Please review my ESP based WLED controller board for WS2815 12V LED Strips - New

[ElecGeek]

I'm designing an ESP-based WLED controller board for WS2815 12V LED strips, with a primary focus on energy efficiency.

To minimize power consumption, my goal is to completely shut down the system when the LED strip is turned off.
However, the current buck converter powering the ESP module draws about 8 watts even when the LEDs are idle, making it unsuitable for low-power operation.

I am using below components:
Buck Converter: TPS562201 (12V → 5V regulated)
Microcontroller: XIAO ESP32-C3 (with built-in USB-C)
LED Strip: WS2815 (12V digital LED strip)
Switching MOSFET: Logic-level N-MOSFET (IRLZ44N)
Level Shifter: BSS138 MOSFET-based (3.3V to 5V data line)
Status LED: Green LED

I've attached the schematic—could you please review it and let me know if I've missed anything or if there are areas for improvement?
I designed it with limited experience, so I’d really appreciate your feedback on whether the setup is correct and functional.


Thanks

[PlainName]

Just a note about the schematic...

Usually, power is up and ground is down. Your chips have power down and ground up. The ground wire above the buck converter is confusing. You would do better to reorientate  or redraw the chip symbols so you connect power to the top of them and have ground come out low down.

Also, concerning ground, usually the ground symbol is used wherever a connection is required (within reason). That would remove quite a lot of confusing wire in your schematic and make the design clearer. (But don't conflate this with saying you should use net labels instead of wires - no, ground (and power) is a special case.)

Connectors... one has pin1 for power a t'other has pin1 for ground. Scope for seriously screwing up there. It's a personal thing but generally I always make pin1 ground (on the basis that ground is always connected, and there will always be a pin1). Then, logically, power will be at the other end.

Also on connectors, it would help if you label what the pin function is. Not net labels per se, just what it is (+5V, 0V, data) so one doesn't have to check where connections go to figure out what they are.

[ElecGeek]

Thanks for the suggestions. I have re-wired everything and made some improvements. Please have a look and review once.

I have one query, the R2 for buck convertor was connected to AGND, I have connected it to Ground, I did right?

Also can you please let me know what would be the best way to extend it, so that it can be used with a 5V strip as well?

[PlainName]

While not perfect it's much better to figure out what's going on 

But it ain't going to work. An obvious issue is U7, the dual-supply bus transceiver. According to the schematic, the intention seems to be to have U7 interface between signal levels on the ESP32 and the LED string fed from U8. In fact, U7 is taking it's voltage source from the U8 data signal, so nothing will happen. Instead, U7.8 (VCCB) should connect to whatever voltage the LED string needs for data (presumably +5V) and R8 connect to  U7.7 (1B).

The DIR pin (U7.5) is superfluous for your purposes since data will only go one way: towards the LED string. Connect it to either VCCA or VCCB to keep it pulled how, so data always goes A->B. Then D4 on the ESP32 is your data pin and should connect to U7.2.

I am failing to understand the intention for Q1.

Anyyay, as to your query about R2: there is no AGND. Is that something you saw on the datasheet and figured it means GND? In this case that would likely be correct.

[ElecGeek]

My bad, my understanding was wrong for level shifter I used here. Corrected it out.

Intention of Q1 is to save energy when LED strip is off. So I added N-Channel Mosfet to turn of completely. Am I doing it correctly here?

For R2, I see it in datasheet. Please have a look.



Updated Schematic:

[tooki]

My bad, my understanding was wrong for level shifter I used here. Corrected it out.

Intention of Q1 is to save energy when LED strip is off. So I added N-Channel Mosfet to turn of completely. Am I doing it correctly here?

For R2, I see it in datasheet. Please have a look.
(Attachment Link)


Updated Schematic:
(Attachment Link)

The TPS62203 datasheet is pretty bad (including the totally stupid symbol with the pins in physical order). Take a look at the documentation for the part’s evaluation board; it has a real-world schematic.

