Issues Wiring Three 5v 100watt PSU to power LED Strips

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
Hello everyone! First-time poster here.

So, I've been banging my head up against a wall with this project for a while now, and I was hoping to get a bit of expert guidance. Instead of me vomiting 1000 words trying to describe what is a relatively complex circuit (for a project like this), I think I can boil it down to this one simple question highlighting a concept I don't quite understand.

Why does this setup (see attached img) make the LED strip blink off/on for a split second a couple of times per minute seemingly at random? If I remove the ground wire from PSU2, the random flashing stops.

I know it doesn't make sense from this diagram why I would want to connect the ground wire of the second PSU. There are several things I've left out for the sake of asking as clear, simple question as possible. Any light anyone could shed on this for me would be extremely helpful. Thank you!
 

Attachments

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
To Mods:

I think I should have posted this in The Projects forum. I've re-posted it there. Feel free to delete this thread. My apologies!

MOD: Moved your Thread.E
 
Last edited by a moderator:

LesJones

Joined Jan 8, 2017
4,174
I am assuming that the AC in (Live, Neutral and earth are just connected to the inputs. Rather than the way it is drawn with an output from PSU3 feeding PSU2 and an output from PSU 2 feeding PSU1.) It looks like only PSU 1 is feeding the load. There is no connection between PSU 3 and the load and there is only the negative of PSU 2 connected to the load. Is this really the way it is connected ?. Am I correct in assuming that you are trying to parallel the outputs of the power supplies to get a high enough current for the LED lighting ? As the total load is only about 44 watts (43 for the LEDs and 0.6 for the controller.) I don't see why you need more than one 100 watt PSU.

Les
 

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
I am assuming that the AC in (Live, Neutral and earth are just connected to the inputs. Rather than the way it is drawn with an output from PSU3 feeding PSU2 and an output from PSU 2 feeding PSU1.) It looks like only PSU 1 is feeding the load. There is no connection between PSU 3 and the load and there is only the negative of PSU 2 connected to the load. Is this really the way it is connected ?. Am I correct in assuming that you are trying to parallel the outputs of the power supplies to get a high enough current for the LED lighting ? As the total load is only about 44 watts (43 for the LEDs and 0.6 for the controller.) I don't see why you need more than one 100 watt PSU.

Les
Thanks for your reply! Your assumptions were correct. I'm not much for drawing diagrams. Sorry about that.

As I mentioned in the OP, the diagram isn't going to make sense because it doesn't reflect the full complexity of the rest of the setup. Rather than trying to go in depth about everything, I really just need to know what's going wrong with the circuit I crudely diagrammed. If I can understand why the LED strip blinks off/on in that setup - what concept I'm either misunderstanding or am ignorant of in the first place - I think I can fix the whole system.

At first, I thought perhaps the blinking was caused by overcurrent protection from the paralleled PSUs (via the ground connection) not being load balanced; the LED strip was drawing its full power from only one of the PSUs (the one with slightly higher voltage) rather than equally from both. But, that doesn't make sense. Each PSU has more than enough capacity to drive one LED strip. So, something else must be wrong.
 

BobTPH

Joined Jun 5, 2013
8,813
You cannot simply parallel multiple power supplies because they will not be at exactly the same voltage.

The best solution if you need the three power supplies is to have each one connected to only 1/3 of the strips.

Bob
 

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
You cannot simply parallel multiple power supplies because they will not be at exactly the same voltage.

The best solution if you need the three power supplies is to have each one connected to only 1/3 of the strips.

Bob
After I adjusted the PSUs output voltage to be within +/- .01v, the random blinking I mentioned in the OP stopped. Before adjustment, one PSU was 5.56v and the other was 5.10v.

I don't really understand why this would have caused the blinking issue when the V- rail of all the PSUs was commoned. Does commoning the V- place the PSUs in parallel? What confuses me is that the loads still look - at least visually - split between each PSU even with the V- commoned.
 
Last edited:

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
My thinking is that, if commoning the V- does place the PSUs in parallel, then the vdroop from load was probably smaller than the voltage mismatch between PSUs, and that's not good. However, I only know about such a configuration causing problems for load balancing between paralleled PSU, the mismatch causing the load to "cascade" from one PSU to the next rather than be shared between them. But, in even then the load I was applying should have been easy for a single PSU to handle even under those circumstances. So, maybe that's not the answer?
 

BobTPH

Joined Jun 5, 2013
8,813
This works until the temperture changes or the load changes, or a component ages or any number of other things thatt make the output voltages diverge.

Now, please tell me how you used two sipplies for the same load without connecting the ground together?

