Extend Farmbot to “infinite” length & “multi-bed”


I don’t have a Farmbot (yet), but I have been digging into all about the Farmbot (reading documentation, blogs, forum posts, etc.)
One thing that seems to be a limiter is the overal size (range & scope) of the robot. Having said that, and given that I have no actual hands-on experience, here is my idea:

  1. What if Farmbot got its power from the two rails? (Or a 3rd rail like a subway train). That would allow Farmbot to be much longer and eliminate many cables.
    Another thought would be to devise a way for the Farmbot gantry assembly to switch from one bed to an adjacent bed. Both of these ideas would allow users to expand the Farmbot in both the x & y directions. Just raw ideas for mulling over - I have no idea if these are feasible.


Hey Tony,

There has been plenty of discussion about the scope and range of FB (of which FB XL was borne). I’ll look for the exact threads and post them here if I find them, but the take away is that your run up against structural integrity and precision constraints, especially when increasing along the y-axis. A point will be reached when the combined weight of the gantry (cables, structural materials, water in the tubing, etc) will compromise the current design, causing stability and movement issues.

I do like your idea of of an electrified third rail. You enter a new realm in which there’s a tradespace b/w power supplied, the efficiency of the operation, the electric shock of people and animals, and weight of the infrastructure. At the power levels needed to run a heavy rail commuter rail, it is expensive to shield the cables. It’s also highly inefficient and practical to ‘bring the cable with you’ given the distances involved. It’s also much more economical to put dielectric spacers after 10 feet or so since the third rail is somewhat self supporting (and does not need to support the weight of the train). Finally, the protection mechanism is pretty simple: curved aluminum.

In terms of the FB, you’re hopefully not at the power levels of a heavy-rail commuter train :slight_smile: However, with the third rail, you could gain enough power to say run a tiller or a dump shovel of some kind. This would drive up the need for heavier materials and more powerful drive motors, and so on. From a safety stand point, you still would have to protect that third rail from accidental contact. Also, you’d have to insulated the live rail from the non-live rails. That usually means some kind of dielectric material spaced as far apart as the structural integrity of the third rail will allow. If you support it and protect it completely along its length, then you just made a sheathed conductor, or a wired cable :slight_smile: .

My take away is that the design would have to radically change in order to necessitate a third rail. Please note, this is a conversation that I had with myself while trying to wake up with my first cup of coffee, so please grant me some grace if some of the ideas are not fully baked. :stuck_out_tongue:



Thanks for the very detailed reply, Fuzzy!
So, I am wondering what the voltage threshold is at which point electrocution hazard needs to be taken into account? Also, as you mentioned, shielding may mitigate this. I would think that the cost would be offset a bit by not having to use a cable carrier and associated cabling.
As for the load that may be incurred on the gantry, would that be remedied by just using a thicker or reinforced extruded al tube?
If one were to make one half of the gantry the positive pole and the other the negative (with insulators between, of course), all tooling, logic and motors would just need to tap the closest connection of each. If you think it worth the time, I could make a very rude entry sketch, but it sounds like this may have already been explored and I don’t want to re-hash a stale topic (but please do send me links to any historic threads to bring me up to speed).


i’ve been having very similar thoughts on electrified rail.

12v seems ‘safe enough’ for me.

I’m not bothered about water, it’s easy enough to run drip irrigation.


My thoughts were to avoid the need for cabling at all and have everything on board the arm. This way it can go as far as you like in the y axis.
Have a home location with a power source to charge an on-board battery. Similar thing for water, fill an onboard container from a source (could be multiple sources along the gantry).


Yes this would also be a good approach.

I guess the limitation would be the downtime for charging. I want to work my machine fairly hard to get the most return on investment. Taking 4-6 hours out of a day for charging may or may not be worth it?


The voltage threshold for electricution has to do with what’s known as the dielectric breakdown constant. Simply put, it’s how far another conductor can be without causing the electricity to jump through the dielectric and complete a circuit. For air, this depend on the humidity and an particulates in the air. (Nerd stuff here: https://www.britannica.com/science/dielectric-constant) Your most practical experience with this breakdown is when you get zapped by a door knob in the winter time. Effectively, it’s mini-lightning. It’s usually measures in the 20,000 volt range. Then why don’t we die? Well, that’s the other side of the equation: the current. The current is in micro amps. Together, the total power is miniscule. Lighting is the extreme case: millions of amps and hundreds of thousands of volts. (Nerd stuff: P = IV, or Power equals current times voltage.)

Back to the third rail: it depends. 12v is definitely safe from atmospheric dielectric breakdown. However, to get the same power output of a higher voltage, you’d need to crank up the amperage. It’s generally the combination of current and voltage that kills, not necessarily just one of them being high. There are some major exceptions to this, however and they are the driving safety considerations. Take for example the heart-restart paddles used by emergency responders and medical personnel: lots of high voltage so that the nerves serving the heart can be uniformally enervated (activated). The whine you hear are the capacitors charging up to a high voltage. Then they’re discharged like targeted lighting. Relatively low amperage though, as the total power needed to excite a nerve is low. They just have to get through the rest of the chest cavity first.

Back to the FB: if you’d want enough energy to move the heavy cabling and do heavier operations, you’ll need more power. More power means more conduit, higher voltages and/or amperage. Remember you P=IV equation. So, I’ll ask the question: what would you like this ‘infinite FB’ to do? More generically, what tasks are you trying to accomplish and how is the FB helping you with that? That should be the starting point :slight_smile: From there, we can help engineer a solution.


In my personal case I have a 50m x 20m area to cover. I grow leafy green crops (mostly lettuce)

I’d like Farmbot to sow the seeds, weed and harvest the crop. I don’t need irrigation as I’m happy to run drip irrigation.


You could always control the water valve for watter dropping irrigation :wink:
20m length for the gate of the FB you should use some kind of suspension.


I’m not quite sure what you mean by suspension?

I’m thinking of running electrified rails, fed by a 19v ish solar array. The charge controller would be on the Farmbot.

The machine would re-calibrate itself at 1m intervals using magnets that were attached to the rails.


Prepare for first cup of coffee ideas!

One idea to help you span 20m is to make a multiple x axes (with supports) and one long continuous y-axis.
This is similar to the massive irrigation rigs in use today for watering large fields. There are multiple trucks along a line with scaffolding spanning the 16m or so between them. The wheels do carve parallel lines (or concentric circles depending on configuration). To save on cabling, use the most center rail to run the umbilical cord(s).

Let’s assume the current FB XL is brushing against the physical limits of the design, then you’d need 7 x-axes spaced 3m apart, each 50m in length.

Here’s another, nuttier idea: the sky cams they use for football games. Have your FB head ride along retractable cables. You’d have to do some experimentation on how best to operate the cabling to get the desired effect (maybe a retractable z-axis just like now.). I think one of the down sides to this design (and the current FB design) is that it’d be inefficient during a harvest cycle. Each plant requires a separate trip back to the bin or whatever is collecting the produce. Another design consideration (not down side) is that your umbilical would have to be able to withstand the stresses of being reeled in and out. Perhaps solved by having a small tank on the bot (say 4L) and have it refill a few times per watering cycle, a Li-ion battery that is charged between cycles, and wireless comms.

The current FB sw could be used almost as is: x and y movements turn into cable winding movements. Z axis might be the trickiest part of this.

The biggest ‘miss’ I think of this design, would be seed sowing, that is, if you had a traditional approach of digging rows, and then backfilling seeds. It’d have to be done very similar (see injection) that FB uses today. That too would need an on-board hopper.


Here is an idea I was thinking of for the next farmbot revision.

Have a guided vehicle robot with a large pillar/pole (a cheap square steel tube would do, you can buy them on mcmaster-carr or other raw material provider websites). The y axis would be as long as the distance the guided vehicle robot (four wheel vehicle) can move.

The tube sticking out of the robot would hold the x axis tube. The x-axis tube would have roller wheels moving the x-axis subssembly.
They can be powered via a solar/battery powered dc geared motor with encoders. The structural support for the x-axis tube would be supported by a suspended steel wire rope preventing it from bending/cantilevering. The rigidity of the x-axis can be enhanced by using a longer cross section tube along the z axis direction.

A third tube can be used for the z-axis.

The tubes would be cheaper than using linear guides. It’s a simpler design and lots of wiring can be eliminated. If the x,y, and z axis had independent tiny solar panels (like 6 or 12v) to power some batteries that can be mounted on each of the axes then they can all wirelessly communicate with the raspberry pi via esp8266 modules or something similar.