Diy Whole Tank Led Lighting Retrofit

[SteveyG]

I would've worn a welding mask if I had one. lol

On the connector part. It is one of those that you can pull apart male/female 2 pins/2holes. The wires won't be coming out once they've been inserted. the m/f parts will come apart instead.

Just wondered why it says AC only No DC?

AC

The connectors will be fine. If you were running higher currents through the connectors and regularly connected/disconnected them under load you'd find the connections would deteriorate quicker than the same current at AC. Another example is when you select relays - the contact material is optimised for the application.

Don't tin the wires before using them in a screw contact. Although screwed connections naturally loosen as the copper softens under compression, the solder softens far quicker and easily, loosens in the terminal and gives you a high resistance contact.

 

[SuperColey1]

These won't be disconnect 'under load' anyway. Its just something so that if I need to remove the hood I can quickly turn the plugs off and disconnect the wires without having to unravel things and get plugs through holes in the cabinet etc.

Many thanks I shall buy a few more.

AC

 

[SteveyG]

Excellent, thankyou I assume to push 7000ma into 10 LED's in parallel at 3.5V I'll have to amplify the output with say a 741 op-amp?

Don't try to run 10 LEDs in parallel at 7A!! It follows on from the discussion about variances in forward voltage of the LEDs. If you have 10 LEDs in parallel and one draws 700mA at 3.5V and the rest draw 700mA at 3.6V, the one with the lower forward voltage will be overdriven. Eventually it'll fail and then you end up with 7A through 9 LEDs. They'll all then be overdriven, but there'll be another one with a slightly lower forward voltage than the rest except it'll now fail quicker than the first. You then end up with 7A through 8 LEDs etc.

Always drive your LEDs in series!

 

[rabbut]

Excellent, thankyou I assume to push 7000ma into 10 LED's in parallel at 3.5V I'll have to amplify the output with say a 741 op-amp?

Don't try to run 10 LEDs in parallel at 7A!! It follows on from the discussion about variances in forward voltage of the LEDs. If you have 10 LEDs in parallel and one draws 700mA at 3.5V and the rest draw 700mA at 3.6V, the one with the lower forward voltage will be overdriven. Eventually it'll fail and then you end up with 7A through 9 LEDs. They'll all then be overdriven, but there'll be another one with a slightly lower forward voltage than the rest except it'll now fail quicker than the first. You then end up with 7A through 8 LEDs etc.

Always drive your LEDs in series!

OK, I can see where you're comming from... I don't particularly want to loose the whole lot, or part of them, to one LED failing, hence why I was (rather daftly now you've pointed out the fault with it) thinking of wireing in parallel...

Could I take the PIC's output and feed it into 10 transistors and then drive all the LED's as seporate individuals from there, as almost individual circuits of one alone?

Sorry AC, I'm running off with your thread again... Mebe I should get updating my own thread and linking to it round about now.....

All the best
Rabbut

 

[SuperColey1]

No probs. This thread is giving the advanced and the simple version together. saves searching for one or the other. lol

AC

 

[SteveyG]

Could I take the PIC's output and feed it into 10 transistors and then drive all the LED's as seporate individuals from there, as almost individual circuits of one alone?

Yes, you could do this if you wanted to control all 10 individually. If you don't need to control them individually then connect as many as you can in series based on your supply voltage.

Having 10 8-bit PWM channels on a PIC may pose a challenge, options would be to move up to a faster PIC, or to reduce the PWM resolution - I suspect reducing the resolution to 6 bit (64 brightness levels) would not cause any problems.

 

[Schmill]

I would've worn a welding mask if I had one. lol

On the connector part. It is one of those that you can pull apart male/female 2 pins/2holes. The wires won't be coming out once they've been inserted. the m/f parts will come apart instead.

Just wondered why it says AC only No DC?

AC

I'm also guessing that one reason for the AC not DC, is the connectors possibly aren't keyed? Once you disconnect them is there anyway you can be sure that you can connect them the right way around again?

 

[rabbut]

Could I take the PIC's output and feed it into 10 transistors and then drive all the LED's as seporate individuals from there, as almost individual circuits of one alone?

Yes, you could do this if you wanted to control all 10 individually. If you don't need to control them individually then connect as many as you can in series based on your supply voltage.

