Variability of coordinates during the year

I am working with a Farmbot XL Genius 1.4 since March, and managed to grow a complex pixelcropping design for a university research. However, I mentioned the location of plants and tools is changing slightly during the year. This is a pitty if you want to water seeds very precice on their correct location. For my experiment I seeded in different time periods. The FB always started from the same home position, but at the end the line of the seeding is a bit different per sowing period. This should be the same. With picking tools it is the same. One week it is working perfect, and the other week I need to adjust the location of the tools again with centimeters. Any ideas where this variability comes from?


Having to adjust tool locations in centimeters sounds quite a lot. I had to adjust by a few milimeters every ~2weeks, when tool pickup / drop sequences started to fail due to inaccuracy. Haven’t found solution to that problem, however I was not homing farmbot routinely / regularly during the day. Maybe that would help, not sure.

Can you post some pictures of your hardware implementation? I used a lot of wood and that explained why I had changes in centimeters to positions and such. I don’t recommend to use wood at least not for the interconnection between the rails. I used another aluminium rail to mount the toolholder and that seemed to work fine.

I did not try for a long time at the moment because I only could use the watering tool, all other tools are quite useless for me atm… :confused:

1 Like

Yeah, I heard the same stories about wood. We made a construction of stainless steel pipes which are connected with concrete tiles and long metal pins to the ground. It is a solid construction although it was hard to make it perfectly straight. So I think the variability of coordinates will have another reason in my case…

1 Like

Uuuuh like to exchange your garden with mine :stuck_out_tongue: hmm good question, nevertheless as you said these are quite some distances even for this small solid steel pipe. Do you have any other means of checking (measuring) different positions of the pipe to rule out that this things “moves”?

Do you home the bot always before any sequence? To rule out a zero shift?

Well, I’m not measuring the exact location of the pipes, but since the concrete construction with steel pins is quite stable, I don’t see this will cause changes in the coordinates. I do think straightness/equallity of the axis will have an influence, because if the bot is missing some steps due to slipping, it is not able to correct this. Every sequence is always starting at home position. So I mentioned the differences also occur when it starts exactly from home position. I have been testing this with seeding. One day I seeded, and the other day I watered them. The home position was exactly the same, but the water was dropped two centimeters next to the seeds…

@Gabriel @roryaronson :wave: any idea how to help @Talis?


hi tails,

impressive construction. Did you create these concrete things which lay on the soil by your own or did you buy it somewhere?

The seed injector needle is offset from the center of the tool/UTM by about 1.5cm (check the CAD model for exact offset), while the watering nozzle spray pattern is exactly centered on the tool/UTM. Perhaps you are noticing this difference?

If so, you can use an offset, or an extra move relative command when planting seeds or when watering to account for the differences in the tools.

Something else that could be happening is your FarmBot’s axes may be moving slightly when powered off due to the forces from the bent cables/tubing and cable carrier. You might try turning on ALWAYS POWER MOTORS for all of your axes to prevent this.


Do you use a belt tensioner? Although I expect seasonal belt variability would cause it to be more than mm out. I also use steel core belts.

Hi Fafi, Those are just conrete tiles of 30x30 cm you can get at garden centers

Yeah, the difference between those centers could be an explaination indeed. I will have a look. Although I observed the Farmbot is in general not always very precise. If it is looking for the home location for example, this can also vary a bit, especially on the X axis.