ULN 2803 Current Max

[Buzzkill]

Before I release the magic blue smoke...

I am building "canon" this time.  It will be a co2 prop with an audio board, the EZ-8 and a whole lotta LED's.  The co2 horn will be surrounded with really bright straw hat leds.  A canon has 2 groups of led's, a red and a white.  Each group draws about 400-450ma.  There are 2 canon.

By spreading the 4 groups of LED's to 4 channels on the EZ-8, can I run them off the EZ-8 and not blow the ULN?  If I did my math correctly, that would mean I would be pulling 800-900ma total (400ma per led group) through the ULN, but only for about 3-5 seconds at a time as that is as long as the prop "fires".  Both colors will not be on at the same time.  The white leds will light for the initial flash, and then the red will light for the "boom".

The prop will be fired infrequently (once every several minutes at most).  I did read the post about the high current driver.  And I am also thinking about a LM317 current limiter to keep everyting from getting fried.  But if I can just drive them off the EZ, that is certainly the path of least resistance.

Thanks in advance.

RD

[JonnyMac]

Have a look at the chart and information in this thread:

-- http://www.efx-tek.com/php/smf/index.php?topic=567.0

It should help you determine how best to allocate your outputs.

[Buzzkill]

Yes.  I saw that too.  But I am thinking that I am not reading it right as that says to me that I can run 2 channels at 475 ma with a 95% duty cycle for a total throughput of about 900ma.  In other words, as long as I only have 2 of the channels on at a time in the scenario I provided I will be fine?  Or am I running too close to the edge and will be ordering 2803's a dozen at a time?

Or plan B is to just bite the bullet, build the Tip120 circuit and the LM317 driver?  I am totally a hack at this stuff and am mostly guessing at what I am doing and how to connect the various modules.  So far I have not released too much magic blue smoke.  I did nuke a handful of white straw hats so far.  I would like to not do that that again. 

Thanks in advance for any pointers.

RD

[bsnut]

You read it right. You can run two LED strings without any problems, but don't run two channels at the same time and the thing is you want harm the ULN.

Or another way is build a TIP120 circuit like you were saying or use relays connected to the each OUTX. With the relay, you can run more LEDs on a string connected on its contacts of the relay and you won't harm the ULN. You also can run an extra channel at the same time for your show when using a relay. Because the relay coil only draws between120 - 200mA this depends on what voltage you are going to use.

[JonnyMac]

I think you'll be fine.  Keep in mind that you must use the 100% values because the controller does not modulate the output; when it's on, it's on (100% duty cycle).

I would suggest separating the two high-current outputs.  It takes less time to heat the chip than for it to cool down.

[Buzzkill]

Alrighty then.  Thanks for the input!  I'll write it up when I get it going and add it to the projects page.  I ran the numbers again last night.  It looks to be 4 groups at 400ma each.  Only 2 groups will be on at a time.  I am going to use the lm317's as a constant current driver for the LED's.  I also found that the wall wart I was using to test with was rather anemic on the ma side of things.  Bumping that to something with more oomph really made the LED's shine.  I am giddy with anticipation. I have never done anything with this many interconnected sub-systems before and with so many discrete components.  But so far the breadboarding has been fruitful.