New open source controller in the works, what features would you like most?

[DrewT]

Smart phones

But smart phones are not plexiglass covered touch screens... most are capacitive and bonded to a very thin transparent contact surface correct? I'm confident I can diy a piece of plexiglass over a display... I'm not confident I can diy smart phone screens. I guess I need to physically prototype something to figure out if it's feasible.

 

[Latitude17]

@DrewT how's your project coming along?

 

[DrewT]

Hey @Latitude17, thanks for checking in! I'm currently working to wrap a REACT ux over a gutted version of tasmota (deployed to ESP-12 and ESP-32 microcontroller packages). This will serve as the base processing unit for a distributed network of horticulture tools. Once this portion of the project is complete I will immediately begin work on the automated nutrient doser. This auto-doser will be compatible with a wide variety of ph/ec pens and support PH control as well as powder or liquid fertilizer dosing. As with all my work, it will be open source. You will be able to DIY this yourself if you have access to a CNC mill and vacuum former. Prototyping through spring and summer, published work near the end of 2019.

 

[blep48]

Greetings fellow DIY'ers! I've just finished building an open source multi-purpose controller. Although I learned a lot, the resulting device is more proof-of-concept then finished solution.

Very cool! I've been tinkering with MCUs and electronics for a very long time. I've tinkered with everything from a crappy BASIC Stamp to PICs and AVRs to ARMs, programming them in assembly, C and various C-like languages.

My latest toys are a PyBoard (clone) and an Adafruit Trinket M0. MicroPython is a nice change from C, though it isn't terribly fast and interrupts aren't well supported (not at all on the Trinket).

Your post has inspired me to do some projects in the grow room. What fun!

Though I'm not a big Arduino fan, I own a bunch of various Arduinos. Good platform for grow room projects - dirt cheap and easy to use.

 

[DrewT]

Though I'm not a big Arduino fan, I own a bunch of various Arduinos. Good platform for grow room projects - dirt cheap and easy to use.

Welcome to the forums, thanks for dropping in! In addition to arduinos, consider some of the newer MCU's that come with BLE and WiFi. Specifically, the Espressif esp12 and esp32 variants. They are much faster, have much more memory, and with the exception of GPIO count are preferable in almost every way (imo). I'm currently working on a core controller framework that uses HTTP REST as a back-end and React on the front end. It's responsive and will scale well as new features are added. Here's shots of the AP setup/access and wireless configuration views.

 

[Aroxol]

A big problem I see in this arena (across multiple forums) is a ton of people are doing really cool stuff like all of you, but not the same stuff against the same design or code. Are you guys too deep into gear/ source code to swap your parts and work off the same design?

Edit: If that would be of interest.

 

[bobinca]

A big problem I see in this arena (across multiple forums) is a ton of people are doing really cool stuff like all of you, but not the same stuff against the same design or code. Are you guys too deep into gear/ source code to swap your parts and work off the same design?

Edit: If that would be of interest.

It is hard to write code that will be everything to everyone. In general we often have different goals too. And if we have random hardware lying around, we will just use what we have.

 

[DrewT]

@Growca2268 indeed what you describe would be ideal. In practice, my experience is that most of us are working these projects for different reasons. Some want to simply solve a problem quickly, other enjoy tinkering as a hobby, while others want to build feature rich systems that are overly complicated for most practical applications. There is no wrong way to dev these things, its just that too few of us have the same goals in mind. If you want a deeper read on where I'm coming from, just take a look at this thread. There is a solid group of really smart folks on this board and their work inspires me for sure. Some are focusing on task/env specific projects that require quite a bit of knowledge to assemble. I'm shooting for something a bit more platform-ish, something that most of us could put to use with straight-forward instruction or guidance. My code is all open, would love to work with others!

 

[odam2k]

Hey @DrewT good to see you back active again You must have been hard at work...

 

[DrewT]

Heya @odam2k ! Yep been busy with the controller framework. Hoping to have the UX wireframes done for GPIO assignment and the task scheduler. When that's done I'll have something to submit for feedback! I'll also get around to catching up on your journal, it keeps on growing! Hope you've been feeling better and have had time to work on you projects.

 

[Dro]

What's going on fellow tinkerers, I have been following most of you on your projects around the forums here and I love the different approaches we all have to accomplish similar tasks.

So I am currently in the same boat as some of you. I am running a DWC set up and my goal has always been to incorporate sensors, controls, etc to make my growing experience easier but not necessarily automate growing. Basically I like to be lazy and enjoy doing other things besides looking after my plants every single day (although I enjoy checking on them daily lol)
So I am currently debating on having a 5 or 7 inch touch screen display outside the tent to show water temp, tent humidity, tent temperature, power consumption, light status, and pH. I can currently do this on my phone through the adafruit IO framework so adding a $80-$100 touch screen LCD doesn't make sense but it does make it convenient.
I thought about using an old tablet and just have it in kiosk mode showing the same data. The problem with this approach is that I need to rely on internet connection or just set up a local server.
So this is where I have hit a design wall... The PCB design would surround the enclosure that I will use and if I use a screen I would have to match that size or use a daughter board approach.
I am currently leaning on just prototyping using some wood, particle board or a bigger enclosure, see if the screen is needed after 1 grow. Anyway, just wondering what guys thought.

 

[odam2k]

Anyway, just wondering what guys thought.

I thought the same thing, replace my phone with a tablet... but my phone is handier... I am using Blynk, with a local server (extremely easy to set up!) and if I want, I could expose it to the internet, but I haven't yet. I have a spare phone I will probably just leave in the grow room on charge, so if I am in there without my phone, I can just grab it. I am using Blynk not only to display data, but also to interact with the system such as calibrating sensors...

If money was not an object, I'd certainly go for a touch screen

 

[bobinca]

@Dro I spent quiet a bit of time coding for my 4 line lcd. After I discovered running blynk on a local server I ripped out support for the lcd and only use blynk now. I check my stuff remotely (open vpn from android then blynk can connect to local server). So I didn't see the need to support more than one interface.

 

[odam2k]

@Dro I spent quiet a bit of time coding for my 4 line lcd. After I discovered running blynk on a local server I ripped out support for the lcd and only use blynk now. I check my stuff remotely (open vpn from android then blynk can connect to local server). So I didn't see the need to support more than one interface.

That's funny, I did exactly the same thing!

 

[DrewT]

Add me to the list of folks who've retired their local displays (LCD/TFT). I'm not ashamed to admit I wrote an entire lightweight MVC framework in C++. While I do see a practical argument for local displays, the cost and complexity for almost any scaled application is difficult to justify imo. I second Odam's and others input. The phone, if you have it, is the way to go for local device management. If you have an integrated setup that aggregates many devices (OpenHab, Hass, NodeRed), consider just remoting into the host with a standard laptop or tab - no custom PCBs or TFTs required. @Dro , looking forward to seeing what you go with!

 

[Dro]

Those are my exact thoughts but not sure why I keep thinking I will need a screen.

I am going to have to prototype something soon as having bare wires on a breadboard for most of humid vegging it's causing some oxidation and false readings. I am probably going to move the controller out of the tent and just run the sensors in for now.

Thanks guys! I needed to talk my self through it with your lessons learned.

Current set up

 

[DrewT]

I am probably going to move the controller out of the tent and just run the sensors in for now.

Yep, been there too! My solution was to run the sensors remote, keeping as much of the electronics in dryer/warmer places. After a few iterations, I stuck with this design. It minimizes sensor exposure to condensation and pots the connectors behind a water tight grommet. I went with smaller HTU21D that communicate over i2c. Enclosures are milled from 1/2" HDPE.

 

[Potastica]

Very cool @GR33NT0Y! Please do let me know if want to sync on code bits. I've published a recent project that has many of the same features that you're building.



Oh yeah, that looks cool as heck! I've been debating the local display (lcd/tft) but just haven't been able to justify the expense and added complexity. I figure most of us have great input interfaces already, in the form of cellular phones we use every day. After pricing out the cost of adding displays to a component module I keep coming back to using my phone as the interface and serving module management features over HTTP and MQTT. What do you like most about your display? Would you prefer using it over using a phone display to manage the device?

Thanks for sharing!

Just finally got a few things posted on mine...Link
I think I paid $7-8 per screen...bought 2...SD card+screen w/touch @ 2.8"

 

[Latitude17]

Espressif esp12 and esp32 variants. They are much faster, have much more memory, and with the exception of GPIO count are preferable in almost every way (imo)

#truth!

The AM2315 is my personal favorite for an i2C temp/humidity probe. Me and my 3d printer aren't speaking to each other right now, so I told her I'm just going to buy a retail-ready packaged probe (most places sell for < $30 USD, but it's on the pricey side even at that). Accuracy, resolution, and drift were my 3 key determining factors - this had all of them for me.

BTW, @Dro I've purchased my fair share of displays, I can tell you that they're more pain than pleasure when all is said and done. Especially for hobbyists.