Looking at the transom, yes 48 would be perfect. As for port or starboard...would need a minimum 140 per side. That's another day.
Need to stay on the transom task first.

Thanks.

Greg

 

Clarification- are all these LEDs controlled independently, or are there lights with clusters of red, green, and blue which are switched together? The larger power packages initially appear more expensive, but come out better in the long run.

What type of LED is an important question. See most on eBay are tightly focused, and don't show much light at all unless you're in the beam path, where it's blinding. The "rope" effect may not show up then without diffusers. There's the less common "strawhat" LED with much wider viewing range, and the "inverted cone" which mainly spreads light in a lateral plane all around the body.

Is efficiency important? The current draw of a linear design with many LEDs can get quite substantial.

__________________
I thought what I'd do was I'd pretend I was one of those deaf-mutes.

 

There are many different types of LEDs. I attached two pictures of viewing angle. 150 and 7 degrees.

This project needs a switching power supply. It will save many watts.

Attached Thumbnails

   

 

Yeah, but when addressing LEDs individually, if you use current drivers, they'll EACH need their own driver. That's cost prohibitive.
You could determine the minimum voltage and ballast for each color, and make a supply for each one. That could easily triple the efficiency at considerably less cost and complexity. But, the regulation for each LED becomes somewhat less consistent. If you chose a 4v source voltage for each green LED (3.6v nominal) with a 20ohm resistance, then the variations of a +/- 0.1v Vf, you get a +25%/-33% variation in current. Might be a good idea to use a switcher to generate 4v-5v and then you can direct-drive the LEDs without external transistors as long as it's under 20mA. That's a good choice for the 3.6v greens and blues. The reds may only need 2v, but that's only 1/3 the power usage anyways. With a 5v source, you might be able to replace each red LED with two reds in series, running on half the current. Makes a 4v Vf which is closer to the others.

If we DO drop the requirement to fewer, more powerful emitters, then we can use constant-current switchers which are the gold standard. But again, you can only drive so many lights controlled independently before the number of drivers becomes impractical.

__________________
I thought what I'd do was I'd pretend I was one of those deaf-mutes.

 

140 LEDs per side, oh my. A modular approach might work. Below, a slightly different controller with a 9-pin buss to multiple "driver" boards, each "driver" with 16-pin connectors for wiring harness to 64 LEDs (8x8 column/row).

You'd use one driver board for the 48 or 50 LEDs on the transom and three driver boards (up to 192 LEDs) for each side of the boat.

LEDs are multiplexed at 12.5% duty cycle so current budget for seven driver boards should be something like 5.6 amps but that's off the 13.8 vdc buss. The regulated 5 volts that the controller and driver boards need should be handled easily by a single 5v/1a linear regulator.

The hardware is easy. It's the software that will take the most effort.

Good luck with your project.

Mike, K8LH

Regards, Mike

Attached Thumbnails

 

 

For big numbers of LEDs i'd be tempted to use shift registers direct, ie 8 LEDs per 8bit shift reg. You save all the multiplexing etc and you can just concentrate on basic driving and PWM.

 

That would work too. You'd only need about 42 shift registers for 330 LEDs (140 left + 140 right + 50 transom) but you might be able to use a similar modular daisy-chain topology.

Mike

 

Years ago I designed a LED advertising sign with about 800 LEDs. I used a shift register. Multiplexed the LEDs. Which brings me to a question about the design so far. If you use 20mA LEDs and 8x multiplex them then the LED current may be 160mA at 1/8 duty cycle.

I think I ran the LED current at 0.5 amp peak at a very low duty cycle. Will your design run the LEDs at 150mA and low duty cycle? Can the drivers handle that? If LED current is 150mA and a column driver handles 8 LEDs then the total is 1.2 Amps.

One of the companies I design for uses 5,000,000 LEDs a year. Yes I am looking for more work.

 

Hi Ron,

The 5.6 amp figure I mentioned was for all 56 LEDs in a single row (spanning 7 driver boards) lighted at the same time (unlikely) at 100-ma or approximately 12.5-ma "average" current with the 1/8th duty cycle. The Vishay Si2312BDS is rated at 5.0 amps (5 seconds) and up to 15.0 amps in "pulsed" mode with the usual junction temperature and heat sink disclaimers. Now even though most of these LEDs have a max' 100-ma "peak" current spec', I suspect they would work just fine at higher "peak" current levels at a 12.5% duty cycle, yes, no? Each N-FET row driver is on 1/8th of the time. The MIC5891 column drivers have 500-ma outputs which only drive one LED at a time so they should do fine sourcing 100-ma or even 150-ma to each LED, yes, no?

Mike

 

Now I see the 5A FET drivers. OK
I looked at a number of data sheets and see the "100mA max". I know I have used 20mA LEDs at higher levels but those were different parts. Probably 100mA at 12.5% will give good light.
Use 2 oz copper and heavy traces to carry away heat. If you keep the PCB about 90% copper the LEDs will be cooler.