[PlainName]

Intention of Q1 is to save energy when LED strip is off. So I added N-Channel Mosfet to turn of completely. Am I doing it correctly here?

That's a low-side switch of ground, which could lead to tears. Take ground out and the 12V may be fed back down the data line into the shifter. Further, if there is any ground around the LED string there's a chance that the still live 12V could short to it.

It would be much better to switch the supply, i.e. 12V.

However, on checking the WS8215 datasheet it seems that the quiescent current is  max 1uA per LED, which is not a lot even if you have a few of them in the string. Of course, a wall panel pretending to be a TV would be a different matter, but then your circuit wouldn't be able to drive even a small corner of that

Also, the minimum high level input (V_IH) is 0.7VDD. For a 12V supply that would be 8.4V, so your 5V shifter likely won't provide sufficient voltage to drive the data line.

[ElecGeek]

I think there is some confusion with WS2815. Please refer to below link. I think 5V data line would work. Mine is WS2815B.

https://github.com/stanleyondrus/LedBoxV2/issues/1

Also, you are suggesting to use P-Channel Mosfet to switch 12V instead of 0V Neutral, which is what I am doing right now?

[PlainName]

OK. The datasheets for these things can be suboptimal

It would be normal to use high-side switching for something like this. P-channel MOSFET would be appropriate.

[ElecGeek]

Thanks for your suggestions. Please find the updated schematic below. Is everything fine now? Any other suggestions/improvements?



Adding PCB as well. Can you please also review it and give me feedback.

[PlainName]

Two things:

1. Please use the forum's tools for attaching images. That way they don't get in the way until one wishes to look at them, and when they do pop up they fit the window. Yours don't fit my window (too big), and I don't want to look at both at the same time not have to sideways scroll to see them.

2. The PMOS needs a pullup. Do a google for how to drive them, or check this link as a for instance:

https://electronics.stackexchange.com/questions/595465/can-you-drive-a-p-mosfet-as-a-high-side-switch-directly-from-a-microcontroller

[ElecGeek]

Here you go. Updated it. Hope everything is fine now. Please provide feedback.

[floobydust]

How does it start up? You need U1 EN +ve to turn it on, yet the ESP32 is powered from that 5V reg - which is off.
Fuse comes first. D2 should be after the fuse.

[ElecGeek]

That is true, I will remove EN altogether as it will not waste too much power. But if I want to drive buck convertor what would be the best way to turn it off completely?

Also is there a way in Easy EDA to run this circuit in simulation, so that I can avoid these type of mistakes?

[floobydust]

I don't see a need to turn off the buck-converter   The WS2815's will draw a lot of quiescent current.
At some point, electronics with modules can be too much for a simulator. I don't know Easy EDA.
Most of us build a prototype to see if it's going to work. Or ask for a review here like you have done, to spot bigger issues.

[tooki]

Intention of Q1 is to save energy when LED strip is off. So I added N-Channel Mosfet to turn of completely. Am I doing it correctly here?

That's a low-side switch of ground, which could lead to tears. Take ground out and the 12V may be fed back down the data line into the shifter. Further, if there is any ground around the LED string there's a chance that the still live 12V could short to it.

It would be much better to switch the supply, i.e. 12V.

However, on checking the WS8215 datasheet it seems that the quiescent current is  max 1uA per LED, which is not a lot even if you have a few of them in the string. Of course, a wall panel pretending to be a TV would be a different matter, but then your circuit wouldn't be able to drive even a small corner of that

Also, the minimum high level input (V_IH) is 0.7VDD. For a 12V supply that would be 8.4V, so your 5V shifter likely won't provide sufficient voltage to drive the data line.

That’s some outdated third-party version of the datasheet. The current datasheet on Worldsemi’s own webpage, attached below to save everyone from having to fill out a form to get to the downloads, is crystal clear on the voltages:

Absolute maximum ratings:
Power supply voltage V_DD: 9.5V … 13.5V
Logical input voltage V_I: -0.5V … 5.7V

Electrical characteristics (for V_DD = 12V):
V_IH: 2.7V … 5.7V
V_IL: -0.3V … 1.5V

[tooki]

I don't see a need to turn off the buck-converter   The WS2815's will draw a lot of quiescent current.