Having 10 8-bit PWM channels on a PIC may pose a challenge, options would be to move up to a faster PIC, or to reduce the PWM resolution - I suspect reducing the resolution to 6 bit (64 brightness levels) would not cause any problems.

So you couldn't have one output and just connect it to transistors in 10 seporate circuits then?

I could reduce the resolution, however, I've just finished to level setting bit of code (PWM wave form maker to go) in Python. It would appear that it can be compiled to machiene code afterall

All the best
Rabbut

 

[SuperColey1]

No they are not keyed.

But the screw head is on one side and the nut the other. lol

So I can put them together with both screw heads facing the same way (That is if the Wilko label comes off. lol)

Maybe I will put some coloured tape on one matching side of each as a very easy visual clue.

AC

 

[nitro7]

so superc been following this tread
think your doing a brill job
but how much is this costing to build and run compared to normal lights
keep up the good work
eric

 

[rabbut]

AC's design is a fraction cheaper to build and lots cheaper to run than normal lights.

My rig will probibly work out more expencive than normal lights to build, but equally cheap to run over time (possibly cheaper due to not needing multiple drivers and transformers to run it. Transformers are a big power waster due to heating, and one controller over 4 should reduce power lost in control, depending upon the PIC I choose in the end)

All the best
Rabbut

 

[SuperColey1]

My fluoros (ballast plus tubes etc would've cost £60-£70 and is running 48W

This LED setup has cost £110 and is running 42W

However how much power is being drawn to turn watts into light I cannot answer.

Usable lifetime of the tubes on electronic ballast for planted tank use would be 2 years and tube replacement approx £5 each

lifetime of the LEDs (barring faults) is in the region of 7 years.

So it cost me more to setup but it is using less watts and should last 3 x as long as the fluoro tubes.

an equivalent LED setup retail would be £400 minimum and probs not be as adaptable as mine is

AC

 

[TheConstantGrower]

Oh god that is so damned cool.

Idea officially hijacked my friend.


Also, thank you for making many mistakes for me, and thank everyone else for their brilliant keys of advice.

I think when its all said and done you should do a complete DIY write up. keep up the brilliant work.

 

[SuperColey1]

This is basically a DIY write up. lol

I think this sort of post is always much better than a cleaned up version. I find it always very useful to see others mistakes as I did whilst researching this one and then learn from more input on top of the mistakes seen.

Therefore you see how I did it and you also take care not to do what I did wrong etc.

I haven't finished yet either. lol

I have now removed lots of the yellow tape as the silicon sealant had dried. Put some more silicon on other parts etc so it looks a lot tidier at the moment.

then I have to fit it this weekend.

I am then doing a single 3 x 3W series to go over a 2USG (10 litre) Nano tank in the next couple of weeks which will use the proper heatsinks, lenses and collimaters (If this one works OK)

AC

 

[SuperColey1]

I have installed the lighting and it is working. Photos don't do it justice in the slightest but then I would say that. lol

In summary I have moved from 48W fluorescent to approx 42W of LED. I have had to raise the 'luminaire' by 6" and therefore the light is now approx 12" above the water level!!! However it is still brighter. lol

Here is the tidied up board all lit up:

And here it is fitted into the 'luminaire'

Finally a look at the right rear corner where the wires come out of the luminaire:

Onto the lighting schedule. forgive the quality of the pics. A mix of poor camera and lack of photography skills here. lol

Moonlight from midnight until 3pm (running on 10V)

Light 1 comes on at 3pm

Light 2 comes on at 3-45pm

Light 3 comes on at 4-30pm

Light 4 comes on at 5-15pm

And light 5 comes on at 6pm

The full 5 series stay on until 9pm and then they turn off from the left so the reverse as in the sun moving across the sky.
Series 1 off @ 9pm, series 2 off @ 9-45pm, series 3 off at 10-30pm, series 4 off at 11-15pm and then moonlights back on at midnight.

The fans are running 24/7 and should make sure any moisture that may get into the sealed section is dried pretty quickly.

You cannot see in these pictures the shimmer. It is absolutely amazing. Wow factor to say the least and well worth the hours I have put into this (maybe as much as 50 hours!!!) The cost was high compared to comparable wattage of fluoro fittings BUT these are way brighter for the watts and the effect is worth the money.

AC