Bob
 

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
This works until the temperture changes or the load changes, or a component ages or any number of other things thatt make the output voltages diverge.

Now, please tell me how you used two sipplies for the same load without connecting the ground together?

Bob
I very much appreciate your replies, but can you just help me understand why in the following diagram PSU1 being at 5.5v and PSU 2 being at 5.10v caused the intermittent flickering issue I described with the V- commoned (as drawn), all other problems aside?
 

Attachments

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
This works until the temperture changes or the load changes, or a component ages or any number of other things thatt make the output voltages diverge.

Now, please tell me how you used two sipplies for the same load without connecting the ground together?

Bob
So you're also saying I need to put in a load balancing circuit?
 

BobTPH

Joined Jun 5, 2013
8,813
Not without testing your power supplies to see how they respond to current flowing in to their output, or to inability to regulate.

Think about it, if the two supplies have different voltages, how can they both be in regulation at the same time? Some power supplies will shut down when they cannot regulate.

If you insist on paralleling them, you will need a resistor in series with each supply that drops enough voltage to acoount for the difference.

Bob
 

LesJones

Joined Jan 8, 2017
4,174
I suspect that your problem could be due to the voltage drop on the negative line cable being superimposed on the data signal. I suggest that the negative line of the controller is connected directly to the negative terminal of the LED strip. That way there will be no pulsating current on the negative line between the controller and the LED strip. What is the nature of the data signal level and the protocol being used ?

Les.
 

BobTPH

Joined Jun 5, 2013
8,813
You didn’t answer my question though.

I think, maybe, you are saying that it works when you connect one supply to one end of the strip and the other supply to the other end. This might work because, though the power lines ARE connected along the strip, there might be enough resistance to stop bad things from happening.

So, what do you do with the third supply when you need it?

And why are you still pursuing an inferior solution to the one I already gave you? Just divide the strips into equal thirds and power each by one of the supples. You will have to continue the data line across the cuts and common all of the grounds so the signal has a common reference.

Bob
 

BobTPH

Joined Jun 5, 2013
8,813
No, the idea is that of one power supply is supplying the bulk of the current, its voltage will drop more than the others. A diode does not do this as well because its voltage drop is nearly contant over a range of currents

Bob.
 

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
I suspect that your problem could be due to the voltage drop on the negative line cable being superimposed on the data signal. I suggest that the negative line of the controller is connected directly to the negative terminal of the LED strip. That way there will be no pulsating current on the negative line between the controller and the LED strip. What is the nature of the data signal level and the protocol being used ?

Les.
I think the controller uses a PWM signal over the data line, but I'm not sure. It's a cheap Chinese thing that barely comes with a manual.

A diagram of my full circuit is included with this post. See the notes below for more info.

You didn’t answer my question though.

I think, maybe, you are saying that it works when you connect one supply to one end of the strip and the other supply to the other end. This might work because, though the power lines ARE connected along the strip, there might be enough resistance to stop bad things from happening.

So, what do you do with the third supply when you need it?

And why are you still pursuing an inferior solution to the one I already gave you? Just divide the strips into equal thirds and power each by one of the supples. You will have to continue the data line across the cuts and common all of the grounds so the signal has a common reference.

Bob
Here is my full circuit. I hope this clears things up.

Note:
  • I'm pushing 5.6v on the DC V+ because it saves me 5 or so amps across the whole circuit. 6v is the LED's limit.
  • Power is run into the front and back end of the LED strips because the manufacturer used wire that is too small for the amount of current that these things draw. Heat is a problem.
  • Though marketed as being able to control over 2000 LEDs, the performance of these controllers falls apart after 250. This makes the multiple controllers necessary
  • Per-controller control over LEDs is desired, but simultaneous control over all LED strips is required at times. The individual controllers cannot be synced to accomplish this, thus the repeater/amplifier and data line switches are necessary.
  • Yes, it would be great to power all of this with a single PSU. However, this is accent lighting in an art installation in a recording studio. I cannot have a 70dB fan screaming in the corner. The 100W PSUs are fanless.
Now, with all that being said, how do I safely and effectively power this setup in the easiest, most efficient way possible? Do I need load balancing resistors or diodes? If so, which ones are recommended? Is it fine the way it is?

Thank you for your help.
 

Attachments

Last edited:

Sensacell

Joined Jun 19, 2012
3,432
You should link to the specific LED strips and controller, which always helps.

Assuming your LED strips are RGB addressable types with a single DATA input, like the WS2812 type?

