RAYSHOBBY.NET

Hi everyone, Christmas is coming in a few days, and soon we will be entering the year 2013. I figured this is time that I post about recent progress on OpenSprinkler. Sorry that I haven’t blogged for a while. Various projects have been keeping me busy. I am excited to post and talk about the details of these new projects in the near future. Anyways, what’s new at Rayshobby shop in the past month or so?

Sale Update. A few weeks ago we’ve just shipped the 600th OpenSprinkler. The sale really starting going up quickly ever since the Bay Area Maker Faire this year. I am hopeful that by the end of spring next year we will reach the 1000 mark. That would be awesome :). With a growing community of users, I’ve also started seeing people contributing to this project, including feature testing, debugging, writing scripts and mobile apps (see below). I am really thankful to everyone who supported and contributed to this project.

Software Revision. The latest firmware shipped with OpenSprinkler is 1.8.2. It has been relatively stable and I haven’t seen or received report of any major issue with it. Some of the commonly requested features are 1) adding back the parallel running mode (which was available in firmware 1.7 but disabled in 1.8.x to make space to other features); 2) adding a weather feature; 3) adding a microSD card for logging and storing Javascripts. I am hoping to get the first request done soon: since the feature already exists in previous versions, I just need some patience to sit down, optimize and reduce the current code and put it back in. As for the second request, a friend of mine has been helping me to implement the weather feature in Python script. The plan is to host the Python code on Rayshobby server. This way, the controller can periodically send requests to the server with the user’s zip code, and the Python script will return a simple number that reflects how the water time should be adjusted that day. This will free the microcontroller from having to carry out the necessary computation. It would be a simple scheme to fit a basic weather feature into the current generation of hardware. The third request cannot be implemented with the current hardware, due to the program memory limit. If you need logging, please use David’s Sprinkler Logger script that I blogged about in a previous post. It is available in the OpenSprinkler GitHub repository.

iOS/Android App. I’ve been holding back on developing an iOS or Android app for OpenSprinkler, partly because I am not familiar with programming mobile apps, and partly because the OpenSprinkler’s web interface is already functioning on mobile devices through their web browsers. But I do agree that a dedicated app would look much more appealing and make the product stand out. Fortunately the folks at FeedFriendly.com have developed an iOS app for OpenSprinkler called SprinklerAce. I’m quite excited and I’ve already seen the pre-release version of it. The app is extended from the previous version, which was dedicated to the IrrigationCaddy. It makes use of the manual operation mode on OpenSprinkler to directly control station valves. It’s currently undergoing the Apple Store approval process. I will make a separate post about it once it becomes available. Stay tuned!

Hardware Revision. The latest hardware version is 1.4, and we still carry both the full-assembled products (1.4s) as well as the DIY kits (1.4u). A while back when the last batch of DIY kits sold out, I was considering discontinuing it and start carrying semi-assembled kits, similar to the OpenSprinkler extension boards. But due to the popular requests from users, I decided to still offer the DIY kits, not only because its price is significantly lower than the fully-assembled versions, but also because many people enjoy the process of building a kit regardless of the price. In October 2012, I made some minor revisions to both 1.4u and 1.4s, mainly to improve the reliability of the power conversion circuit. Again, the hardware design is pretty much stable and I do not anticipate any major changes to it.

Also, I’ve been working with SeeedStudios to make a injection-molded case for OpenSprinkler. I’m quite excited because this is the first time I’m working on an injection-molded case. The folks at Seeed are awesome and very helpful. I will certainly learn a lot through the process.

