Greenhouse Watering Controller, Pi Zero W, Automation pHAT


And did you shut down & power off the pi each time? Or is it considered safe not to bother?

I soldered things up this evening, ready for prototyping.


I did not shutdown the pi just removed it from the circuit (took of the power and ground lines) and that seemed to have saved my pi’s from short circuits.

By the way that setup looks great.


@PaulRB … well the u.FL connectors arrived on Monday (if I remember correctly) and a certain Mr @sandyjmacdonald managed to stick one on and get it up and running. And then rip it off :D

It seems the u.FL landing on the Pi Zero is… uh… tenuous at best.

I might have a go myself tomorrow, giving the board a good soak (as hot as I dare heat it, will have to check some datasheets for temp thresholds ha!), and attempting to get good adhesion to the larger (and more troublesome) main pads. I’m betting solder tends to stick to the signal pad waaaaay easier than it does to the probably-connected-to-the-ruddy-heatsink-ground-plane pads- leading to the problem with them ripping off really disturbingly easily.

My other, possibly impossible, solution would be to pigtail some copper enamel wire off the signal pad and mount the connector onto a daughter PCB. where it wont be such a pain. We’ll see!


@gadgetoid, @sandyjmacdonald thanks for your help with this. Maybe we should feed this back to the foundation. I for one would pay 33% extra for a Pi0W with the connector fitted and ready. I think they would have to do a production batch like this, because modding the standard Pi0W outside the factory would make it prohibitively expensive I suspect.


Just add my 2 cents, for what its worth, after playing with a similar project, but in a small scale (large tote). I used an automation hat, on a RPi2, with a am2302 sensor for temp/humidity readings, and used two relays off the automation hat to turn on/off pump if RH hits a % and another relay to trigger lighting. All data is logged to influxdb which is made for metrics logging and grafana as a web accessible graphing interface.


So I had a go, and I got the connector to stay put even through a few deliberate disconnects/reconnects. In order to do this, I tinned each pad on the PCB, placed the connector down, secured it gently by remelting the solder and then just blasted it with the hot-air gun until everything was stuck, in technical terms, “good-n-proper.”

What I couldn’t do was the resistor. Hot damn it’s small. Granted I tried doing it at home with a terrible soldering iron which I thought had blown up but chanced turning on again. I’m going to need to clean up those pads with some de-solder braid and have another go.

WiFi still works despite having two antenna connected. shrug maybe I’ve not even connected it properly.


In this post I will explain a bit more about the Mk 1 project.

The picture below is the main board with the Picaxe 08m2 mcu chip. It also has a 12V-5V 78L05 regulator and a P36NF06L MOSFET to control the pump.

Here is the soil moisture sensor. Two lengths of galvanised steel wire. Sophisticated!

And here is the pump. When installed, it was connected to the end of a length of hose and sunk into a water butt that is fed by rainwater from the greenhouse roof.

Power is from a 12V SLA battery, charged by a small-ish 12V solar pannel. No charge regulator was used, because I had read that SLA batteries are reasonably resistant to overcharging, especially with small pannels.


I appreciate this is a bit of a bump on a year old project (apologies in advance!), but I’m very interested to see the final result if it was successful?

I’m looking to partially replicate this for a 300L water butt / greenhouse system and this is the first project I’ve found based in the UK. In the past I’ve only been able to speak to folks across the pond and I cant exactly “pop down to radioshack” for the parts.


Yes. bump OP must deliver :D