Author Topic: Best way to bypass individual LEDs in a matrix of unknown design  (Read 2459 times)

0 Members and 1 Guest are viewing this topic.

Offline tallusTopic starter

  • Newbie
  • Posts: 2
  • Country: us
Hey there!

Glad to be part of this community. Dave's videos brought me here, probably like many of you. I'm a beginner, but not complete beginner. I've got about half of the basics concepts, but don't know how to connect them yet. 

Objective:
  • Selectively remove RED high power LEDs in a 10x6 60 LED matrix.
    Replace red leds with components (resistors and diodes, I think) which allow the remaining blue and yellow LEDs to draw their rated power. No ghosting. No flicker. No burnout.
(Images if you scroll down!)

Details and specifications.
The exact device is this one: https://www.amazon.com/MarsHydro-Spectrum-Certification-Hydroponic-Greenhouse/dp/B00XC3LBI2 Yep. Yes. This is a growlight for growing you know what. So why am I mutilating my grow light? Because I'm not growing what you think I'm growing! I want to use this as supplementary sunlight for my houseplants here in NYC in the deep winter. I'm removing 5/6th of the red LEDs, because currently the matrix emits the tell-tale shockingly bright, acerbic, pink color light that is exactly what will get my door kicked in by the NYPD. I'll run it during the day and with 5/6th of the red LEDs removed the color balance will be much closer to daylight and unobtrusive.

I cracked open the device and it'v very simple. The PSU splits and goes to two identical transformers, each with 110-240 VAC input with an output labelled "DC 55V-86V (0.665-0.68A)"

The wires then disappear into a heat-sink sandwich, probably the front facing Al plate the LEDs are through-holed to, and the rear heatsink which appears to be machined from a wide bar. I have not opened the sandwich. Doing so is difficult for a few reasons I will explain only if asked.

I've cracked open the glass to access the LEDs on the business end. I do not touch ANYTHING while the device is running. It could be safe, sure, and I me and idiot, definitely. But I see two transformers back there the size of bricks so I'd rather be safe than sorry. After powerdown I bleed the terminals outs with a probe connected to a 10 megaohm resistor connected to earth ground.

A few befuddling things when device is unplugged: with a DMM, if I read the voltage across extant LEDs I get a few microvolts each in the strips which I've removed red LEDs, and about 2X that value across LEDs in the lower strips I haven't clipped one out of yet. However, I get no current. Further, all LEDs on the left and right columns (which I think, if they are wired as horizontal rows, correspond to the first and last LED in a row) read a negative voltage. Most but not all of the rest of the LEDs read positive, with some exceptions. I think but do not know that the yellow LEDs read negative voltages.   

Problem and attempts so far
When I clip and remove one or more red leds, the remaining diodes in a horizontal strip suffer. They illuminate less brightly, and flicker. Pic:

https://i.sli.mg/yj5y0a.jpg

Here's a picture where you can see the first four rows which have had LEDs clipped out differ in brightness from the two bottom rows where I've done nothing.



In one horizontal strip in the matrix, I've placed 1000K oHm resistors across where each red LED used to be.



What is the component directly below each LED? Something that enables parallel config perhaps? So that when one LED blows all the others won't go out. 

1000K ohms rationale: because I figure based on the output (55-85 VDC), and number of LEDs, 60, assuming 60 VDC, each LED is going to be about 2.3W. 1000K is probably the wrong value, but it shouldn't be TOO far off.

It now occurs to me what the "D" in LED stands for, diode, and wonder if I should place a diode in series before the resistor. I'm basically trying to "trick" the device into thinking there is still a LED there sucking the same amount of power (dissipated ohmically not luminously, w/ resistor), and behaving the way the LED that was in there before behaved.

Thanks for any help!

-T
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #1 on: January 08, 2017, 07:00:03 pm »
The simplest solution is to replace with another color led, For example white color led, Or you may replace with 5 1N4004 in series


http://www.ebay.com/itm/10pcs-1W-high-power-LED-Star-White-Warm-White-Red-Green-Blue-Amber-140-Deg-Light-/351676642980?var=620650081018&hash=item51e18fb6a4:m:mGj0AqyUgpfojjZdIlcfvnA
 

Offline NottheDan

  • Frequent Contributor
  • **
  • Posts: 281
  • Country: gb
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #2 on: January 08, 2017, 07:54:04 pm »
What is the component directly below each LED? Something that enables parallel config perhaps? So that when one LED blows all the others won't go out. 
At first glance I'd guess a zener diode?
 

Offline tallusTopic starter

  • Newbie
  • Posts: 2
  • Country: us
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #3 on: January 08, 2017, 10:04:44 pm »
The simplest solution is to replace with another color led, For example white color led, Or you may replace with 5 1N4004 in series



ebclr, thanks!

Why 1W LEDs instead of some other power? (The 55-85 VDC output on the two transformers is confusing to me, which alters how I calculated the resistor and power of the LEDs (I got 2.3 watts, closer to 3 watts). Is there a way to test if some LEDs are pulling more current than others? ie. is it safe to take DMM measurements across an LED while the device is on?

Zener diode! Haven't heard of them, but clearly I should know them . And now I do.
 

Offline Brumby

  • Supporter
  • ****
  • Posts: 12298
  • Country: au
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #4 on: January 09, 2017, 03:35:00 am »
It now occurs to me what the "D" in LED stands for, diode, and wonder if I should place a diode in series before the resistor. I'm basically trying to "trick" the device into thinking there is still a LED there sucking the same amount of power (dissipated ohmically not luminously, w/ resistor), and behaving the way the LED that was in there before behaved.

Why try and "trick" the device?  If you don't like the colour, change the red LED for another colour - but make sure the new LED has roughly the same voltage drop and at least the same power handling capability.  (If we knew exactly how the LEDs were connected, we could offer some specific alternatives.)  IMHO, it's better to get usable light than waste that power as heat.

As for the NYPD - why not just go up to them and tell them what you are doing ... even invite them to check it out if they want.  Maybe tell a couple of neighbours, so they don't call them.  If some passer-by gets anxious and calls them - they're more likely to knock on your door than kick it down.
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #5 on: January 09, 2017, 04:45:06 am »
Normally those led driver are based on current , the Led you have can be 1 W or 3 W, those are the unique possibility for this package, It's a high probability that your current  is between 260 to 300 ma, that is stadard for this kind of Led. Just replace with another color LEd it's easy and quick
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 19527
  • Country: gb
  • 0999
Re: Best way to bypass individual LEDs in a matrix of unknown design
« Reply #6 on: January 09, 2017, 10:10:02 am »
Is this really necessary?

The grow light will be designed to output the spectrum required for growing most plants, legal or otherwise. Modifying it is only going to reduce its efficiency and reliability. If the plants are in public view, then the police should clearly see you're not doing anything illegal and leave you alone.

Another thing to note is, if you can keep the plants dormant, it may be possible to overwinter them, without any additional light. The problem is most often a combination of high temperatures and low light levels. If you can keep the temperature between 10oC and 15oC then it will keep them dormant and the low light levels won't be so much or a problem.

What are you growing?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf