Choosing Between PFCv2.1 and MVP


#1

Hello everybody!

I am interested in building a Personal Food Computer this summer and am considering between implementing the PFCv2.1 and the MVP. Ideally I would like to build a PFC but am concerned as it seems to be unsupported and have an uncertain future. Can anybody comment on it being wise to invest my time and money into building a PFCv2.1 when it sounds like a new version will be coming out of the OpenAg team sometime in the near future, possibly breaking backward compatibility between new and old systems?

Thanks for all the help,
Colin


#2

A few thoughts:

  1. Do some research on hydroponic system design if you haven’t already. If you poke around the forum a bit, you’ll find some great links to videos, books, papers, articles, etc. Once you see what’s out there, maybe you’ll get a feel for what type of system you’d like and where you want to be on the spectrum of designing your own stuff vs. following pre-made plans.

  2. It’s not wise to start building a PFC2 unless you intend to extensively modify the hardware design, roll your own software stack, and do a lot of your own troubleshooting. Key components of the OpenAg design–lights, nutrient dosing pumps, refrigeration, and software–are either not generally available or have been reported to have consistent problems. People have reported major delays and difficulties with trying to closely follow OpenAg’s design and run the openag_brain software.

  3. On the other hand, some folks sound happy about their experiences with building systems that are loosely based on the PFC2 design: 80/20 aluminum frame, Atlas pH & EC sensors, custom software, etc. If you like the idea of designing your own system, this could be a good option for you.

  4. People have been reporting good results with the MVP design. Peter and his team are committed to supporting it. The one tricky part is the 120V wiring with the relay and light fixture. You need to be careful with the 120V stuff to avoid shorts, water ingress, or exposed wires energizing parts of your enclosure. For example, you might want to add a junction box for the light fixture wiring so it doesn’t have the opportunity to short out against the Reflectix and burn holes in the mylar (that happened at least once).

  5. Based on what I’ve seen, I don’t recommend making plans about OpenAg’s new designs being ready at any particular time. The original post in Update from OpenAg - Please Read said Caleb would be writing a Medium post about it by the end of March, but then the 60 Minutes thing happened. So, I guess that timeline didn’t work out. I’m curious to see what they come up with, but I expect it will take time.

    It looks like the new stuff will be really interesting. But, the hardware is totally different than the PFC v1 & v2 and they’re rebuilding the software stack from scratch. The new hardware is a modular system of custom circuit boards that would need to be manufactured–either with really tricky hand soldering or perhaps pick & place with a reflow oven. I’ve seen no signs that OpenAg has a manufacturing and distribution pipeline that might be able to provide the new circuit board modules to the community. They’ve made a bunch of the new light controllers for their hazelnut tree thing (see Calab’s instagram), but it sounds like (see github & instagram) those were slowly and painfully hand soldered with a microscope.


#3

Thank you for your informative reply, @wsnook. I wondered about the same thing. I just found out about the PFC and want to use time this summer to build one or to prepare for the new design (PFC 3.0?).

Following your thoughts, I will:

  1. Start reading a little bit more about hydroponics. I will also start experimenting with growing plants together with sensors, Arduino, RPI, and so on. If a new design is released, I don’t think the pre-made plans will be easy to follow in the beginning considering the size of the community. A lot of experience is probably needed/helpful.

  2. I will not start building the PFC2 since I have no time or skills to extensively modify the hardware or software.

  3. Building a highly customized system seems to take away the point of following an open source design. I want to contribute by replicating, does that make sense?

  4. Is the MVP a minimalistic version of the PFC 2.1? Do you think the MPV is easier to make/modify?

  5. It looks like you have been following the community for some time. Do you have any guesses of when the new design will be ready or announced? I heard in an interview with Caleb Harper that they have a team of 15 people, and it seems that they are working on multiple projects. Do you think there is a change that the PFC-project will be abandoned?

I love the concepts and ideas around OpenAg, and hope I can contribute and learn from the community some day. Do you have any other recommendations for me? Should I just start with something?


#4
  1. Yeah, I think lots of people have similar expectations when they first hear about OpenAg and the food computer concept. The catch is that making hardware with embedded firmware to run it is a lot more slow and difficult compared to the pure software projects that people are familiar with. What’s been happening so far in the OpenAg community is mostly early stage prototyping.

