Invitation/project suggestion: An automated potted plant

I’ve uploaded a new version of the pot, but I haven’t made a final test-print of it yet. Will do one of these days.

@konig I started a similar project at the end of May this year because I wanted to use Flaura with HA. I haven’t modified the design of the pot as much as you have yet. I just made some changes to fit a TTGO T-Energy (ESP32) and a 5x7 double-sided prototype PCB. My esphome code/config is almost working (having issues with the pump not turning off before deep sleep around 20% of the time).
After seeing this thread for the first time today I decided to publish my progress in case you could use some of it. You can find it here: GitHub - mathgaming/FlauraHomeAssistant

I agree that it takes too much time to assemble the pot with screws and silicone to make it watertight. I don’t like your approach of making it one big part because it makes the print take too long (I want to be home and awake while printing so I can stop it in case of a failure) and the current version requires a lot of support (The lastest version I have seen is “flaura 2.0 no holes in bottom v5”). I am considering making two “bowls” (one for the soil and a larger one for the reservoir) and a foot for the electronics and stacking them on top of each other. I will post an image of it when I have made a draft, so you easier can visualize my idea

Here is the draft: Fusion

Hey guys.

I’ve dropped my work-in-progress ESPHome config for my potted plant irrigation project here.

Perhaps you can reuse some parts, get some ideas, or suggest improvements.

I’ve implemented a watering regime that runs it dry based on a lower threshold then wets it to a higher threshold.

I had the idea to use a hygrostat type component for that (like a temperature contoller but for humidity) but haven’t progressed that idea further.

I still need to do more testing and refine it but most of the guts of it is there.

Cheers.

https://community.home-assistant.io/t/solar-powered-self-watering-plant-no-plumbing-no-power-wip/464263/14?u=mahko_mahko

Uh, that is nice. Thank you for joining and sharing!

I think you are right about the support. If you use the fusion file to print from, the thing is actually chopped up in separate parts that you can print from for that particular reason. You can also join them back together if you would rather like them to be bigger parts. It takes ages to export to STL’s, so I didn’t bother.

If you print with OctoPrint, there is a plugin for failure detection using a camera. It can detect spaghetti quite effectively I’m told and also be setup to notify home assistant as well as pausing the print. I haven’t gotten the camera to work with OctoPi, so I haven’t really tried it.

Btw, are you still working on the pump-issue you mentioned and what are your hopes that it will work?

Yes. I just solved it. I did some experiments, and it turned out it worked when powered by USB, but not when powered by a battery. GPIO21, which I used to control the water pump, goes high when the ESP goes to deep sleep when powered by a battery, but stays low on USB power. I changed it to GPIO23 and now it works.

I realized that I didn’t want the seam where the parts go together. Another issue with the original design is that is hard to swap the battery when it is at the bottom of the pot without accidentally pouring water out of the reservoir. Therefore I decided to do a complete redesign of the pot. You can see the progress here: Fusion

That is awesome news :slight_smile:

I like where your design is going, nice job! The drawer approach is a good one. Just now, I also came to think that the original has a high Centre of gravity and a small base. That is a toxic combo that your design does not have. Here’s some assorted feedback:

  1. The original project on Thingiverse has 3d models of pump battery, pcb ec. It will make it easier getting a perfect fit for the compartment. I have added them in fusion format to the Git repo for you to use.
  2. Are we back to the old design for moisture and water-level detection? (so not using the capacitive sensors made from copper tape…?) It may have changed, but 80% of more than 100 ordered of the black moisture sensors used in the original are faulty according to the creator of the original Flaura. They also add to the price per unit.
  3. If you make sure the edges of the chamber for electronics are always above waterlevel, that will decrease the likelihood of an accidental floodding. It can be achieved by adding a few holes near the upper edge of the planter. With them, the plant will flood before the electronics. A separate lid for the electronics compartment would also help.
  4. Both PLA and PETG can be polished to remove the lines or even to get a glossy look by simply sanding it with increasingly fine sanding paper. It can be done quite easily by mounting the thing in a lathe I think, but I think an electric drill could work too, if you ad a mounting point in the bottom of the lid and pot. The lid may need a punch-out in the hole to better support the sanding. Polishing will definitely make the thing look nicer in the windowsill.

Hope the above helps :slight_smile: Let me know if you’d like a hand with the Fusion-work.

  1. I will add some hardware models
  2. I bought the hardware I wanted to use before I saw this thread and your discussions about hardware. Therefore, I will be designing it for the hardware I have. I will release the f3d file to my GitHub repo once I am a bit further, so you can modify it for the hardware you prefer. I was also a bit skeptical about the moisture sensor, but I took the chance and ordered 10. I have tested 3 of them and they work. I also verified all of them had the component that sometimes was missing. I have the link on GitHub to the seller I bought from in case you want to use the same moisture sensor as in the original
  3. That sounds very similar to what I had in mind
  4. I don’t think I will optimize it for sanding, but as I said earlier, I will release the f3d file and then you have the option to do it

I’ve had some trouble with turning the pump on reliably, and it turns out that the MOSFET transistor, that the was recommended in the original isn’t a great fit because the threshold to turn it completely on is 4 v, which is less than the output of a pin on an ESP.

Which transistor do you use to switch on the pump?

I use a NTF3055L108T1G MOSFET N-CH.

Thanks. Got it :slight_smile:

Any updates?

I got Mathgamings code to work with the original PCB. My version isn’t perfect, but the plant gets water.

I made some minor modifications to used Mathgamings design for the physical parts to acommodate the original PCB and eliminate the need for glue for the inner part. I like it, at least for the first version because the electronic bits are so easily accessible. Really smart design @mathgaming :slight_smile: The first versions I’ve printed were not watertight, however and then my printer started with all sorts of issues. I think maybe it has reached puberty or something. Not sure.

think I have it sorted now, though. I’ll make some test-versions for watertight designs shortly.

1 Like

Keep us updated on this, thx! (maybe nothing has been done since, but when you have time to, please keep up updated) :smiley:

1 Like

I am still working on getting the damn thing watertight. There is progress, but I am not yet there.

Any update?

Yeah, actually there is. I’ve made several updates to the model posted by mathgaming in december:

-The inner pot is screwed in with a big thread so that it can easily be removed if you need to replant or something. I found out, that if you get dirt in the watertank, it’s not that easy to remove.
-The lid is changed so that it doesn’t pop off. Hopefully. I have not tested it.
-The electronics compartment is printed separately to make it easier to get it tight (the join between the pot and the compartment tends to leak in the original version from mathlab).

There is also a jig for putting the outer pot in a lathe so you can easily sand it. You may want to print it 1mm thicker if you want to sand. Also, check if the holes in the jig match,

If you need to, the whole thing can be painted with a sealant. I use varnish for boats. 3d prints sometimes have tiny leaks that will let out a drop or two. The varnish helps with that and also gives the whole thing a very nice shine.

You can find the new pot here: GitHub - MartinVinter/Flaura2: An ESP-home based version of Flaura

It would be nice if someone would print it and report back if it works.

I could try and print it throughout the week. What are the exact files you need testing? I could also try and assemble the whole pot if you want to, but I would need some sort of materials/parts list. Would it in theory work if I got everything? I mean, does the software-part work?

I saw the link with all the files, just want to make sure I print the ones needed. What material would you reccomend? I was thinking about PETG as it is strong and “food safe” (good enough for what it would be used for.)

I can’t find your updates. The last commit on the repo you posted is from 2022-12-12. Did you remember to push the update?