Hacking an Illumagic light for reef-pi control

mawhins · Jul 5, 2019

Hi all, I'm from Australia and I'm new to reef keeping and trying to keep costs down. I bought a 4 ft Illumagic light that's suffered damage to the LCD on the controller. The LEDs all work but without a screen it's impossible to navigate the menu and change the settings. My plan was to either a) fix the LCD and cut the light in half to fit my tank, or b) take out the controller and control the light with a reef-pi (my preferred option) and cut the light in half to fit my tank.

So I've got the pi and reef controller software all up and running and I've searched high and low and can't see any information on anyone having done this with this light.

I've looked at the PSU and it's outputting 15v. There are only 2 wires coming into the light from the power supply. The controller board has a 5 pin connection:
- Vin 15v measured on multimeter
- gnd
- channel 1 white
- channel 2 blue
- channel 3 violet

It's tricky to see what's going on over these channels because I can't make changes to the settings to observe the effects.

How else might I work out what voltage I need to use to control these channels. Is that the forward voltage?

Any help would greatly appreciated.

Crabs McJones · Jul 5, 2019

@Ranjib might know

oreo54 · Jul 5, 2019

Looks sort of constant currenty.. What are the numbers on those square bigger chips?
Likely driver chips and may help to understand whats going on..

most of these are usually Aduino based type lights w/ either using 5V PWM or 5V pwm to MOSFETS (constant voltage but doesn't look right for that)..

All really just guesses at this point..

mawhins · Jul 5, 2019

Thanks oreo, shot of the chip attached. Am thinking about setting up the circuit and creeping up the voltage until it works. This of course seems inherently stupid. Not sure what else to do though.

oreo54 · Jul 5, 2019

http://product.ic114.com/PDF/H/hs711.pdf

PWM Dimming Control A711 provides PWM dimming control function through OE pin. Basically, by changing the duty cycle of the PWM control signal, the ratio of LED ON-time to OFF-time can be changed. Consequently, larger duty cycle results in larger average current for driving LED and thus higher luminance.
Enable Voltage, VOE13.2

got you 1/2 way there..
Since it now ha exceeded my paygrade it "looks" like 15V PWM is th key..
Need others more qualified to confirm..

You'll need a converter to "upgrade " the pwm output of the GPIO ..AFAICT.
to 13.4V.. maybe other voltages work but there isn't a range listed afaict..

The voltage 100% between the say "white" and ground will tell..
REMEMBER this is just a logical guess.

mawhins · Jul 5, 2019

Ahh well done on finding that and thank you. I just cut the fixture up and rewired at 100% over the tank. I hate the Disco ball effect of it so if it gets too hard I may abandon it.

oreo54 · Jul 6, 2019

Looks like your violets lenses are browning out as well IF those are the violets .. Fairly easily replaced..

I have an experiment you could try..
https://www.ecolocityled.com/product/pwm_led_light_dimmer_knob/pwm_led_light_dimmer

These little manual dimmers will drive a PWM signal (well meant for power PWM)
All you need to do is use a 15V power supply (possibly just a tap off the one you are using) and if need be a DC/DC cheap step down board to drop it to 13.2 volts or whatever I said above.
Turning the pot sends a 5v pwm signal to the internal MOSFET.
That will be mirrored on the 13.2 volt feed.
You could add a resistor to current limit it but prob not necessary (no guarantee)
At least you could get manual channel dimming out of it and you'll know how it works..

More than likely if you use a second power supply you may need to common the grounds..

Oh one per channel obviously..

mawhins · Jul 6, 2019

Do you think it's a lens on the led chip or is it the lens itself?

oreo54 · Jul 6, 2019

I can't tell from here. but those "egg" violets have a tendency to brown on the chip..
From what I understand it's really "impurities" that break down and brown.. like oils and organics in the plastic, not the plastic itself really ..though there isn't a practical difference..
Best ones use glass lenses on chip. Unfortunately prob. not in your current footprint on the board.

Michael Lane · Jul 7, 2019

The datasheet for the A711 states a maximum pwm voltage of 13.2v, so I don't recommend signalling with 15v. It also states that OE (pwm) pin can be left floating for normally on. I would bet that you could use a standard 5v pwm signal to drive these lights. The datasheet also provides some guidance on pwm frequency, so we could also figure out a good pwm frequency if you get it going.

Ranjib · Jul 7, 2019

Sorry I do not have any prior experience with this ic. Hopefully Michael or Oreo will get you sorted

mawhins · Jul 8, 2019

Michael Lane said:
The datasheet for the A711 states a maximum pwm voltage of 13.2v, so I don't recommend signalling with 15v. It also states that OE (pwm) pin can be left floating for normally on. I would bet that you could use a standard 5v pwm signal to drive these lights. The datasheet also provides some guidance on pwm frequency, so we could also figure out a good pwm frequency if you get it going.

