yes, we will be open-sourcing the hardware, though not in quite the same way as with the OT-One. All the OT-One parts are off the shelf, 3D printed, or laser-cut, so putting all the files on github makes a lot of sense. But the OT-2 is a lot of custom parts injection molded / extruded / CNC'd by our production team in Shenzhen, and it is a lot harder to both share those designs on our end, and to make anything out of them as a hacker. So we'll be releasing a lot of high fidelity CAD files of the full assembly, as well as deeper dive documentation like this whitepaper on our new pipette designs: https://s3.amazonaws.com/opentrons-landing-img/pipettes/OT-2...
sorry if my answer is confusing. the short answer is: yes, there will be more (and more in-depth) open-source documentation of OT-2 than with OT-One, and not just limited to dumping everything in a github repo. This is rolling out in the coming weeks as things get finalized.
For those who are un-aware of laboratory automation, this is not a new industry. There are MANY automated liquid handlers on the market. Beckman, Eppendorf, Hamilton, and Tecan are the larger names.
What's new is the way Opentrons is offering value: thru low-cost hardware and open-source software. With the OT2 they ditched the use of a manual pipetters with dedicated 8 and 1 channel pipetters (a good move). The software looks more polished... but not caught up the industry yet. Definitely catering to the DIY crowd.
> The software looks more polished... but not caught up the industry yet.
most industry software in this area is garbage. i've seen the tecan software in operation; those folks should be embarrassed.
> Definitely catering to the DIY crowd.
i'm tempted to get one and see if we can't get it into production. the up front cost is only a bit more than an annual service contract for a tecan machine which doesn't do much more, and is less flexible.
My god I couldn't agree more. I used a Tecan liquid handler for years and wow did I hate it. It was very easy to fuck something up or make a change in the software without an error being thrown or the user being notified; the GUI was absolutely a black box to new users and I had to write my own manual to walk labmates through what to do to run known protocols; there was 0 attempt at optimizing movements to save time; the machine required manual calibration every single time it was used in a shared environment to make sure that someone hadn't changed tip alignment, plate holder position, or plate specs manually or accidentally by crashing a tip; it was trivially easy to break the arm or tips if you weren't paying attention and they cost a fortune to replace; there was no version control etc etc.
Their hardware was also total crap. Both the main pump and tip cleaning pump failed on me two times without any notification being thrown, the arm failed once, tips broke countless times. Our regional Tecan rep was on a first name basis with me for awhile and I wasn't even the owner of the machine.
I almost ended up taking a Tecan engineer certification class just so I could use the equipment as a knowledgeable user. Tecan also wanted to charge me something like 15 grand for an arm and tip solution which would handle 1536 well plates which I was able to replicate with 1 cent worth of plastic shims and a few of their black 10 uL tips -_-
I will absolutely recommend that my lab looks into buying one of these.
This is exactly the pain we are trying to help in the lab -- your story is all too common!
It is crazy to me that we are the first lab automation company to bring best practices of consumer tech UX and engineering to the wet lab, but its true. We're the first to use continuous integration / delivery, Agile development, usability testing, easy APIs / DSLs -- the list goes on. Would love to work with you and restore your faith in what lab automation can be!
A lot of software in science and engineering R&D seems pretty bad. Even the firmware/user interfaces to some instruments are terrible, like ease of operation is an afterthought.
Then you have giant turds like LabVIEW by National Instruments who really should know better. I really don't know why it is so popular.
> A lot of software in science and engineering R&D seems pretty bad. Even the firmware/user interfaces to some instruments are terrible, like ease of operation is an afterthought.
nobody shops for a $250k scientific instrument on the basis of the control software, unfortunately. it's all about the machine's resolution or speed or something compared to the competitors.
or even just what you're familiar with from your previous laboratory.
> Then you have giant turds like LabVIEW by National Instruments who really should know better. I really don't know why it is so popular.
they dump a lot of money into undergrad lab/education. or did, anyways. and 20 years ago if you just wanted to pull some one-off analog sensors into a PC for an industrial process (without being an EE), NI hardware/software might very well have been the cheapest/easiest/fastest way to do it.
Thanks! Through user research we learned that most labs dont have robots today because 1. they are too expensive and 2. they dont have space in their lab for the huge machines currently available. Thats why we focused on low-cost and compact; the OT-2 is designed to fit on 1/2 of a standard lab bench so biologists can work alongside their robot.
OT-One was deff for the DIY crowd and there are lots of great hacks folks have done on it with our open API and hardware like plasmotron.org or this Alexa integration: https://youtu.be/s4WgCs-tH3o
On the other hand, the OT-2 is designed for the mainstream biologist, the folks that dont want to have to write code or hack hardware, just want something plug-and-play to do their pipetting for them. So far in our usability tests and for beta users, its working! :)
If I could add a personal number three to your list: most robots are far from robust. Running a robot becomes a full time job and require a dedicated guru to maintain it. The activation energy to use the equipment means that you must be desperate enough to deal with the pain rather than do the job manually.
totally agree. lowering that "activation energy" is one of the main focuses of our UX design team, and making sure that the machine doesn't break at the smallest bump is a core part of our hardware engineering philosophy. thanks for the input, very validating for us :)
Thank you! Would love to work with you! We are already shipping OT-2 robots, and right now with the amount of demand we have there is a 6 week lead time (though it is increasing fast).
