DIY 12v Dimmable Power Compact Driver

[huscre]

Abstract: This thread is a supplement to my nano buildoff thread to specifically cover the mods done to my lighting system. I've put it in the DIY forum since I hope this info will be useful to people with both nano and not-so-nano tanks.

WARNING: This project involves potentially hazardous voltages!

AIM: In this thread I will document the changes I've made to the lighting hood of my AquaOne AR-320 tank. My aim was to change the fittings in the hood to support better quality tubes, and to replace the magnetic ballast with an electronic circuit to drive the tube from a 12 volt supply.

Here's the kind of tank I'm talking about.. In reality mine's a lot older and in need of a good cleaning

RATIONALE: These small tanks generally come with crappy tubes with low power output and an algae-friendly colour spectrum. There are numerous ways to improve the lighting such as replacing the hood with a metal halide pendant, or dropping in some strong LEDS, but I chose to stick with the tried and tested 10,000K/Actinic 50/50 flouro tubes since I've used them before and know they work.

I also wanted to be able to control the brightness of the lighting from a separate DIY controller (ie. simulated dawn/dusk). This isn't very easy to do with mains-powered designs, but I found that many 12v circuits are inherently suited to varying the tube's brightness by simply regulating their supply current. This project is one of the simplest solutions I could think of to achieve my goals, and as an added bonus there is no "mains" wiring, so it's probably legal for us non-electricians to do as well.

I'm open to any suggestions and feedback, and I hope this thread will inspire further development in this area.

[huscre]

Stage 1 - Background research.

To anyone planning to undertake a project like this I would recommend you check out the following 2 websites. Their authors took the time to write much better descriptions than I ever will, and it was their projects that formed the basis of this one.

http://www.talkingelectronics.com/pr...tInverter.html

http://ludens.cl/Electron/Fluolamp/fluolamp.html

I also researched into filament heating for the tubes, but eventually decided that it only added unnecessary complexity (in fact later in testing it proved nothing but troublesome). I eventually decided to base my circuit on the 2 watt driver schematic presented in the link directly above this text. It is so simple I was almost certain it wouldn't work, especially since I wanted to run it at 10 times it's rated power...

[huscre]

Stage 2 - Prototyping

The only changes I made to the aforementioned circuit were the transformer (i wound mine on a ferrite rod with different windings), the mosfet (i used a 30n20 which was a free sample from ON Semiconductor but similar to the one used in the example), and a mosfet driver chip (MAX4420, another free sample) which I added to the circuit to switch the mosfet fast and keep it from getting hot.

Here's the schematic I came up with.

Winding the transformer is a crucial part of the project. I found that with too great a turns ratio between the small and long winding the circuit runs at a lower voltage (otherwise current draw will get out of hand) and with too small a ratio, the tube may never fire at all on a 12v supply. 53:360 worked well for me, but I started with a ratio of more than 80:360 then gradually unwound the primary coil until the circuit would fire at 11 volts (which happened to be at 53 turns).

The 2 Ohm resistor is in there for testing or running from a constant voltage source. The tube starts glowing at around 5 volts and fires at around 11 volts creating a bright light. I found it's best to have the circuit running from a constant current source if you're not using the resistor, since after the tube fires, the voltage likes to sit around 9 volts @ 500mA and 10 volts at 2A. If you were to supply 12 volts to the circuit without the resistor in place the current would go through the roof and the tube will fry (remember i'm only using an 18watt tube).

Running the circuit from a constant current source, it starts to glow at around 300mA and gets progressively brighter as you ramp up the current, except there's a bit of a jump in brightness when the tube fires at around 700mA. For controlling purposes I think varying the current instead of the voltage is the best way to go. This will take some more circuit designing methinks...

Another design note is the frequency. I found that running the circuit at lower frequencies has the same effect as increasing the turns ratio. Ie. @ 30KHz the tube may fire at 13volts and run at 11, whereas reducing the frequency to 20KHz (by altering the 47K resistor) caused it to fire and run at a lower voltage.

Here's a pic of my test circuit running:

Note the transformer in the background (green and black)

[huscre]

Stage 3 - Installation

Installing the new circuitry in place of the old was pretty straightforward. I've taken a few photos to illustrate the processs.

1. Unscrew the light cover and take out the old crap. (this involved drilling out the pop rivets and cutting a few wires)



2. Pop rivet the new tube base and plastic support then stick in the new tube.



3. Stick in the circuitry with velcro tape (this was all i had on hand...)



4. Wire in the switch, put the covers back together,



5. ...And test it!

Here ends my description of my hood mods. Please feel free to ask questions and make suggestions for improvements. I hope someone out there in reefland finds this useful!

[Geoff]

That is a very nice upgrade!!!

can not wait to see it fully up and running!

G~

[kcress]

Certainly impressive.
What are you using for the supply?

[huscre]

I'll be using a 12v DC 4A plugpack for the supply. They are pretty cheap on ebay for LCD screens. I'll control the brightness with a separate circuit as part of a digital controller which I'll write up in another thread.