Posts made by Devilstower

I've tried going to 85. Also changed the minimum to 1. It didn't make a difference. No idea what's going on. One of the sticks gets "streaky" and the screen shows some kind of jumbled pixels before completely locking up, if that provides any sort of clue.

That same code locks up EVERY stick I have. Both M5StickC and M5StickC+. All burned to 1.8.7.

If I take the Screen Brightness calls out, it will run fine. Put the screen brightness calls back in, and I can run it 3-4 times before everything comes to a halt.

You really don't need much to demonstrate this issue.

Create a function called "ScreenOff" that contains just the command "Set Screen Brightness 0". Create a second function called "ScreenOn" that contains the command "Set Screen Brightness 100." Try calling them a few times.

I've literally reduced the code to that, where the functions alternate on button A press. It locks up within the first ten calls.

By adding LED On before the screen brightness call, and adding LED Off immediately after, I can see that the LED goes on and stays on when the stick locks up, indicating that it's failing directly on the Set Screen Brightness call. Once the stick is locked, it has to be reset with both side buttons before new code can be loaded.

In my app, I want the screen to go off after three seconds, then turn back on for another three seconds when someone presses the button. I accomplished this by using "Set Screen Brightness 0" and "Set Screen Brightness 100" at the appropriate points in the code.

However, if I cycle through this code more than a few times, either the M5StickC or M5StickC+ lockup with screen off. Using the LED for testing, I can tell that they freeze exactly on the "Set Screen Brightness 100" line. There is also erratic behavior in other parts of the code when these lines are implemented.

Take just those two lines of code out, and leave everything else in place, and the program runs just as expected, even when tested over hours. It's definitely this one command that's causing issues.

Is this happening again? Everything was going along well, but now ... even though my device shows as connected to the Cloud, and even though the API key checks out fine, every attempt to download code gives me the "failed" message. It was working earlier, but now I've gone through it again, and again, and again. Burned the latest version to the device. Tried both desktop and online versions. Can't connect.

I'm a bee keeper in the U.S. in an area that has cold winters, hot summers, and problems with pests all year round. So I'm trying to build a "smart beehive" that keeps track of a series of items:

Minimum feature set

• Temperature and humidity outside the hive
• Temperature and humidity inside the hive
• Orientation of the lid
• battery / solar panel
• Option to connect to local WiFi

Better feature set

• Volume / pitch from the microphone
• Weight of the hive box
• Camera + red light LED strip

I have a version of this up and running already using an ESP32 (a couple of different versions, but the latest is an Adafruit S2 "feather"), a triple-axis accelerometer, OLED screen, 2 DHT11 temp/humidity/pressure sensors, a little MEMS microphone, and an ArduCAM Mini. It works, but it's far from ideal. Battery management is messy. The camera doesn't do well (i.e. barely registers anything) in dim light, and mostly the whole thing is a whopping mess of wires and breadboards that lasts about a day before the bees turn it all into a goop-filled mess.

The M5StickC and it's ecosystem looked like a way I could neaten all this up and, along with some 3D printed covers and a hot glue gun, give myself a system that would stand up to at least a few months of bee-monitoring.

What I had imagined was a system that looked like this:

• M5StickC
• M5STICKC-18650 battery cap
• PbHUB
• ENV II sensor (x2)
• MEMS Microphone
• Mini Weight Unit (x2) + load cells

On top of the hive, I placed a solar panel that feeds into the 18650, then connected the hub to the M5StickC. I can connect then two ENV II units to the hub and ... then I'm suck.

Here are my current problems:

1. I can't figure out how to physically connect the Weight sensors without removing the battery hat, which ... doesn't work for me.
2. I had hoped to use UI Flow rather than the C I was using on the existing system, but there are two immediate problems:
   2A. I can't see a way to bring the system up in WiFi_Sta ("host mode") so that I can either offer to detect available WiFi networks or run disconnected (a critical feature).
   2B. There doesn't seem to be any support for any form of microphone in UIFlow.
3. I'd really like to connect a camera and LED strip to provide lighting, but this seems to introduce problems both in the "how do I connect it" and the "how do I control it" categories.

I can always solve the software problems by going back to C. But that's a shame, because UIFlow is so easy to follow and looked like it would be nicer to maintain. But that doesn't solve my hardware connection problems.

Any tips on how to make progress appreciated.

(NOTE: I've started this a couple of times, but can't find my previous posts. Unsure if they've been deleted, or if the search just isn't bringing them up. So ... apologies if you've seen this before.)

And sorry. I meant to put this topic in “projects” but somehow hit the wrong spot. That’s what happens when me and my giant fingers try to do something on a phone.

I’ve previously put flown a couple of high altitude weather balloons with a collection of instruments and cameras. In the past, I’ve used a variety of arguing-compatible controllers and sensors to collect GPS position, speed, humidity, barometric pressure, temperature, etc. I’ve also included 3 or more el-cheap action cameras (Go Pro clones) which I’ve activated and controlled either directly, or in one case, using servos.