The 96-well head wont be an addition to this robot because it would make it too top heavy and risk tipping it over -- best we can fit on this guy is the 8-channel pipette already available. We'll start considering things like a 96-well head when we decide to move up-market and start making bigger robots, but for now we are very happy to be making lab robots accessible to labs with smaller budgets that today have to do all their pipetting by hand.
yes, we love smoothieware! it's existence is one of the main reasons we're able to offer this tech at such an unheard-of low price. we've forked it (available here: https://github.com/Opentrons/SmoothiewareOT) to work with our OT-2 motor controller, which has more stepper drivers than the smoothieboard and is designed to fit on the head of the robot to minimize cabling. full board design (as well as all of our electronics designs) are going up on github in the next few weeks!
Great to hear. I work very closely with the project (hardware QA), as well as RepRap. I have followed your machine for quite some time. I have mentioned it to pretty much everyone I know who is in the field of chemistry/biology or does fluid handling.
I bet things are a lot easier now that Smoothie has true 6 axis :)
It is great to see so many OSHW projects get going in the last few years. Excited to see what the next few years brings.
Congrats! We've had a good amount of fun playing with the OT-One (and also some frustrations). I think the large range of volumes is great, along with what seems to be more streamlined/accurate pipetting. I think the enclosure is also a solid feature.
One of my favorite features is the ability to use a simple Python API to define a protocol. Will this still be the case or do people have to use the GUI?
1 - yes, our python API can still be your main tool for creating protocols, no one needs to use the GUI if they dont want to. In fact, OT-2 runs a jupyter notebook onboard so you can just connect over wifi / USB, pull up the notebook in your browser, and live code to the robot (as well as run a python script through the App same w/ OT-One).
2 - yes, all OT-One users get a $2000 discount on the OT-2, so starting price for you is $2k :)
28 comments
[ 22.1 ms ] story [ 737 ms ] threadWhat's new is the way Opentrons is offering value: thru low-cost hardware and open-source software. With the OT2 they ditched the use of a manual pipetters with dedicated 8 and 1 channel pipetters (a good move). The software looks more polished... but not caught up the industry yet. Definitely catering to the DIY crowd.
The larger, highly capable robotic liquid handlers cost in the $100-500k range (examples: https://www.beckman.com/liquid-handlers, https://www.hamiltoncompany.com/products/automated-liquid-ha..., https://lifesciences.tecan.com/products/liquid_handling_and_...).
This little guy is the smallest of the batch, but interesting to keep an eye on.
most industry software in this area is garbage. i've seen the tecan software in operation; those folks should be embarrassed.
> Definitely catering to the DIY crowd.
i'm tempted to get one and see if we can't get it into production. the up front cost is only a bit more than an annual service contract for a tecan machine which doesn't do much more, and is less flexible.
Their hardware was also total crap. Both the main pump and tip cleaning pump failed on me two times without any notification being thrown, the arm failed once, tips broke countless times. Our regional Tecan rep was on a first name basis with me for awhile and I wasn't even the owner of the machine.
I almost ended up taking a Tecan engineer certification class just so I could use the equipment as a knowledgeable user. Tecan also wanted to charge me something like 15 grand for an arm and tip solution which would handle 1536 well plates which I was able to replicate with 1 cent worth of plastic shims and a few of their black 10 uL tips -_-
I will absolutely recommend that my lab looks into buying one of these.
It is crazy to me that we are the first lab automation company to bring best practices of consumer tech UX and engineering to the wet lab, but its true. We're the first to use continuous integration / delivery, Agile development, usability testing, easy APIs / DSLs -- the list goes on. Would love to work with you and restore your faith in what lab automation can be!
nobody shops for a $250k scientific instrument on the basis of the control software, unfortunately. it's all about the machine's resolution or speed or something compared to the competitors.
or even just what you're familiar with from your previous laboratory.
> Then you have giant turds like LabVIEW by National Instruments who really should know better. I really don't know why it is so popular.
they dump a lot of money into undergrad lab/education. or did, anyways. and 20 years ago if you just wanted to pull some one-off analog sensors into a PC for an industrial process (without being an EE), NI hardware/software might very well have been the cheapest/easiest/fastest way to do it.
OT-One was deff for the DIY crowd and there are lots of great hacks folks have done on it with our open API and hardware like plasmotron.org or this Alexa integration: https://youtu.be/s4WgCs-tH3o
On the other hand, the OT-2 is designed for the mainstream biologist, the folks that dont want to have to write code or hack hardware, just want something plug-and-play to do their pipetting for them. So far in our usability tests and for beta users, its working! :)
The 96-well head wont be an addition to this robot because it would make it too top heavy and risk tipping it over -- best we can fit on this guy is the 8-channel pipette already available. We'll start considering things like a 96-well head when we decide to move up-market and start making bigger robots, but for now we are very happy to be making lab robots accessible to labs with smaller budgets that today have to do all their pipetting by hand.
Are the newer machines still using Smoothieware?
I bet things are a lot easier now that Smoothie has true 6 axis :)
It is great to see so many OSHW projects get going in the last few years. Excited to see what the next few years brings.
One of my favorite features is the ability to use a simple Python API to define a protocol. Will this still be the case or do people have to use the GUI?
Is there a reasonable (cheap) path to upgrade?
1 - yes, our python API can still be your main tool for creating protocols, no one needs to use the GUI if they dont want to. In fact, OT-2 runs a jupyter notebook onboard so you can just connect over wifi / USB, pull up the notebook in your browser, and live code to the robot (as well as run a python script through the App same w/ OT-One).
2 - yes, all OT-One users get a $2000 discount on the OT-2, so starting price for you is $2k :)