Cant figure out whats wrong with my lights

Hi all. I have built a fish tank LED unit controlled via an arduino mega and the PWM pins within. Ever since the start I have had a problem with only two of the channels. I have a driver board that holds 7 mean well drivers. I have 6x 700mA and 1x 1500mA driver in it. My problem relates to the 700mA channels.

On channel 4 I have 12 cool white 3w LEDs.
On channel 5 I have 6 Cool white and 6 Royal blue (Which I believe are similar in spec both Fv and Current needs. I don't have the specs no more so cant really check. But I'm 90% sure I calculated them correctly)

I can have both on individually not a problem, but if I put both on together, the higher one takes precedence and the other switches off or turns very low.

I have tried changing the channels but even through different drivers it still happens.

I can have literally all of the other channels on and one of the offending two and it works fine so I don't believe it to be a dropout of the power supply.

Does anyone have any suggestion as to what can cause this? I believed that each driver will run each channel completely independently so why is one effecting another? Because of this I thought it might be the power supply but as mentioned above, I can turn ALL of the others on and nothing else is effected. Its just these two channels effecting one another.

Attached are pictures of the unit and drivers.

Thanks in advance.

I have tried changing the channels but even through different drivers it still happens.

So this points to your wiring, not so much the actual connections but the way the wires interact with each other inducing signals in the other channels. You could try making the wiring of the two channels a lot neater so they are not so tangled up with each other and are kept separate as much as possible.

This effect is also known as cross talk.

Grumpy_Mike:
So this points to your wiring, not so much the actual connections but the way the wires interact with each other inducing signals in the other channels. You could try making the wiring of the two channels a lot neater so they are not so tangled up with each other and are kept separate as much as possible.

This effect is also known as cross talk.

Thanks for your reply. Does this 'cross talk' happen even if there is no display of it when one is off and the other is on? Ie. when one channel is on, none of the other channels LED's come on showing it leaking across channels.

I have just measured the channel the voltage across the channels. The two channels that I am having problems with are measuring 46.2v each when on full. The other channels are maxing out at only 36-39v so quite a bit lower.

The power supply I am using is a 48 volt, 6.25a one of which a few volts is required to power the drivers so the two channels are drawing close to the max voltage available.

I have just tried removing an LED from each channel (3.2-4v each) to try and bring the draw down but anything over 30PWM the LEDs all just turn off.

Does this 'cross talk' happen even if there is no display of it when one is off and the other is on? Ie. when one channel is on, none of the other channels LED's come on showing it leaking across channels.

Yes it gets into the controller and can do some weird stuff.

Grumpy_Mike:
So this points to your wiring, not so much the actual connections but the way the wires interact with each other inducing signals in the other channels.

The trouble is there are so many LEDs and connection points i found it hard to do it any better.

The trouble is there are so many LEDs and connection points i found it hard to do it any better.

Then the project is beyond you.

Physics dosn't care if you tried your best and coudn't do it right.

Grumpy_Mike:
Then the project is beyond you.

Wow, you chose the right username. Thanks anyway.

I can’t think what is remotely grumpy about telling you what is wrong with your project, you telling me you haven’t got the skill to do that, and so I tell you you can’t do your project.

If you think you can rectify the error in wiring layout you made then maybe you can do it. Either way it is your choice, I am just telling like it is.

Seems to be a LED wiring problem.

Are the LEDs on a star base?
Asking, because I once had a short between the LED and the base of a 3watt LED chip (normally isolated).
Leo..

Edit: just saw the image.
Check if mounting screws are not touching/shorting the LED wiring.

Can your 48volt supply be turned up one or two volts (most open frame supplies have a pot).
Or, temporary short one LED in those strings, to lower total Vf.

Wawa:
Seems to be a LED wiring problem.

Are the LEDs on a star base?
Asking, because I once had a short between the LED and the base of a 3watt LED chip (normally isolated).
Leo..

Edit: just saw the image.
Check if mounting screws are not touching/shorting the LED wiring.

Can your 48volt supply be turned up one or two volts (most open frame supplies have a pot).
Or, temporary short one LED in those strings, to lower total Vf.

Thanks for your reply.

Yes they are mounted on star PCBs and I have fitted nylon washers between the screw heads and the star but yes still there is a possibility that a washer has twisted and there could be a short, however I have checked continuity on all the LEDs, both +ve and -ve terminals against both other LED's near it, and against the heatsink but nothing has shown up.

To be honest I've left it as it is and screwed the perspex on etc. I had a channel of 12 Ice whites and channel of 24 Royal blues (amongst many others) so missing out a white and blue mixed channel won't be making much difference. But would be nice to crack it.

I still think its got something to do with those two channels, and only those two channels pulling such a high voltage. 46.2v each to be precise, especially when the PSU is only rated at 48v and the LDD drivers need a few volts to run. The LEDs are rated at 3.2-4v MAX so did my calculations on 12x3.2= 38.4... plenty to play with. However they are pulling more. Even when I took one of the LED's off of one circuit (which I assumed would have reduced the load) they wouldn't come on at all past about 5 PWM.

LEDs do not pull voltage, they pull only current. Those drivers you have are constant current drivers and they increase the output voltage until the set current flows. The fact that they are having to provide more voltage than you expect can mean one of two things.

  1. The impedance of the LEDs is higher than you expect. This could be due to faulty LEDs, or a bad joint in the wiring, not a short. Or higher than expected forward voltage on the LEDs, you could try removing a couple of LEDs from those two chains and seeing if things change.

  2. Something interfering with the current measuring part of the circuit like interference or cross talk between the other circuits and drivers.