I’d like to recreate this as a “kit” of M5 components that I can package with a small styrofoam container, keeping the total weight down so I can include the smallest practical balloon.

So, I’m thinking a parts list that would include something like: M5 Stick C, 2 ENV III sensors (one for inside, one for outside of capsule), Mini GPS/BDS, and a fisheye camera module.

But I’m having trouble with the most basic thing: how to hook them together. I had assumed I could use a hub, but haven’t been able to get even the ENV + GPS working… and no clue how to integrate the camera. Any advice to get this “off the ground?”

After all this time, and all the other units added, is there still not even basic support for the MIC unit?

@rob_d I've also had no luck with the PbHUB, even when using a single unit. Switching to a PaHUB things worked as expected.

In UIFlow, if I take an ENV sensor and hook it up to a port on the PbHUB, the program just gives me an error saying the unit wasn't found (and yes, I indicated the sensor was on the hub and selected to correct port). However, take the same unit and hook it up directly, or through a PaHUB, and everything is fine. I understand the PbHUB isn't suitable for using multiple sensors of the same type, but this is happening even with just a single sensor attached to the hub.

Looking at the latest set of units available in UIFlow, there seems to be good representation, except that the MIC unit seems to be completely missing. Considering that other more complex units are covered, is this just an oversight? Is there a reason the MIC unit is excluded from support? I picked up a stack of them for projects, but using this unit seems to immediately move me out of using UIFlow.

Thanks.

Thanks. I was hoping I'd missed some way to do this in UIFlow, as it is fun to knock up some code that way. But ... back to C for me.

It's disappointing to see no replies to this. It's such a central feature of the whole ESP32 family, that not seeing it supported in UIFlow and other examples here makes me seriously concerned if the M5Stack products are somehow more limited than the vanilla ESP32.

I’ve previously coded up ESP32 devices using C that boot the first time in AP mode, allow the user to connect and provide network information, then reboot in station mode. However, there don’t seem to be any examples in either Blockly or Python for the M5StickC.

This is pretty vital behavior when prototyping a consumer product. Does anyone have an example? Thanks.

I'm working on a project that involves monitoring bees inside a hive. Bees can't see red light, so it's great to use red LEDs to light up the inside of the hive. It doesn't make things too hot, and it doesn't disturb the bees -- it may even help cut down on disease.

However, all the cameras seem to have a red filter that makes it almost impossible to see anything, even when there are a lot of red LEDs. Is there any way to make the cameras work better in red light? Some what to remove the filter, or some version of the camera that works better in red light than others?

Thanks.

I created a number of programs about a month ago, some of which got quite complex. At the time the files (created between 5/30 and 6/5) saved and loaded without an issue, and after developing a prototype of an app, I ordered a dozen more units so we could put some some trials in the field.

But not none of these programs will load completely. When I try to load them, I get just a small part of the original program, sometimes part of the main loop, and even that is missing many commands. These programs were created using 1.5.4, but it doesn't seem to matter which version I use in attempting to load them.

I had put this StickC aside to work with another (I have a dozen or so of them). And then I ran into the same thing about ten StickC's later.

With both of these, it seems to be a screen failure. I can download a program, get the LED to flash, and even connect via network. But the screen remains dead -- even if I just run the burner and start the device over from scratch. I've also tried shorting BAT to GND, just in case, and get the same results. Another 8 StickC's working dandy (I was making a batch of prototype devices) but these two appear to have simply lost their screens.

So it's not really dead. It's just gone dark.

Yes, it sounds like this is what will be needed.

It would be great if UIFlow would just roll up the whole sequence of switching to AP mode, collecting the info, and reconnecting. It's such a standard behavior for ESP-based devices, that it seems like it should be baked-in.

I want to be able to take it anywhere and detect the local networks. The initial setup of the StickC allows this, and I've set up a similar process on an ESP8266, but but I don't see any way to trigger that kind of sequence in UIFlow.

Since arriving a couple of weeks ago, the M5StickC has worked like a champ. Together with the ENV and weight sensors, it's proven to be perfect for a project that was in the works and (except for some issues around swapping networks) has performed like a champ. I've had the StickC connected to a small (2600mah) power bank which was in turn connected with a small (2w) solar panel. The combination had it running for days, recording data and sending it along to Thingspeak.

But last night I disconnected the StickC to make some code changes. It tested out just fine, but when I went to power it up this morning ... it wouldn't start. Not when connected to the battery, not when connected to the PC, not when connected to a USB power source. I've tried all the familiar long press and left+right options that have helped with past issues, but the screen is stubbornly dark.

Is there something I'm overlooking? Anything else I can try to revive a StickC that seems to be ... Sick-C?