These strips work well as long as you pay careful attention to grounding and voltage drops caused by careless wiring practices.
If you find that everything works fine until you start lighting up lots of lights, then glitches start- this is a sure sign of data corruption due to ground voltage drop issues.


Here is the solution:
Wire the controller to the LED strip with 3 wires, connect the 5V power DIRECTLY to the LED strip, with HEAVY wire.
Run HEAVY power wires to both ends of the strips.

This way there is very little current flowing in the wires to the controller, only the small current to run the controller. The controller 'floats' at the ground potential at the beginning of the strip.
This wiring scheme ensures that the controller ground and the LED strip ground share the same voltage potential, this prevents voltage drops in the wiring from corrupting the data signals. Wire your 3 power supplies to 3 separate circuits, DO NOT connect the +5 together, you can connect the grounds, but leave the +5 as a totally separate circuits.
 

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
You should link to the specific LED strips and controller, which always helps.

Assuming your LED strips are RGB addressable types with a single DATA input, like the WS2812 type?

These strips work well as long as you pay careful attention to grounding and voltage drops caused by careless wiring practices.
If you find that everything works fine until you start lighting up lots of lights, then glitches start- this is a sure sign of data corruption due to ground voltage drop issues.


Here is the solution:
Wire the controller to the LED strip with 3 wires, connect the 5V power DIRECTLY to the LED strip, with HEAVY wire.
Run HEAVY power wires to both ends of the strips.

This way there is very little current flowing in the wires to the controller, only the small current to run the controller. The controller 'floats' at the ground potential at the beginning of the strip.
This wiring scheme ensures that the controller ground and the LED strip ground share the same voltage potential, this prevents voltage drops in the wiring from corrupting the data signals. Wire your 3 power supplies to 3 separate circuits, DO NOT connect the +5 together, you can connect the grounds, but leave the +5 as a totally separate circuits.
THIS IS VERY HELPFUL! THANK YOU!!

This is actually similar to the setup I have now except I'm running data wires only out of the controller to the front of the strips. I am running separate power rails from the PSU directly to the front and back of the strips like you suggested.

The issue is that I have multiple controllers and a repeater/amplifier that get switched into and out of the circuit depending on what functionality I need. I'll attach a diagram I drew below.

It's kind of complicated, but the short version is that there are 4 controllers that control 4 separate groups of LED strips. This allows for independent control of each group. When simultaneous control is desired, a switch sends one controller's data line over to the input of the repeater and one of the group of LEDs' data input lines to one of the repeater's four outputs. Another switch moves the other three LED groups' data inputs to the remaining outputs of the repeater. This allows one controller to modify all groups of LEDs while thinking it is only controlling one.

You'll notice there is no ground or V+ out of the controller moving around with the data wires. Each strip draws current from parallel runs of V+ and V- at its beginning and end that begin/end directly from one of the 3 PSUs. This worked but created ground loops.

The ground loops were solved by connecting together all the V- terminals of the PSUs, but this caused the LEDs to blink a couple times a minute randomly. Turns out one of my PSUs was outputting nearly half a volt higher than the other two. Fixing this solved the blinking.

Right now, I just need someone to go over my circuit and tell me if I messed anything up. Internet commenters and PSU literature have me stressed out about load balancing, but I'm not sure how much that really matters at such low voltages (5V). The PSU outputs are tuned within 0.01V of each other and Vdroop at full load is something like 0.04V. Plus, right now, the PSUs only have their DC grounds commoned thus aren't really paralleled anyway...which is partially why I am so confused that the mismatch of output voltages caused that blinking issue. But, it did.

Anyway thanks for your help! Would love some feedback on this circuit!
 

Attachments

Last edited:

Thread Starter

JazzMac251

Joined Apr 24, 2019
25
I forgot to mention in my reply above that the strips are 144LED/m SK6812 RGBW and the controllers are SP108E. Thanks!
 
Last edited:

Sensacell

Joined Jun 19, 2012
3,432
Your drawing shows the ground for the controllers far away from the end of the LED strip.
What are the physical distances involved between the stips and controller? between strips?

You want the ground of the controller connected to the ground of the LED strip, RIGHT WHERE IT CONNECTS TO THE STRIP.
Think of it as a star ground concept, the point where the LED strip begins should be the center of the star.

The LEDs need to see a clean digital signal that is referenced to the same voltage at the LED ground, not some wire 10 feet away, which will never be at the same potential.

Since these LED strips use high speed, single ended TTL signalling, where and how you connect the ground is EVERYTHING.
The inductance and resistance of even a few feet of wire with 20 A flowing will totally bugger any concept of ground.
 
Top