Result of the next test.
Disconnecting the signal wire allowed the board to boot.
LEDs on the strip did not turn on at all (as expected).
While the board was powered up, I connected the signal wire and it worked normally.
Conclusion. The signal wire is either creating feedback or drawing from GPIO2 before boot is complete.
Yah, it’s looking more like some kind of “protection” between the ESP and the LEDs is needed.
As I said, the diode trick might work, but levelshifter is the preferred approach.
The abridged version:
Your ESP8266 (or ESP32) processor works with a supply voltage of 3.3V, that’s set by the manufacturer.
That means and data coming out of the chip will have a “Hi” digital output of at most 3.3V.
Addressable LEDs are typically designed for 5V systems. Common mfg specs say the LEDs expect a “Hi” level digital input to be at lease 70% of that 5V power supply or 3.5V.
The levelshifter is a “black box” that watches for a 3.3V input and sends a 5V output to match.
When the input goes Lo (no more than 1.0V) the output goes Lo as well.
This has the effect of massaging the ESP8266 data stream into something that meets the LEDs data specs.
So how come people report that sometimes it “just works” without a level shifter at all?
The answer is, different mfg. have different allowances as to what constitutes a Hi voltage level. So the 3.3V input might work in some senarios, but if you want guaranteed operation use a levelshifter.
As to what your 12V (or even 24V) LEDs do to this discussion?
Nothing at all, the design of higher voltage strips still use the same 5V based LEDs, just in a different layout that lets them handle the higher power supply voltages.
The data line however is always 5V based.
Thanks for the links.
I still don’t understand how they work but it looks like I can figure out how to set one up.
It might take me a bit to incorporate it into my design (as well as secure some SN74AHCT125 chips).
It also dawned on me, during a nap, why my original test strip has been working flawlessly for 5 years. Simply put, it’s a 5v, 3-pin
New design is complete.
I have GPIO1 and GPIO2 going through the level shifter. I have GPIO3 undisturbed and an additional 6 GPIO possibilities on the board I’m designing.
Now the part to decide. How many boards to buy? I can get 3 proto boards for around $12-15 or 100 boards for $75.
I’ve already purchased the IC, resistors and capacitors for 100 boards. I have all the rest I need to get started.
As soon as I get the IC’s I will put it together on a breadboard and see if it works.
There are a few reasons.
And I sometimes get wordy so hold on.
I started designing a board for the NodeMCU a few years ago. Then a lot of life happened and I decided to “finish the project” before I die. (I don’t plan to die soon).
During the life that happened, I lost a job and got a new one. Had a learning curve there. Then I nuked my computer and didn’t have a lot backed up but managed to scavenge a few projects from the HD. One was my LED project here.
I also ended up dealing with the rona. Not getting the rona (which I did but it wasn’t bad) and had the reaction to the shot and ended up with heart failure (It’s not debatable in my case).
Then the company I was working for moved and I had to deal with all that stuff (I was a department manager) and after that they decided to restructure which eliminated my position so I had to get a new job.
This new job is so easy, and so boring, that I have time to think about the projects. It also pays better so I have the finances to spend money on stupid projects that are out dated or might not even work.
So, that’s why.
And if I get something working with the design I have, I will get bored and probably take on the Esp32.
But, also, I’m not getting younger, and I’m pretty sure I lost a step when I went through the heart stuff so my learning curve (time) isn’t unlimited. So who knows if I will do the Esp32 or not.
Also, I don’t know what the Esp32 cost but I’m getting the NodeMCU for under $3 each so it’s not an expensive venture.
P.S. It looks like what I'm doing will be fully compatible with the ESP32 version of the NodeMCU as well as the 8266.
Ah gotcha. Sorry to hear of all your troubles. It just seems like it could be way less of a headache to migrate to the 32 and see if it works vs all the effort you have been putting into the 8266. I’m not saying it will function any differently, but it does open up so many more possibilities with pin choices and processing power and very well may solve your issue (or not).
Meh, everyone has troubles from time to time. Makes for a life experience.
I think I’m just going to place the order for the boards. If needed, I can use the boards without the level shifter and then it would have pins that might be helpful in installing other types of components.
I will still put it all together on a breadboard before the full set of components arrive.
Well, I bit the bullet last night and ordered my new boards. I hope they work.
My design is such that even if they don’t work, as designed, they can still be used.