  2. They’re different. It’s difficult to explain how and why. I think it helps to consider the prototypes so far in terms of different people getting inspired by Caleb’s big picture vision and exploring how to apply it to their own areas of interest. It helps to consider who designed which plans and what their motivations and constraints were. One big area of difference is budget and access to tools. A related concept is hardware prototyping traditions: university robotics (big budget & low volume, machine shops, heavy duty metal frames, heavy duty wiring), Shenzhen style (3D printed plastic, injection molding, aluminum extrusion, breadboard prototyping, custom PCBs, cost sensitive mass production), and Maker style (all the stuff in Make magazine, focused on home hobbyists with limited tools plus people with access to community shops).

    Tim from OpenAg likes to make things using machine tools to shape metal and acrylic. The hardware for MIT’s PFC designs is made in Tim’s style. This is a good fit for people at universities who have access to machine shops and the training to use them. The PFC electrical wiring and use of ROS software appear to be influenced by people who had exposure to building robots.

    The software for the v1 and v2 PFCs was built in loosely coupled modules by several different people. Again, it makes more sense if you consider it as the result of different people exploring technologies they were interested in with limited coordination between their areas of responsibility. The new software is being written mostly by a new group of people. Rob is doing a lot of the work, and he’s mentioned having other engineers helping him at times. I’m not sure how that will unfold.

    For the MVP, the name and concept originated in this thread: Growing food: I just ordered a MicroGrow Kit from Hamama. In particular, it came from Caleb’s comment on 28 March 2017 in that thread.

    Peter and Howard, along with some other folks in St. Louis, collaborated to make what’s now known here as the MVP design. One of their big goals was to make the hardware design practical for school kids to build with simple hand tools. The electronics are a mix of Raspberry Pi and Shenzhen style prototyping. Howard did most of the software in a classic Unix command line automation style–cron, bash scripts, python scripts, and a bit of CouchDB. It’s very different from the PFC v1-v2.1 software, but it shares the ambition of crowd sourcing climate and phenotype observation data from a network of food computers. It’s like people have been trying to scale the same mountain from different directions. Nobody’s made it to the summit yet.

    A couple other big things are happening here on the forum that aren’t necessarily oriented around crowd sourcing phenotype data. To me, this stuff is the most interesting part… Some folks are here to seek or share share information about custom hydro and aero systems for growing food at their house. There also seem to be a fair number of people working on data collection and automation systems for growing at a larger scale in a corporate or university context.

  3. I don’t know what will happen with the older PFC stuff going forward. If you look at what OpenAg has announced in recent years that they wanted to do, and you compare that against what they’ve released publicly, it looks like they have more ideas than they have resources to act on those ideas. Starting this past winter, it seems like they have been narrowing their focus. So, that’s a good sign.

    As far as I know, Rob and his team are only making software for the new hardware designs. He told me that he hopes to have a beta software release ready some time this summer. Other than the survey announcement in the OpenAg Update post from this March, I haven’t heard anything about plans for doing anything with the older PFC hardware designs. But, I can point out that the v1, v2, and v2.1 designs are already unmaintained and officially unsupported.

As for recommendations, I’d say it’s a good idea to start by educating yourself about what other people are doing. It’s useful to read about plant science. It’s useful to understand how light is measured for architectural vs. horticultural applications. There is a lot going on in industry and at universities. Some of that is public. Peter has linked to a lot of great stuff.

Beyond that, find some things that you can try that won’t take a lot of time or money. Try stuff and see what you find interesting. Then, follow that trail and see where it leads.


#5

Hiya gang!

I’ve been out for a bit. Looks like some things changed and didn’t.

I’ve got plants growing in my PFC1 and am experimenting with passive and active cooling as that’s one big flaw of the insulated PFC1 casing. Atlas sensors coming soon too.

I never built the v2 system because it didn’t work. Maybe that’s changed but the v1 system is dumb simple. I’m interested in learning about the mvp platform.


#6

Great to have you back @juhnke! There’s been quite a few attempts at V1 with few successes (that I’ve seen) and am curious what you’ve done regarding the UI and climate control.

I’m curious to know what you’ve been growing and what your goals are for the future.

https://wiki.openag.media.mit.edu/mvp_1

Is the best place to start regarding MVP. The $300 Food Computer thread is also a good history of the project (and MVP - Design will give you context) I’d love to hear your feedback!