I’ve installed 572 LEDs on 1 channel (with WS2815 ribbon LEDs) and I’m having slowness problems on an ESP32 (more precisely, a WT32-ETH01). The first 100 LEDs are fine, but they deteriorate rapidly until the last few are very slow (they update every second or half-second).
The WLED is in its default configuration, I’ve only deactivated the brightness limitation and configured the number of LEDs.
The LEDs are connected directly to the 12v power supply with multiple injection points.
The WT32-ETH01 is powered in 5v via a 12v to 5v converter. The LEDs are connected to pin 2 of the WT32-ETH01. The power supply, the converter and the WT32-ETH01 are in a box, there is 10m of cable between this box and the LEDs.
Could this slowdown be due to the length of the cable?
Some people use a module to increase the voltage of the data, can this help?
Have you measured your injection points to see what voltages you get down the length of the strip at 100% White?
What GPIO pin are you using to drive this?
Might be worth trying with a different pin.
Have you tried a segment or ledmap that drives the end of the strip first and then the first LEDs last?
You could even try a simple Reverse and/or mirror on a Sweep or Wipe effect.
Possibly could use a levelshifter depending on the distance from the ESP to the strip.
Lastly, you could try to split the strip onto 2 (or more) GPIOs.
I did some tests as advised. I put them in reverse mode but it’s the same. I tested the voltage and apparently I have an injection problem but I don’t think that’s what’s causing the slowdowns.
I tested with 2 other whole led strips, so 600 led (because the 572 are on my gutters) and another esp, with the same power supply (1200W) and I have problems. The leds only light up when I only have one ribbon to connect and if I only configure 100 leds. With another less powerful power supply (120W), I don’t have any more problems and I’ve got the 600 functional LEDs.
As a result, I suspect there’s a problem with the power supply, but I can’t really understand why or what it is, as it supplies 12v.
I still have to test the power supply on the 572 LEDs to check if that’s the problem.
So I tested my 572 LEDs with the 120W power supply and the other ESP which worked on my 600 LEDs, but I’m still having slowdowns.
So to sum up:
It is not related to injections (because on the 600 leds, I had none and no slowdowns)
It’s not a problem with the gpio, the esp or the configuration of the leds on wled because the same problem exists with another one.
It may not be related to the power supply, I think I had problems during the tests because of a problem with the ground.
I think the only remaining option is the data level, which isn’t powerful enough because of the 10 metres of cable between the power supply and the ribbon. I’ve ordered CYT1076 modules to convert the 3.3v to 5v and we’ll see what happens.
Those are I2C based shifters and are typically too slow to work well with addressable LEDs
Take a look at the suggested shifters in the KB: Compatible Hardware
10m is a rather long data run from the ESP32 to the strip (even with a levelshifter)
You might want to check out the differential pair section in the KB: Long Data Lines
BTW - if you go the TxRx pair route, you don’t need the data shifter, that’s handled automatically.
At the moment, they’re connected by a very short cable and my 10m cable is behind it. The aim is already to test the modules.
Some of the first leds light up in white whereas I have a colour-changing effect (pride 2015).
I have the feeling that the data sent isn’t good and that the leds are doing what they want.
Anything I should know about these modules?
I followed this diagram (A+ on A+, B- on B-, Tx of one module on esp, Rx of the other module on strip led, VCC on 5V, GND on ground, Earthing on earthing)
Those modules are very simple to use.
Vcc should be 5V on both ends, TXD goes to your GPIO at the MCU end and RXD goes to data in on the strip side.
As in your diagram, Tx A+ to RX A+, and Tx B- to Rx B-.
The only other potential gotcha is the termination resistor.
Labelled R9 on the board besides the A+/B- terminals, its a 120R resistor that should be across the A+/B- for both boards. By default it’s not active, you have to short the 2 pins labelled R0 on each board.
I reproduced the experiment with another strip led without the 10m of cable and it works, so the modules are ok. I did the same thing again with 10m of cable and the LEDs behave as they do on the gutter. So all I have to do now is test the modules on either side of the 10m of cable, but I think I’m nearing the end.
Thank you for your help! I tell you quickly if it’s ok or not.
When I use the modules with my 572 leds, I still have the problem. So I tested it directly (without the 10m of cable) and it was the same. The same setup on 2 rolls of 300 leds (so with 600), I have no problem.
My 572 leds are mounted on aluminum bars. Could this interfere with the data?
These bars are also fitted with 12v and neutral. I understand that the neutral and the datas side by side over long distances can disturb the datas too. What do you think?
If you’re using WS2815, what are you doing with the backup data wire? It should be connected to the ground as close to the first LED as possible, from what I’ve researched. The further you get from the first LED with the backup data line, the more problems you will get.
DI and BI are connected together. I’ve just seen that it is necessary to connect BI to the ground as this can cause interference. Thanks for the info because I didn’t know.
I did the test with my 2 rolls and I have no problem connecting them this way (DI and BI together) but I realize that I have wired my 572 leds wrong because I connect DI and BI in 5 different places. So that’s certainly the cause. I find a way to do this test and I tell you that.
