I am planning an installation of several thousand 3V LED seed pixels (1.6cm pitch, several tens of meters) on a bipolarity 2-wire string. Because of power considerations, I am thinking of splitting the long LED string into segments plugged into daisy-chained amplifier modules - the LEDs of each polarity all need to be synced, receiving the same signal. I would like to ask for advice on designing the amplifier PCBs.

[Bipolarity strings meaning: LEDs on 2-wire strings with alternating polarity. The polarity is switched at several KHz so that both polarities of LEDs can be lit. Further, each direction is a PWM signal which is frequently varied in order to generate patterns of varying brightness on the LEDs. Controllers are cheaply available commercially, although I am working on my own version using an MCU]

Each amplifier takes the incoming signal, amplifies it and splits the output it between an attached LED string segment and downstream to the next amplifier. For the amplifiers, I am thinking of using an H-Bridge like the DRV8876. which seems to have enough power and fast enough switching. I am thinking of using DC 5V power, relying on PWM to keep the LEDs from frying.

Does this setup sound feasible? Is there any advice for signal conditioning, splitting the H-Bridge output, etc.? TIA!

 

Just to close this off, I managed this simply chaining the amplifier input to output, and using an h-bridge to feed the pixel LED string attached to each amplifier. Works like a charm, although this is only a meter at most between amplifiers.

 

I can't answer your question. However, the mental picture developing in my mind is LED's powered via PWM. At the same time polarity is being reversed to light certain LED's while holding others off. I don't like being the one to find problems with the solution but in this case what I'm seeing is a strobing effect. With PWM and polarity reversal at some frequency there will be times when everything will sync up and work and then will go out of phase and not work. Then back in phase and back out - repeating continuously. It's like tuning a piano. A single note has 2 or 3 strings that need to be tuned to the same frequency. If one of the strings is not matching the other the resulting sound will be a Wah Wah sound where the strings harmonize then cancel each other out. I see your lights doing the same thing. It's like having two speakers working on very close frequencies. One at 600Hz and the other working at 602Hz. There will be times when they work together and times when they work against each other.

Perhaps a more simple explanation of what you're trying to achieve may be able to shed more light on a solution to your goal. As for MY skills at this - I'm not the one to be asking. I know too little about such things. I'm a very basic electronics guy. VERY basic.

 

I can't answer your question. However, the mental picture developing in my mind is LED's powered via PWM. At the same time polarity is being reversed to light certain LED's while holding others off. I don't like being the one to find problems with the solution but in this case what I'm seeing is a strobing effect. With PWM and polarity reversal at some frequency there will be times when everything will sync up and work and then will go out of phase and not work. Then back in phase and back out - repeating continuously. It's like tuning a piano. A single note has 2 or 3 strings that need to be tuned to the same frequency. If one of the strings is not matching the other the resulting sound will be a Wah Wah sound where the strings harmonize then cancel each other out. I see your lights doing the same thing. It's like having two speakers working on very close frequencies. One at 600Hz and the other working at 602Hz. There will be times when they work together and times when they work against each other.

Perhaps a more simple explanation of what you're trying to achieve may be able to shed more light on a solution to your goal. As for MY skills at this - I'm not the one to be asking. I know too little about such things. I'm a very basic electronics guy. VERY basic.

Four different colors and an old smartphone camera may account for that. The commercial controller that I am using currently outputs at about 6Hz (which should be visible, but not to me), and the H-bridge switches are much faster than the human eye.

 

If the 2-wire PWM signal is the same PWM for both polarities, it will not be possible for the opposite polarities to have different frequencies. THE puzzle, though, is that the LEDs are not damaged by the reverse breakdown, Is each LED junction protected by a reversed silicon diode? or is the actual mechanism more complex than that?? Or is it a dark secret hidden from some of us???And as for controlling separate segments over a single pair of wires, the drivers will need to be not amplifiers, but separately addressed controllers, and more probably the basic pair will need to be AC with much higher frequency digital commands also present. Not new technology, but certainly technology still in use.

 

It isn't the same PWM for both polarities. The signal can be thought of as frames with a 6Hz rate, where the first half of the frame is for the one polarity and the second half is for the other polarity. Within each half-frame, the duty-cycle can be set independently. I've emulated this with suitably toggling two pins on an MCU (and driving an H-Bridge to boost the power).

The LEDs come as strings, apparently using 0603 chips, but I don't know whether there are additional components. When I ordered some strings with custom pitch and colors, I didn't get the impression that there are any additional bits. One type of string used a color I requested for which I was given the LED datasheet; IIRC the reverse voltage was either 4.5 or 5V.
In any case I am taking them as fit-for-purpose as-is. And they work.

 

OK, so it is not a constant polarity system, but rather a much more complex stream providing a rather complex combination. In fact, it may not really be PWM control at all, but rather a pulse width modulated string of digital commends that just looks like the standard PWM power control wave form.
Quite brilliant, streaming both power and data over the same pair of wires. It also means that each control will have an address decoder and a command decoder function as well as some sort of driver circuit. Not so simple as hoped.

 

OK, so it is not a constant polarity system, but rather a much more complex stream providing a rather complex combination. In fact, it may not really be PWM control at all, but rather a pulse width modulated string of digital commends that just looks like the standard PWM power control wave form.
Quite brilliant, streaming both power and data over the same pair of wires. It also means that each control will have an address decoder and a command decoder function as well as some sort of driver circuit. Not so simple as hoped.

I don't follow what you are thinking about. Those fairy light controllers that come with the blinking lights for a couple of dollars are doing exactly this. I have used them with a pair of standard 5mm LEDs as lighting inside small Christmas ornaments, and as I mentioned you can emulate those controllers with a 8-bit MCU. Or in my case I use an ESP8286 ESP-M2 module, which allows me to control the lights wirelessly using ArtNet DMX.

 

I was thinking that each group was controlled separately, which would mean they had to be individually addressed. Suddenly that requires a more complex communication scheme. Entirely possible, but more complex.

 

I was thinking that each group was controlled separately, which would mean they had to be individually addressed. Suddenly that requires a more complex communication scheme. Entirely possible, but more complex.

They are not as flexible as pixel LED strings, but they are about 1/10th the cost, and easier to control.

 

OK, a misunderstanding on my part.
The BIG challenge is power. An amplifier always requires it's own power, separate from the signal being amplified. That would add two more wires to the string, but the added wires would just be providing the DC power for the packages controlling each segment. The concept of an amplifier supplying one set of LEDs and also sending power down stream would require an efficiency level not possible. So the additional power will need to be provided by the second pair of wires.
The good news is that the "H bridge" driver concept seems reasonable.

 

OK, a misunderstanding on my part.
The BIG challenge is power. An amplifier always requires it's own power, separate from the signal being amplified. That would add two more wires to the string, but the added wires would just be providing the DC power for the packages controlling each segment. The concept of an amplifier supplying one set of LEDs and also sending power down stream would require an efficiency level not possible. So the additional power will need to be provided by the second pair of wires.
The good news is that the "H bridge" driver concept seems reasonable.

These things also have very low power requirements. The repeaters I made are more isolation and learning experience than necessary for power reqts.

 

It does make sense to have separate power supply wires because then they do not need to be concerned with the signal integrity. Considering the distance the string may extend, signal integrity at the far end might become an issue. A separate pair of power wires could support an actual AMPLIFIER part way thru the string..