Aquarium LED Lighting Project

[ShaneB]

Hi I have very little electronic experience and I'm starting a new project to build an aquarium light.

I've purchased 5 LED's from eBay:

Operating Voltage:9.0-11.0V
DC forward current: 1000MA max
Rated power: 10W

Each LED consists of 9 small LED's pre wired into a single light, I'm not sure how many of these LED lights I will need to light my aquarium but I plan to test with 3 and increases or decrees as necessary. The LED supplier has informed me that an "LED Driver" is required for these LED's but I have a 12 Volt DC 10A power supply laying around that I would like to use for this project.


Can I use my power supply, or do I need to purchase the suggested LED Driver?

I've done some research on wiring LED in series using resisters but I'm not sure if I'm applying the formulas correctly or what the diagram would look like.

Any advice will be greatly appreciated, Thanks.

 

[flippineck]

Hi Shane I'm doing something very similar, I have blown up plenty of gear so far so you can benefit from my learning curve:

Watch the current through the LED itself, like a hawk.

Don't worry about the voltage, it's the current that's paramount.

You get a runaway situation very easily with the current, one secon things seem hunky dory then with almost no change in voltage, the current goes through the roof & -bang- you have a dead LED (and probably also a dead whatever-it-is you're using to feed power to it)

look up 'constant current power supply', there are a few really really good links on here about the subject that I really wish I'd read before blowing up a stack of gear ;-)

was it one of the mods 'Steve' who did a sort of big FAQ page on the subject?

IIRC I started out using a 50 Watt LED and a constant *voltage* source off ebay - whilst it was working, the light output was fantastic for growing plants etc at something like half - three quarter brightness. But I think next time I deffo will go more your way, use lower power LEDs but more of them.. and a constant current not constant voltage, source

I was using a sheet of 3mm thick aluminium as a lid for the tank, and was using arctic silver thermal epoxy to glue the LED to it so that the whole lid of the tank could act as a heatsink.

 

[(*steve*)]

flippineck (just what is a flippi neck?) has provided some good advice.

The "big sorta FAQ page" is here: https://www.electronicspoint.com/got-question-driving-leds-t256849.html (whilst I maintain it, it has had lots of input from other people).

It's worth having a read of it.

The problem with driving those LEDs from a 12V supply is that the driver may not have sufficient overhead, and that you'll need one driver per LED. However this problem also means that you can probably get away with a simpler linear current source.

 

[flippineck]

Flipping heck: Northern UK colloquialism.
Translation: OMG

I return you to the discussion at hand

^ yep that's the page I meant, it's really illuminating on the subject, all my missus heard that evening was me going "ahhhhhh.. so *that's* why..." over and again

 

[(*steve*)]

Flipping heck: Northern UK colloquialism.
Translation: OMG

Yeah, I knew that, but I still read it first as flippi-neck, and that's all I now see when I read it

You know what heck is? And who rules it? (I think you've just earned yourself an avatar!)

^ yep that's the page I meant, it's really illuminating on the subject, all my missus heard that evening was me going "ahhhhhh.. so *that's* why..." over and again

That is a hugely satisfying thing to hear. Thanks.

Whilst I don't take the full credit for it, I will accept blame if it is not readable or understandable.

 

[davenn]

And who rules it?

haha that's funny ... " You can call me Phil ... Prince of insufficient light"

Thanks Steve

Dave

 

[ShaneB]

first i would thank everyone for your feedback.

OK, so i read "Got a question about driving LED's?". Although i didn't understand it all i now understand i need to control the current...and because i don't have the tools necessary to make my own circuit boards i went back to eBay and found this:

item # 370769730395, or this item # 231114115641 on eBay.ca.

the LED in the description is the same as mine, just different color. If i were to use one per 10w LED array in conjunction with my power supply, will this solve my issue?

 

[flippineck]

231114115641 has a lot more info - from that one I'd hazard a yes (as long as your LED does use the voltage / current specs mentioned)

the other one doesn't seem to give much tech detail but, the circuit board does look quite similar & it's a cheaper option - better if you were going to go into production with the whole idea. might have to just buy one and see what happens!

wait for others answers as I'm no expert

 

[(*steve*)]

They both seem to match your LED.

Be careful that you heatsink your LED. Even with a constant current driver, they can still overheat and be destroyed.

 

[flippineck]

Datasheet for the chip they both use

PT4115:
http://www.datasheet.in/datasheet-html/P/T/4/PT4115_PowTech.pdf.html

possibly a slightly different variant - PT4115E:
http://www.micro-bridge.com/data/CRpowtech/PT4115E.pdf

 

[ShaneB]

Thanks again for your help, I'll be ordering the necessary current regulators ASAP. Heat sinks and timers are the other elements I've been giving a lot of thought to.

As far as timers go, should I put one timer on the main power supply for the entire system?

Or should the system be turned on in steps, main power on and then perhaps a 5 to 10 second delay on the lights? If this is the suggested method, what is the best way to do this without having to keep 2 or more manual timers synchronized? Again I'm not equipped to make my own circuit boards. Does anyone know if there's an "off the shelf" delay timer, that will simply delay the lights each time the main power is turned on?

 

[flippineck]

Just had a look back at the specs you posted for the LED

Operating Voltage:9.0-11.0V
DC forward current: 1000MA max
Rated power: 10W

Looking at the 231114115641 driver board, it says

Output Voltage: Constant 12V DC - 10W
Output current: 900mA

So it's one volt higher than the voltage range specified for the LED

However as previously discussed it's more the current that's the issue - does that extra volt matter too much? I'm a bit confused here by the driver board description - if it's a constant 900mA out no matter what, then surely you can't also have a constant 12V output? I understood that the voltage would be allowed to vary slightly whilst the current was held still. Mind you I might be making the mistake here of thinking too much in terms of Ohms law when LEDs seem particularly fond of deviating from it any chance they get!

