The Aliqbot, a DIY Liquid Handling Robot

Published: October 17th, 2017   Last Modified: January 31st, 2018

Pipetting hundreds of tubes worth of reagent can be hard on your hands, and on your soul. In this video I try and make my life a bit easier by throwing together a DIY liquid handling robot in the cheapest/rickety-est method possible.

This is part 1 of a multi-part series where I make this beasty do my bidding. As I mention in the video the hardest part of a liquid handling robot is not the mechanics, but the programming and control. Pt. 2 will cover expanding this concept further.

I got in touch with the people at Opentrons (https://opentrons.com/) to see if I can get their software running on my machine, which they are willing to do, yay! From what I can gather from their github page that they are running their bot from a smoothieboard of some sort, which is a very capable motion controller.

A build log, of sorts

The 3D printer I re-purposed is the Tronxy X1. As is mentioned in the video, this printer is too small to be practical for a pipetting robot, go with something bigger with more Z-axis height.

The 3D printer comes with an all-in-one controller board, I swapped that out for a CNC arduino shield that can run GRBL, which is a cut-down Gcode interpreter. It is in theory possible to run GRBL on RAMPS boards which sometimes come with cheap 3D printers, but you’d have to do a fair bit of firmware tweaking to get it to work.

The Biohit ePET was pretty easy to re-purpose and is actually sturdy considering what I put it through. At a certain point the pipette was completely out of sync and throwing an Error 1, but turns out I had damaged a connection near the hall effect sensors while soldering a pad. The actuation pad is connected to a small transistor which has it’s gate controlled by the spindle enable pin of the CNC shield. FYI, different versions of GRBL change the pinout of the board, be careful of that.

It seems that the electrical noise from the motors will sometimes trigger the actuation of the pipette. I fixed this problem by using relay modules instead of transistors, which you will see in the part 2 video, yay!

Part 2 Stuff 

Let me just say up front, when I filmed the original video, Opentrons had somehow completely evaded my google search to see who had done this kind of stuff before. So, my goal of bringing pipetting robots to the masses was a bit premature. The fact that Opentrons can offer a pipetting robot with open source software at 3000$ is incredible, there are worse built 3D printers out there for double the money. Sure, the parts cost for the Opentrons One is likely only ~1000$. However, factor in how much your time is worth. If you’re not the tinkering type, building a pipetting robot from scratch (just the mechanics, mind) would take weeks, what with sourcing parts, manufacturing parts, troubleshooting etc. After you’ve done that math, 3000$ is a steal for a pre-built unit WITH techinical support.

So, what will I bring to the conversation? Well, I will be describing how I’m building the Aliqbot-Mini, and how to get the Opentrons software running on it, which should be of interest to the tinkering biologists out there.

Since Opentrons was kind enough to help me out with running their software, it would be pretty foolish for me to keep writing my own code. Their code, however, is run on smoothieware, meaning my dinky little CNC shield won’t be up to the task. A 5 axis smoothieboard costs between 200-300$ CAD, which is literally more than the all the nuts/bolt/structural stuff already in my machine. I decided to purchase the knock off MKS sbase 1.3 for about 60 bucks. It seems to run the Opentrons firmware just fine, however since there is very little support availible for knockoff boards this may come back and bite me. We shall see though, could be fine, right? right??

While I wait for parts, I decided to knock together a bigger frame to house the Aliqbot-mini. The 2020 extrusion I had should be more than sufficient, the robot won’t be handling the same loads as a CNC.

I just mocked up how many 96 well plates this thing could potentially hold. 2×4 plates is probably more realistic, I think I will have trouble reaching those edges.

Also, to use the Opentrons software I will likely have to convert to their stepper motor driven plunger pipetting system. The ePET could be made to work with their software however there isn’t an unlimited amount of used electronic pipettes out there, which is part of the reason why they probably went with manual pipettes. If someone was mass producing electronic pipette modules, things could be different.

UPDATE!
>>> Working Opentrons compatible robot <<<