Michael, you're a legend, thanks for your interpretation of the datasheet. I'll try 5v and let you know the results!

mawhins · Jul 8, 2019

Ranjib said:
Sorry I do not have any prior experience with this ic. Hopefully Michael or Oreo will get you sorted

No worries Ranjib, thanks for reading! I'll no doubt have plenty of pi specific questions for you in due course [emoji16]

mawhins · Jul 12, 2019

@Michael Lane
Last night I tried the 3.3v PWM directly from the Pi on one channel and it did have an effect, it did dim the channel, but the modulation was observable and setting the duty cycle to 0 or very low didn't reduce the intensity to the expected levels. I tried a range of frequencies but the flicker didn't really stop. So next I'll try stepping up the voltage somehow. I have a couple of motor drivers, one is an H bridge. Maybe I can use on of those.

oreo54 · Jul 12, 2019

https://www.ebay.com/itm/DC-DC-2A-2...349586?hash=item48ce27a392:g:5dgAAOSwQHBdAw90

Michael Lane · Jul 12, 2019

That's a shame. I use a little pwm signal generator to mess around with stuff like this.
It's super cheap on aliexpress if you don't mind waiting. Otherwise you can usually find something similar on amazon for <$10. It can work with a variety of voltages and you can change the frequency too.
https://www.aliexpress.com/item/32867909072.html?spm=a2g0s.9042311.0.0.27424c4dy23wp2

The result you describe makes me wonder if the pwm frequency was too low. Keep us posted if you discover any other behavior.

I'm sure you also connected the pwm ground to the ground pin on the light, but I figure I should mention it just in case.

oreo54 · Jul 12, 2019

Michael Lane said:
That's a shame. I use a little pwm signal generator to mess around with stuff like this.
It's super cheap on aliexpress if you don't mind waiting. Otherwise you can usually find something similar on amazon for <$10. It can work with a variety of voltages and you can change the frequency too.
https://www.aliexpress.com/item/32867909072.html?spm=a2g0s.9042311.0.0.27424c4dy23wp2

The result you describe makes me wonder if the pwm frequency was too low. Keep us posted if you discover any other behavior.

I'm sure you also connected the pwm ground to the ground pin on the light, but I figure I should mention it just in case.

Yea the frequency thing was my first thought.. and the odd (ONLY cursory) understanding of the pi and it's frequency/duty cycle thingy.. (only way I know of to explain it.. forgive me) but anything above 500Hz doesn't demonstrate strobing at low duty cycles normally..
They did mention changing the frequency though if it "stuck" is ????
Sorry all beyond my pay grade so forgive the "simpleness of this all....
don't remember spec sheet mentioning frequency at all and not familiar w/ the orig lights and how it functioned either (some drivers don't dim well to zero).

If the orig light dimmed to zero w/out a cutoff at say 10% then should be able to repeat it..

OH and forget my cheap little dimmer I keep forgetting they pulse the negative side. Some drivers don't like that..

Boost boards seem the easiest but believe one needs to remember that you can get current cut..

Again. I really only understand this enough to.. as they say..be dangerous..

https://www.aliexpress.com/item/32867909072.html?spm=a2g0s.9042311.0.0.27424c4dy23wp2

Need to get me one of those..

mawhins · Jul 13, 2019

Thanks for the links. Might get one of each of these. There's no issue putting the DC-DC step up in front of the PWM signal is there?

Re the ground, I did connect the ground but am wondering if I need to connect to a specific ground pin on the pi that is associated to the gpio I'm using.

Ranjib · Jul 13, 2019

@Michael Lane I'm curious if we can use a snpn transistor circuit to generate 12v pwm output from pi's 3.3v pwm. Employe a voltage divider circuit to generate 12V from the 15V input (or another buck converter)
The boost converter will draw more current from Pi, if I understand correctly, which is a concern.
I'll be updating you all with my pi pwm frequency verification on 2.4 releases. I had tested it on my end, but I'm hearing same issues fr0m new builds, so I'll double verify.
@Bigtrout @theatrus any advice?

mawhins · Jul 13, 2019

Ok so I checked the ground and I had lost continuity, replacement results in some interesting outcomes.

The dimming now works. Changing duty cycle from 0 to 100% dims accordingly. However, the leds now strobe, not flicker. Full on and full off multiple times a second. I've tried frequencies from 10 to 3000. Took me a while to realise you have to reload reef-pi for the changes to take effect!

Should I take this setting higher? How does frequency in reef-pi translate to Hz? Could the strobing be the result of the wrong signal voltage? I'm thinking not.

Getting closer! Thanks for the input all.