW. Will be warm, but should survive.

It is 55ma per channel, 440ma total.

My conclusion was based on paying attention to what the op is saying...

Bob, could you explain the 880mW figure please? Even with Vce of 1.1 I make it less than 500mW.

Like I said:
100mA * 1.1V Vce = 110mA, x 8 = 880mW
which is worse case for the lowest current shown in the datasheet.
with 55mA, would be less.
Looks like I tend to ramble on in my late night postings!

with 55mA, would be less.
Looks like I tend to ramble on in my late night postings!

No problems my friend. You spotted 2 mistakes I made today!