Robotic arm for feeding plate reader

[Dakota Hamill]

Anyone ever purchased or re-built old robotic arms?  I've seen a few professionally built ones on ebay and a company donated some to a local college once a year ago but they disappeared.

I also came across this tonight, https://www.thingiverse.com/thing:3327968

Unsure if you need proprietary software to control them or if an old serial port and an arduino or RaspPi would work. 

Would like to be able to just leave stacks of 96 well plates and have it feed them into a plate reader, take readings, and put in the next one.  Simple for a human but a lot of sitting and clicking and it happens fast enough (1-2 minutes / read)  there isn't really enough time to do anything else in between reads. 

Thanks. 

[jlund256]

I can't address your question directly.  I've used commercial robot arms, and even with the installed easy-to-use software that knows about different plate types, has a grid for plates with known positions, etc., they are a pain to use.  Useful only for tasks that will be done the same way many, many times.  I would imagine even after the hardware is made or setup and working, the software end would take at least tens of hours to get set up (measuring plates, measuring the platform, test, repeat, etc.).  So if you want to do experiments, do them by hand, and after a few tens of repeats, when you know you'll be doing a lot more, then think about automation.

On the other hand, these automation robots are generally too big and heavy to ship, so they tend to sit around when labs are done with them.  I've seen older ones come up on surplus sites fairly often, and they tend to be inexpensive.  There are various different types of lab equipment that are basically a two axis positioning arm + something on the end, so you don't need to limit your search to just automation robots.

Cheers,

Jim L.

[John Griessen]

On 2/19/19 9:20 AM, jlund256 wrote:
> I've used commercial robot arms, and even with the installed easy-to-use software that knows about different plate types, has a
> grid for plates with known positions, etc., they are a pain to use.

<snip>

> On the other hand, these automation robots are generally too big and heavy to ship, so they tend to sit around when labs are done

> with them.  <snip> they tend to be inexpensive.

> There are various different types of lab equipment that are basically a two axis positioning arm + something on the end

I met a robotics programmer the other week who said ROS was a real thing that he uses all the time. ROS stands for robot
operating system and is used to connect little computer/microcontroller driven things via messages rather than the typical
motherboard only controls of an operating system. Something that connects via ROS is going to be good for lab automation.
Most of the old big heavy stuff won't though.

Another trend I found out about is how transcriptic.com uses robots -- they ask machine developers to come up with controls that
are automatable, then using whatever method the lab gear vendor builds, they communicate and control steps and transfer trays with
their own robot arms and gantry systems. They would not respond to questions about do they like ethernet or USB or radios or
what, just that it had to be established selling and they would adapt to it. So I guess they have to have their bots press lab
gear buttons sometimes for some hot equipment that does steps a researcher wants. They have rooms with many refrigerators with
stackers for trays in them, so they can address probably 0.5 million wells like a RAM (Random Access Memory) each containing
separate sterile samples of anything, and move it through clean rooms to do process steps on it..

To get some control of costs and not depend on rented rooms of equipment, but lab-benches/desktops of equipment, I think custom
made machines like my culture shock electroporator, and evolver systems of bioreactors, will be good, and especially if they run
python and talk between with ROS. Asking transcriptic for a quote would be interesting to hear about. I got as far as, "No we
won't deal with your half finished lab gear", and stopped talking with them.

One category of equipment that is not super huge and heavy or expensive used is stackers. There are some of those on ebay. Those
and some tiny blind bots based on microcontrollers might glue together to be useful. But their blindness is limiting.

Another tool that I think will become part of some good lab automation soon is OpenMVcam, a camera with computer that runs python
image processing codes to "do stuff". Instead of drive a complex human-like arm, an OpenMVcam could easily see so as to tell when
to throw switches on conveyor belts and start stackers doing their things, and input commands to a gantry bot with a "claw" or
bucket to transfer trays, and samples to be measured or processed.