What about vines

I was weeding the (outdoor) garden today and was looking at the peas and hops. So if I want to incorporate bigger plants, I need bigger spaces. Starting at the idea of the cubes, but I remove a lot of material, which will probably cut costs as well. Uprights holding the floor and the ceiling. Matched holes are drilled through those. Add on a pop in side wall. The back should still be solid.

So this isn’t 100% what I want, but it’s a start. I still need to figure out a lot of the other details. How big are the cubes? How much power do I want on those rails? How do I want to integrate watering? How am I going to do the communication? How do I want to pop in the sides, magnets?

I should get the first one built soon. Even a rough idea is better than no idea at all. Also, the (indoor) garden would like it.

BOM

Preliminary notes. All of these are based on my not finding cheaper or easier substitutions. The threaded rods (and matching nuts) are probably a given, but I’ll need to think about the motors.

  • https://www.pololu.com/product/2823/
  • https://www.grainger.com/product/GRAINGER-APPROVED-1-4-28×3-ft-4RDC7
  • https://www.grainger.com/product/FABORY-1-4-28-Hex-Nut-4FAZ8
  • https://www.sparkfun.com/products/12211

I’m really on the fence about the motors.  Assuming I don’t care about speed, within reason, the only real driving forces are a minimal amount of torque and overall price. By overall price I have to include not only the motor but controller and any encoder. Price is telling me to get some Chinese stepper motors off of Amazon, but that seems like a good way to have to replace some motors. I’ll have to continue looking at this.

 

Messaging

Or how to ghetto build an extendable sensor network.

Well really, other people have fine this before me, I’m just trying to come up with a quick and dirty hot despair system. It’s more of a brain exercise than how I’d really go. Chances are I’ll just drive all these with a stack of Pi Zeros with hacked together Ethernet. But what if I don’t. This would be for the planter cubes. Each one would report moisture, temperature, illumination (to show if there’s a problem) maybe some other low grade sensors.  The data wouldn’t be much. Now I want this to be cheap and plug and play. Any time I plug in a cube, I want it to give me a couple LEDs. Red is bad, flashing green is coming online. Green is go. So what happens when it’s coming online?

I’m thinking we have a power rail. Next to this we have…

Sorry, just realized I’m a tool shed who should learn about parallel and series circuits.

…master line. This will be used to send a pulse down to all cubes on the rail. This will start the calibration for internal timing, and provide everyone a chance to become acquainted. So when someone comes online, it listens for the first pulse and then listens to the data line. During this time all onlined critters are sending their data. If we listen to that, we find out how many other cubes are on. After the next pulse, I now can calibrate my internal timing to everyone else’s timing which means that I don’t have to have high quality parts.  After a quick sanity check I’ll use one of the sensors to seed an RNG. This will provide an offset from the back end off the pulse. At this time it’ll give a shout out. The rest of the time it’ll be listening to ask the established cubes. If it detects another new cube, the first one wins. Now since we can’t know if there’s another cube on the same time track, we keep on going with more RNG action all the way down the line until our newest cube is right after the last established cube. Once we’re there, we start transmitting actual data.

End the day, assuming I have a large enough current source, this will work. I could probably limit my current needs fit the boxes by having signal diodes on the data lines, and then… oh duh. 

Master gives start. New boxes do their thing. When the new boxes transmit, they transmit a unique identifier. The units identifier gets passed back by the master. If the master passes back a corrupted identifier, we have cross talk. Do another round of rng hangs. If this happens three times, start flashing the red LED which is the pop out and plug in signal. This should be enough to get it back on the rails.

I should start putting together a BOM.

Notes to Self

Self:

  • Start arrow order
  • Research constant current driver for LEDs
  • SD cards for stuff
  • More LEDs, or at least prices

I’ve declared a new word

Joink – The vebification of janky.  to make janky.  Usage: Can you joink that?  Tenses: joinked, joinking.

When I have a minute I’ll clean this up, just wanted to make sure documentation standards were being met.

CAD and sustainability

Mostly unrelated. Mostly.

I started the tutorial three more times last night for FreeCAD. Each time, I must have clicked something wrong, and no dice. Tonight I’ll try the tutorial again with the video rather than just the website instructions. I don’t feel super loving toward it though.
I just had a thought pop up.  On the planter boxes I want a window front, so I can look and stuff.  This is easy enough, have a hinged door with a plexi panel. Oh. Plexi. Right. Have a hinged door with a glass panel.  it seems like such a small difference, but a little plexi is a big environmental commitment over time. Also, I’m not really sure if plexi starts leaching out whatever at some point. Either way. Glass.  Now, money wise, am I better having a wood door with a little window or a little wood with a big glass.

CNC

Since I’m thinking of starting to build a CNC machine, based on some of HomoFaciens.de’s (amazing) work, I realized I need to be able to layout parts for cutting. This means that I need to learn some rudimentary stuff in 3d CAD. I’m cheap though, so where to start?

Since it’ll be running on Linux, I started with an incredibly broad Google search, choose names that I liked, did Google searches looking at their comparisons, and I brought the list down to two. I’m looking at FreeCAD and BRL-CAD. So I installed them both and have started to poke around.

FreeCAD has a lot of great tools that I’ve seen (though never used), but that interface isn’t really jelling with me. With time, I’m sure that I’ll be able get used to it, but right out the door, it’s not me. BRL-CAD has all sorts off good things, but the first bit I tried was even more painful. The silly thing is, that at least what I’m starting with, I could totally do in Inkscape as it’s all 2d anyway. So tonight I’ll try to finish the FreeCAD tutorial, and then try to make the same part in BRL-CAD.

Just I’m about to go to bed

Peristaltic pump(s) with only two motors. Have one motor drive a carriage with the other motor. That motor is driving a great that interfaces with whichever other gears to go to each of these pumps.

Now, I’d need a way to verify the gears were going to mesh, so probably multiple photo diodes and a laser? I could dive as many pumps add I want, but only one at a time.

More elegant solution could be a motor turn a wheel, on the wheel is a motor. The motor messages with the pump gear. This one would only be limited by the great ratios (ie how many gears I could get around a gear.

The first idea is better.

All for better gardening.

Pin isolation

So I’m worried about the risk of esd versus a Pi if I go with a Pi to grow my garden. I was thinking about this while walking up the hill to meet the Wif when she gets done with class. And this is why beer is the solution to all of life’s problems. I have a Teensy lying around that I never got around to building a keyboard around. Cheap, small, durable.

#blessed

One other minor issue

I’m fairly sensitive to high frequencies, whether it’s noise or light flicker. Which is to say that CRTs are the bane of my bleeding headache.

So the lights have an imperceptible flicker, and the regulator has a little hum. I think… after telling a coworker about my project, that I’ll want to change out from just going high, to going almost high via PWM, thereby I’ll new able to adjust the frequency (maybe).