Datasheet says 1 μA, so probably not a concern until into the hundreds of LEDs?

[PlainName]

That’s some outdated third-party version of the datasheet

Er, this was resolved and corrected up-thread. Is there some reason you need to pick over it again several days later?

(At least I provided a link the one I used, to be helpful.)

[tooki]

That’s some outdated third-party version of the datasheet

Er, this was resolved and corrected up-thread.

Huh? Where? The GitHub page linked earlier has screenshots with the same incorrect values.

Is there some reason you need to pick over it again several days later?

Because I was scrolling through my Updated Topics list and it showed up? And it didn’t look like the voltage question had been conclusively answered?

(At least I provided a link the one I used, to be helpful.)

Huh? So did I, I attached the PDF, since I can’t be sure if a link would even work, since it required filling out a form to download it.

Stating the unequivocal values is also helpful to anyone who stumbles upon this thread in the future (for example, from a web search on the topic).

[PlainName]

Huh? Where? The GitHub page linked earlier has screenshots with the same incorrect values.

It shows the erroneous one and then the correct one (for the B).

Huh? So did I, I attached the PDF

Sorry, I missed that. Too used to checking in-text links, I guess 

[tooki]

Huh? So did I, I attached the PDF

Sorry, I missed that. Too used to checking in-text links, I guess 

I even said “attached below” in the text!

Huh? Where? The GitHub page linked earlier has screenshots with the same incorrect values.

It shows the erroneous one and then the correct one (for the B).

And just as you overlooked my attachment, I overlooked the last screenshot in the GitHub thread!! 

However, the thread does not include the voltage logic thresholds table at all, and the last screenshot there does not agree with the current datasheet (the one I attached).

[floobydust]

I don't see a need to turn off the buck-converter   The WS2815's will draw a lot of quiescent current.

Datasheet says 1 μA, so probably not a concern until into the hundreds of LEDs?

What I've seen on the datasheets, Input Current is +/-1uA (I take to mean D_IN) but quiescent current (all LEDs off) is 2.1mA or say 210mA for a string of 100 LED's.
I think you would use a switch to disconnect power from them, if OP's board is battery powered. It's high parasite drain compared to the buck-converter.
There are also variants WS2815B, C.

[tooki]

I don't see a need to turn off the buck-converter   The WS2815's will draw a lot of quiescent current.

Datasheet says 1 μA, so probably not a concern until into the hundreds of LEDs?

What I've seen on the datasheets, Input Current is +/-1uA (I take to mean D_IN) but quiescent current (all LEDs off) is 2.1mA or say 210mA for a string of 100 LED's.
I think you would use a switch to disconnect power from them, if OP's board is battery powered. It's high parasite drain compared to the buck-converter.
There are also variants WS2815B, C.

Oh, I think you’re right about the input current meaning the data pin. But in that case I don’t see quiescent current anywhere on the datasheet.

[floobydust]

It's kind of a mess with the Worldsemi Co. datasheets. Even timing, LED current has moved around.
WS2815B-V1_V2.2 seems to be the latest and says "Quiescent Current＜2mA".
Older WS2815 V1.1 gives 2.1mA
WS2815B-V1 "added power consumption" at V2.1 {which appears to be the datasheet version}

[tooki]

It's kind of a mess with the Worldsemi Co. datasheets. Even timing, LED current has moved around.
WS2815B-V1_V2.2 seems to be the latest and says "Quiescent Current＜2mA".
Older WS2815 V1.1 gives 2.1mA
WS2815B-V1 "added power consumption" at V2.1 {which appears to be the datasheet version}

Ok, I finally did a text search in the datasheet for “quiescent” and found it as a header in the table of the LED emitter characteristics (where the light output and wavelengths are listed)… wtf? (I mean, what is the “quiescent current” of an LED? Not the control chip, but the LED itself…)