I've only scan-read the datasheet but I don't think I saw any mention of output voltage anywhere in it at all.. wondering if in fact the ebay listing has put '12V' spuriously.. maybe they have been badgered 'what volts is it mate' that sort of thing

The current specified for the driver output would seem to indicate a 100 milliamp safe margin. By how much would that 100mA below max, dim the LED below it's max brightness?

Also.. noticed from the spec sheet that the controller's chip has a dimming feature that works on pulse width modulation. I was wondering, on the boards we're looking at, is the dimming hard-wired for a 100% duty cycle?

I was worried that the 100mA below max, coupled with (say) the board being hard wired for 50% duty cycle, would cause the LED to be too dim.

Looking at the reverse side of 231114115641, it looks like pin 8 of the IC (the DIM pin) is not connected to anything although it's hard to see if there's a track that goes back under the chip. From reading the datasheet it looks like this would cause the chip to just keep the output 'on' all the time i.e. 100% duty cycle

Also looking at the same board, it looks like R1 is the current sense resistor Rs mentioned on the datasheet. You might be able to replace this with a trim pot to be able to vary the constant current output level? Current out = 0.1/Rs (Rs > 0.082 ohm)

Just a bunch of wee small hours thoughts

Regards the delay, are you thinking here of trying to protect the main power supply from damage? or is it an aesthetic thing.. I'm not certain but I don't think any timers would be necessary at all unless there was some specific effect or function required? Might be wrong







All in all, speaking with limited knowledge but a good pair of specs & peace and quiet in the house and time to concentrate on looking deeper at the kit, you're probably good to go with one of those boards without much further ado..

 

[(*steve*)]

By how much would that 100mA below max, dim the LED below it's max brightness?

Not a lot. Probably hardly noticable.

Also.. noticed from the spec sheet that the controller's chip has a dimming feature that works on pulse width modulation. I was wondering, on the boards we're looking at, is the dimming hard-wired for a 100% duty cycle?

Yep, it would be hard on.

Also looking at the same board, it looks like R1 is the current sense resistor Rs mentioned on the datasheet. You might be able to replace this with a trim pot to be able to vary the constant current output level? Current out = 0.1/Rs (Rs > 0.082 ohm)

The resistor carries the full current. You wouldn't use a trimpot.

 

[ShaneB]

The delay was purely to protect power supply of potential power peak at start up. If it's not necessary I'll exclude it and just use one timer on main supply for switching the lights on and off.

This is all great advice, thanks.

 

[flippineck]

Trimpot unsuitable - ok. Could you use a suitably specified, full size potentiometer though - maybe one with a linear characteristic?

I was thinking about the dimming by use of duty cycle - seems a nifty way of doing it but, in the aquarium application, would it actually be bad (I'm kinda wondering about effect of the flicker on plant growth etc)

Maybe flicker wouldn't be a problem for that, maybe it would be bad at low frequency but no prob at high frequency.. don't know

I'm probably thinking into it all far too deeply now lol

 

[(*steve*)]

Trimpot unsuitable - ok. Could you use a suitably specified, full size potentiometer though - maybe one with a linear characteristic?

No, and for several reasons:

1) the resistance required is very low and you just don't get pots in that range.

2) The current is quite high and pots are almost always low current devices.

3) Even if you could, it would be expensive (possibly very expensive)

The best solution is to use the PWM input. It's there for a reason.

I was thinking about the dimming by use of duty cycle

That's what I'd do.

Seems a nifty way of doing it

It is, for all sorts of reasons.

but, in the aquarium application, would it actually be bad (I'm kinda wondering about effect of the flicker on plant growth etc)

Don't say it would be bad and then say you're wondering. Ask if it would have any effect.

My gut feeling is that it would not.

Maybe flicker wouldn't be a problem for that, maybe it would be bad at low frequency but no prob at high frequency.. don't know

I'm sure someone has done some research on the effect of PWM on grow lights. Why not google and see if there is anything?

I'm probably thinking into it all far too deeply now lol

Yeah, I think so. Just remember, it's all conjecture until you have evidence or proof.

 

[flippineck]

I'm hoping to provide evidence and proof of a really cool, fully working fishtank hood and it's only my apathy that's going to make Shane the winner Back on topic ;-)

 

[ShaneB]

Although being able to dim the LED's would be cool I don't have the tools to modify the circuit boards. If the drivers will work without modification with my 12v power supply for my application, then that will be fine.

Thanks for all the feedback.

 

[ShaneB]

Although being able to dim the LED's would be cool I don't have the tools to modify the circuit boards. If the drivers will work without modification with my 12v power supply for my application, then that will be fine.

Thanks for all the feedback.

OK so I know I said the it won't be necessary to dim d the LED's. But, I did Google it and now have more questions...is the voltage used to control the duty cycle also the voltage used to power the LED driver? Or is it just the control voltage? Will I need to power the LED driver as well as supply the control voltage? So far I have my 12v power supply, LED drivers & obviously the LED's. To do this, what would I need to add to my application and how?

 

[flippineck]

sorry Shane I wonder if I might have over analysed things & thus confused you.. I think the dimming function on these boards has just been left switched off as it were.. all you need to worry about is just supplying sufficient power to the board via the 2 main power leads & all should be ok

no need for any seperate control voltage or anything

The dimming function is accessible on one pin of the IC if you wanted to start hacking the board, but you could just use it as-is & get full brightness