Another bit of news is that I’ve started prototyping OpenSprinkler 2.0. The current plan is to simply upgrade the microncontroller to ATmega644 in order to fit in features like microSD card and more sophisticated weather-based control. Because I am planning to design it to fit the new project case, it will not be available until the case design has been finalized. So most likely it won’t be released until some time in late spring and early summer next year. I do have two prototype images to show off here. I was curious to experiment with different PCB colors, so I ordered a white version as well as a black version:

They both look gorgeous, but I personally prefer the black version because the dark background looks really cool. Feel free to let me know your preference/opinion. In addition to upgrading the microcontroller and adding the microSD card slot, I’ve also changed more components to surface mount packaging to facilitate manufacturing, and adding support pillars for the 16×2 LCD. Again, this is just a hardware prototype, I haven’t started working on the software code yet. The final version will also have to be designed together with the new case. So a lot of things need to be orchestrated before its release.

Upcoming New Products. As I mentioned above, I’ve been working on several new projects. Among which there is the OpenSprinkler Bee, which is an open-source controller for battery-operated sprinnkler valves. I already have a prototype and have demonstrated it at Maker Faire. I am working on making the kits ready soon.

Another product is a controller for watering indoor plants. It has built-in humidity, temperature, and light sensors, and will interface with a compact water pump to automatically water plants for you. I’ve temporarily named it OpenSprinkler Sip. This is a very exciting project that I hope to blog about soon.

All right, so much for the end-of-the-year OpenSprinkler Update. I hope everyone have a Merry Christmas, and see you again in 2013!

Logging is a useful feature which allows you to find out how OpenSprinkler has worked say, in the past week. While you can check the scheduled water programs using the ‘Program Preview’ feature, it does not account for dynamic changes like rain delay, power loss etc. This is where logging comes in handy: you can find out what OpenSprinkler has actually done in the past. Unfortunately I have not been able to include logging in the firmware because of the memory space that it would require to store the log data.

Recently, David Gustavson, an OpenSprinkler user, successfully implemented logging by using a PHP script that he wrote. The way it works is to have a computer that periodically polls the OpenSprinkler status say, every 20 seconds, and the script checks the differences between every two consecutive polls to record status changes. This is quite clever, as it does not require any modification on the OpenSprinkler side. All that it needs is for OpenSprinkler to return the status of each station, which it already does. Of course this does require a computer running on the side (or even remotely), but it’s pretty easy to set up, and you can use a plug computer or a Raspberry Pi which has very low power consumption.

David is kind enough to share his scripts and I’ve checked them in to the OpenSprinkler GitHub repository under the script/PHP/ directory. He also wrote a blog post about OpenSprinkler, which you can read about here. Thanks David!

Recently I received a request to customize OpenSprinkler for a 12V DC external relay board. It turns out to be relatively easy to do. Since it may be useful for other folks, I am writing a post here to briefly describe the modifications. First of all, the relay board has built-in opto-couplers and flyback diodes. It is designed to work directly with microcontroller pins. This particular relay board I received is an active low type. This means normally the relay control pin is pulled high; when the control pin is set low by the microcontroller, the relay will be activated. So before you go ahead and make changes, make sure to find out the type of your relay board.

OpenSprinkler can run on 12V DC power supply without any modification. As long as you keep in mind the polarity, it should work right away. Even if you connect the power supply in the wrong direction, it shouldn’t cause any damage because there is a protecting diode D1 that prevents reverse polarity. The only modification I needed to do is to replace the triacs (used to drive AC solenoids) by NPN transistors. Any common NPN transistor is fine. For example, 2N3904. The remaining step is to connect the station wires to the corresponding relay control pins.

The way it works is that normally the shift register outputs low. So all NPN transistors are turned off, and the collectors are pulled high by the relay board. When a station is open, the shift register pin outputs high, turning on the corresponding NPN transistor, and driving its collector low (to ground). This in turn activates the relay. That’s it.

Here is a video showing the relay board in action. The OpenSprinkler running a self test that turns on each station for 3 seconds.

It is also possible to use the NPN transistors to directly drive the relays. However, in this case, you need to add an additional flyback diode (between the collector of the transistor and the +12V line). The flyback diode is important to protect the transistor under inductive load. To accommodate this, I am planning to add PCB holes for these diodes in the future, which will make it easy to solder these extra components.