I have 3 Pixelblaze (up to date firwmare) all connected as clients to my wifi router (1 floor lamp and 2 handfans). Maximum distance is 5 meters (my living room) without anything blocking. I assume my wifi is pretty stable as I never encounter any issues with other devices.
2 Pixelblaze are “followers” to the “main” one which I drive from the browser of my computer.
I have very erratic behavior on these devices:
- Sometimes impossible to connect even just after starting electrically the lamp. The “loading” rotating cube shows up, but rotates forever or does eventually connect after too long time (20 seconds or more).
- Sometimes the master Pixelblaze changes pattern, but the “followers” don’t, like they are not connected anymore to the the master. If i open the settings, it shows the “leader” as offline even though it’s up, running and controllable.
Both problems can be solved by restarting everything, but this is not an acceptable solution for what I want to do.
I’m working on pieces I aim to commercialize to people who are not “makers”, and connecting to the Pixelblaze to change settings is not reliable as is and I don’t understand where is the problem…
Is there a checklist somewhere to investigate why this is so hard to connect to it ?
- Might be the heat but i don’t feel it’s too hot when i touch the ESP32, and it can happen even when just starting cold
- Power supplies are clean and deliver 5V
After a few additional tests, i noticed that the device struggling to stay “in sync” to the “master” was “added as follower” from the master, not configured “as follower” from its own page.
I think there is a glitch/bug with this function to add/remove followers directly from the leader. Will continue to do tests.
Hi @hololit ,
I have seen a few cases were sync can get weird and a power cycle fixes it. Some of the changes in the next update might help.
It does sound like connectivity troubles that might be caused by any number of things. The physical install of pixelblaze is the biggest. If it’s buried the signal can get blocked. Metal can cause all kinds of problems but some plastics will too.
Power noise can also impact things. Would need an oscilloscope to see that. It’s not going to show up on a multimeter. Sometimes a power supply can still output a lot of current and keep around 5v but have high frequency noise that can interfere with WiFi.
The configuration is stored the same way no matter the method of configuring leader/follower. Doing it from the follower just means you have to load the app, and the websocket will load patterns and previews, but the end result is the same. If you have connectivity problems the websocket might get hung up through either method. That can take a while to clear on its own or a power cycle would fix it.
Can you show some photos of how PB is installed? Have access to a scope?
The electronics are buried in the metallic base of the lamp. It might indeed shield against the wifi but I think it’s unlikely because my wifi is only 3-4 meters away without other obstacle. I’ll do a counter test by setting one of handfans (no obstacle) as master to see if i have sync issues with the other fan.
However you might be right about the power supply. I don’t have access to a scope, do you think i can mitigate this issue (if it comes from there of course) with a capacitor (what size ?) ?
Here is below the power supply and cabling (view from the bottom). I tried to put the PB as far as possible from the power supply.
That is definitely going to drop your signal strength and bandwidth considerably. The only way it’s getting signal out is down through the floor.
You can run a WiFi scan on the WiFi tab and look at the rssi. Higher numbers are worse.
I got -49 and -76 depending of room placement. How high is bad ? Is there a recommended range ?
76 isnt great. 30 or lower is excellent, 80+ is very poor.
Ok thank you this is going to help for future products but i’m a bit concerned about this one.
Any way I can improve ?
- Can we even plug an external antenna to the PB ?
- holes in the metal base would help ?
- put the PB in an external case (but then he would it next to the power supply…)